The University of Chicago Pressgratefully acknowledges a subventionfrom the National Science Foundationin partial support of the costsof production of this volume,

Contents

Part 1

A Guide to the Interpretation of BoneAccumulations in Afrícan Caves

Part 2

Fossil Assemblages from the Sterkfontein ValleyCaves: Analysis and Interpretation

Acknowledgments ix

l. Introduction 3

2. Parts of the Skeleton: Survival andDisappearance 11

3. Food Remains of Primitive People inSouthem African Caves 30

4. Food Remains of Camivores in AfricanCaves 56

5. Porcupines as Bone Collectors in AfricanCaves 109

6. The Contribution of Owls 118

7. Some Compressional Effects on BonesPreserved in Cave Breccia 134

8. Sumrnary: Bone Accumulations inSouthem African Caves-A Search forInterpretive Criteria 138

9. The Fossil Animals 147

10. Sterkfontein 190

11. Swartkrans 220

12. Kromdraai 248

13. A Note on Taung and Makapansgat 262

14. Who Were the Hunters and Who TheHunted? 266

Appendix: Tables 275

References 347

Index 363

vii

Acknowledgments

Paleontological work is usually time-consuming, and thisproject has beco no exceptíon. In my particular circum­stances 1 have beco able to complete it only through pro­longed misuse of time intended for recreation: the personwho has borne the brunt of this misuse is my wife Laura.who has supported me beyond measure with her particu­lar blend of loyalty and laughter. Our four children,Rosemary, Virginia, Timothy, and Conrad, have allhelped me in many ways, from sorting innumerable bonesto surveying the caves they carne from. In addition, Vir­ginia has drawn many of the diagrams and charts thatiIIustrate the text.

TIte investigatíon described in this book has beco, forme, an adventure of the mind, prompted by the imagina­tive concepts oí Professor R. A. Dart. 00 this adventuremy elmost daily companion has been Dr. Elisabeth Vrba,who has freely given me the benefits ofher lucid mind andunbounded enlhusiasm. She had also added a good deal tothe significance of this book through the results of herown research.

At the Transvaal Museum I would not have been ableto remain scientifically active without the help of Mrs. M.C, Erasmus, who has willingly shouldered much of whal Iwould normally have been expected to do.

Other museum cotleagues have also aided me in manyways, particularly Mrs. Elizabeth Voigt, who helped withthe analysis of bone accumulations during the early stagesof the project and is now continuing her studies of faunalremains from archaeologícal sites. Dr. Atan Kemp hasaided me wíth research in the Kruger National Park, andthe companionship of Mr. O. P. M. Prozesky was ap­preciated during fieldwork in Soulh-Wesl Mrica. Ms. Im­ogen Chesselet has prepared many of the drawings in thisbook; Mrs. Ronel Goode has tracked many obscure li­brary references for me, and Mrs. EIsa Kirsten typed themanuscript with patience and precision.

For seven years my field team at Swartkrans has beensupervised by Mr. George Moenda, and the services oCMr. Absalom Lebelo and Mr. Jack Sepeng have not geneunnoticed.

The Swartkrans site was acquired by the University ofthe Witwatersrand in 1968 and, since then, the Board ofControl ofthe Bemard Price lnstítute for PalaeontologicalResearch has generously aUowed me to continue my in­vestigations there. 1 am much indebted to Professor S. H.Haughton and Professor S. P. Jackson for their interest inthis work.

In the Anatorny Department of the University of theWitwatersrand Professor P. V. Tobías remains anesteemed colleague and warm friend, and he and Mr.Alun Hughes have been my companions during manydays of fruitful discussions at the caves. On sorne suchoccasions we have been joined by Dr. James Kitching,Mr. Brian and Dr. Judy Maguire, and Dr. Tim Partridge. 1have benefited greatIy from the experience and kindnessofthese people.

It would not have been possible to initiate tbe re­search described in this book without the personal interestof Mrs. Lita Osmundsen and the generosity of theWenner-Gren Foundation. The symposium this founda­tion sponsored in Austria during 1976helped to crystallizethe new science of taphonomy and to chart its futurecourse. My appreciation is due to fellow taphonomistsDr. Andrew Hill, Dr. Kay Behrensmeyer, and Dr. AlanWalker for their active guidance in this venture.

On geological aspects of this project 1 have had thebenefit of Professor Karl Butzer's wide experience andcritical appraisal. My research has been the better for 11.I have also appreciated the wise counsel of Professor F.Clark Howell and the fruitful cooperation of ProfessorRichard Klein. At the South African Museum, Dr. BrellHendey has generously helped me in a variety of ways.

It is a pleasure lo acknowledge the help and hospitalityof Mr. Attila Port and his wife Karen during productiveperiods of fieldwork in Soulh-Wesl Africa. I am alsograteful to Mr. C. K. Cooke, who was my companion andguide during many happy days spent in Rhodesian caves.Col. J. Scott has generously allowed me access to hisUitkornst nature reserve where so many of my observa­tions have been made.

Many thanks are due to the following friends and col­leagues without whose help the work described in thisbook would not have been completed: the late ProfessorW. W. Bishop, Professor C. S. Churcher, Dr. R. 1.Clarke, Mr. C. G. Coetzee, Professor H. B. S. Cooke , Dr.O. H. S. Davts, Professor H. J. Deacon, Mrs. J. Deacon,Dr. N. J. Dippenaar, Mr. W. du Plessis, Professor L.Freedman, Dr. C. E. Gow, Professor R. F. Holloway,Professor G. L1. Isaac. Professor T. Jenkins, Or. M. O.Leakey, Mr. R. E. F. Leakey , Professor A. E. Mann,Professor R. J. Mason, Professor H. McHenry, Mr. M.G. L. Milis, Dr. U. de V. Pienaar, Dr. I. L. Raulenbach,Professor J. T. Robinson, Dr. B. H. Sandelowsky, Dr. M.K. Seely, Miss V. SCOll, Professor J. D. Skinner, Mr. F.

ix

x Acknowledgments

van den Broek, Dr. W. E. Wendt, Professor M. H. Wol­poff, and Professor A. Zihlmann.

Finally my thanks must go to the Board of Trustees ofthe Transvaal Museum, under chairmanship of ProfessorF. C. Bloff the University Research División of theCouncil for Scientific and Industrial Research, headed by

Mr. W. J. Weideman, and the National MonumentsCouncil for their sympathetic support of my research.

Drawings initialed 1. M. C. were done by Imagen Ches­selet. AH photographs are by the author except thoseotherwise acknowledged in the legends.

Part 1A Guide to the Interpretation of BoneAccumulations in African Caves

1 Introduction

This is a detective story, but a rather odd one. The cluesare bones, and the airo of the investigation is lo establishcauses of death, but the evidence ís ancient and no wit­nesses survive to relate their experiences. More normaldetective stories often show sorne confident and efficientinspector systematically unraveling the evidence, withthe professional expertise of Scotland Yard al his elbow:in this case the investigator is a zoologist who found him­self in ao uncharted field where guidelines were few and iIldefined. What was needcd was sorne kind of paleodetec­tive ' s handbook-a guide to the interpretation of bonyc1ues found in African caves. Part J ofthis book forms therudiments of such a guide , and in Part 2 evidence from thecaves of Sterkfontein, Swartkrans, and Kromdraai is pre­sented and interpreted in terms ofthe guide's entena.

The title of this book , The Hurüers or the Hunted?, hasbeen used befare, in the same contexto After attending theThird Pan-African Congress 00 Prehistory at Livingstonein July 1955, S. L. Washbum visited the Wankie GameReserve to study baboons. Here it was not only baboonsthat c1aimed his attention; he also made observations 00

bones left by a variety of camivores, his interest havingbcen sparked by an extremely provoeative paper de­livered at the Livíngstone eongress. It was called "TheMakapansgat Australopithecine Osteodontokeratic Cul­ture," and in it R. A. Dart (1957a) drew sorne remarkableconclusions about early hominid behavior. From ananalysis oí more than 7,000 bones from Makapansgat,Dart concluded theat the collectíon represented food re­mains of Australopíthecus, who had apparently been ahighly effective hunter, capable of killing the Jargest andmost dangerous animals of the times. The unusually highproportion of cranial remains among the fossils was takento indicate that australopithecines had been headhunters,sometimes practicing their art on their own kind.

Looking at the remains of kills in the Wankie GameReserve, Washbum pondered Dart's far-reaching cIairns.He observed Ihal many of the kills retained their skullslong after other parts had disappeared; could it be thathyenas had taken such residual remains to the caves?That the australopithecines had in fact been the hunted ,rather than the hunters? Such a suggestion had alreadybeen made by K. P. Oakley (I954a,b) of the BrítishMuseum after his study tour of fossil sites in southernÁfrica. From [he point of view of understanding earlyman, the question was significant, as Washbum explainedin his paper .•Australopithecines: The Hunters or theHunted?" (1957, p. 612):

Thc taste for meat is one oí the main characteristicsdistinguishing man from the apes, and this habitchanges the whole way oí Iife. Hunting involves co­operation within the grnup, división of labor. sharingfood by adult males, wider interests, a great expansiónof territory, and the use of tools. It is therefore impcr­tant to date the beginning of hunting in order to inter­pret the origin ofhuman behaviour. Did roan take to thegrasslands because he was a hunter, or did he becomecamivorous long after leaving the forests? The answersto these questions may Iíe in the earlíest australo­pithecine deposits, those at Makapan and Sterkfontein.

Washbum was not the only delegate at the Livingstonecongress to be stirred by Dart's challenging c1aims. Insubsequent vears 1 foHowed the development of his thesisconceming "the predatory transition from ape to man"with great interest and in 1965 returned from Rhodesia tothe TransvaaJ Museum specifically to pursue the tapiefurther. My intention was to analyze bone accumulationsfrom the other australopithecíne sites of Sterkfontein,Swartkrans, and Kromdraai and lo see if they could shedmore light on man's predatory beginnings. It soon becameapparent that interpreting these collections would be pos­sible only after a good deal of background research onbone-accumulating agencies in African caves. Sorne ofthis work had now been done and is reported on in Part t.Results of the bone accumulation analyses thernselves,with an attempted interpretation, are presented in part 2.That the product oí my ten years' research 00 this tapieappears under the same title as Washbum's original paperis deliberate. OUT motives for writing were the sarne and Igratefully acknowledge the ímpetus bis rhoughts havegiven to my work.

Over forty years, many paleontologists have describedfossils from the cave breccias of Sterkfontein, Swart­krans, and Kromdraai, and the fossil animáis are nowreasonably well known. In the past, however, it hasnot been possible to view the collection from each site asa complete assemblage, to estímate minimum numbers ofaH the animaIs whose remains are involved, or to drawoverall interpretive conclusions. This is partIy becausethe bones typical1y occur in solid rack that has the con­sistency of concrete. They must be partially or whollyfreed from this enclosing matrix before they can beidentified as to skeletaJ part or animal taxon. This is anextraordinarily tedious business, so that the analysis thatcan now be presented, involving 19,487 bones from a

3

4 Inlroduetion

minimum of 1,331 animals is the result of an enormousamount oí painstaking work by rny paleontological col­leagues, rny helpers, and myself.

Dart's Predatory Hypothesis

The fossil animals slain by the man-apes al Maka­pansgat were so big that in 1925 1 was misled intobelieving that only human beings of advanced intelli­gence could have beco responsible roe such manlikehunting work as the bones revealed . . . . These Maka­pansgat prolomen, like Nimrod long after them,were mighty hunters.

They were also eallous and brutal. The mosl shock­ing specimen was the fractured lower jaw of a 12-year­old son of a manlike apeo Tbe loo hOO been killed by aviolent blow delivered with calculated accuracy 00 thepoint of the chin, either by a srnashing fist or a club.The bludgeon blow was so vícious that it had shatteredthe jaw on both sides of the face and knocked out aIl thefront teeth. That dramatíc specimen impelled me in1948 and the seven years following lo study furthertheir murderous and canníbalistic way oí life. [Dart1956b, pp. 325-26]

In a remarkable series of thirty-nine papees publishedbelween 1949 and 1965, Dart developed his hypothesis ofpredatory, cannibalistic australopithecines, who prac­ticed an osteodontokeratic (bone, tooth, horn) culture.Looking baek over his research in 1962, Dart (1962e j con­cluded that the evolutíon of his concept had passedthrough seven distinct stages,

The first stage involved the realization that many of theprimate skulls from the australopithecine-bearing cavesshowed damage that, in Dart's estímation, could have re­sulted only from purposeful predalory behavior on thepart of the early hominids. The sample consisted of 58baboon skulls, endocranial casts, oc other specimens­21 from Taung, 22 from Sterkfontein, and 15 fromMakapansgat-many of whieh showed depressed frac­tures oC their cranial vaults. In addition lo the baboons, 6australopithecine specimens were selected as showingevidence oC interpersonal violence. Dart went so far as toeone1ude that 64% of the skulls hOO reeeived blows deliv­ered direetly from the front and 17% from the left side;only 5%appearto have received blows deliveredfromtheright side. The australopithecines were, it seems, mainlyright-handed. But wíth what weapon were these lethalblows deJivered? In Dart's opiníon ít was a bludgeon con­sisting of the shaít and distal end of an antelope humerus,whose double-ridged extremity eaused characteristic de­pressions in the skulls of the prey. Recognition of thesebludgeons represented the second stage in the evolutionof Dart's hypothesis, and the concepts were presented intwo papers. "Tbe Predatory Implementa! Teehnique ofAustralopítbecus" (19490), and "The Bone-BludgeonHunling Teehnique ofAustralopíthecus" (1949b).

Stage three of the process consisted of "running thehyena myth lo earth." Ever sinee Dean Buekland (1822)attributed the very large bone accumulation in theKirkdale cave, Yorkshire, to the actívities of spottedhyenas, it had been usual to regard hyenas as importantcollectors of bones in caves. 8uch a concept has beenvigorously ehal1enged by Dart in bis paper "Tbe Myth ofthe Bone-Aeeumulating Hyena" (19560), his monographon the osteodontokeratic culture (1957b), and elsewhere.

Sorne views on the current status of hyenas as bone ac­cumulators are given in chapter 4.

Stage four in the evolution of Dart's hypothesis in­volved the recognition that "the transition from apehoodto manhood" had been conditioned by hunting behavior.In his paper "The Predatory Transition from Ape toMan" (19530, p. 209), Dart wrote: "On this thesis man'spredecessors differed from living apes in being conflrmedkillers: camivorous creatures, that seized living quarriesby violence. battered them to death, tore apart their bro­ken bodies, dismembered them limb from limb, slakingtheir ravenous thirst with the hot blood of victims andgreedily devouring livid writhíng flesh."

The fifth stage in the progress of Dart's appraisal ofAustralopithecus carne with the statisticaL analysis ofthe osteodontokeratic fragrnents from Makapansgat(l9570,b). Here sorne startling facts were broughl lo lightthrough the analysis of 7,159 fossils extracted from thegray breccia: 91.7% of the specimens were of bovid ori­gin, while 4.0% were from non-bovid ungulares and 4.3%were from other animals. Fossüs in the last category werealmost excluslvely cranial, leading Dart to suggest thatthe australopithecines were "headhunters." That verte-

I/

Introduction 5

Seven basic techniques thought by Dart (I964a) to havebeen practiced by the Makapansgat australopithecines.

1. The "crack and twist" technique, in which a longbone was given a blow 00 the shaft and the two ends werethen twisted aparto This resulted in a spiral fracture of theshaft, providing a perforating tool or blade as well as salidends that could be used as pounders (l9570,b, 1959g,1960e,e, 196tb, 19620, 1964b,e),

2, Sp1itting of bones longitudinally by forcing one boneinto the cavity of another. One example was a gazellehorn that had been thrust into the cavity of a spirallybroken femur of a larger antelope (l957e). Other exam­pIes have a1sobeen described (l965b) and compared withan Iron Age skin-preparing tool from the western Trans­vaal (Mason 1964), which consisted of a proximal radiusof a large bovid, surrounding a distal tibia shaft into whichhad been foreed a rib fragment as the cutting blade.

3. Localized or precisíon battering to produce scoopsor spatulate tools. The bone most frequently used was abovid metapodial, which was subjected to repetitive bat­tering on one surface with a pointed object (l957a,b,195gef, 19600, 19620J, 19640). Experiments were de­scribed in which as many as 140 blows were required(l95geJ, 19600)to produce a metapodial "scoop," simi-lar to "apple-corers" that were still in use during the pastcentury (Campbell 1959) and are known as paleolithic fos­

.~ sils from Egypt (Arkell 1957).

4. Making of composite tools, in which the cleft be­tween the two articular processes at the distal end of abovid cannon bone was used as a slot into which a tooth,bone, or stone fragment could be wedged, serving as areplaceab1e cutting blade (1957b, 1959[, 19600,c, 19620,1964e).

6 Introduction

5. Splitting of antelope skulls transversely into anteriorand posterior halves by striking them with a blade such asa scapula. The skull pieces were thought to have servedas bowls, mortars, platters, and saucers (l957a. b, 1961h.1962/, 1964c).

6. Detaching of the tibia at the talotibial joint by a blowthat severed the malleolus and gave access to the distaltibia, talus, and calcaneus-bones thought to have beenof cultural significance to the australopithecines (l957a,b, 1961b, 1962d).

7. Splitting of tendons and cutting of skins to formthongs. The tools thought to have been used for this pur­pose were specially prepared bone ñakes and sharp endsof distal tibial pieces. Smooth indentations on the ends oflong-bone shafts and flakes were thought to have resultedfrom the preparation of thongs of varying widths. Trans­verse grooves have also been observed 00 antelope limbbones, resulting, according to Dart , from abrasion byleather thongs (l957b, 1962e, 1964b).

In addition to these specific techniques, Dart visualizedthe australopithecines as using many other parts ofskeletons as tools or weapons-mandibular toothrows assaws, maxillae as scrapers, hom-cores as points and dag­gers, and so on. In fact, a use could be visualized for eachof the many thousands of bones preserved in the graybreccia al Makapansgat.

brae were almost entirely absent suggested that thehominids had been "professional decapitators." The bulkof the fossils were found to have come from 293 individualantelopes , the remains of which showed remarkableskeletal disproportion. Vcrtebrae again were scarce. butmost of those present carne from the upper neck region,suggesting that the antelope heads had been severed fromthe bodies high up. Dart interpreted the almost total ab­sence of tail vertebrae to mean that the tails had beensystematically employed as signals or flags and whipsoutside the cave." Interesting disproportions were alsopresent among limb bones; while 336 humen were found.for instance, only 56 femurs were present. Within Iimbbones themselves , disproporticns between proximal anddistal ends were marked: 336 distal ends of humen werefound associated with only 33 proximal ends. Dart inter­preted these disproportions as meaning that the aus­tralcpitheoines purposely selected certain parts of theskeleton for use as tools. These were brought back to thecave while others were left at the sites of the kills. Sornesupplementary information about these dispraportions isgiven in chapter 2 of this work.

Stage six of Dart's appraisal Involved the "direct anddetailed comparison between indisputably human and aus­tralopithecine bone splitting, and through eliminating theporcupine gnawing confusión." The most detailed com­parison was between the Makapansgat bone assemblageand that from Kalkbank, an open site 64 km northwest ofPietersburg in the central Transvaal (Mason 1958; Dartand Kitching 1958). An excavation in this old pan pro­duced 3.619 animal bones associated with 50 stone ar­tifacts and 38 waste fragments, dated at about 15,000years B. P. The bones. 25% of which had been gnawedby porcupines, were interpreted as human food remainsand were comparecí directly with specimens fromMakapansgat.

An assessment of the role of porcupines as bone col­Icctors in caves was made in DaI1's osteodontokeraticmonograph (1957b) and in his paper "Bone Tools andPorcupine Gnawing" (l958b). Further observations 00

this topic are presented in chapter 5.The seventh phase in Darts interpretive pracess was

reached in 1962 and consisted of "finding out more aboutwhat the man-apes did with the bones." In his review ofthe Makapansgat investigations from 1925-J963, Dart( 1964a) listed seven basic techniques he visualized theMakapansgat australopithecines as having practiced. (Seepp, 5-6.)

Analysis and Interpretation of Bone Assemblages

Dart'.s study of the Makapansgat fosslls was a pioneeringproject in that it represented the first analysis and inter­pretation of abone assemblage from an African cave.Before this, a great deal of paleontological work had beendone on fossils from caves. but this did not involve com­plete bone assemblages: it was concemed with evaluatingspecific fossils isolated from such assemblages.

Tbe analysis and interpretation of complete bone es­semblages or representative samples of them has come tobe known as taphonomy. The term was coined by EC­remov (1940) in a paper entitled "Taphonomy: A NewBranch of Palaeontology" and means lite rally the "Iawsof burial." It concerns itself with what happens to animalremains between death and fossilization. The aims oftaphonomic work are often very similar to those ofpaleoecological reconstruction-c-maklng use of the fossil

Introduction 7

assemblage to draw conclusions about the living commu­nity that gave rise to the remains and about the interactionofthat cornmunity with the environment of the time. Un­fortunately. the fossilized remains , or death assemblages ,are very different from the living communities that pro­duced them. Between death and fossilization a variety ofbiases opérate, and the evaluation of tbese biases pre­occupies a good deal of taphonomic research (Hill andWalker 1972;Hill 1975).

The aims of taphonomists are varied. Fossil assem­blages may be used 10 reconstruct the faunal compositionof the original cornmunity: the nature of the environmentin which the community lived: the process of communitysuccession: or the determination of relative ages ofdifferent cornmunities. AH these are valid objectives,but they do not coneern us particularly here. Mytaphonomic aims in this study are quite speciflc: toanalvze the fossíl assemblages from the caves 01Sterkfontein, Swartkrans, and Kromdraaí in arder lo de­cide how these bones may ñave found theír way truo thecaves and to draw conclusions about the behavior of thehominids and other animals that tnteracted with them,

The Caves and Their Fillings

The three well-known cave sites of Sterkfontein,Swartkrans, and Kromdraai occur within 3 km of oneanother in the Bloubank River Valley, 9 km north­northwest of Krugersdorp in the Transvaal. The countryrock is Precambrian dolornitic limestone of the Trans­vaal supergroup, dipping north at approxímately 300.Although the form oí the caves and their fillings may becomplexo as described in subsequent chapters, the siteshave presumably passed through the same generalizedstages of formation, as is shown by the idealized verticalsections in figure 1 IBrain 1975a). In stage 1 the cavernhas formed by solution of the dolomite in the phreaticzone beneath the water table; its contours may well havebeen determined by planes of weakness in the countryrack.

By stage 2 the water level has dropped through valleyincisión at sorne point in the general area. and the cavemis now air ñlled. It may be ventilated by a distant andindirect connection with the surface , and secondary cavetravertines such as stalactites and stalagmites havestarted to forro within it. Stage 3 shows aven formation inthe dolomite overlying the cavern. The avens representjoints, or other planes of weakness in the rack, enlargedby rainwater passing down through them. The passage ofthis water is accelerated by the presence of the cavem.since the water can drip freely fram the roof ofthe cavemas rapidly as it percolates down through the joint.

By stage 4, one of the avens has broken through to thesurface , providing the first direct link between the cavernand the outside. A talus cone begins to form beneath theaven, containing bones of animals living on the surface.This cone may be calcified by lirne-bearing solutíons drip­ping from the roof, in which case the resulting deposit isknown as a cave breccia.

Stage 5 shows the cavem almost fiIJed with breccia andthe roof further dissected by aven formation. In stage 6,surface erosion has removed much of the roof, exposingthe bone-bearing breccia on the surface. The sítes ofSterkfonteln, Makapansgat, Swartkrans, and Kromdraaiare currently at this stage. though in sorne cases the cavehistories have been complicated by floor subsidences orsecondary decalcification of the breccia masses.

8 Introduction

SUITABLE

consisting of bones of small animal s such as rodents, in­sectivores, birds, and reptiles that were almost certainlyderived from the regurgitations of owls roosting in thecave (chap. 6), and the macrovertebrate component made

6 up from the bones of larger animals whose remains foundtheir way into the caves in a variety ofways (chaps. 3.4.5).

The rnain concern of this book is with the macroverte­brates, althcugh samples of the microvertebrate fractionhave been prepared and analyzed wherever possible.

Stratigraphic work at the sites has resulted in subdivi­síon of the deposits into discrete members or other siteunits. The following is a summary of how the macrover­tebrate bone assemblages used in this study are relared tosuch subdivisions:

a

2

I~,',':l BONE DepOSIT

Fig. 2. Vertical sectkms through a hypothetical cave where tbe rate of4 inftux oí bones is equal in each case. Jo (a) the inftow of sediment is

minimal, resulting in a dense bone concentraríon: in (b) tbe bones aremuch diluted with sedimentary matrix. From Brain t975a.

30!

1»1 BRECCIA

o 15I

METRESSUITABlE SCAlE

1l]]]rRAVERTINE

3

.00lOMITE

Fig. 1. Vertical sections showing typicaJ slaICa in tite formation ofTrans­vaal dolomitic caves and their fossillfernus deposits. Details are gíven inthe text. FrOID Brain 19750.

The concentration oí bones in the cave breccias variesgreatly and is regulated by the rate of bone infiow relativeto that oí the enclosing matrix. Figure 2 shows two cavesituations where the number of bones passing down thetwo avens per unit oí time is the same. In case (a) the rateoí sediment accumulation is low, as a result oí the smallcatchment area round the cave entrance, so that the re­sultant deposit shows a high concentration oí bone. Bycontrast, the sediment accumulation in case (b) was com­paratively rapid, "diluting" the bone breccia and result­ing in a low concentration oí the bones themselves. Withthe exception oí the gray breccia al MakapansgatLimeworks, where the bones show remarkable concen­tration, the other cave breccias referred to in this studyshow varying degrees of sediment dilution,

The Bone Assemblages

As will be outlined in subsequent chapters, the bone ac­cumulations from the Sterkfontein valley caves fall natu­rally into two groups: the microvertebrate cornponent,

Sterkfontein Member 4 (ST 4) 1,895 specimensSterkfontein Member 5 (ST 5) 1,202 specimensSterkfontein Member6 (ST 6) 454 specimensSwartkrans Member l (SK 1) 2,372 specimensSwartkrans Member 2 (SK 2) 5,895 specimensSwartkrans Channel fill (SK C) 837 specimensKromdraai A,faunal site (KA) 1,847 specimensKromdraai B,australopithe-cine site (KB) 4,985 specimens

Total 19,487 specimens

Sterkfonteín

As will be described in chapter 9, this cave deposit hasbeen divided into six members. Although fossils areknown to occur in the lower levels, coUections from onlythe upper three members were available for this analysis.

Al! the specimens from Member 4 carne from Iime­mining and paleontologícal excavations in the Type Sitearea that occurred between 1936 and 1948. The recentwork of P. V. Tobias and A. R. Hughes al Sterkfonteinhas produced a wealth of other fossils from this and otherbreccia members, and their evaluation of this material ís

anticipated with much interest. Microvertebrate remainsin Member 4 breccia are not abundant, and acetic acidpreparation of the other fossils failed to pro vide me withan adequate sample for analysis.

Member 5 was explored paleontologically after stoneartifacts were found within it during 1956. Two seasonsof excavation (1957-58) by J. T. Robinson provided thebone assemblage analyzed here. The member containsa remarkably rieh mierovertebrate concentration, andRobioson sampled this extensively in the eourse of hisexcavation. Subsequently 1 dissolved a quantity of theseblocks in acetic aeid and analyzed their bone content. Thissarnple contained remains from a minimum of 644 indi­vidual animals, details of which are given in ehapter 9.

Member 6 was formerly koown as Robínson's "upperbreccia." It is an insignificant unit appreciably more re­cent than the others. Fossil bones are present, but micro­vertebrate remains are rareo

Swartkrans

The original fíeldwork conducled by the late RobertBroom and J. T. Robinson between 1948 and 19521ed lothe recovery of 3,600 fossils, including a very large andsignificant hominid coUection. My subsequent work at thesite between 1965 and 1975 added 5,504 macrovertebratefossils to the total. Formal separation of the Outer Cavefilling into Members 1 and 2 has occurred fairly reeently(Butzer 1976; Braín 19760), and the fossil samp1e hassince been divided between these members 00 thecharacteristics of the enclosing matrix, as is described inchapter lO.

lo addition to the two breccia members, channel fillsare present in the Outer Cave, varying a great deal in ageand degree of calciñcation. The bone sample from theSwartkrans channel fin was exeavated in 1973 and isappreciably younger than the main mass of Member 2breccia.

Microvertebrate bones are very abundant in theMember 1 sediment, and two samples were preparedand analyzed in the course of this study. A minimum of527 animals was found to have contributed to the twosamples.

Since I eompleted the manuscript of this book, newexcavations at Swartkrans have uncovered a hitherto un­recorded earlier component oC Member t. This extensivedeposit is older than that previously designated Member 1and is now being investigated.

Kromdraai

Australopithecine remains from Kromdraai are knownonly from site B, but it has been customary in the past toassociale the fauna from Krorndraai A (the "fauna! site")with the hominids from Krorndraai B (the "ape-rnansite"), although this association, in strict temporal terms,is now koown to be invalido For historical reasons,therefore, Kromdraai A is inc1uded in this study.

The maerovertebrate fossil assemblage from site Acarne from Broom's excavation of 1947; the mícroverte­brate sarnple, containing bones from a minimum of273 in­dividual animals, was obtained by dissolving blocks ofbreccia, recently colleeted at the site, in acetic acid.

The type specirnen of Paranthropus robustus carnefrom Kromdraai B in June 1938; it was foLlowed in 1941by a small number of other rernains, but the buIk of the

Introduction 9

sample analyzed here was derived from an excavation Iundertook there between March 1955 and May 1956.Most of the bones carne from the decalcified fringe of thebreccia mass along the north wall; they included an ex­tremely large microvertebrate collection that is beingstudied by D. H. S. Dav¡s and T. N. Pocock.

Sorne Methods and Procedures

My objectives in this study ofbone accumulations were toestablish what animals had contributed to the assemblageand by what skeletaJ parts these individuals were repre­sented. The first step in such an analysis of a fossil as­sembJage involves removing all bone pieces whosespecies can be identified with certainty. These form thebasis of the species list and generally consist of cranialpieces or other skeletal parts with diagnostic characteris­tics. For this stage of the work, a complete and well­organized osteological reference collection is indispens­able, for it is on this, as well as on the competence of theinvestigator, that the relíability of the investigations wilIdependo lf speciaJists on particular groups of animal s areavailable, problematic specimens will naturally be re­ferred to thern,

After removal of specífically identifiable specimens, asecond sorting is aimed at removing bone pieces referableto broader taxonomic categoríes such as "suid." "carni­vore," or "bovid." Special mention should perhaps bemade of the bovid or antelope groupings, since southernAfrican bone accumulations from Quatemary sites areoften dominated by antelope remains. Where these arefragmentary, or where many species are involved, it isgenerally not possible 10 do more than group the an­telopes from which they carne in size c1asses. For thispurpose, four antelope size classes have been proposed(Brain 1974a), based on the liveweights of the animalswhose remains are present in the bone accumulatíons.

Thirty-four extant species of antelope are currently ree­ognized in the southem African region, and in table 1these are arranged in order of mcreasing weight from dik­dik to eland. The list has been divided into four arbi­trary, partly overlapping categories as follows:

Antelope class 1, (}...23 kg liveweight, with the upperlimit represented by a large female common duiker.

Antelope class 11, 23-84 kg: upper Iimit, large maleblesbok.

Ante/ope class III, 84-2% kg: upper limit, large wil­debeest or roan antelope.

Antelope class IV. more than 296 kg: very Iarge animal ssuch as eland or buffelo.

Yarious investigators have used variarions of thisscheme of bovid size classes to suit their particular needs.In his study of bones from southem Cape caves, RichardKlein (19760) has employed a fifth size class to accom­modate the extinct giant buffalo, Pe/orovis antíqtuts, ananimal apprecíably larger than the living Cape bufTalo oreland. In a recent consideration of southern AfricanBovidae relative to habitat and camivore predation pat­terns, Elisabeth Vrba (l976b) has used four weight classesseparated al liveweights of 27, 125, and 343 kg. Thesefigures have cube roots of 3,5. and 7. which aids graphicrepresentation.

On completion of these sorting procedures, the samplewill have been reduced to a residue of fragments thatcannot be placed with confidence in any taxonomic cate­gory. One often finds that a large pan of this residue

10 Introduction

consists of pieces from the shafts of long bones, particu­larly of bovid origino In primitive human food remains,the long bones will generally have beco smashed toextraet marrow. resulting in characteristic bone frag­ments. These are referred to as bone fíakes ir they con­form to che following requirements: that they come fromthe shafts oC long bones such as the femur, radius, ormetapodial; that they lack complete articular ends: andthat they do not preserve more than half the circumfer­ence ofthe long-bolle shaft. In cases where more than halfthe circurnference of the shaft has been preserved, thespecimens are called sh aft pieces. Fragments showingrecognizable articular ends are classiñed according toanatomical parts and would therefore not form part of theresidue.

After removal of the bone flakes and shaft pieces, theremaínder of the residue is Usted as consisting of mis­celtaneous fragments.

In any interpretation of abone assemblage, it is impor­tant ro know how many individual animals of variouskínds have contributed to the sample. The estimation ofindividual animal numbers on the basis of abone as­semblage has been attempted in a variety of ways by dif­ferent workers (Clason 1972). The usual method is toestímate the minimum number of individuals from a countof the skeletal element occurring most frequently. If, forinstance, the assemblage ineludes 50 left distal humen ofa particular species, 45 right ones, and al] other elementsin lesser nurnbers, then the minimum number of individu­als for that species is 50. As Chaplin (1971) emphasized,skeletal elements should be matched against a1l others inthe sample by age, sex, and size in an attempt to decidewhich bones could belong to one individual. Reservationsabout the reliability ofthe "minimum number ofindividu­als" estimate have been expressed by Perkins (1973), whosuggested instead that the "relative frequency" of eachspecies should be determined. I have little doubt, how­ever, that for the purposes of this study the minimumnumber method is the more appropriate.

In the course of the sorting procedures outlined here,the bone pieces should be examined for special features.

\r,¡ ,

Bone ñakes

These include evidence of use as tools. surface abraision,cut-marks, carnivore damage, porcupine gnawing, andpathology. The incidence of such features in an as­semblage will greatly inñuence the result of any ínter­pretation that is attempted.

FinaHy, it is remarkable how much information can beobtained from the study of bone accumulations--oftenfrom those parts of the assemblage that, in the past , havebeen ignored or discarded by paleontologists. It is desir­able that al1 bone fragments from an excavation be re­tained, as the seerningly uninteresting fragments oftenprovide clues vital to the interpretation.

f,

2 Parts of the Skeleton:Survival and Disappearance

Occasionally, to the delight of paleontologists, the entireskeleton of a long-dead animal is preserved, with everybone present and in place. Such an event occurs only inspecial círcumstances: the body of the animal mus! havecome to rest in a place where it could lie undisturbed for agreat many years ; it must have been rapidJy covered witha suitable enclosing matrix, and it mus! have been pre­served and strengthened by percolating solutions ofappropriate chemical composition. More usual1y, the ani­mal' s body is subjected lo the destructive inñuences thatcharacterize any natural environment. The skeleton be­comes disarticulated-broken clown ínto its individualparts , each of which has then to contend with the atten­non of camivores and with the forces of decay anddisintegration.

A typical mammalian skeleton consists of parts thatdíffer greatly in size , shape , and ability to withstand de­structíve treatment. Compare, for instance , the astragalusand the scapula of an antelope shown in figure 3. One is acornpact, subspherical object, enormously resistant todamage , while the other is essentially a thin, bladelikestructure . fragile and delicate, apart from its head, whichis somewhat more robust.

Such difTerences may not interest a11 paleontologists­sorne specialists are concerned with individual fossilsonly-but when a study is made of a fossil assemblageand an explanation is sought for how the bones foundtheir way into a cave, then the presence or absence ofcertain parts of the skeletons can provide vital inter­pretative clues.

­•

Fig. 3. Variery of form in mamma!ian skeletal parts: a fragile bcvidscapula and a robust astragalus.

Very early in this study it became apparent to me thatthe consistent absence of certain skeletal parts from theSterkfontein valley fossil assemblages could well be re­lated to their original .delicacy and inability to survivedestruetive inñuences/Direct observation suggested thatsorne skeletal parts were more robust than others, bur anexperimental situation was clearly requíred to assess suchrobusticity objectivelY~JAt that time, in 1965, an interest­ing aceount, ..SkeJetal Durability and Preservation," hadjust been published by Chave (1964). It described the re­sutts of laboratory experiments aimed at testing the re­sistanee of various skeletal materials to destructiveforces, but it unfortunately restricted itselfto invertebrateanimals. The ealcareous skeletons of various marine in­vertebrates were tumbled in a porcelain barrel containingwater and either chert pebbles or sand. From time to timethe fossils were examined and weighed. It was eoncludcdthat the major factors controHing the resistance of shel1sto physícal destruction were the rmcroarchitecture of theshells themselves and the disposition of organic matrixamong the crystals of carbonate. Dense, fine-grainedskeletons like those of Nerita, Spisula, and Mytiluspreved most resistant to destruction-surely a result to beexpected.

During 1965 I was. plaAaiDg a similar series of experi­ments on bones from mammalian skeletons, and part ofthe projeet involved observationsofthe eñeces-eé-wee­thering 00 bonesin vaneas, env~ts. To this end [visited the Namib Desert Research Station, 96 km inlandfrom Walvis Bay, in South-West África, lo set up abone-weatbering experiment in that extremely arid envi­ronment. After ten years, the experiment has yieldedsorne interesting results (see lig. 120 in chapo 5). fj.116)

While in Ibis par! of lhe Namib. 1 happened to visitsevera! Hottentot villages and was struck by the abun­dance of goat bone fragments that lay about among thehuts. Making a small collection ofthese, llaid them out atthe research station and sorted thern into skeletal parts asan exercise in osteology. lt was immediately obvious thateertain parts were well represented, while others wererare or absent. Distal humen, for instance , were common.but, search as 1 might, 1 eould not discover a single prox­imal humeros. Tbe explanation was not dífficult to find-the sample represented the resistant residue of goatskeletons able to survive the treatment they had received.

BUI what was this treatment? Inquiries and observa­tions during the following week showed that goats werevirtually the only source of meat for the Hottentots; when

II

12 A Guide to the Interpretation of Bone Accumulations in African Caves

Fig. 5. The age and sex structure of the Hottentot populatíon in theKuiseb River villages.

fig.4. Map ofthe lcwer Kuiseb valley in South-west Africa: the riverbedforms the división between sand cunes to the south and bare grave! plainsto the north. Positions ofHottentot vüíages are shown scattered along thenorth bank; occupied villages are numbered. deserted village sües aremarked with crosses. 1, Roofbank: 2, Ururas: 3. Itusib: 4, Swartbank; 5,Klipneus: 6. Soutrivier: 7, Natab: 8. Osseweter.

'"

f>-'.::'~:-J males

_ females

~,..,s.....~

2() 30 40 50 60 70 -+Years 01 age

I10

50

10

A village is composed essentially of a farnily group , andfigure 6 shows a typical village genealogy. representingthe inhabitants of Soutrivier. In thc past marriagepartners have been found in neighboring villages. bUI asmigration to Walvis Bay has increased the selection ofpartners farther aficld has been possible.

Spacing of the villages along thc riverbank is de­termined by the number of gcats kept al each , becausegrazing is restricted to thc riverbed and the extent of a

t

i

The Kuiseb is one of several rivers that arise in the cen­tral plateau región of South-West Africa and make theirway across the Namib plain to the Atlantic Ocean. On tbeeastern fringe of the Namib plain the Kuiseb passesthrough a spectacular canyon beforc continuing almost160 km to its mouth in the vicinity of Walvis Bey. Thelower part of the Kuiseb course runs through highly aridcountrysidc where no more than a few centimeters of raincan be expected in a year. In the south is the Namib sandsea, with sorne of the highest dunes in the world; to thenorth are gravel plains with occasional outcrops of thePrecambrian country rock. Except for a few water boles,the Kuiseb river course is nonnally dry, although sub­surface Ilow sustains a dense growth of trees, sorne oftbem extremely large, Once or twice ayear, rains on theescarpment bring the Kuiseb down in ñood, scouring theriverbed of the dune sand that has encroached upon itfrom the sourh.

a goat was slaughtered, íts skeleton was treated in a tradi­tianal manner and those parts the Hottentots ccnsideredinedible were tossed to the dogs. When the dogs in theirturn had finished. a residue of parts unchcwable by Hot­tentot or dog was left to bleach on the desert surface.Here recovery W3S easy, since the ground was devoid ofvegetaüon.

Por sorneone in scarch of an experimental situationwhere the comparative robusticity of rnammalian skeletalparts could be tested, the activities of the Hottentots andtheir dogs proved a godsend. Only goat bones were in­volved, and these were being subjected to treatment thatcould still be observed. Here was, in fact, an experi­mental sltuancn that could hardly have been bettered­short, perhaps, of substituting australopithecines for theHottentots.

After the initial reconnaissance in 1965 (Brain I967b ,1969b, 1976b), I returned in March 1966 lo finish col­lecting all available bones and to investigate the cir­cumstances in greater detail, On tbis occasion 1 wasaccompanied by Trefor Jenkins of the South AfricanInstitute for Medical Research, who undertook a thoroughgenetic study of the Hottentot population (Jenkins andBrain 1967),

The Kuiseb River Environment

The Hottcntots and Their VilIages

At the lime of our 1966 study, the total population of thelower Kuiseb valley was 133 persons,living in eight sepa­rate víllages, all en the north bank of the Kuiseb, asshown in figure 4. Eighteen villages are known to haveexisted in recent years, but ten ofthese are now deserted.the people having moved to other centers in South-WcstAfrica. Table 2 lists the numbers of people, dogs, andgoats living at each of the eight occupied villages. A totalof61 males and 72 females were counted, together with 40dogs and 1,754 goats. The histogram in figure 5 shows theage- and sex-distribution of all human inhabítants of theviUages when our study was made. Very few individualsin the 20-40·year age group were found because in th.eirlate teens many leave home to seek employrnent in WaJ­vis Bay and elsewhere. Sorne girls marry while they areaway. hut many become pregnant, and in this event thebaby ¡s brought horne to grow up in the grandparents'care while the mother retums lo her work.

Parts of the Skeleton: Survtval and Disappearance 13

• • "-{] ,--- ,*, ,*, 8$Oe"'~

<! 9~1l'f1JI«Oo O " IIESIOE"t A' VlllAGt:

Q ~ ORESIDE"' hSE'WtE1IE

.,. SlJ8JEtl DEtEA5EO

óh I1 'rEAl! OF IIAIOIn~ g--o uru, .....Il"...l UOIle-

"III '.' PlIIEGlUoIH

'*. ,., ,- ,., ,- ,-'HE $OJTRIVIER VILLAGf f'C:DlúREE

Fig. 6. Genealogy of the people of the Soutrivier village, Jower Kuiseb valley (after Jenkins and Brain1967).

village's pasturage ís measured linearly along the Kuisebbed. The main stockowner at the Ossewater village ex­plained how, as the size of his original goat herd at Sout­rivier increased, he was obliged to move his home to anew village to avoid disputes over pasturage.

The villages thernselves are typically situated on thebanks overlooking the riverbed, exposed to the glaringsun. The cooler. shady riverbed is avoided, forfearoftheoccasional ftoods that sweep down the valley. The layoutof a typical village, in this case Soutrivier. is shown infigure 7. Three wells, each about 3 m deep, have beco dugclose to the riverbank; the upper part of eaeh ls lined withtree trunks , and the water is raised in a bueket suspendedfrom the end of a long eounterbalanced pole running overa fulcrum. Living quarters eonsist of beehive-shaped hutseovered with long strips ofbark fromAcacia albída trees,whieh grow in the riverbed (fig. 8). Occasionally the indi­vidual huts are eonneeted to form c1usters with ínter­leading doors.

Goats and lambs are kept in sepárate enclosures madeof dry tree trunks and branches embedded upright in thesand. A similar palisade had been eonstructed around asmal1 tobaceo garden on the riverbank and also protecteda few maize plants-the only attempt at agrieulture foundin any of the villages when the study was made.

The goats were found to subsist very largely on thefallen pods of Acacia albida trees (fíg. 9), and herding issimple because the onIy water usually avallable to theanimaIs is at the wells that have been dug at the villages.The goats therefore retum to the viJIage for water and,after their evening drink, are herded into kraals for thenight,

Most of the people in the villages regard themselves asTopnaar Hottentots, and a srnaller number considerthemselves Bergdamara. One c1aimed to be a Herero, andtwo men were of mixed origin--one had had a Hottentotrnother and an English father, the other a Bergdamaramother and a German father. The Hottentots were ofslight build and had light brown-yellow skin, in contrast tothe Bergdamara, who were more robust, darker, andmore negroid.

Sir Francis Galton, the famous English traveler andanthropologist. traveled a short distance up the Kuíseb

River at the start of his journey from Walvis Hay toDamaraland in 1850, He stayed for a while with mis­sionaries at Schleppmansdorf, now called Rooibank(Galton 1889), then moved up the Swakop River to Bar­men, where he deseribed the physical eharaeteristic:s ofthe Hottentots he saw there. In a letter to his cousin,Charles Darwin, he wrote on 23 February 1851:

1 have just left the land of the Honentots. I am surethat you will be curious to learn whether the Hotten­tot ladies are really endowed with that shape whichEuropean milliners so vainly attempt to imitare. Theyare so, it is a fact, Darwin. I have seen figures thatwould drive the females of our native landdesperate-figures that eould afford to scoff atCrinoline, nay more, as a seientific man and as a Ioverof the beautiful 1 have dexteriously even without theknowledge of tbe parties concemed resorted to actualmeasurement. Had 1 been a proficient in the language 1should have advanced, and bowed and smiled likeGoldney, 1should have explained the dress ofthe Jadiesof our eountry, 1 shouJd have said that the earth wasransacked for iron to afford steel springs, that the seaswere fished with eonsummate daring to obtainwhalebone, that far distant lands were cver-run topossess ourselves of eaotchoue-that these 3 productswere ingeniously wrought by eompeting artists, te the

Parts of the Skeleton: Survival and Dlsappearance /5

utmost perfection , that their handiwork was displayedin every street comer and advertiscd in everypcriodical but that on the other hand , rhat great as isEuropean skill , yet it was nothing before the handiwcrkof a bounteous nature. Here 1 should have blushed,bowed and smiled again , handed the tape and requestedthem lo make themselves the necessary measurementas 1 stood by and registe red the inches or rather yards.This , however. 1 could not do-there were none butMissionaries near to interpret for me, they would neverhave ente red into my feelings and therefore to them 1did not apply-but I sat al a distance with rny sextant,and as the Iadies retumed themselves about, as womenalways do, to be admired, ] surveyed them in everyway and subsequently measured the distance of thespot where they stood-worked out and tabulated theresults at my leisure. [Galton, in Pearson 1914]

A short while before Galton's visir, the Hottentots ínthe walvis Bay area had been described by the earlytraveler 1. E. Alexander (1838) as the talles! and best-builtofthe Namaquas; but when Mrs. Hoemle studied them in1923 she described them as "the most miserable remnantof all." ascribing their decline to the fact that they hadsold their stocks to sailors in Walvis Bay. in return forbeads. brandy , and "other worthless articles" (Hoernle1923).

Processes Involved in Building up the Bone Accumulation

Fairly detailed information 00 Hottentot butchering tech­niques and eating habits became available from directobservations and from questioning of local people. Anapparently typical goat-processing procedure will be de­scnbed, based on observations made in the Soutriviervillage during March 1966. The goat, a young maleestirnated to be one year old, in which the second molarswere about te erupt , was led to a particular tree whereslaughtering is normally carríed out (fig. 10). SeveralHonentots he Id the goat down on its left side whileanother cut íts throat with a pocketknife. The blood wascaught in an enamel basin and fed to two waiting dogs,who lapped it avidly. Once dead, the goat was suspendedby its hind feet from an overhanging branch and the skinwas removed complete, being split along the midventralline , along the insides of the limbs, and around the neckjust behind the horns. It was salted and pegged out in theshade. The abdominal cavity was opened next, and theviscera were removed; the stomach was slit open, itscontents emptied out, and íts lining washed. Thís. to­gether with the liver and kidneys, was said to be a del­icacy. The intestine, once the contents had beensqueezed out, was kept for making sausage. Other ab­dominal organs were fed to the dogs.

The front legs were then removed complete with thescapulae; lhe hind Iimbs were laken off wilh the in­nominate bones attached, by cutting through both thepubic symphysis and the sacroilíac joints. The feet weresevered from the legs at the metapodiaVphalangeal jointsand were taken by sorne children. who cooked themthemselves over a tire.

The ribs on one side of the carcass were separated attheir vertebral articulations, and finalIy the head was re~

moved. a knife being used to sever the axis from the thirdcervical vertebra. The atlas and axis vertebrae remainedattached to the ocdpu!.

AH "'e8t,i~}I G00ked befOlcit~eft. eitbeJ:llyboiling"H~etalpotsor b,.eiFcGt.....m:asting over.thefire. The head was dealt with as follows: the horns werebroken off at their bases by sharp blows from an ax andwere discardcd. Thc dogs chewed the horn-core basesbefare rejecting them. The complete head was then boiledfor several hours in a pot standing over the fire. AHediblemeat was picked from it and eaten, after which the braincase was smashed in the occipital region with a ham­merstone for removal of the braín. The skuU and mandi­bles were then passed to the dogs.

As eating progressed, all marrow-containing bones werebroken. They were held on a rock anvil and hammeredwith another stone , Neither the anvil nor the ham­merstone is an artifact in the usual sense of the word­they are simply suítable pieces of rock that happen to beal hand (fig. 11). The Holtentots seemed lo habitually ealwhile squatting on the ground, and their utensils werepocketknives, rock anvils, and hammerstones. Theirfeeding behavior seemed to be a mixture of longstandíngtradition and European ínñuence.

Once discarded by the Hottentots, the goat bones weregnawed sporadically for many days by the dogs, all ofwhich were jackallike in size (see fig. 10). Jackals them­selves were extremely rare in the vicinity of the villages atthe time of the study , and spotted hyenas had not beenseen in the area for sorne years, It therefore seems likelythat these scavengers do not enter the picture.

Pied crows are fairly common aJong the Kuiseb Riverand, when they can, wiU carry off scraps of meat, some­times with bones adhering to them. 00 one occasion in1966 a crow was seen ftying from the Soutrivíer villagewith most of a goar's tail in its bill.

When Iying fully exposed on the gravel surface, bonefragments become bleached and degreased within aboutthree months. Exposure to the SUR weathers the bonesurface, and a soñ, chalky superficial (ayer develops.Gnawing of the bones by gerbils of the genus Des­modillus, whose bUlTOWS are often concentrated aroundold goat kraals, is not uncommon.

While collecting bone fragments from the vicinity of theHottentot villages, 1 was surprised to find many piecesthat appeared to be bone tools. They tapered to points(fig. 12) and showed wear and polish that had surely re­sulted from human use. In reply to my queries, the Hor­tentots denied that they made use of bone tools at all, andI had to find a different explanation for the remarkablysuggestive appearance of these "pseudotools.." Furtherobservations showed that the worn and polished boneswere special1y abundant in areas regularly used by menand animals, such as around the Ossewater water hole,where 4{j() goats converge daily lo drink (fig. 13), in theimmediate vicinity of goat kraals, and along paths used bythe Hottentots and their goats in the riverbed. In pro­tected areas among rocks, for Instance, the bones woulddeveJop their characteristic chalky sulfaces but wouldlack signs of weaT and polish. The mechanism ofpseudotool production was therefore clearly related to thedisturbance of the sand in which the bones lay by the feetof anímals and men (BTain 1967c). The process maytherefore be surnrnarized as follows: bones come to reston the sand, and their sulfaces weather to a chalky con­sistency. Regular disturbance of the sand by the feet ofanimals abrades the chalky surface as it fonns, leading tobones that are both wom and polished. If the whole pieceof bone ís Iying in the disturbed sand zone, it is likely to

Parts of the Skeleton: Survival and Disappearance 17

..~ .

,- o," " •• • ,' ',-;"' ... \. 4 _ " ...

, . -::.~. ~ .'. ~.-, ".' ,. .-..' ~:.,. ;:_:..:, ~:~.;.- • .?, ~ ~ • -

Fig. J l. A Hollentot in lhe Soulrivier village breaking a goat limb bone with a qnartz hammerstone. The bones arebroken to extract Lhe marrow before being tossed lO Ihe dogs. From Brain 1976b. Courtesy C. K. Brain: copyright ©1976 by w. A. Benjamin, [nc.

acquire wear and polísh on al! surfaces (fig. 12a), but ifsorne part of it is buried deeper thís will remain protected,and only a part of its surface wíll be converted into apseudotool. Selective abrasion of this kind has been ob­served on a number of metapodial and other límb-bonepieces that had been buried with their long axes vertical,or at 'least inclined. Thís meant that parts of such boneswere buried too deep to be affected by superficial sandmovements, so that wear and polish occurred on one endonly (fig. 12b). Pseudotool production is not restricted toarid environments like that of the Kuiseb River, and themechanism should be borne in mind when any interpreta­tion of abone assemblage is undertaken. Sorne of theworn bones showed numerous shal!ow striations, usual!yat right angles to theír long axes. 1 suspect that these weremade either by the hooves of goats or by chewing bygoats. It rnay seem unlikely that noncarnivorous un­gula[es should regularly chew bones, but they certainlydo, as Sutcliffe (I973b, 1977) has documented in detail(see also chapo 7).

It seemed advisable to be able to separale the darnagedone to goat bones by the Hottentots themselves fmrnthat caused by their dogs. Consequently 1 bought a goatfrom one of the inhabitants of the Soutrivíer village andthen gave ít back to the people of the comrnunity. Overtwo days they consumed, in their traditional manner, al!that was edible of the goat and returned the bones to mewithout allowing theír dogs access to them. The goat wasa subadult male, and the following is a summary of dam­age to its skeleton as a result of Hottentot feeding:

Skull: The 18 cm horns were broken offat their

bases to allow cooking of the head; theocciput was smashed to allow removal ofthe brain; snout and palate were brokenoff as a unit; the mandible, in two halves,was undamaged.

Vertebrae: The head was removed by choppingthrough the axis; the atlas and part ofthe axis remained attached to the occi­pul. Very little damage was done to theother cervicals. Thoracic vertebrae suf­fered fairly extensive damage to theirdorsal spines and transverse processes.Lumbars showed slight damage to theirtransverse processes; the sacrnm wasundamaged. Of the caudals, only thefirst survived, the rest having beenchewed and eaten.

Ribs: SIight damage was suffered by the distalends only.

Scapulae: Undamaged.

Pelvis: This had been chopped through thepubic symphysis and across the ace­tabulums. There was no other damage.

Humeri: Both shafts were broken transverselyfor the extraction of marrow. Oneproximal end was completely chewedaway, the other was complete; bóthdistal ends were undamaged.

18 A Guide to the Interpretation of Bone Accumulations in African Caves

Fig. 12. "PseudotooJs" colJected around the Ossewaler water hoJe in theKuiseb Ríver bed: (a) specimens slJowing overaIJ wear and polish; (b)

bones with localized wear at points indicaled by ¡he arrows.

First molar unerupted: under 6 monthsFirst molar in use, second unerupted: 9-12 monthsSecond molarinuse, thirdunerupted: 15-30 monthsThird molar in use: more than 30 months

stuple Hottentot diet, apart from occasional meat, is~meaJ porridge. which very likely results in accel­erated dental decay. Jt is to he expected thut Stone Agepeople would have done even greater damage [O boneswith their teeth than do Kuiseb River Hotlentots (thistopic is mentioned again in chapo 7).

The Composition oC the Bone Accumulation

The collection made in the Hottentot villages duríng 1965and 1966 consisted of 2,373 goat bone pieces; the compof2:J/sition in terms of body parts is shown in table 3, The 7'C

minimum number of individual goats that contributed tthe sample, estimated on horns, is 190. Since the boneaccumulation was original1y described, however, I havefound that this figure is deceptively high. The reason is asfollows: in the extreme aridity of the Namib environment,horn is almost indestructible and lasts for many yearsafter the last trace of bone has disappeared. Part of theoriginal sample came from two deserted village sites thathad not been occupied for more than ten years. Theseyielded horns to the almost complete exclusion of otherskeletal parts. The average annual rainfaJI in the Kuisebstudy area is less than 2.5 cm per year; but in more normalenvironments, with more than 25 cm of rain annually,horn disappears rapidly, exposing the core, which iscomposed of easily destructible spongy bone. It is nowapparent that if data from the Kuiseb sample are com-pared with those for other bone accumulations, the in­cidence of horos wilI appear deceptively high. In thisdiscussion, therefore, horns will be omitted.

After horns, the most numerous single skeletal partsare mandibles. 1 found that the 188 fragments could bedivided into 53 left and 64 right half-mandibles. This in­dicates that a minimum of 64 individual goats contributedto the sample. Initially, 1 used aging criteria quoted" byCornwall (1956) for s.heep, but J. P. White kindly pointedout (pers. comm.) that these figures were almost certainlylow. Revised age estimates, based on Silver's (1969)figures for "rough goats" have therefore been made; they.,allow three months for first and second molars to comeinto wear after eruption. The aging cnteria are thereforeas follows, even though these estimates may be on thelow side according to recent information on known-agegoats given by Noddle (1974):

On this basis, indicated ages for the left and right half­rnandibles are given in table 4. The figures suggest thatthere was one goat in the sample under 6 months of age,23 between 9 and 12 rnonths, 7 between 15 and 30 months,and 35 more than 30 months of age. The goats had there­fore been slaughtered largely when either just under ayear of age or when fully mature. The Hottentots con­firmed that this was their usual practice, the yearlingsusually being the surplus males.

The combined feeding action of Hottentots and dogshas resulted in the disappearance of sorne parts of theskeletons and the survival of others. It has also resulted insome very charactenstic damage to certain parts. Suchdamage may be summarízed as fol1ows:

Skull: The braincase has been broken open by stoneimpact to allow removal of the bruin. In most

,!'

"h i

'\,#

a

b

Radii and ulnae: Both were severely shattered by stoneimpact.

Femurs: Heads and trochanters had been re­moved and proximal shaft ends chewed;both shafts were broken through themiddle; both distal epiphyses were re­moved and the distal shaft ends werechewed.

Tibiae: Both shafts were broken through therniddle, and there was sorne damage toeach end.

Metapodials: All four proximal ends were complete,but aH the distal epiphyses had beenremoved and the shaft ends' severelychewed back.

Carpal and tarsalbones: Undamaged.

Phalanges: Undamaged.

Apart from the results of stone impact,)t was surprisingto find that the Hattentots were capable of inflicting con­siderable damage on bones with their teetlí: Fifteen tailvertebrae were chewed and swallowed, aríd limb bonessuch as femurs and metapodials suffered severely attheir ends. It is doubtful that the condition of Hottentotteeth is as good as that of hunter-gatherer peoples. The

Parts of the Skeleton: Survival and Disappearance 19

Fig. 13. The Ossewaler water hoJe in Ihe Kujseb Rjver bed, where "pseudotools" have been colJected. Hcre boneshave been wom and polished by the movement of sand around Ihem. The movement has been caused by the feet of460 goatS Ihat converge on the water hole daily.

cases the occiput or floor ofthe skull has beenbroken out, producing a receptac1elike frag­ment (fig. 14). In most cases the palates havebeen detached from the braincases complete,and damage to mandibles is usually confinedto their lower margins, angles, and ascendingramio

Vertebrae: These show damage particularly on theirspines and processes.

Ribs: These have generally been chewed at bothends.

Scapulae: Extensive damage is usually present on theblades.

Pelvises: These have characteristically been gnaweddown to little more than their acetabular por­tions.

Damage to limb bOlles is best reflected by the presenceor absence in the sample of their ends, to be discussedshortly. Shafts have typically been broken through byhammerstone impact, and spiral fractures are common

(fig. 14). Carpal, tarsal, and phalangeal bones, when theyoccur, are generally undamaged.

Figure 15 diagrams the way different ends of the longbones are numericaIly represented in the sample. In thehumeros, for instancc, 82 distal ends were found, but nota single proximal end has survived. Again, the proximalend of the fused radius/ulna is represcnted by 62 pieces,compared with 19 distal ends. In the femur, the proximalend occurs twice as commonly as the distal end, but theposition ís reversed for the tibia, where the distal end isabout six times as common as the proximal end. Withboth metacarpal and metatarsal, proximal ends are more.commonly found than distal ends.

Su{'vival and Disappearance of Skeletal Parts

The survival of parts of these goat skeletons is clearlybased on their durability. Certain elements disappearwhen subjected to the chewing of Hottentots and theirdogs, others do noL The percentage survival of differentparts is therefore a measure of their resistance to this kindof destructíon.

Working on a mínimum number of 64 individual goats,

20 A Guide to the Interpretation of Bone Accumulations in Mrican Caves

, ,

cm.

Fig. 14. Examples of goat bones lhat have survived me feeding ac!ivity ofHotlenlots and lheir dogs. Distal humenare shown aboYe. cranial pieces below.

Parts of the Skeleton: Survival and Disappearance 21

The Makapansgat/Goat Comparison

lf we compare the plots of the Kuiseb bones (fig. 18a)with those ofthe Makapansgat fossils (fig. 17), we see thatthe form of the two histograms is similar. In both, partswith the highest percentage survival are mandibles,followed by distal humeri. At the lower end of thesurvival curve in both collections are such parts asthoracic and caudal vertebrae. But, in spite of a broadsimilarity of the two histograms, the detailed order of

Survival of Parts in the Makapansgat Bone Sample

Darts (l957a) analysis was undertaken on remains from293 antelopes. His estimation of minimum numbers ofindividual animals of different sizes was as follows: Iargeantelopes, based on 74 radius fragments, 39 individuals:medium antelopes. based 00 238 humeral fragments, 126individuals: small antelopes, based 00 191 mandiblepieces, 100 individuals; and very small antelopes, basedon 53 mandible fragments, 28 individuals, Using the totalnumber of 293 individuals, it has been possible tocalculate the percentage survival of different parts of theskeleton as was done for the Kuiseb River goat bones.Skeletal parts, lised in descendíng order of survival, aregiven in table 8 and plotted graphically in figure 17.

that is easily broken. making the shaft vulnerable to dam­age. This means that, when a year-old goat is caten. thedistal end of hume rus will be fully ossiñed and unchew­able. while and proximal end remains cartilaginous.Epiphyseal fusion times are known for domestic animals ,but al the time of rny earlier writing about the KuisebRiver goats (Brain 1967b, 1969b) data on goats were notavailable, and so figures for sheep were used. In themeantime, a study on ages of epiphyseal closure in feraland domestic goats has been carried out by Noddle(1974), whose figures are used here and are listed intable 6.

In addition to fusion times, structural considerationsare very important. The proximal end of the humeros iswide, thin-walled, and filled with spongy bone; the distalend is comparatively naITOW and compact. Such qualitiesmay be quantitatively expressed in terms of specific grav­ity of each end of the bone. The experimental procedurewas as follows: the shaft of a dry, defatted humeros wascut through at right angles to its axis, midway along thelength of the bone. Each end was weighed individually,and the cut ends of the hollow shaft were filled with Plas­ticíne. Any other openiñ"'gs were sirnilarly filled. The vol­ume of each end was then measured by submersion inwater, and specific gravities were calculated. [ found thatthe proximal end of a goat humeros had a specífic gravityof approximately 0.6, and that of the distal end was about1.0. There is a clear and direct relationship between thespecific gravity of the end of a long bone and its percent­age survival.

Table 7 gives figures for percentage survíval, specificgravity, and fusion time for each end of the goat limb­bones Iisted. These figures are plotted in figure 16. Per­centage survival of a part is related directíy to the specificgravity of that part, but inversely to the fusión time ex­pressed in months.

The conclusion to be drawn is sirnply that survi~snot haphazard but is. related l.<> the.ioberent.~theparts,

7~I " ~:" " ~; ' ~I

, ..0 ...... - __ IH"'ERuS

'd'"'' - ----,J>«>O.'''OI- -- •..... -- -

~.pIUS

'.",,, •• __ o' - __o

.~,",O""" -

l'E'-.c.::,~L _el_---'~

The Predictable Pattern of Survtvat in Limb Bones

It is clear that the parts of the goat skeletons that survivebest are the unchewable ones. Nevertheless, for limbbones, percentage survival can be related in quantitativeterms to particular qualities. In the course of this study itoccurred to me that, in a sample derived essentially fromirnmature animals, the survival of limb-bone ends couldbe related to the time at which the epiphysis of that bonefused to the shaft. Consider the humerus for Instance, inwhich survivaJ of the proximal end is.!!i1 but that of thedistal end amounts to 64%. The proximal epiphysis islikely to fuse to the shaft at about 36 months, whereas thedistal epiphysis is fully fused by 12 months. An unfusedepiphysis is linked to its shaft by a cartilaginous interface

Fig. 16. The percentage survival of proximal and distal ends of goat limbbones correlated with the specific gravities ofthe bone píeces and timesofIusion of the tpiphyses 10 the shafts. -

Fig. )S. Diagram of a goat skeleton showing numbers of each eOO of theprincipallimb bones found in lhe Kuiseb River sample.

it is possible to calculate thc original number of eachskeletal part that must ha ve existed, and from this onemay estimate the percentage survival of the part in thesarnple. In the case of ribs for, instance , 26 of which arefound in a single goat skeleton , thc original numbcr con­tributed by 64 goats must have been 1,664. Only 170 havebeen found, indicating a 10.2% survival. Table 5 showsdifferent parts of the goat skeleton arranged in descendingarder of survival, and these results are plotted in figure180. The parts most resistant to destruction are mandiblesand distal ends of humeri, and these are the most numer­ous. Proximal ends of hurneri and caudal vertebrae havepreved so vulnerable as to have disappeared completely.

22 A Guide to thc Interpretation of Bone Accumulations in African Caves

HUMERUS PROXIMAL

RIBS

FEMUR: PROXJMAl

bones left after a known attritional agent had acted uponthcm. They also aimed to measure characteri stics 01'bone s relevan¡ 10 their survival. Significant observationswere made on the food remains of Nunamiut Eskimos andtheir dogs as well as on thosc of Navajo Indians inArizona. Binford and Bertram were able to.conñrm thatborre density was a vital factor in survi~aL but thatbone densification was not an allometric growth processand that bones changed their density at different rateswith age:.- Applying their results to the Makapansgatasse mblage , they wrote 0977, p. 148): "We concludethat, based on the survivorship of anatomical parts atMakapansgat, there is absolutely no basis for theassumption that the hominids present played abehavioural role in the accumulation of the deposit.'

R. R. Inskeep (pers. comm.) has drawn my attention toan interestíng, and much earlier, parallet to the KuisebRiver goat-bone study. In his book Ear!y Man in Europe,Rau (1876, p. 111) described the conlents of Stone Agerefuse dumps, or kitchen middens , in Denmark , pointingout that the bones were the discarded food remains of thepeople and their dornestic dogs. To verify thisassumption , Rau described how Professor Steenstruplocked up sorne dogs, restricting them to a dier of bones,and thereby "ascertained that all the bones rejected bytbe dogs were the same that are present in thekitchen-middens , while the bones or portions of bonesdevoured by them are correspondingly missing there."

These observations were also reporteé on by Lub­bock (1865). Quite clearly. the characteristics of bonesdescribed here as influencing the survival or disappear­ance of such parts are not the only significant ones,The strength of attachment of one skeletal elernent toanother will also affecr the potenual survival of eachparto The strength and durability of interpart attach­ments wíll be indicated by the natural sequence ofdisarticulation of a skeleton after the animal's death.Detailed information on such disarticulation sequences isnow available from observations made by Andrew Hill(975) 00 skeletons, particularly those of topi(Damaliscus korrigum), in northem Kenya. Hill hasdescribed twenty-one stages in thc disarticulatíonprocess, starting with tbe first, where the forelimb,including the scapula, separates from the ríb cage , Secondto detach are the caudal vertebrae, and thereafter all theother parts of the skeleton separate in an orderdetermined by the nature of their attachments. Last toseparare are the cervical vertebrae, whose survivaldoubtless is enhanced by the durability of theirligamentous connections.

It Is not unusual for the separate specimens in aboneaccumulation to have been transported by water beforetheir deposition, and a factor of importance here is theease with which different skeletal parts can be carried inan aqueous current. A remarkable fossil bone concentra­tion of Pliocene age has been studíed in the Verdigrequarry of northeastem Nebraska by Voorhles (1%9)..Inan attempt to understand the reasons for the presence andabsence of different skeletal parts, there. Voorhiesundertook stream-table experiments that ,suggested thatcurrent sor:ting was probably responsible for the scarcityof elementi,such as ribs, vertebrae, sacra, and phalangescompared with rami, metapodials. and libiae.

Important taphonomic studies have recently been madeat the East African hominid localities by Kay Behr­ensmeyer (I975a,h). She has pointed out that bones are

80o

CAUDAL VERTEBRAE

THORACIC VERTEBRAE

LUMBAR VERTEBRAE

PHAlANGES

SACRUM

ATLAS

...xrs

FEMUR : OISTAl

ASTRAGALUS

CERVICAL 3·7 VERTE8RAE

CALCANEUS

PElVIS

METATARSAL: PROXIMAL

METATARSAl: DISTAL

RAOIUS , ULNA: DiSTAL

TIBIA: DISTAL

SCAPUlA

METACARPAl: PROXIMAL

METACARPAl: DISTAL

HUMERUS: DISTA.l

HALF MANOIBlES

TIBIA: PROXIMAL

HORN CORES

RAOlUS & UlNA: PROXIMAL

survival of parts differs. For direct comparison , thepercentage survival figures for the Makapansgat sampleare replotted in figure 18h so that they follow the orderlaid down by the goat bone s, Although the two histogramsare not identical , we can see that the trends in survivalarder are broadly similar.

When comparing these results one should bear in mindthat the Makapansgat sample is made up of borres fromanirnals ranging in size from eland to steenbok. They havealmost certainly been subjected to a variety of destructivetreatment , including feeding action of primary predatorsand scavengers. By contrast, the goat-bone sample ismade up ofbones from one species of small bovid. subjectonly to feeding by meo and domestic dogs. In view ofthis,the overall similarity in composition of the bonecollections is remarkable. It reñects the predictablepattem of survival that manifests itself when whole bovidskeletons are subjected to destructive treatment.

__The differential survival of anatomical parts inarchaeological contexts has been further explored in animportant contribution by Binford and Bertram (1977),whose aim was to obtain information on assemblages of

20 LO 60PERCENTAGE SURV1VAl

Fij. 17. The percentage survival of part~ of bovid skeleton~ fromMakapansgat. The sample consisted of bones from a minimum of 293individuals. From Brain 1976b (data reworked from Dan 1957h). CourtesyC. K. Brain: copyright © 1976by W. A. Benjamín. Inc.

Parts of the Skeleton: Survival and Disappearance 23

HALF MANDIGLES

HUHERUS DISTAL

TIBIA DISTAL

RADIUS 8. ULNA: PROXIMAL

METATARSAL PROXIMAL

SCAPULA

PELVIS

MET ACARPAL PROXIMAL

AXIS-----

ATLAS

METACARPAL DISTAL

RADIUS & ULNA DiSTAL

METATARSAl DiSTAL

FEMUR PROXIMAL

ASTRAGAlUS

CAlCANEUS

sras----

TIBIA· PROXIMAL

LUMBAR VERTEBRAE.

FEMUR DiSTAL

CERVICAL )·7 YERTEBRAE

THORACIC VERTEBRAE

PHAlANGES

SACRUM

HUME RUS : PROXIMAL

CAUDAL YERTE8RAE

OTHER PARTS

o 2o "'o 50 60 lOO 20PERCENT AGE SURVIVAl OF PARTS

a4.

b.0 ••

Pig. 18. (a) Percentage survival of parts of goat skeletons from the Kuiseb River villages. Calculationsare based on a minimum of 64 individuals. (b) Percentage survival of parts of bovid skeletons fromMakapansgat. arranged in the sameorder as for (e). From Brain 1976b. Courtesy C. K. Brain: copyright© 1976 by W. A. Benjamin, Ine.

not necessarily buried where an animal died, and thatthey are even less likely to have been buried where theanimal lived. In fact, it ls important to be able to distin­guish between autochthonous and alfochthonous as­sernblages: the former consist of bones that have not beenmoved far from the general environment where the ani­mals died, whereas the latter have been significantIytransported. in the course. of water transpon, cel18ift.parts of bominid skeletons are transported more readilyand farther than others, as is indicated by flume experi­ments (Boaz and Behrensrneyer 1976). Completecraniums pro-ved to be the fastest-moving elements, whilecranial fragments and isolated teeth fell into the lag group,to become buried in aggrading parts of a fluvial system.Behrensmeyer (1975b) suggested that such taphonomicsorting of skeletal parts could have accounted for thelarge concentration of hominid teeth recorded from the

Shungura Formation of the Omo basin in Ethiopia byHowell and Coppens (1974). Selective transport of differ­ent stone artifact types in water is also an important fac­tor. as experimental work by Glynn Isaac (I967b) hasshown.

When assemblages of hominid skeletal rernains fromOlduvai beds I-IV are compared with those from theShungura and Koobi Fora formations (Behrensrneyer1975b), it is found that the composition of the collectionsrelate to the degree of taphonomic sorting to which eachhas been subjected. The Omo fossils appear to have suf­fered the most from transport, sorting, and reworking, asis to be expected in a large, active river situation. Olduvaifossils have been affected the least , having been buried inlake-margin and deltaic environments, while the fossitsfrom the Koobi Fora Formation indicate effects inter­mediate between these two extremes.

24 A Guide to the Interpretation of Bone Accumulations in African Caves

There are at least two other ways in which skeletal-partdisproportions are often introduced into bone as­semblages. These are human butchery practices, to bediscussed in chapter 3, and selective camivore feedingpatterns. sorne of which are mentioned in chapter 4.

Comparatlve Vulnerability of Bovid and Primate Skeletons

A striking feature of the Sterkfontein valley fossil as­semblages is that antelope are represented by a far widerrange and abundance of skeletal parts than are primates.In fact, postcranial remains of horninids, baboons, andmcnkeys are rare in comparison with cranial ones,whereas bovid postcranial fossils are comparativelycommon. Dart (1957a) made a similar observation for theMakapansgat assemblage, and, since he assumed that thebones were horninid-collected, he likewise inferred thatthe australopithecines had been "professional de­capitators" and headhunters.

I will now discuss sorne observations relevant tobovid/prirnate skeletal disproportíons, made incidentallywhen I had intended to gather information 00 anothertherne. Early in 1966 1 decided lo examine the kind ofdarnage done lo prey skeletons by cheetahs, in the hopethat Ibis rnight throw sorne light on the way saber-toothedcats treated bones. The postcaníne dentition of a cheetahis specialized for slicing meat and includes a very smallbone-crushing component. In this respect cheetahs andsabre-toothed cats are similar, although the former aresmaller than most of the known southem Africanmachairodonts.

During February 1966 the Natal Parks Board wasacclírnatizing six adult cheetahs in a large enclosure in theUmfolosi game reserve, before their release there. Withlhe kind cooperation of the Nalal Parks Board staff, par­ticularly R. C. Bigalke and John Clarke, 1 was able lomake detailed observations of the cheetahs' feeding be­havior and food remains. It was soon apparent that thecheetahs did very little damage to the skelelons of theantelopes provided as food. The remains of an adultfemale bushbuck, Tragelaphus scriptus, are shown infigure 19a. As can be seen, the whole skeleton is there,even though six adult cheetahs had been feeding on it formosl of one day and part of the succeeding night. Damagewas restricted lo the distal ends of the ribs, the blades ofthe scapulae, and the vertebral processes.

Fieldwork in the Kruger National Park duriog 1967-68aIIowed further observations on cheetah food remains innatura) circurnstances, aJthough thesewere continuallyhaJilpel'éll"1)y"lllleifCiU__fro....;ac:kalS'",,<Hty.mas. Infact, it was extrernely difñcult lo find remains that had notbeen modified by scavengers. I was fortunate 10 have thehelp of A. C. Kemp, who was workiog in the field close loSalara al that lime, studying hombills. He look me lo thefew cheetah kills we were able lo locate. Remains of anadult male impala that had been killed and eaten by sev­eral cheetahs in the bed of lbe Timbavati River duringSeptember 1968, which we studied before scavengers ar­rived, are shown in figure 19b. As wilh lhe Umfolozibushbuck, lhe skelelon was virtually undamaged.

The problem oC interference by scavengers in thecheetah field observations suggested to me that controlledfeeding experiments would. afier alt, provide the mostsatisfactory results, and the opportunity to undertakethese presented itself at Valencia Ranch, in South-WestAfrica, lhrough lhe kindness and cooperalion of Attila F.

Port, the ranch owner. The Valencia Ranch environmentis described in sorne detail in chapter 4, since studies werealso made there of leopard food remains. As is shown infigure 20, two enclosures had been constructed clase tothe Valencia Ranch farrnhouse. The smaller enclosure,55 m by 28 m, contained a six-year-old female cheetah thathad been in captivity for four years; the larger, 55 m by 37m, housed three cheetahs: two adult males caught on 20and 23 January 1%8 and a female, somewhat smaller andcaught on 25 January 1968. These three had fonned partof a natural group and had beco trapped at a particulartree clase to the Hakos River leopard breeding lair de­scribed in chapter 4. This tree had an inclíned trunk, andthe cheetahs visited it regularly, climbing up the trunk anddefecating from an elevated position. 1 am not sure if thisis characteristic cheetah behavior, but the regular visitsthe cheetahs made to the tree certainly aided their cap­ture. In fact the two remaining members of the group,bothadult males, were caught 00 13 and 20 May 1968 andadded lo the three already in the enclosure.

The cheetahs were fed Iargely on karakul sheep, andbetween March and September 1%8 many observationswere made, confirming that cbeetah damage. lo .aduhsheep skeletons was minimal. On 22 March, delailed ob­servations were made as the three cheetahs in the largeenclosure fed 00 the complete carcass of a freshly shotspringbok, the following being an extract from rny fieldnotes at the time:

22 March 1968: an adult male springbok weighing 103 lbwas shot on the farro Friedental and placed in thethree-cheetah enclosure at 2 p.M.-a sunny and ratherhot day. It was immedialely dragged into the shade bythe two males while the female threatened us. All threecheetahs started to feed on the ventral surface of thebody (fig. 2Ia): one worked from there onto the innerparts of the hind legs, one chewed away the sternal ribattachments, and the third entered the thorax. The in­testines were partly eaten and partly discarded a fewmeters to one side. The two maJe cheetahs were obvi­ously dominant over the fernale, who was threatenedseveral times. One male left after 1 hr of feeding butreturned 15 mio. later, AII activity had ceased by5 P.M.

The remains were left in the enclosure until Sunday,24 March, when lbey were removed and photographed(fig. 21b). A good deaJ of the skin was still intact, butthis was cut away befare the photo was taken. Damagewas restricted to the ends of the ribs and the vertebralprocesses: the right scapula and humeros had beendamaged when the springbok was shol.

The pattem of damage lo this skelelon was typícal ofwhat 1 had observed previously, but lwo days earlier Ihad obtained sorne very dífferent results in a feeding ex­periment. Although we had spent a considerable timesearching for a suitable antelope to feed to the cheetahs,on that occasion we found none and instead shot a malebaboon who had been taunting us from the top of a nearbycliff. The baboon was offered to the three cheetahs, withquite unexpected results:

20 March 1968: The body of an adull male baboooweighing 29.5 kg was placed in the enclosure at 9:05A.M. It was immediately taken by the two malecheetahs and carried by its arms to the shade of a tree.AII lhree cheelahs slarted lo feed on lhe venlral sur-

26 A Guide to the lnterpretation of Bone Accwnulations in African Caves

Fig. 20. The Valencía Ranch environmeol io Soulh-West Africa where the cheetah-feedíng experiments were under­taken. The cheetahs were confined in the enclosures visible in lhe middle foreground.

íace of the abdomen; the viscera were removed andpart of the intenstine eaten. The rib cage was quicklychewed away and the vertebral column símplycrunched up and swatlowed-quite unlik:e the antelopesituation. As the vertebral column was destroyed, thepelvis and both hind limbs were removed by onecheetah and carned a short distance away. The sac­rum was eaten so that the femurs, still articulated intothe innominates, were separated. Qne cheetah left thebaboon after 1 hr, 10 min, the others remalned 15 minlonger, then left, but atl three retumed intel1l1ittentlythroughout the day.

The remains were removed and photographed the nextday (fig. 22a); the entire vertebral column, from atlas 10first caudal, had disappeared, as had most ofthe ribs. Theinnominate bones showed damage round the edges, andboth knee joints had been disarticulated and chewed.

The disappearance of the vertebral column in this ba­boon carcass carne as a complete surprise and suggestedthat a primate backbone was less resistant to carnivorechewing than its bovid counterpart. To test tbis suspicion,an adult sheep of almost precisely the same liveweight asthe baboon was fed.JQ,-~e cheetahs when they showedequivalent signs O~~g:y The result is shown in fig. 22b;although the innominales had been separated from thesacrum, the vertebral colurnn was intact except for thetall, which had been consumed.

During 1968, many controlled feedíng tests were dónewith the cheetahs on baboons and bovid prey. A charac­tenstic feeding sequence is shown in figure 23a-<:, inwhich the prey was a subaduIt male baboon whose ver­tebral column was also entirely eaten except for the tall.

1 do not intend to suggest that cheetahs would everhave eontributed food remains to bone accumulations inAfrican caves-they are simply amenable camivores onwhich experimental work can be done. Under controlledconditions they are able to do far more damage to theskele10n of a baboon than to that oí a bovid of equivalentliveweight. Most striking is the fact that a primate's ver­tebral column is far less resistant to camivore damagethan is an antelope's. The reason is clearly the compara­tive structure and robusticity of the two baekbones­individual boyid vertebrae are comparatively resistantstructures with more robust spines and proeesses. This isof the- greatest significance when bone accumulationscomposed oí primate and bovid remains are interpreted.

LPrimate vertebral columns will-suffer morefrom whatevercarnivore action they are subjected lo Ihan will bovidones_ In the same way, erimate hands and feet are -farmore palatable to a carnivore than are bovid equivalents.

In view ofthese observations, it is in no way surprisingthat antelopes should be represented in a fossil as­semblage by more abundant and more complete skeletalparts than are primates of similar size. The key lies indifferent degrees of resistance to equivalent camivoreaction.

3 Food Remains of Primitive Peoplein Southern African Caves

To date, the huoting way of life has been the mostsuccessful and persistent adaptation man has everachieved. Nor does this evaluation exclude the presentprecarious existence under the tbreat oí nuclearannihilation and the population explosion. It is still aoopen question whether man will be able to survive theexceedingly complex and unstable ecologicalconditions he has created for himself. If he fails in tbistask, interplanatary archaeologists of the future willclassify OUT planet as one in which a very long andstable period of small-scale hunting and gathering wasfollowed by an apparently instantaneous efftorescenceof technology and society leading rapid1y to extinction... Stratigraphically," the origin of agriculture andthermonuclear destruction will appear as essentiallysimultaneous. [Lee and Devore 1968, p. 3]

Hominids are changeable cre~tures"~d. in eomparisonwith most carnivores, are verY'tiresom~) A leopard oftwomillion years ago not onIy looke¡nike one of today, butprobably behaved in the same way. Giveo equivalentprey, it would presumably have left the same food re­mains then as now. Not so the hominids; Horno habiliswas a very different animal from Horno sapiens, not onlyin the Oexibility of its behavior, but also in the traces of itsformer presence. From the point oC view oC tbis in­vestigation, the cbaracteristics of early hominid foodremains are more relevant than those of their late de­seendants, but in paleontology one has to be content withwhat one can gel. Food remains resulting from the earlyphases of human existence are understandably less com­mon than those from later phases, and so it is that thetraces of Middle and Later Stone Age people are the bet­ter documented. In an attempt 10 gain sorne insights intothe nature of primitive human Cood remains, I haveanalyzed extensive bone samples from four caves or rockshelters that, on arehaeological grounds, appear to havebeen intensively occupied by Stone Age peoples. Thesites have a wide geographic scatter in soutbem Afriea,one being in Zimbabwe, one in tbe Transvaal, one in theCape. and one in South-West Africa. Sorne details ofthese sites and bone samples will now be considered in anattempt to discem consistent pattems between them.They will then be examined in the context of results fromother sites. obtaioed by other investigators.

Pomongwe Cave

The cave is in the Matopo Hills south of Bulawayo, anarea of rugged granite hills that will be discussed further

30

in connection with leopards and black eagles (chap. 4) andCape eagle owls (chap. 6); the cave has formed near thebase of a very large granite dome, whose convergingsides form a natural tree-filled amphitheater at thecave's mouth (fig. 24). The Dame Pomongwe appears tobe of Karanga origin, perhaps derived from the wordmamongwe. the term for a small wild melon of whoseshape the granite dome is reminiscent (C.K. Cooke (963).

The cave, which has resulted from negative spheroidalweathering of the granite, is dome-shaped and dry-aplace exceptionally suitable for human habitation and onethat has been used for many tbousands of years. In plan(lig. 25a) it is roughly triangular with a gently sloping ñoor(fig. 25b) descending into a delightful tree-Iilled glade .Two springs arise on the graníte hill and form pools closeto the cave's entrance almost throughout the year. Withthe abundance of wildlife in the area, 1 can think of fewmore desirable plaees to live.

Three separate excavations were conducted by C. K.Cooke and his helpers, the ñrst commencing in June 1960(e. K. Cccke 1963). Bedrock was reached at a depth of 13ñ, 6 in, or 4.12 m, and the profile was fouod to be made upto twenty-seven definable layers, the upper ones rich inash of various colors. while tbe lower part of tbe sectionconsisted of granitic sand with a certain amount of bonebreccia. A striking feature ofthe profile was a layer rich ingranite spalls, typically occurring l m below the surface.The spalls had become detached and had fallen from tberoof and walls of tbe cave, perhaps in response to sornechanging conditions of temperature or humidity; accord­ing to currently available dates, this happened between16,000 and 20,000 years ago.

00 cultural grounds the profile was originally dividedinto ten stratigraphic units (C. K. Cooke 1963). Readingfrom the bottom up, these were Proto-Stillbay, layers27-21; Stillbay, 20-13; Scraper 1, or Magosian, 12-10;Scraper 2, 9-7; and Wilton, 6-surface. The culturalphases have undergone many changes in nomenclature(e.g .• e. K. Cooke 1966, 1968, 1969), which 1 am notcompetent to assess, and the subdivisions should prob­ably oow read: Bembesi, layer 27; Chararna, 26-22;Bambata, 21-I3B; Tshangula, l3A-IO; Pomongwe, 7; andMatopo, 6-1. Figure 26 shows subdivisions of the profileinto levels and cultural uruts, linked to radiocarbon datesand to the occurrence of the 17.756 bone fragments and39,032 stone artifacts recovered from the excavation.

lo recent years, radiometric dating has suggested thatmany of the cultural phases in the southern AfrieanPíeistocene are far more ancient than had been previously

32 A Guide to the Interpretation of Bone Accumulations in African Caves

5001 llIllllO ISllllO o looa..-o 1IF AIITEFAtTS flllMlEl llJ lllItai

.., SIDIa lMaI IIlI1

Fil. 26. Profile through the ftoordepoait in the Pomongwe Cave showin¡:tbe relationship of artifact numbers per jevel lo those of bcnes.

people walking around in the cave. 1am inclined to thinkthat thc inñuence of trampling would be les s for thePomongwe deposit than for others, such as Wilton Shel­ter, since the soft Pomongwe ash layers would presum­ably have cushioned and protected the fragments. Theoverall size distributionof bone fragments in an accumu­lation will indicate fragmentation but will of course beinfiuenced strongly by the ioitial sizes of the completebones and by the sizes of the animals that contributedthem. More meaningful as an index of fragmentation arethe sizes ofbone ftakes derived from the limb bones oftheprey, particularly antelope. Body size of the originalanimals will of course also constitute a factor here, butto a lesser extent than with the whole bone sample.

As mentioned earlier, of tbe 17.756 bone pieces in theentire sample, 9,549, or 53.8%. consisted of bone ftakes.Each of these was individually measured, and the resultsare given in table 10. As figure 28 shows, the length dis­tributions of the bone ftakes from the various Pomongwelayers are remarkably constant, though the samples werebroken by difIerent people whose combined lives spannedaD immense period of time. It seems clear that theefficient use of prey skeletons as food by Stone Agepeople was not random and haphazard. Long bonesfor instance, were systematically broken with hammer­stones so that no marrowwas wasted. This procedure atPomongwe has resulted in 3D abundance of bone tlakesbetween 1 in and 2 in (25-5.2 cm) long. As will be dís­cussed in chapter 7, the production of these bone ftakescauses extremely few marks 00 the bone píeces that in­dubitably indicate humanaction-the same is true of boneftakes that I have made experimentally with a stone ham­mer and anvil. However, theirassociation with very largenurnbers of stone artifacts and their burial in ash fromartificial hearths suggests that these are indeed food re­mains of people.

Although many of the bone fragmenta from Pomongwewere charred lo sorne extent, the actual frequency ofcharring was not determined. To do this, each piecewould have had lo be broken so that a fresh surface couldbe examined. This was not thought feasible or justifiable.

On the basis of all bone fragments that could be placedtaxooornically, mioimum oumbers of individual anirnalsthat must have contributed to the sample were estimatedfor each level. The list for the combined Pre-MSA levelsis given in lable 11, forlbe MSAlevels, in lable 12. and forthe Later Stone Age horízons, in lable 13. Obviously thevarious animals on these lists, depicted in figure 29, var­ied considerably in the quantity of edible meat lbey wouldhave provided. An attempt has lberefore been made toestimate the percentage contribution each would havemade if the sample were a true reftection of the overallmeat diet of the hunters. Quite clearly, such an assump­tion is dangerously untrue, owing to particular butcherypractices, portability of kills, and similar factors that willbe discussed shortly. But, accepting the existence ofthese biases, we can pío sorne rough indication of whatthe people ate inslde the cave. Mean liveweights for eachof the animal laxa listed are given in the tables, and theassumed contribution to tbe diet is then calculated,working on the assumption that 70% of each animal'sweight would have been edible. In bis paper "A Methodof Calculating the Dietary Percentage of Various FoodAnimals Utilized by AboriginaJ Peoples." White (1953a)concluded that 50% by weight of animals such as bison,elk, deer, and antelope would have been edible, as op-

iPlen

2 , 6

metre.

o 5 10

lIltl".

1"" cner Wlfholt .,.-,01 Illu. lIt1p

,*, Co."'•.. ".."...... .-IUIIO n.......11.,..

ITlllUY

1.......1

4Ueo

appear to have served as convenient seats. The mainfire-making areas were in the central area of the cavefloor.

The most striking feature of the bone sample fromPomongwe was the extreme fragmentation of the individ­ual pieces, there being virtually no complete bones (fig.27). Of the 17,756 pieces, 9,549 were found to consist oflong-bone ñakes, as defined in chapter 1. and there were981 fragments too incomplete to be placed skeletally ortaxonomically. As listed in table 9, 5,364 fragments couldbe identified, with varying degrees of conñdence, as toskeletal part; there were 1, 357 pieces of tortoise carapaceand plastron; 366 ostrich eggshell pieces, 136 land-snailfragments, and 3 mussel shells.

The fragmentation of the bones at Pomongwe can beattributed largely to two factors: purposeful breakage byStone Age people to extract marrow, and trampling by

F'i&, 25. (a) PIaD of thcPomonawe Cave witb thc positioa of thcexcavaM

tions sbown. (b) Sectiootbroulh tbe cave andtalus hcap. RedrawnalterC. K. Cookc (1%3).

Food Remains of Primitive People in Southern African Caves 33

posed to 70% for others like pigs, raccoons, and badgers.In view of the intense fragmentation to which thepomongwe people subjected their prey skeletons, 1 havetaken 70% edibility for al! animals on (he lists.

Table )4 gives figures for the percentage contributionsto the diet of the people at Pomongwe made by the vari­ous animal groups, and this information is presentedgraphically in figure 30. Most of the meat eaten carne frommedium-sized antelopes (c1ass III) and nonbovid mam­mals, particuIarly das sies, zebras, and warthogs.Although many tortoises were collected, their contrib­ution to the overall diet was very small.

In an analysis of this sort, information on the bodyparts by which each kind of animal is represented wouldbe useful in reconstrocting something of the behavior ofthe hunters. Unfortunately, the extreme fragmentation ofthe remains makes this impracticable at Pomongwe for aHthe larger animals. With dassies, however, the situation is

more hopeful. Two species, Procavia capensis andHell'rolzyrax hruceí, occur very commonly among thegranite boulders in the Matopos. and both were obtainedin large numbers by the Stone Age people, ei[her byhunting or by snaring. The two forms may be separatedeasily by cranial remains, but this is not troe for post­cranial bones. For the present discussion the remains areconsidered together.

The Pomongwe sample was found to contain 1,192pieces of dassie' skeleton, as listed in table 15. On the }lU?¡fbasis of the most commonly occurring part, the distalhumeros, at least 96 individual animals are represented inthe sample, but the minimum number demanded by partsseparated into discrete leveIs is 114 animals. It appearsthat the dassie bones were not broken up with stonehammers as were those of larger prey, but that they weresimply chewed by the people, presumably after the wholecarcass was roasted over a fire. The chewable parts were

Hg. 27. Bone fragments representing Stone Age human food remains from Pomongwe Cave. The extreme fragrnen­tation appears typical oC bones broken up with harnmen;tones for lhe extraction oC marrow.

,M>rJ&.,,-~~ > ~«~ ,e-~34 A Guide to the Interpretation of Bone Accumulations in African Caves

Fig, 28. Lengths ofbone nates from various leveb in the Pomcngwe Cavedeposito As shown in the histogrems, che majority of tlakes fmm aU tevetsare between 2.5 and 5 cm in length,

The samples from the two lowest layers contained solittle radioaclive carbon that only a mínimum age could begiven, but the measurements iodicated that MSA culturesexísted before 51,000 R.c.-an unexpecledly early dale althat time. According lo Mason's (1969) inlerpretation ofthe evidence, the cave was unoccupied for a considerableperiod al the end of the Middle Stone Age before LaterStone Age peoples again made use of it, In lhis regard hewrote (1969. p. 57),

Layers 27 and 28 are apparently al the contaclbetween Later and Middle Stone Age occupations ofthe cave. Layer 27 is the bottommosl Iayer containingL.S.A. artefacts, bul a1so contains M.S.A. artefacts,Charred wood GrN 4815 from layer 27 dated 10210B.C.like the wood GrN 4816 from Iayer 28 daled 10560B.C. may come from Later Stone Age occupanls ofBushman Rack who walked ínto the surface of MiddleStone Age gravel layer 28 and dislurbed it, adding thewood GrN 4816 lo the gravel and a1so churning up theM.S.A. tools found in layer 27.

Further excavations should try to find evidence onthe Later Stone Age-Middle Stone Age contact alBushman Rock, but for the moment it seems thatM.S.A. occupation ended in Iayer 28 al about51000-45500 BC and L.S.A. occupalion began in layer27 al c. 10560 B.C.

although what is eaten or rejected appears to depend onthe traditions of a particular people. At the same time. wemust remember that the staple diet of African hunter­gatherers has probably always been vegetable-a matterto be discussed shortly-and that meat has represented amuch sought-after, but often unessentiaL. bonus.

The Bushanan Rock Shelter

The site is in a south-facing dolomite ridge (fig. 32) closelo the Echo Caves. 30"38'E, 24'35'5. in the OhrigsladDistrict of the eastem Transvaal. A gentle slope leads upto the shelter from the valley floor, as shown in the .sec­tion (fig. 33). and the shelter ítself is rougbly triangular inplan (fig. 33), aboul 52 m wide, 23 m deep, and 14 m high.A few paintings in poor condition still decorate its walls.

The first systematic excavation of the floor deposit wasundertaken in 1%5 by A. W. Louw afier the discovery ofMiddle Stone Age artifacls in pils dug by the farmer whowas using the shelter as a tobacco bam. The excavationcovered an area of 15 ft. by 5 ft. (4.6 m by 1.5 m) anddescended lo a depth of 8 ft., or 2.4 m. On the basis oflíthology, Louw (1%9) divided the profile inlo forty-threelayers, 1-2 resulting from Bantu occupation, 3-26 fromLater Stone Age, and 27-43 from Middle Slone Agehabitation. Charcoal was found to be abundant in thesediment, and seven radiocarbon dates were obtained(Vogel 1%9). These read:

Layer 9: 9.510 ± 55 R.P.12: 9,940 ± 80 R.P.21: 12,090 ± 95 R.P.27: 12,160 ± 95 R.P.28: 12,510 ± 105 R.P.38: 51.000 R.C.41: 45,500 R.C.

Tbe unexpected antiquily of the terminal phases of theMiddle Slone Age, indicated for the firsl lime by theBushman Rock evidence, has since been confirmed by

PRE-MIDDLE STOME AGE

(L....... 7&8&9)

na1283

PRE IIIDDLE STOME AGE(Level 10)n-1I3

,: IUDDLE STONE,: AGE

:(L .....I. 5&8) n_4020,

,,J LATER 5TONE: AGE•¡(L...... 4' n_3387

,: LATER STOME

: AGE,: u, • .,.1 3) n.393

.,,I LATER STONE: AGE:O-e....I. U2) n_ 403

3 • 5 inCh8S7.5 10 12.6 cm

OF BONE FLAKES

22.5 •

LENGTH

o

SO=%

O

so:%

O

50:'%

O

50:0/0

OOO

chewed and swallowed, and the unchewable pieces weredíscarded into the ash, where many became charred. ThePomongwe sample therefore appears to consist of thoseparts of the dassie skeletons that the people rejected asinedible. The nature of these parts is further discussed inchapter 7, but it is clear from the table that those partsmost commonly rejected were rnaxillae, mandibles, anddistal humen. These show consistent patterns of damage(fig. 31). The cranium lypically has been broken away althe back for the removal of the brain, and this damagecould have beco done either with a stone tool or with theteeth. The angles of the mandibles and ascending ramihave frequently been damaged during rernoval of thetengue, and the proximal end of the humeros has simplybeen chewed away. The distal humeros, perhaps oftenarticulated with the almost equally resistant proximalradius and ulna, was then discarded. In addition to thesecommonly recurring parts, a scatter of other skeletalfragmenta, as indicated in table 15. is likely to result fromhuman feeding. The nature of these food remains islherefore very differenl from those that lypically resultwhen leopards and other large felids feed 00 dassies, as isdiscussed in chapter 4. Such consistent differences infood remains are valuable indicators as to collectors offood remains in caves.

Tbe Pomongwe bones suggest that, tbrcughout theperiod of cave occupation, the people were successfullyhunting large and small game and were also collectingmany slow-moving animals such as tortoises. Sorne of theanimals, like the occasional leopards represented, wereperhaps hunted for their skíns rather than their meat,

Food Remains of Primitive People in Southern African Caves 35

,

I¡

LATER

STONE AGE

Horse 2Elltinct

--Tortoi.. SJftBuahplg 1

j. ~.~ ~ ~h~~~ "~_.~

Waterbuek 3 Sable 2 WlldebHat 1 Kudu 1 Roan 1

.""," ~,.. ..",. ~W,'lhog 4 ~~ "o,. 2

~~Sprlnghare 1

STONE AOE

PRE - UIDDLE

MIDDLE

STONE AGE

Roek Plgeoft 1

Fig. 29. Animals represented by the Pomongwe human food remains. Numbers of individuals are índicated.

36 A Guide to the Interpretation oí Bone Accumulations in African Caves

many dated sequences elsewhere in southem África. Thismatter has already been touched upon in the discussion ofPomongwe Cave. After the preliminary work by Louw, anew excavation was cornmenced in July 1967 by J. F.Eloffand has been continued every year since then. Afterlwo seasons' work J. F, Eloff (1969) was able lo reportthat his new excavation was already slightly deeper thanthar of Louw, but that bedrock was not in sight. TheMiddle Stone Age material appeared to represent a south­eastem expression ofthe Pietersburg Industrial Cornplex,with close resemblances to artifacts from Bed 4 at theCave of Hearths.

Faunal remains from the new excavations have beenanalyzed by Mrs. 1. Plug, who reports (pers. comm.) thatthe hiatus between the end of the MSA occupation andthe beginning of the LSA occupalion, observed in Louw'ssection, is not as marked as had beeo expected. Publica­tion oí the results of Plug's study, and oí Eloff's excava­tion, is awaited with interest.

A sample of 4,819 bone pieces carne from Louw's initialexcavation, and this was kindly made available to me forstudy by Louw and Mason. Each piece of bone was indi­vidually marked as to its level before sortiog proceeded.

It was found that, of the 4,819 fragments, 1,775 con­sisted of recognizable skeletal parts, albeil broken, andtbere were also 2.723 bone ñakes and 341 indeterminatefragments. As at Pomongwe, fragmentation of the boneswas extreme, there being very few unbroken skeletalelements except tbose of tbe smallest anirnals. As withthe Pomongwe sample, an attempt was made to assessfragrnentation by measuring tbe lengths of each of tbe2,723 bone ñakes. Results by individual level are given intable 16, and these are combined inlo MSA, LSA, andBantu horizons in table 17, with the figures presentedgraphically in figure 34. Fragmenlation in the MSA andLSA horizons is very similar, with the majority of ftakelengths falting between 1 in and 3 in. In the uppermostBaotu layer fragmentation ís more extreme, but this couldhave resulted from a great deal of modero activity whentbe shelter was in use as a tobacco bam.

"'''ULO''IIEel.... I

ALL OTHEII .. NI.....L.

o H ~ ~ o a ~ o ~ ~

PIlIIE-M.S.A. ".S.A. l.S....

PEItCENTAGE CONTAI8UllOHTO IIEAT etar

Fig. 30. Histognuns showina; the: percentage contribution made by variousanimal groups lo the diet of Stone Age people in Pomongwe Cave.

Table 18 lists the animal taxa identified frorn the bone,tooth, and shell fragments, with 30 indication of occur­rence per level. The minimum numbers of individualanimal s occurring in the Middle and Later Stone Age hori­zons are given in tables 19 and 20 respectively, whilethese animals are depicted in figure 35. In contrast to thePomongwe Cave situation, very few dassíes were beingeaten at Bushman Rock, although tortoises were gatheredin fair numbers. As indicated in tables 19and 20 and figure36. the main contributions to the meat diet of the peoplecarne from medium-sized antelopes (class III) and 000­

bovid mammals, particularly zebras and warthogs. Re­mains of smaller antelopes are surprisingly absent fromthe Middle Stone Age sample, although sorne were huntedby the Later Stone Age people.

The Willon Rock Sheller

The site is known as the Wilton Large Rock Shelter todistinguish it from the Wiltoo Cave. a deeper sheJter oc­curring on the same farm, wilton. 3)019'5, 26°5'E. in tbeAlicedale area of tbe eastem Cape. The two sites aresignificant in tbat it was frorn tbem that the Wilton Cul­ture, one of the expressions of tbe Later Stone Age, wasoriginally described by John Hewitt (1921), at that limedirector of the Albany Museum in Grahamstown. His de­scription followed the discovery of pygmy crescents inthe large rack shelter by C. Windsor wilmot, postmasterofQumbu.

During 1921 Hewitt, assisted by the Reverend Mr.Stapleton, the Reverend Mr. Kilroe, and W. W. Wilmot,owner of the farm Wilton, spent five days on an excava­tion in the rock shelter. They found that the deposít con­sisted largely of ash, devoid of stratífícation, and reacheda depth of 4 ft (1.2 m). Four burials were uncovered, ineach case the skeleton having been covered by ftat stones,painted on their undersides with red ocber. The burialsproved lo be offtal-faced "shortbeaded bushmen," ratherdifferent in appearance from those found in the nearby"Wilton Cave." whicb were of prognathous peopIe andwere thought to be more recento

In view of the significance of tbe remains recovered byHewitt, a subsequent excavation at the large rock shelterwas undertaken by Dr. and Mrs. H. J, Deacon during1966 and 1967 (J_ Deacon 1972). Tbe posilions of theirexcavations relative to the plan oftbe shelter are shown infigure 37. and the appearance of the shelter al the time ofmy visit in April 1969 is shown in figure 38. The rocksheller has formed in a steep-sided valley composed ofWitteberg quartzite cut by a tributary of the New Year'sRiver. Tbe valley itself is heavily wooded, but areas ofgrassland occur aboye It, and there are areas of bushveldbelow. Inhabitants of the shelter lherefore had access loseveral habitals in which to gather food and hunt.

The Deacons' excavation provided a litbic sample ofabout 34,000 pieces, which included 1,353 tools, a fewpotsherds and fragments of ostrich eggshell, and 43,629bone pieces and invertebrate fragments. The profile wasdivided into four layers, oumbered from the surfacedownward as follows (J. Deacon 1972, p. 13):

Layer 1: Layer 1 varied in thickness from 25 mm inthe south to 130 mm in the west of the grid. Leaves,twigs and other organic debris accumulated largelyfrom natural sources were variably preserved. The onIyplant-food remains were sorne corm scales of Watsonia

Food Remains of Primitive People in Southern African Caves 37

sp. One or possibly two oval areas more twiggy incompositíon than the leaf fall were noted as possiblyrepresenting decayed bedding heaps, but there is noother evidence to support this suggestion.

Layer 2: Below a transitional interface (2A), Layer 2is developed as a white compact horizon (2B). Thiswhite ash covers the whole of the excavated area andvaries in thickness from 150 mm to 300 mm, becomingthicker towards the dripline. Although the deposit hasthe superficial appearance of ash, grading analysisshows it to be a medium-grade sand, much the same asLayer 3.

Layer 3: Layer 3 is a somewhat oxidised, red-brown

deposit 50 cm to 70 cm thíck. Oxidisation has not beenuniform and towards the back wall the colour grades togrey-black. Subdivision of Layer 3 was partly on thebasis of thin, lenticular, interdigitating hearths andpartly on arbitrary spits. None of the subdivisions ismore than 8 cm thick. Layer 3D marks a general but nota defined change between the deposit aboYe (3C-3A)and below (3E-3I), and inc1udes three major white ashy¡enses (Hearths 1, 2, and 3) which interleave, makingvertical separation difficult. For purposes of analysis,the lithic material from the hearths was groupedtogether with the material from the rest of the layer, butthe faunal samples from the hearths were kept separate.

\)

Fig. 31. Dassie remains fmm Pomongwe Cave, apparently represeming human foo<! refuse. The maxillae, mandibles,and distal humen appear lo represen! unchewable remnants of the dassie ske1etons.

38 A Guide to the Ioterpretation of Bone Accumulations in African Caves

Fig. 32. The Bushman Rock Sheller in a dolomite ridge near Ohrigstad, easlern TrnnsvaaJ.

Layer 4: The top of Layer 4 is marked by an abruptincrease in roof rock spalls. There was an observeddecrease in the density of spalls fram the back towardsthe front margin of the rack shelter. The gradinganalysis supports this observation as there is lesscoarse material in the sample from the west wall in 4A

N

1

10 20,

Fig, 33. Plan (above) and section of the Bushman Rock Shelter.

than in the south wall (ie the back of the shelter). Layer "4 varies in thíckness from 35 cm to 50 cm and issubdivided into 4A aboye and 4B below a dark markerhorizon. There was very titde cultural material in 4B.The coarser composition of the Layer 4 deposit and theincrease io roof spaIls probably refIects the naturalbuild-up on the shelter fIoor during initial occupatioo ofthe shelter.

Three radiocarbon dates have beeo obtaioed as follows:Wood charcoal fram Layer 2B: 2,270 ± 100 years B.P.

Wood charcoal from Layer 3 F: 4,860 ± 115 years B.P.

Sample from the vertebral column of a burial, whichprabably relates to the base of Layer 3: 8,260 ± 720 yearsB.P.

The whole deposit appears to have accumulated duringLater Stone Age times.

The faunal remains fram the 1966-67 excavations werekindly made avaílable to me for analysis by Dr. aod Mrs.Deacon. Bone fragmeots from each stratigraphic level orsubleve! were studied as discrete units. Initial sortiog ofthe 43,629 píeces separated the unrecognizable flakes and'fragmeots from píeces that showed at least sorne diagnos­tic features. The unrecognizable flakes and fragmentswere placed in size groups aod counted. Recognizablepieces were sorted ioto skeletal parts and, where possi­ble, specific ideotifications were made.

The most characteristic feature of the Wilton bone ac­cumulatioo is its extreme fragmentation, which makesidentificaton of skeletal parts and animal taxa difficult. Ofthe 43,629 pieces, 33,891, or 77.6%, eonsisted of shortbone ftakes or unrecognizable fragments less than 1io (2.5

Food Remains of Primitive People in Southern African Caves 39

lOO

Fig. 34. Lengths oCbone ftakes from various levels in the Bushman RockShelter deposito

that they contain Iittle marrow; nor is much meat to behad from bovid feet.

Nonbovid mammalian remains were not nurnerous, ex­cept for microfaunal fragments such as bones and teeth ofsmalJ rodents and shrews. These have almost certainlycome frcm owl pellets, and evídence of bam owts in twolevels suggests that the shelter was sometimes used as aroosting place, as one would expect.

Among the reptile rernains, tortoise fragments areabundant in all layers. The tortoises were probably col­lected and ealen by the Willon people and appear to havebeen a favored food throughout the occupation periodoThe heads and necks of the tortoises have largely dis­appeared, suggeating that these parts may have been con­sistently chewed and swallowed. Two forms, Chersineangulata and Geochelone pardalis, appear to have beeninvolved.

Snakes are well represented by isolated vertebrae inmany layers, Sorne of these may have come from owlpellets, but they are usuaUy abundant in relation lo therodent remains, suggesting that snakes may have featuredregularly in the diet of the Later Stone Age people or thatthey were used for other purposes, as discussed below.

There is strong evidence that a freshwater streamexisted clase to the shelter throughout the Wilton occu­pation periodo Widespread crab and molIusk fragmentsoccur, as well as fish vertebrae at two levels.

Wherever possible, 1 have tried to identify the speciesof animals involved and to estimate the minimum numberof individuals that must have contributed to the bone ac­cumulation; results are given in table 25 and presentedpíctorially in figure 40. Table 26 provides an invenlory ofskeletal parts upon which identifications were based.

Blue duiker has been positively identified in five layers00 the basis of isolated teeth, hom-cores, and a metatar­sal fragmento The species oceurs in the area today and hasclearly done so throughout the Wilton time span. There isan indication of gray duiker in the uppermost layer only,!he diagnosis being based on a single lower molar.

Gray rhebuck has been identified in two layers on thebasis of a few teeth only. Although Ihese conform welI lothe teeth of Pelea in the comparative collection, a firmidentification requires more complete material. The sameis lroe of the bushbuck, idenññed from Layer 3B on thebasis of an isolaled tooth, Large bovids were poorly rep­resented at the site, A domestic cow is tentativelyidentified from the uppermost layer on the basis of asingle terminal phalanx, and a large alcelaphine, almostcertainly a wildebeest, is present in Layer 4.

Remains of common dassies were found in nine layersand carne from a mínimum of 12 individuals. Tbese arerepresented by the same skeletal parts as were the dassiesin the Pomongwe Cave deposito Almost as numerous areremains of scrub hares, likewise represented by a scatterof resistant skeletal parts. Three small camivore individ­uals are represented by isolated teeth only, one of whichwas positively identified as coming from a striped skunk.These animals were perhaps not hunted for food, thoughtheir skins may well have been used. Baboons have beenidentified in two layers on the basis of wom incísors. It isnot known whether these primates were eaten. The sarnecan be said of the single puff adder, idenlified from Layer3D, on the basis of a maxilIary bone complete with fangoThe possibility exisls that the Wilton people used puffadder venom as an arrow poison; during historie timesBushmen in the eastern Cape are reported to have usedthis venom 00 their arrows, perhaps mixed with plant

6 inch

15 cm4 S10 12.5

FLAKES

3

7,5

eONE

2

2.5 5

lENGTH OF

olOO

cm) long. The length distributions of these, by strati­graphic unit and by layer, are given in tables 22 and 23. Inview of the very short lengths ofmany of the bone flakes, itwas not practicable to separate these pieces from the mis­cellaneous unreeognizable fragments , as has been done inanalyses of the other three site assemblages. The frag­mentation at Wilton is more extreme than at the othersites and may reflect more intensive trampling of the caveñoor, supertmposed upon the usual breakage for thecxtraction of marrow. The degree of fragmentation is re­markably consistent in eaeh of the four layers, as isshown in figure 39.

1 found that, 6,702 píeces in tbe collectíon were com­plete enough lo allow identification of the skeletal par!involved. After sorting into parts, these were allocated tothe following groups, according to the kind ofanimal fromwhich they came: bovid, mammalian other than bovid,bird, reptile, fish, and invertebrate remains. Results ofthis analysis are given in table 24, bUI it must be stressedthat very few of tbe pieces listed consisted of completebones; rather, most were fragments with enough diagnos­tic features to allow recognition.

It will be seen that 1,454 pieces came from bovid an­telopes, and, as would be expected, the most resistantparts of the skeletons were best represented. Tooth frag­ments were numerous, and pieces of the lower legs suchas carpals, tarsals, metapodials, and phalanges were wellrepresented. The reason such parts survived is doubtless

'YoM.S.A. 50

%L.S.A. 50

.%BANTU 50

40 A Guide to the Interpretation of Bone Accumulations in African Caves

poison such as Euphorbia juice (Shaw. WooUey, and Rae1963). Part of a marine limpet (Patella sp.) was found inLayer 3A and is certainly indicative of human transportoThe shell may have had sorne cultural sigmñcance.

As with rernains from Pomongwe Cave and BushmanRock Shelter, 1have attempted to estímate the percentagecontributions to the diet of the people made by each ani­mal taxon, Results are given in table 27. and the figures

for broader animal groupings are listed in table 28 anddepicted graphically in figure 41. It is immediately obvi­ous that, in terms of meat yield, class II antelope havebeen the most important resource, followed by larger andsmaller bovids, other rnammals, and finally reptiles,birds, fish, and invertebrates. Although crabs were eatenin large numbers, they constituted a delicacy rather than astaple food.

OSTRICH EGG FRAGMENTS•~

SMALL CARIVORE 1

TSESSEBE 1

~

~PYTHON

SABLE 3

DASSJE 5

~m." n. ~..REEDBUCK e

." ~L. S. A.

WILDEBEEST 5

~KUDU 3

~ZEBRA 8 WAATHOG 6

.- ~TOATOISE 22 VARANUS 4

"MUSSEL 1 LAND SNAIL FRAGMENTS

,4'lLANO SNAIL fRAGMENTS

EXT1NCT HOASE 1

'~r\\ ~I ;'ll ~

\

•eSTRICH EGG

FRAGMENTS

PYTHON 1TORToaSE 10

~ "HARTEBEEST 1 WI LDEBEEST 1

KUOU SABLE 1

~ ~ ~M. S.A. OASSIE 1

WARTHOG 2ZEBRA 3

Fig. 35. AnimaJs represeated by the Busbman Rock Shelter food remains. Minimum numbers of individuals areindicated.

Food Remains of Primitive People in Southem ACrican Caves 41

PERCII!:NTAGE CONTR,8UTION

TO IIEAT DIET

F~. 36. Histograms showing the percentqe contributioa made by variousaRlmal groups te the diet of Stone Ase people in the Bushman Rock.Shetter.

upper part of the deposit: from 28-36 cm, 2,980 ::!: 120y~ars R,P., and, from 9-1 l cm, 2,770 ± 120 years D.P.

Since aH the bones are derived from the upper par! of lhedeposit, the assemblage was certainly built up within thelast few thousand years.

The bone sample was found to consist of 3,098 frag­ments, among which were 1,588 bone ñakes whoselengths are I.isted ~n table 29 and whose distribution is /i.qOsbown graphically In figure 44. It is apparent that the greatbulk of the ftakes have lengths less than 2 in (5 cm).The frágmentation is likely lo be due partly lo deliberalebr~akage for extraction of maITOW and partly to the tram-pling ofpeople's feet on the rather hard and grítty graniticsand surface. Fragmentation of the other skeletal parts isalso extreme, so much so that many of the identificationsare necessarily tentative. A Iist of animal taxa and theparts by which they have been ídentified is given in lable30, and the animals involved are shown pictorially infigure 45.

When we consider tbe numbers of individual animalsrepresented in the bone accumulation, we find that thesample is dominated by small antelopes, hyraxes, bares,tortoises, monitor lizards, and birds. The inhabitants ofthe shelter c.learly bunted small and medium antelopes aswell as dassies and bares. Many tortoises were gatheredand monitor lizard~ were favored food, as were rack pi~geons and larger birds such as francolins and geese.

Tooth fragments of a single black rhino were found toaccur in square B5 at a deptb of 15 cm. These remainshighlight an effect inherent in any attempt to calculatethe contributions made by individual animals in the dietof primitive people, If the meat yield of the rhino is in­cluded in the calculations, its contribution to the diet ofth~ people: as indiealed by the available bone sample,will outweigh that of the numerous smaller animals thatpresumably constituted a more regular and predictabJecomponenl of the meal supply. Apart from the tooth frag­ments, no other parts oCthe rhino's skeleton were foundamon~ the excavated remains, and it is highly Iikely thatthe rhmo was largely eaten where it lay rather than beingtransported to the shelter. The dietary contributions ofthe various animals and animal groups, expressed as per­centages of diet indicated by the remains, with and with­out the single rhino included, are lisled in tables 31 and 32and shown graphieally in figure 46. When the rhino isincluded, the contribution of its group of "other mam­mals" outweighs that ofthe various antelopes, whereas ifthe ~no is excluded from the calculation, most of themeat lS found to have come from steenbok, springbok,and gemsbok.

Regrettably, little information can be derived from theFackeltráger sample about butchery practices or aboutwhieh parts of prey skelelons were brought back lo thecave. One gains the impression, however, that parts ofthe entire skeletons of steenbok and springbok are pres­enl among the fragmented remains, ineluding lower legelements, lt is likely that these smaller bueks were earriedback whole lo the shelter, Remains oflarge antelopes aretoo sparse to be infonnative.

A minimum of 18dassies, one ofwhich was a juvenile,were represenled by 68 bone pieees. The skeletal partsinvolved bere are very similar lo those found atPomongwe, there being 9 maxillary pieces, 19mandibularpieces, 16distal humeri, and a scatter of other parts, con­srstent with the pattem of unchewable remains built upfrom the Pomongwe sample.

20 40 ea aol.S.A.

o 20 40 &OII.S.A.

ALL OTHER ANI ....LS

~HE.~:A""ALS

fflANT~':.':. ,~H---+-TEL¡'''-ITA elASS 11

Faekellriiger Sheller

The site is on the farro Omandumba West 137, in theOmaruru District of South-west Afriea, 21"34'S, 15"32'E.The country rock in this northwestem part of the ErongoMountains is Precambrian granite that forms prominentbare domes and pites of boulders (fig. 42) in an arid envi­ronment of sparse thorn scrub.

As is shown in the plan and section (fig. 43), the shelterconsists of an extrernely large granite boulder resting onfive smaller ones. The undersurface of the roof block isdecorated with many paintings, and the shelter was one ofsevera! selected by W. E. Wendl for exeavalion as part ofhis research program, conducted between 1968and 1971,00 archaeological deposits associated wíth rock art(Wendl 1972). The bone sample from this excavatíon hasvery kindly been made available lo me for study by Dr.Wendl.

The excavated area extended over 22 m-, as indicatedon the plan, and reaehed a maximum depth of 165 cm. Althe base of the excavation, in some of the innermostsquares, a layer of coarse granite grit was found, contain­ing Middle Stone Age artifaels but no lraee of bone.Above this was a deposil of brownish gray, gritty sedi­ment witbout clear stratification except for sorne ashlenses and pockets of ehareoa!. This upper layer, whiehea~ be regarded as Later Stone Age, yielded 26,707 sloneartifacts as well as 82 grinding stones. The artífacts in­eluded 663formal tools, a number ofbone tools, ineludingpoints, awls, a scoop, and various disks, pendants, andtubular bone beads. In addition, there were several pen­dants of mica schist, sorne ostrich-eggshell beads andpendants, a complete ostrich eggshell worked into a con­tainer, a fragmenl ofengraved ostrich eggshel1,and a pos­sible rattle made from a cocoon.

Two earbon dates are available from ehareoa! from the

Food Remains of Primitive People in Southern African Caves 43

Pig. 39. Lengths ofbone fragments from various tevels in the Wilton RockShelter deposito

Very little in the way of recognizable remains has SUf­

vived to attest the presence of at least 10 hares and lhesame number of monitor lizards. Apart from teeth andjaws , the skeletons of the se animal s appear to have beenhighly vulnerable to human feeding action. As at the othersites, tortoises are represented by abundant fragments ofcarapace and plastron and a few limb bones.

InformatioD 00 Skeletal Disproportions from SoutherDCape Caves

Seven caves and shelters have been excavated in theKlasies River Mouth complex, on the southern Capecoast about 130 km west of Port Elizabeth. Excavationswere organized by R. Singer and undertaken by J. J.Wymer between December 1966 and July 1968. Thefaunal remains have been studied in detail by Klein, whowas able to discuss three aspects of'the body-part data thebone samples provided (Klein 1976a. pp. 85-97).

Effects 01 Species Síze

The Klasies bovids were placed in fíve size groups, vary­ing from "small" to "very large," and it was apparentthat the pattern of body-part representation differed ac­cording lo the size class being considered. To clarify thissituation, Klein drew up a table in which the minimumnumbers of animals in Ihe '"small bovid" category de­manded by each skeletal part were arranged in descend­ing order of abundance. Thus, al the top of!he co1umn 54small bovids were demanded by scapulae, foUowed by 46on the basis of mandibles, right down lo nil on the basis ofcarpals and larsals. Following Ihe body-part arder de­termined for the small bovids, the same was done foranimals in !he four o!her size classes, and histogramswere plotted for eaeh. The form of these histograms wasfound lo vary a great deal; for instance, it was only in thesmaller animals that scapulae were the most abundantparts-in the larger ones mandibles tock over the positional the lop. Various statistical tesis that Klein applied con­firmed that the pattern of body-part representation inlarge and very large bovids was very similar, but !hat itwas strikingly dífferent from that in the small and mediumbovids. In general, as the size of the bovid decreased, Iheratio of cranial to postcranial parts increased, whereas theratio of limb bones lo foot bones decreased. What dothese difIerences mean? Klein suggested that they arerelated lo the portability of!he prev-s-to what Perkins and

Although the bovid remains frorn Klasies River Mouthcarne from animal s differing widely in size and in age atdeath, Klein found that the conclusions reached on thebasis of Kuiseb River goats, described in chaper 2, con­cerning the relationship between bone density and survi­val held good for those bones as well. He wrote: .. As inthe case of the Kuiseb goats, those ends of the bovid longbones at Klasies which are denser and fuse earlier tend tobe more common than their less dense opposite ends thatfuse Iater" (p. 86).

tive and provide detailed insights I was unable to obtain.Asevere drawback of the collections 1 studied was thatthe bone pieces had been very heavily fragmented, and asa result 1 could deduce little about the butchery practicesof the Stone Age people or about the parts of animalsbrought back to the caves. Fragmentation seems to be anattribute of bones from Stone Age cave or shelter sitesthat have been intensively occupied. Fortunately, re­mains from the Klasies River Cave 1, analyzed by RichardKlein (1975b. 1976a), proved more complete than usualand have provided sorne extremely valuable data.

Effects 01 Dtfferences in Bone Density and EpiphysealFusion Times

3 4 inches

7.5 10 cms

FRAGMENTS

2

5

BONE

o 1

O 2.5

LENGTH OF

50

1

100

100

o

Discussion

It would be very naive to imagine that the results of thefour analyses presented here could do more than providea glirnpse oí the remains of the animals eaten at thosespecífic sítes. CoUections from other areas wíll give a verydifIerent picture and will serve to emphasize that Horno isa rernarkably adaptable animal. making use of whateverfood resources happen to be available. Fortunately, theresults of sorne excellent studies 00 food remains fromother caves in southem Africa have been published withinthe last ten years, and these place my results in perspec-

100

LAVER 4

~n= 8252

LAVER 3

%n-24768

LAVER 1 50

%n= 223

LAVER 2 50

~n= 2648

44 A Guide to the Interpretation of Bone Accumulations in Afríean Caves

Daly (1968) have ealled the "schlepp effect." In this eon­neetion, Klein wrote (1976a. pp. 87-88):

Basically they postulated that hunters were likely tobring horne smaller animals intact, but they wouldprobably bring back only selected parts of largeranimals. This is beeause larger animals would bebutchered at the place of the kili and the less usefulparts would be left there. In documenting the operationof the "schlepp effect" at the early Holocene("Neolithic") hunters' site of Suberde in Turkey,Perkins and Daly showed specifically that larger bovidstended to be represented disproportionaIly well by theirfoot-bones versus leg-bones, just as al Klasies. Theypostulated that the Suberde people discarded manylarger bovid limb bones at the kill sites, but broughtback the feet either as handles in the skins (used ascarrying containers for the meat?) or because the feetwere particularly valued, perhaps as sources of sinewsfor sewing.

It seems that heads of larger bovids were selectivelybrought back and that mandibles have survíved betterthan maxillae as a result of their structural strength.

Effects o/Site Type: "Occupation Sites" versus"Kill-Butchery Sites"

Klein has pointed out that, if the "schlepp effect" hadbeen in operation, ít is reasonable to expect that the

skeletal parts of antelope prey found at "base campa" or"occupation sitcs" like Klasies would differ from thoseencountered al "kili or butchery sites." He has comparedthe body-part information from Klasies with that from anopen-air kili and butchery site he investigated at Duine­fontein 2, near Melkbosstrand in the southwestern Cape{Klein 1976b). Those parts, such as vertebrae, that wereunderrepresented at the Klasies "base camp" wereabundant at the Duinefontein 2 "butchery site." Kleinhas pointed out that a rather similar effect may be ob­served for bison remains at American Indian village andhunting sites, He compared body-part information fromfour village sites deseribed by T. E. White (l953b, 1954a,b) with that for thc Casper kili site investigated by Frison(t974) and found that, aIthough many differences existed,it was reasonable to assume that the "schlepp effect" hadbeen responsible for the low axiaUappendicular ratio inthe Klasies remains.

Observations on contemporary hunter-gatherers suchas those of Yelten (1977) are very valuable in showingwhat kinds of activities led to the fossíl traces as we knowthem. Working with the !Kung Bushmen in northwestemBotswana. Yellen was able to observe butcheríng of vari­ous large antelopes that had been killed with poisonedarrows. He described twenty-eight typical steps in theprocess and found that the cannon bones were charac­teristically cooked and eaten at the site of the kili, as weresorne of the ribs. Such parts would then be missing from"base camp" food debris.

Fig.4O. Animals represented by (he bones in various levels ofthe Wilton Rack Shelter deposit. Minimum Rumben ofindividuals are indicated.

Food Remains of Primitive People in Southern African Caves 45

l S •lEk;__l------2 _

A Classification of Bone-Bearing ArchaeoJogical Sites

In his paper on the die! of early mano Isaac (1971) pro­posed a classification of bone-bearing archaeoJogical sitesthat is particularly relevant lo this discussion. It is basedon the relative abundance of stone artifacts and bonerefuse as indicated in figure 47, and is properly applicableonly to undisturbed siles where preservation has beengood. The diagonal defined by increasing artifact and bonedensities also represents increasing intensity of occu­pation.

Transitory camps, characterized by a low density ofstone artifacts and of bone refuse, where the traces haveresulted from brief occupation-perhaps overnight---of amoving human bando Examples suggested by Isaac areKoobi Fora site FxJj 1 (Isaac, Leakey, and Behrensmeyer1971), parts of Olorgesailie (e.g., Isaac 1968), and Peninj(Isaac 1967a).

\."'"-._--._--- -'6020 40

L.S.A.

PERCENTAGE CONTRIBUTION

TO

MEAT DIET

o

ANTELOPE CLASS

OTHER MAMMALS

ALL OTHER ANIMALS

~NTELOPE\n J CLASS 1I

Pig. 41. Contribulions made by various groups of animals lo the diet ofSlone Age inhabitanls of!he Wilton Rock Shelter.

Fig. 43. Plan and seclion of lhe Fackeltrager Sheller. Granile boulders andbedrock are shown in black.

oo

I

2.5

2

537.5

•10

5 inches

12.5 <:m

Fig.42. A view of lhe Fackellriiger Shelter under a large granile boulder inthe Erongo Mounrains. PhOIO by W. E. Wendl, from Wendt 1972.

Fig. 44. Lengths of bone flakes from lhe Fackellrliger Shelter. SOulh-WeSIAfrica.

46 A Guíde to the Interpretation of Bone Accumulations in African Caves

Ki/l or butchery sites. characterized by a high density ofbone but small numbers of artifacts , these sites may cen­ter on the remains of a single animal or contain remains ofmany that were attracted to sorne resouree, sueh as awater hole, and were killed there.

A well-doeumented instance of a single eJephantskeleton, surrounded by Aeheulean tools that had clearlybeen made on the spot for butchery, has been deseribedfrom Mwanganda's Villge in Malawi (Clark and Haynes1970). A rather similar situation, but later in time andinvolving two elephants, has been reported from the zoogrounds in Windhoek, South-West Afríca (Sydow 1961,1963; MacCa1man 1967). Other elephant butchery placeshave been recorded from Olduvai Gorge in the top of Bed1, level 6, of site FLK NI, where the skeleton of anElephas recki was associated with 123 artifacts. Nearby,in the base of Bed II at FLK, the skeleton of a De­lnotherium was preserved with 39 artifacts (M. D. Leakey1971). Farther afield, sorne fine exarnples of elephant

butchery places have been described from the middleAcheulean sites of Torralba del Moral and Arnbrcna inSpaio (HoweIl1966; Freemao and Butzer 1966), where, inalluvial beds from the Rio Ambrona, remains oí elephantsand other large mammals were preserved, as they wereleft, by the Stone Age hunters,

Butchery sites oí híppos are also known, for instance,at Koobi Fora FxJj3 and Olorgesailie (see aboye) and atIsimila in southern Tanzania (Howell 1961; Howell, Cole.and Kleindienst 1962). Mention has already been made ofthe Duinefontein 2 síte near Cape Town where many ani­mals, inc1uding several buffalo, appear to have been sys­tematically butchered by Middle Stone Age peoples,

FinaUy, it is remarkable how much detailed information

ALL OTHEIt ANltUlS

o :lO.DeoO 2Q"Oeo

W'TH ItHt"O l.S.A. WlTHOIIT IIHINO

PERCENTAGE CONTR'Buno,", ro

MEAT DIET

Pig. 46. Contributions made by VariODS groups of animals to the diet ofLater Stone Age inhabitants of tbe Fackeltriger Shelter. Tbe histogramson the left show contributions wbeu a single rhino is included; tbose on theright reñect the sttuetíon when tbe rhino is excluded.

1

•high

of bone refuse

moderateDensity

low

TRANSITORY Klll OR

CA MPSBUTCHERY

SITES

- CAMp- OR

~WORKSHOPSITES OCCUPAT ION SITES

oC

""oC

..­..~.....O

E

me=w

weme

j.IIAIl~ 2

~ELEPHAIIT _W :1

~~VIlR"IIUS 10 LIZAItD 2

RH'"D •

~HIlIIE 10

TORTOlSE H

!>ASS1E 1.

Fig. 45. Animals represented by the bones from the Fackeltriger Shelter.Minimum numbers of individuals are indicated.

Fig. 47. A scbeme for the claesíñcatíon oí bone-bearing archaeologícalsítes. proposed by Isaac (1971).

Food Remaíns of Primitive People in Southem ACrican Caves 47

about the behavior oCStone Age hunters may be derivedfrorn meticulous excavation of butchery sites , as has bcenpointed out by Desmond Clark (1972). For instance, workat the Olsen-Chubbuck site in Colorado has allowed thereconstruction of a bison hunt that took place about 8,500years ago (Wheat 1967). lt seems that about two hundredbison were stampeded and driven into a valley, wheretheir remains have beeo found. Investigation of these re­mains has made it possible to reconstruct, with reason­able certainty, the month in which the hunt took place,the wind direction 00 that particular day, and the direc­tion of the hunters' dríve, along with details of the butch­ery process and of whích cuts were eaten on the spot.

Quarry ar workshop sites, charactertzed by an abun­dance of stone artífacts but a low density of bone refuse.Isaac mentions the "chert workshop site" from lowerBed II at Olduvai as a possible Lower Pleistocene exam­pIe (Stiles, Hay, and O'Neil 1974), as well as several morerecent ones, including parts of the Cave of Hearths ac­curnulation al Makapansgal (Mason 19620).

Camp or occunation sites, typified by high densities ofstone artifacts and bone. The cave sites discussed earlieraU qualify for this definition, as do cpen-air localities Iikethe "Zinj" floor al Olduvai FLK 1, the DK floor, somelevels al FLK NI and the DE/89 horizon B alOlorgesailie, where the remains of at least 40 adultand 13 juvenile baboons, of the large extínct speciesTheropithecus oswoldí, together with 15 kg oí bonepieces, were found among very numerous Acheulean ar­tifacts (Isaac 1968, 1969).

As mentioned earlier, evolutionary changes in theHorno Iineage during the course oí the Pleistocene Jed tomarked behavioral changes that must inevitably havebeen reflected in the traces left by the people. Sorne in­formation 00 progressive changes in the nature oí suchtraces from Bed 1 al Olduvai through upper Bed n hasbeen collected by M. D. Leakey (1971) and further dis­cussed by Isaac (1976). It is especially significanl that,with the passage of time. a rise in the ratio of artifacts tobone refuse is discernible. while the fragmentation thebones have suffered has Iikewise increased. Larger ani­mals are slightly better represented in the later levels, andthe early preponderance oCbovids is replaced by a widerprey spectrum. including equids and hippos. In upper Bedn, al Localilies SHK and BK, lhe earliesl evidence ofherds oC antelopes having beeo successfuUy hunted isCound. Antidorcas recki remains are found at the forroerand PeJorovis at the latter.

Any interpretation oí food remains associated with thesouthern Afriean australopithecines should take into ac­counl lhe fael lhal lhese early hominids probably did nolbreak up banes as intensively as did subsequent membersof the genus Horno.

The Galhered Componenl iD lhe Diel 01 Slone Age Pooplell

Tbe people we have beeo discussing were, presumably.hunter-gatherers. who derived part oC their sustenancefrom gathered Coods. Sorne oCthe bone reCuse at the sites1 have consídered. such as remains of tortoises. camefrom aoimals that were gathered rather than hunted, butthe main gathered component in the diet surely carne fromplant foods. except at certain coasta! sites where gatheredshellfish were extremely important. In fact, in his surveyof the diet of living hunter-gatherers, Richard Lee wrote(1968, pp. 42-43):

Although hunting is rarely the primary source of food.it does make a remarkably stable contríbution to thediet. Fishing appears to be dispensable in the tropics,and a number of northern peoples manage to do withoutgathered foods. but with a single exception, al!societies in all latitudes derive at least 20 per ceot oftheir diet from the hunting of mammals. Latitudeappears to make little difference in the amount ofhunting that people do. Except for the highest latitudes,where hunting contributes over half ofthe diet in manycases, hunted foods almost everywhere else constitute20 lo 45 per cent of the diet. In fael, the mean, themedian, and the mode for hunting all converge on afigure 0135 per cenl for hunler-gatherers al all latitudes.This percentage of meat corresponds closely to the 37per cent noted in the diet of the !Kung Bushmen of theDobe area. It is evident that the !Kung, far from beingan aberrant case, are entirely typical of the hunters ingeneral in the amount of meat they consume.

From his study of the !Kung Bushmen in oorthwesternBotswana, Lee concluded that these people "eat as muchvegetable food as they need, and as much meat as theycan." Gathered plant food constituted the stable dietarybase, and, in comparison with hunting, gathering provedmore productive. In the Dobe area, Lee found that oneman-hour of hunting typically produced about 100 ediblecalories, while gathering brought in 240. In fact, huntingproved to be a high-rísk, Iow-retum subsistence activity,while gathering represented a low-risk, high-return one.This may well have been true for African Stone Agehunter-gatherers as well.

In the Dobe area, Lee observed that the Bushmen madeuse of a wide range of plant foods, but that mongongonuts from the tree Richinodendron rautanenii were themost important dletary ítem, the average daily per capitaconsumption being 300 nuts, which yielded about 1,260calories and 56 g of proteln. Such a helping of nuts,weighing about 213 g, eontaíned the calorlc equivalent of1.1 kg of cooked rice or the protein equivalent 0097 g oflean beef. Although the habitat in northwestem Botswanawas harsh and arid, Lee established that a Bushman's lífe

.'~ ....

48 A Guide to the Interpretation of Bone Accumulations in African Caves

October, Tuesdav the 2nd, 1685.They are lean and slight of frame, due to the great

hunger and hardship which they endure. They eatnothing but the bulbs of flowers which they [we?]call uventjes, tortoises and a eertain Iarge kind ofcaterpillar, rogether with Iocusts which are foundhere in abundance.

The uyentjes referred to here appear to have been animportantitem in the diet of hunter-gatherers ín the Cape.Parkington (in Parkington and Poggenpoel 1971) hassuggested that the chief task of the women in Capehunter-gatherer bands was coUecting roots, conos, bulbs ,and tubers with a digging stick and kaross. Early Euro-

hotnotsbroodedible cetecear

September. Tuesday the 4th, l68j.They said they were Sonquas or Obiquas, and they

had come here lo look for an eland which tbey had shotwith poisoned arrowe the day before. Tbey carriedarrows, bow, and assegai. They have no cattle, livingon honey and tbe wild animals they shoot, Their skin isvery rough and scurvy owing to the frequent hungerthey endure and the laek of fat with whieh to smearthemselves. The honourable Commander made them apresent of a sheep and altbough tbey are not men of anybreedíng they were nevertheless polite enough to givehim in return three tíger-cat skins. They irnmediatelycut the sheep's tbroat, flayed it and then cut off hothforelegs, allowing nothing to go waste except fourglands in the legs, whieh they cut out and tbrew away.When we asked why they did so, they couId give noother reason than that they oever ate such things. Theyplaced the mea! under ashes for half an hour and feasted00 it until there was none left, gnawing the legs likebeasts. As we stayed here the whole day and took ouraltitude in the evening and discovered our latitude to be32°58' and our longitude 38"57'.

was not the precarious and arduous struggle for existencethat most people thought was the unavoidable lot ofhunter-gatherers. The Bushmen of Dobe devoted onlytwelve to ninteen hours a week to food-getting; the rest ofthe time was spent in leisure and relaxation. Despite tb¡s,they were well-nourished, cnnsuming an average of 2.140calones and 93 g oí protein per day-quantities greaterthan those known to be required by people oí small stat­ure participating in vigorous activity.

The mongongo nut tree does not oceur everywhere,and in other areas hunter-gatherers are known to dependon a variety of different food plants. After his study of theG/wi Bushmen in central Botswana, SiIberbauer (1972)listed thirty-five species oí plants regularly used as foodsources. These included four kinds oí melon, seven otherfruit specíes, two types of seeds, eight types of leaves,and fourteen root or storage-organ forms. Oí these ,twelve species represented food scurces oí major im­portance in the economy oí the people. A rather smallernurnber of plant foods is regularly used by the Hadzahunter-gatherers oí the Lake Eyasi area of Tanzania(Woodbum 19680). The bulk of tbeir vegetable foodcomes from only ten species oí plants the edible partbeing the root of four species, the berry of five, and thefmit and seeds oC the rest.

Over most of southem Afríca, hunter-gatherers havedisappeared from the scene. and thelr activities can nolonger be observed. But when the first Europeans startedto explore the interior of the subcontinent, they fre­quently encountered bands of people who praeticedneither agriculture nor animal husbandry. In the westemCape sueh people were known as the Sonqua or San,and a particular study of tbem has been made by J. E.Parkington. In addition lo rus archaeologícal studies,Parkington (Parkington and Poggenpoel 1971) has in­vestigated the accounts of early travelers who describedcontacts with hunter-gatherer bands. One source of suchinformation is Simon van de Stel's joumal ofhis expedí­tion to Narnaqualand, 1685--1l6, lranslated and edited byWaterhouse (1932). The following extracts from this jour­nal indicate the kind oC information available from suchhistorical sources (Waterhouse 1932, pp. 117-18, 128):

Food Remains of Primitive People in Southern Afrieao Caves 49

I,

1I

pean settlers used the Dame uíntjiestok, "onion stick."for the digging tool and uintjiesak, "onion sack." for thekaross. The records suggest that the corms were groundand roasted befare being caten, the common Dame forseveral species of Watsonia being Hotnotsbrood or Hot­tentot's bread.

The importance of the corms of Watsonia and similarplants in the Stone Age human diet has beco confirmed byplant remains preserved in various southem Cape caves.The first such evidence carne to light al the Melkhout­boom Cave, Alexandria District, in the course of an exca­vation by Hewitt (1931) during May 1930. The site wasreexcavated in 1967 by Deacon; one of his aims was torecover further plant remains for analvsis (H. J. Deacon1969, 1970). Hewitt's original list of plant species wasextended, and the main food plants were shown to beHypoxis sp., Watsonia sp., Freesia sp., Moraea sp., Bu/­bine alooídes, OxaJis sp., Cvperus usitatus, Schotia afra,and Harpephyllum caffrum. By the time H. J. Deacon(1972) wrote his revíew of Ihe Postpleistocene in SouthAfrica, information on food planta was available from fivemore rack shelters in the eastem and southern Cape.Farther afield, sites such as Border Cave (Beaumont1973), Kruger Cave (Mason, Friede, and Pienaar, 1974),and Mirabib Shelter (Sandelowsky 1974) have also beenyielding valuable insights ínto the plant foods of theirStone Age inhabitants.

In addition to plants, Stone Age inhabítants of thesouthem African coast had beco using the resources ofthe sea fcr a very long time. In facl the Middle Stone Agelayers in the Klasies River Mouth Cave, ranging in timefrom about 125,000 lo 55,000 years s.e., contain c1earevidence for the earliest systematic exploitation of marineresources known anywhere in the world. These layers arepacked with mollusk shells and a1so contain bones ofseals and penguins that the people had discarded. Klein(1977) has pointed out that at more recent coastal sitessuch as Nelson Bay (Klein 1972a,b) and Elands Bay(Parkíngton 1976), where the levels with marine food re­maíns are al) younger than ]2,000 years, remains of sealsand penguins, comparable to those from Klasies, are ac­companied by numerous bones of físh and flying birds,which are Tare at KIasies. It seems that fishing and thehunting of airbome birds were beyond the technolugical

capabilities of the Middle Stone Age people , but thatLater Stone Age people had acquired these skills.

The evidence of shell middens around the southern Af­rican coa st, and of shell-packed deposits in coastal caves.restiñes to the importance of gathered shellfish in theStone Age diet. The first analysis of the faunal content ofa soutbem African shell midden was carried out al theBonteberg Shelter on the westem sirle of the Cape Pen­insula (Maggs and Speed 1967), where excavations con­ducted between 1962 and 1964 revealed a midden deposit104 cm deep consisting of two layers dated at 2,OSO±95and 4,505±IOO years O.P. (Grindley, Speed, and Maggs.1970). In the report on lbe Bonleberg material. ElizabethSpeed (now Voigt) laid down some guidelines for tbeanalysis of molluscan shells in middens and subsequently(Voigt 1975) further defined criteria for the estirnation ofminimum numbers of individual mollusks in an as­semblage. Using such methods, it was possible to showthat in the shell midden at Bonteberg Shelter the lowerleve! contained 55 percent limpets (Patella sp.) as well aslarge numbers of welks (Burnupena sp.), while in theupper level limpets decreased and whelks and winkles(Oxystele sp.) increased.

The analysis of mollusks from the Klasies River MouthCaves (Voigt 1973), showed equally clear variations indietary preíerence over an enormous span of time. Therepresentation of limpets in the shellfish diet of thepeople varied from 71% to 17% at different phases of theaccurnulation, and winkles also fluctuated in popularity.By her layer-by-layer analysis, Voigt was also able lo

1=:_11II=::::1' em

Painted burial stone

Robberg Cave

03:¡¡stele

50 A Guide to the lnterpretation of Bone Accumulaüons in African Caves

show which parts of the intertidal zone were being ex­ploited at each cultural phase ,

The extent and thickness of sorne of the southern Afri­can coastal shell middens is remarkable, and it would beinteresting to know how many people were involved intheir accumulation. Most of the authentic "strandloper"bands have disappeared, but coastal tribesmen of theTranskei still exploit shellfish in a traditional manner(Bigalke 1973), and so direct observations can be made onthe cate of midden accumulatíon. For instance, Voigt(1975) has shown that a shell midden covering an area of24 m2 was accumulated by four persons over a period of40 years.

This discussion has made it c1ear that the gatheredcomponent has always been extremely important in thediet of Stone Age peoples. In a recent review entitled"Gathering and the Hominid Adaption," AdrienneZihlman and Nancy Tanner (n.d.) have gone so far as tosuggest that gathering rather than hunting exerted thegreater influence 00 the development of typically humanbehavior and thus on the course of human evoiution,

Sorne Evidence for Seasonal Occupation of Southern Afri­can Caves

It is c1early important for hunter-gatherer peoples to beknowledgeable aboul lhe availabilily of natural foodsthroughout the year, and it is by moving about on a reg­ular seasonal round that the bands will make the fullestuse oftheir resources. Hilary Deacon (1969)has recordedan oral tradition concerning the movements of the lastBushman band in the Long Kloof of the southeasternCape. This group, of about sixteen hunter-gatherers, wassaid to have occupied a series of three caves--one on thecoast for about two months in summer, another inBaviaanskloofduring the winter, and a third in a tributaryof the Kouga River for the rest of the year. In this waythey presumably made use of both marine and terrestrialresources to their best advantage,

Evidence, sorne of it ingenious and interesting, issteadily accumulating 00 the seasons during which Capecaves were occupied, and the evidence seems to be atvariance with the spoken tradition about the Long KloofBushmen. Por instance, plant foods preserved in the in­land Melkhoutboom Cave deposít and excavated by Hil­ary Deacon (1969) include inflorescences of Themedatríandra and Heítpterum mUleflorum, which could havebeen galhered only in spring or early summer, and in­dications provided hy shells in Willon middens al the Nel­son Bay Cave indícate that the mollusks were collecledonly in the colder months, so that this particular coastalcave must have been occupied in winter. The conelusionis based en the ratios of oxygen-16 lo oxygen-18 in the lastgrowth increments of the shells concemed (Shackleton1973).

Mention has already been made of the importance ofthe corms of Watson;a and other lridaceae in the diet ofStone Age people in lhe Cape. Parkington (1976) has pro­vided histograms for ñowering times of lridaceae in thesouthwestem Cape that show that August, September,and October are tbe months when most f10wers are pres­ent. It is known that the corms of plants like Watsoniaattain their maximum size after ftowering in about Octo­ber; thereafter the leaves wither and the plant becomesmuch more difficult to find. This strongly suggests that acave such as De Hangen in lhe Clanwilliam area (Park-

ington and Poggenpoel 1971), where abundant remains ofIridaceae corros are preserved, was occupied in earlysurnmer. The De Hangen deposit also contains a gooddeal of plant material that the prehistoric inhabitantsbrought in for bedding. This was made up of four maincomponents, Helichrysurn sp., Ehrhartu sp., Pennesetumsp. and a sedge. either Cyperus sp., or Mariscus sp. AHthese plant rernains had f1orescences, and all are known toñower between October and December.

Faunal rernains from De Hangen confirm the summeroccupation of the cave. The bones inelude pieces from atleast 313 tortoises that had presumably been gathered insummer. During the winter, the tortoises in this part ofthe country tend to hibemate. Parkington also recoveredremains of at least 64 dassies, many of wbich were im­mature, but which could be aged on the basis of tootheruption, since it is known that the young are bom therebetween Septernber and November. This study showedthat all the dassies al De Hangen could have been killedbetween November and February, but that very fewcould have died in the winter.

An equally delailed study has been carried out by JohnParkington (1976) al sites such as Elands Bay on the adja­cent west coast. Here the evaluation of seal bones dis­carded by tbe cave's inhabitants appears to bave greatpotential for establishing the season of the year when thehunting took place. Al Elands Bay Cave the sizes of sealmandibles, compared with those of known-age animals,strongly suggest occupation between July and October.

Inhabitants of these west coast caves and shelters atevery large numbers of marine mollusks, which are fre­quently toxic in summer months owing to contaminationby "red tide" dinoflagellates (Grindley and Nel 1970).The people presumably avoided being poisoned by eatingthe shellfish only during the winter.

The accumulating evidence certainly seems to suggestthat Stone Age hunter-gatberers living within reach of thesouthem Mrican coast rnade seasonal rounds wherebysummer months were spent inland and the winters werespent on the seashore.

In the interior, Stone Age bands presumably movedseasonally in response to migrations of the game herds. Amodel for Later Stone Age seasonal movements betweenthe Drakensberg escarpment and the Natal coastal areashas been suggested by Carter (1970), who pointed out thatduring spring and early summer the Highland Sourveldprovides attractive grazíng, In late summer the grazingvalue declines, and in winter tbese highland areas experi­ence very low temperatures, Game herds used to respondlo such seasonal changes by migrating up and down theescarpment edge. They were presumably followed by thehunter-gatherer bands who were responsible for the spec­tacular Drakensberg rock paintings. On this topic Carterwrote (1970, p. 57):

Acceptance of the model would have important re­percussions on the interpretation of the many rock­paintings that occur in tbe Drakensberg. The virtualrestriction of painting to the summer months could beinterpreted as evidence of periodic, possibly annual,ceremonies in whicb painting was 3n integral and im­portant parto Such an interpretation would be in accordwith findings in which the ritual aspects ofthe paintingsare emphasised. It is tempting to suggest that the sum­mer occupation of the Drakensberg was a time ofplentyfor lhe hunler-galherers of lhe Lale Slone Age. Wilh

Food Remains of Primitive People in Southem African Caves 51

adequate supplies of animal and vegetable food it wouldbe economically possible for band size to be larger andceremonial gatherings lo take place. During the warrn,dry , winter months in the Thom Veld, band size wouldof necessity be reduced, camp sites would have been inthe open and ceremonial gatherings less Iikely to occur.

Observations 00 living hunter-gatherers (e.g., Silber-bauer 1972; Lee 1972; Woodburn 1968b) emphasize theloose social structure of bands and their constantlychanging size. In view of this there could welJ have beentimes when large numbers of people carne together forceremonial purposes. In bis paper on group size duringthe Laler Stone Age, Maggs (1971) has drawn evideneefrom various sources, including rock paintings, and hasconcluded that bands of eight to twelve members or oftwenty to twenty-fivc members were most common, butthat rnuch larger aggregations were by no means unusual.Figures for human group sizes in Drakensberg rockpaintings have also been given by Pager (1971); thesevaried from two lo thirty-two.

The evidence discussed here comes very largely fromthe c!osing phases of lbe Stone Age, but il is likely thatseasonal movements charaeterized the way of life ofhunter-gatherer bands throughout the Pleistocene, Anattempt has recently been made to document seasonaloccupation of Olduvai Gorge living sites through com­parisons of food remains left 00 these floors with those ofcontemporary Bushmen bands. In thís attempt, Speth andDavis (1975) listed animals eaten by !Kung and G/wiBushmen in Botswana and correlated tbese with tbe sea­sons when they were hunted. They worked out percent­age representations of three groups of animals-Bovidae(the various antelopes), Carnivora (largely jackals andfoxes), and Chelonia (tortoises and lerrapins) and com­pared these with percentages derived from Mary Lea­key's data on bone food refuse from various ftoors andlevels in tbe Olduvai sequenee (M. D. Leakey 1971). Acomplication in the comparison is that the percentages oCBushman prey were based on numbers oí individual ani­mals, whereas those from Olduvai were calculated onnumbers of bones identified in each taxon. Serious dis·crepancies could resulto particularly for tortoises and ter­rapios, since when their carapaces are broken up verynumerous fragments resulto Nevertheless, Speth andDavis did find an interesting correspondence between theproportions of the three animal taxa in the Bushman preyand the Olduvai remains. They were able to establish that

Bushmen ate the tortoíses and terrapins almost exclu­sively during the summer rainy period, since they werchibernating in the dry winter months. Assuming thatclimatic seasonality was comparable between the üld­uvai and Botswana habitats , Speth and Davis conctudedthat all but four of the twenty-two Olduvai coJlections hadresuIted from dry-season occupation. The four levels thatsuggested wet-season occupancy were DK (aJI Ieveis},FLK NN level 3, FLK NN level r, and the MNK skullsite. Three of these are among the earliest occupations inthe gorge and may have eoineided with a period of slightlyreduced aridity.

If there is validity in these interesting deductions , itappears that Olduvai Gorge, or rather the lake that oc­cupied the position of the present gorge, was one of thestopping places on an early hominid seasonal round thattook in the adjacent plains and mountains.

It seems likely that the early hominids whose remainsare preserved in the Sterkfontein valley caves would alsohave made regular seasonal movements to exploit the re­sourees of the highveld grassland and adjacent bushveldareas. The Sterkfontein valley is close to the northemedge of the highveld plateau, and within 20 miles lo thenorth, other habitats are available that would have beenmore productive in winter. The open highveld was almostcertainly a rich and desirable habitat in spring and sum­mer but may well have been largely avoided in winter.Hunter-gatherers, unlike their sedentary successors,were locked into an ecologícal framework that they couldiofluence but little. Like other animals, they would havetried to make optimal use of a patchy environment byexploiting each part of it during its most prcductive sea­son (MacArthur and Pianka 1966).

Sorne Consistent Features in Primitive HumanFood Rernains

In the following brief discussion 1 will restrict my com­ments to the animal component of Stone Age human diet.represented by skeletal remains.

The Nature 01 (he Prey

Study of food remains from habitation sites in southernAfriea has made il abundantly clear that the people wereopportunists, feeding on whatever animal protein sourcewas available to thern, although from time to time andfrom place lo place their diet may have been restricted bycustcms and taboos, For instance, in his study of theHadza hunter-gatherers, Woodburn (19680) found lbalthese people rejected civets, monitor lizards, snakes, andterrapins, although they would eat lions, leopards, ser­vals, wild cats, hyenas, jackals, and vultures. Reptilessuch as monitor Iizards, snakes, and terrapins are cer­tainly eaten by contemporary Bushmen, and they alsofeatured regularly in the diel of Stone Age people, as didtortoises. In faet, 1would go so far as to say lhat abundantfragmented remaios of tortoises in abone accumulationstrongly suggest human involvement.

Evidence from the southem Cape caves suggests lhatregular and effective fishing was beyond the capabilitiesor Middle Slone Age (or eartier) peoples, although lhetechnique was mastered during the Later Stone Age. Thesame appears lo be lme ror lhe hunting of ftying birds,although there are sorne bird remains in the bone ac·cumulations from almost aH the sites studied.

52 A Guide to the Interpretation of Bone Accumulations in Afriean Caves

Invertebrates such as land snails, freshwater mollusks,and crabs appear to have beco eaten whenever they wereavailable, and large-scale exploitation of marine mollusksgoes back well into the Middle Stone Age. if not further.

Comprehensive data are available on the animáis thatcontributed to bone accumulations in thirteen southemAfrican cave sites known to have been occupied byStone Age peoples. Tbe relevant collections come fromPomongwe, Bushman Rock, Fackeltráger, and Willon,which I have discussed in this chapter, from Nelson Bay,Die Kelders 1, Klasies River Mouth caves, and Redeliff,described by Klein in various publications, from Scott'sCave, described by Klein and Scoll (1974), from An­drieskraaI 1 (Hendey and Singer 1965), from De Hangen(Parkington and Poggenpoel 1971), and from Elands Bay(Parkington 1976). Information from olber irnportant sitessuch as Beyeneskrans, Boomplaas, and Border Cave willbecome available shortly.

In lable 33 1 have listed the contributions made lo eachcollection by individual animals of different marnmalianorders and have calculated percentage representations ofeach order. The ranges and mean percentage repre­sentations per order are given in table 34 and plottedgraphical1y in figure 48.

Primates Other Than Horno. Human remaíns, occasion­ally derived from burials in the caves, have becoexcluded. Sparse remains of baboons were found in eightof the thirteen collections studied, suggesting that theseanimal s were occasionalIy hunted. Elsewhere in Africa,baboons may have beco hunted 00 a larger scale fromtime to time. This is suggested by evidence al locality

~,~"_\a-------

~~"'...

20 "O 60 80 KlOPERCENT_~

Fig, 48. lbe percentage rcpresc:ntatioR of rnammals belongins to variousorden in the remains from thirteen Stone Ase cave-habitation sites insouthern Africa. The range for each arder is indicated in black, and (hemean value is shown as a while line.

DE/89 of Olorgesailíe. where remains of at Ieast 50 adultand 13juvenile Theropíthecus oswaldi baboons were pre­served along with Acheulean tools. Glynn Isaac (1968)suggests that these baboons may have been huoted, in thesame way that the contemporary Hadza people occasion­ally huot them, by surrounding a sleeping troop al night,dislodging them with arrows, and clubbing them to death.

Carnívora. Small quantities of carnivore rernains are aconsistent feature of the bones from all thirteen sites:they come typica1ly from fairly small anirnals, althoughleopards are represented at sorne sites, It is difficult to besure whether the carnivores had invariably been eaten bythe people, or whether they had been hunted for theirskins. In two collections, those from Scott's Cave and DeHangen, several mongooses and genets appear to havedied naturally al the sites; their remains were thereforeexcluded from the calculations.

Artiodactyla. These even-toed ungulates, the bucks, pigs,giraffes, and hippos, were unquestionably the most im­portant meat source for the Stone Age people at most ofthe thirteen sites. As was shown earlier in this chapter,where meat yields were calculated for prey animals fromPomongwe, Bushman Rock, Wilton, and Fackeltréger,medium-sízed antelope eootributed most of the protein toMíddle and Later Stone Age diets. The species huntedvary wilb habital and age,

Perissodactyla. Remaíns of zebras and rhinos are far lessprominent in the food refuse than are those of antelopes.When large animals such as rhinos or the extinct Capehorse, Equus capensis, were eaten, it is probable that thehunter-gatherer band moved to where the carcass lay,rather than transporting much of the meat back to thecave. Zebra remains were particularly abundant in theRedcliff deposit, perhaps owing lo local abundance ofequids there during the Rhodesian Stone Age.

Hyracoidea, As was diseussed earlier in this chapter,dassies were much favored Stone Age prey and werehunted or snared in large numbers. Al asile like DeHangen they were brought back lo the sheller more fre­quently than were other mammals, although they were faroutnumbered by tortoises.

Lagomorpha, Remains ofhares have been found al all thesites under review. These animals apparently made asmall but eonsistent contribution to the Stone Age meatdíet,

Large Forms 01Rodentia. Oceasional remains of por­cupines, springhares. and cane rats are found in the foodrefuse at the sítes, but at a cave like Die Kelders 1, on thesouthem Cape coast, remains of ineredible numbers ofdune mole rats, Bathyergus suíííus, have been found.Klein (I975b) showed that their remains characteristicallylacked fool bones, suggesting that the pelts had been usedand that lbe feet were left attached lo the skins.

If it were not for the large numbers of mole rats repre­sented at Die Kelders and Klasies River Mouth, largerodents would have featured insignificantly in the preyspectrum,

Pholidola, Tubulidentata, and Proboscidea. Pangolinsand ant bears were apparently eaten on rare occasions.

Food Remains of Primitive People in Southern African Caves 53

Sorne rneager traces of elephants are present, but theselarge animal s were probably eaten at the kili sites.

Pínnipedia, The evidence of the coastal sltes indicatesthat seals were regularly exploited. At Elands Bay, andpossibly elsewhere, seal-hunting appears to have been aregular seasonal actívity.

cm

5 ¡nches4

10

8uahman Rock

n_ 2723 • 56.6 96

Pomongwe

,,- 9549- 53.6 %

Wlllonn-33891-n.6 %

3

Fackeltrlger

"_1588_51.2 %

"'50:;o

%so=

o

%so=

o

%50=

o o

oFil. 49. Percentagc eepreseraetíon Dfbone ñekes of various Jengths fromthe four Stone Age cave sites discussed in tlle texto

antelope lirnb bones has typicaHy involved smashing thebone shafts with stone choppers, which produces verynumerous bone flakes. The ratio of bone flakes to otherpieces in the Middle and Later Stone Age collectionsstudied is surprisingly constant. At Pomongwe , BushmanRack, and Fackeltráger, the percentages were 53.6, 56.9,and 51.2 respectively. At Wilton, fragmentation fromtrampling makes it difficult to distinguish bone f1akesfromother fragments; here 77.6% of the sample falls into thisnondescript category. As discussed earlier, lengths ofbone tlakes in the collections tend to be remarkably con­sistent, probably because efficient extraction of marrowtypicaUy results in bone flakes with a mean length ofabout 5 cm. Lengths of flakes from the four sites wheremeasurements have been made are shown in figure 49.

As wiU be described in chapter 4, hyenas produce boneflakes when brealdng up long-bone shafts with their pre­molars. 1 am not confident of my ability to distinguishthese hyena-rnade bone flakes from flakes produced withstone choppers. This matter will betouched upon again inchapter 7 (see also figures 146 and 147).

Stone Age people were undoubtedly capable oí chew­ing and swallowing a good deal of bone , and the pattem ofunchewable skeletal parts from animals such as dassies ishighly díagnostic, as mentioned earlier and in chapter 7.

Chopping and cutting with sharp stone tools produceoccasional marks on bones that could not have beencaused by the teeth of carnivores or by other agencies.When these are found. human involvernent is conñrmed.A study 1carried out on bone food refuse from the Iron Agesite of Zimbabwe (Brain, 1974<:) suggested thal cut marksand chop marks are far more common on bones wheniron, rather than stone, tools have been employed. Onceprimitive people acquired containers in which to cookmear, the pattern of carcass-preparatícn and of bonedarnage appears to have changed in various ways. Porinstance, horns were often chopped from antelope cal­variae so Ihe head could be boiled in a pot. Long boneswere also smashed into very small fragments for boiling toprepare "bone grease"-a practice described among

1.72.84.25.4

5.87.19.0

Il.O11.111.113.315.032.2

Bushman Rock, I : 59Willon, 2: 72Pomongwe, 4: 95Redcliff, 35 : 643Klasies (excluding Cave 1),

13: 223Scott's Cave, 3: 42Klasies Cave 1, 38: 424Die Kelders 1, 28: 254Fackeltráger, 3: 27Nelson Bay, 38: 347De Hangen, 2: 15Elands Bay, 18: 120Andrieskraal 1, 10: 31

The Nature of the Damage

lt will have become abundantly clear, from the dis­cussions in this chapter, that a striking feature of foodremalns from human occupation sites is the damage thebones have suffered, both during marrow extraction andfrom trarnpling by people's feel. Extracting marrow from

The mean of these thirteen percentages is 10.0. 11Ieratio for Andrieskraal 1 is 32.2-a figure appreciablyhigher than those for!he other sites. It is conceivable thatfrom time to time this site was a camivore lair in additionto beíng a human occupation site.

Cefacea. Occasional remains of dolphins and whales arefound in the coastal caves. They presumably eame fromstranded animals that were used by the people.

00 aceount oí the apparentJy extremely eatholic tastesin meat of Stone Age people in southem Africa, 1 amdoubtful if human food remains could be recognizedsolely on the basis of the prey specíes involved. RichardKlein (l975a. 1977) has suggested rhat hominid food re­fuse míght be separated from carnivore refuse by the pro­portional representation oí camivore remains themselveswithin them, The assumption is that camivores typicallyfeed 00 other carnivores more frequently than dohominids. Kiein has proposed (l975a, p. 284) that thecarnivore-ungulate ratio may be used as indícator oí thebone-accumulating agency. Higher figures would suggestcamivore involvement, lower figures, hominid activity. 1am inelined to Ihink that the lerm "ungnlate" needs to besomewhat restricted to be really useful in thís context. ltsconventional usage embraces five orders of marnmals­Artíodactyla, Perissodactyla, Hyracoidea, Proboscidea,and Sirenia (Wender 1948). In my opinion it would beexpedient to exclude the Hyracoidea, sinee the presenceoí dassies at a southern African site is usually detenninedby the presence of a rocky habitat nearby. Inelusion orexclusion of elephants and sea cows would be immaterialto the present discussion. 1 suggest that, in this context,"ungulate" rnight be taken lo mean the artiodactyl plusperissodactyl component, On Ibis basis 1 have calculaledcamivore-ungulate percentages for the thriteen sitesunder discussion:

54 A Guide to the Interpretation of Bone AccumuIations in African Caves

North American Indians (Leechman 1951) but probablyemployed by African Iron Age peoples as well.

A good deal has beco written about the human "crackand twist" technique for breaking long bones. The shaftfrequently breaks in a spira1 manner, and sharp-poíntedbone pieces result. It is true that no animal other than manis capable of grasping a long bone in its two hands andtwisting the shaft in opposing directions. If spiral frac­tures resulted only from this proeess then the presence ofspirally fractured bones in an assemblage would be ex­cellent evidence for human ínvolvement. Regrettably,this is not true (see fig. 144). Sueh bones are found amongfood remains of hyenas and leopards, and the fact thatshafts have broken in a spiral manner is more the result ofinherent properties of the bones than anything else (fig.145). This Is further discussed in chapter 7.

The Presence o/ Biases in che Bone Sample

It is obvious that the feeding ectíon of people, like thatof other camivores, will resulr in the consistent dis­appearance of eertain skeletaJ elernents and the persis­tence of otbers. The relevant factors here are the dentalperformance of the eaters and the robusticlty of the bonesbeing chewed. Other more typically human biases likelyto be introdueed into collecttons of food remains resultfrom culturally determined butchery practices and tradi­tions. The disappearance of foot bones from mole-ratskeletons at Die Kelders suggests, for instance, that thepeople used the pelts of these animal s and typically leftthe foot bones attached to the skins. Likewise, the pauc­ity of large antelope vertebrae at human occupation sltesis probably because the axial skeletons of such prey werecharacteristically left al the kilI sites,

Assocíation with Artifacís

Perhaps the surest and most dírect evidence of humaninvoJvement in abone accumulation is the associationof artifacts with the remains. Studies at caves likePornongwe indicate that, at least for Middle and LaterStone Age assemblages, an abundance of bone pieces inany particular level is assoeiated with an equivalent rich­ness in artifacts. A further indication of sucb an associa­tion is given in figure 50, where weights and numbers ofartifacts in each of the Middle and Later Stone Age levelsal Redcliff are correlaled with weights and numbers of

bone pieces found in association (Brain and Cooke 1967;Brain 1969a). Such associanons were probably charac­teristic of human occupauonal debris throughout thePleistocene, although evidence from Olduvai, discussedearlier, suggests that the very early occupation sites weretypified by fewer artifacts per unit of bone than the laterones.

Association with Evidence DI Fire

Bone fragments from the Míddle and Later Stone Ageoccupation sites 1 have discussed were frequently pre­served in layers rieh in ash, and charring of the piecesthemselves was not uncommon. Sorne simple experi­ments with bones in the ashes of campfires has demon­strated to me that there are two distinct stages in thecharring orbone. The bone starts white then. as the colla­gen within it is carbonized. turns black. Finally, withcontinued heating, the black carbon is oxidized and thebone reverts to a white color and a chaJky consistency. Itis quite possible for a single piece of bone to show aUthree stages, as does the one shown in figure 51.

Befare the introduction of cooking pots, meat presum­abJy was cooked directLy over the fire or beneath thecoals. This could have charred exposed bones, but Iimagine charring occurred more frequently when inediblebone pieces were discarded into the hot ash or when anew fire was made over the refuse from a previous meal.

In open sites, bone eould be charred by natural veldtires; for instanee, in the Pliocene deposit at Langebaan­weg, tortoise remains frequently show the effects of tiresthat may nol have been delíberately slarted (Hendey1974a). However, consistent evidence of ñre in layer afterlayer of a cave site is highly suggestive of purposefulhominid actívity.

There is apparently no acceptable evidence of fire insouthem Mrican caves predating the late Aeheulean,where undoubted traces oeeur in the Cave of Hearths atMakapansgal and al Monlagu Cave in the Cape (Oakley1956). Al one lime it seemed likely that the blackenedbones in the gray breccia at Makapansgat Limeworksshowed evidence of having been burned, and more thanfifty years ago Professor Dart wrote (1925b, p. 454):

As the deposit seemed to be of the cave variety andsorne of the bone fragments had a blackened, charredappearance, the agency oí man in its formation was

ro roQI "1'11&<:15

,n kitograln$

10 20 30 o 2000 10 12000We,gtll ot b" ....a Number ot boolI f.agm'lrU$

in k¡tog.am&

Oeptl\ ",prolit'l

18ft! metera

ca~e

Q(:cUl)aI;on..ta.ls

,

'--..J Fig, 50. Data from the excavated profik: in the Redcliff Cave (Brain 1969a) in Zimbabwe. Weight5 and numbers ofbone fragments per level are correlated with weights and numbers of artifacts found in association.

Food Remains of Primitive People in Southern African Caves 55

highly probable. For this reason sorne promising bonefragments were taken for chemical examination todetermine if any free carbon were present in thebone. This examination was brought to a slIccessfllIconc1l1sion through the courtesy of DI' Moir, of theGovernment Chemical Laboratories, and DI' Fox, ofthe SOllth African Institute for Medical Research.After the bone had been treated with acid and thesoluble material had been washed away, a residue oíyellowish dirt containing numerolls black particleswas revealed. These particles hao an appearancesimilar to that of carbon particles under the micro­scope, and by means oí transformation into carbondioxide a considerable percentage of the elementwas demonstrated chemical!y.

This evidence certainly seemed reasonable enollgh,and it prompted Dart (1948a,h) to name the first aus­tralopithecine remains from Makapansgat Australopithe­cus promefheus. Regrettably,' the first indieations nolonger seem as unambiguous as they did, and in 1956Kenneth Oakley wrote (1956, p. 103):

First one naturally looks for eonfirrnation that therereally is free earbon in the bed eontaining the remainsofAustralopilheeus. Unfortunately it is not forth­eoming. A number of new samples have been testedand no free earbon can be found in any of them. Theblackness of the bone fragments in al! the samples

unburned

carbonized

calcined

Fíg. 51. A piece of Jong-bone shaft showing lhe charaeteristics of charredbone. The upper paje seclion is unbumed; in the middle blackened part theorganic component of ¡he bone has been carboni:red. and in the lowerwhite segment the heat was sufficienl to oxidi¡e the carbon completely.

tested recently proves to be due to manganese di­oxide. Yet the fact remains thal competent chemistsdetermined considerable quantities of carbon in thepieces collected at the same site in 1925. One cannotavoid entertaining a doubt whether the carbon in theoriginal samples was indigenolls. There is always thepossibility to be considered that material from recentfires had, by sorne strange chance, infiltrated, 01' beenintroduced through blasting. So long as no carbon canbe found in samples of the Allstralopithecine bedcollected under test conditions, one feels bOllnd tosay that there is no valid evidence that Australo­pithecus was a fire-user.

Food Remains 01 Precultural Hominids

The recognition oí hominid food remains in the absenceof both artifacts and traces of fire would be extremelydiffieult, and 1 would certainly not be competent to makesuch a diagnosis. It presumably would depend on thedamage done to bones by hominid teeth-a subjectrequiring a good dea! of further study.

4 Food Remains of Carnivores in African Caves

lfone had to give a general answer lo the question,"What do the Carnívora eat?" it would be a verysimple one-"What they can get." [Ewer 1973]

A variety of camivorous animals use caves as retreats,feeding places, and breeding sites. In such placesbony food rernains may accumulate, and in certaincircumstances these may become fossilized. From thetaphonomic point of view it would be useful if boneaccumulations built up by different camivores could bedistinguished. This chapter will consider mainly livingcamivores other than hominids that may take bones toAfrican caves; the role of certain extinct carnivores thatperhaps were importanl in this respect will be touchedupon briefty.

The living camivores that appear lo be of specialsígnilicance are the three species of hyena and theleopardo Several other species may contribute occasionalbones lo an accumulation in a cave; tbese ínclude thecivet, ratel, wildcat, wild dog, serval, and caracal, as wellas certain species of rnongoose, all of which rear theiryoung in holes or rack crevices that may fall within thecatchment area of a cave entrance. Theír possible con­tributíons should be noted but will not be consideredhere, Mention will also be made of the black eagle as anaccumulator oí hyrax bones.

Tbe Spotted Byeno, Crocuta crocuJa (ErxJeben)

As will be mentioned in chapter 8~ there is sorne argumentabout the geographic origin of lbe spotted hyena, whichloday has a wide African distribution (lig. 52). During thePleistocene the species was also abundant in Europe andAsia. According lo Kurtén (1968), C. crocuta firsl ap­peared in Europe in the I-Günz II al Süssenbom andGombasek and then steadily increased in body size lobecome the tme cave hyenaC. crocuta spelaea Goldíuss.At the end of the lasl glacialion it became extinct both inEurope and in Asia.

The first suggestion that spotted hyenas may have beenresponsible for collecting bones in caves was made wellover a century ago, An excavation conducted in 1821 bythe Reverend (laler Dean) William Buckland al KirkdaleCave, Yorkshire, revealed vast numbers oí hyena re­maíns, together with teeth of hippopotamus and otheranimals. He concluded that the cave had served as ahyena lair during antediluvian times and that the layer ofmud that covered the remains had been laid down in thecave by the waters of the bíblical Delnge (Buckland 1822.1823).

56

Sutcliffe (1969) has suggested that rhere may have beentwo kinds of hyena lairs in Britain. The first type is exem­plilied by Kirkdale Cave and by Tomewton Cave inDevon. in which the bone contents consisted mainly oísplintered and gnawed remains of the spotted hyenasthemselves, especially teeth and foot bones. The hyenashad been of all ages, including many juveniles. They wereassociated with rernains of other anirnals, including hip­popotamus, which indicated an interglacial age for thedeposito and very numerous hyena drcppings were alsopresento

As an example ofthe second kind oflair, Sutcliffe citesLevaron Cave. Torbryan, less than a mile from Tornew­ton Cave. A low chamber clase to the entrance containedmany gnawed bones of woolly rhinoceros, reindeer, andother animals índicatíve oí last glaciation times, togetherwith less common remaíns of tbe spotted hyenas.

One explanation for the remarkable hyena concentra­tions in the Kirkdale and Tornewton caves is that thehyenas were forced to take refuge in such caves duringsevere winters and that many died and were eaten by theirfellows en these occasíons. Ifthis were so, however, whyis such evidence not also forthcoming from a glacial­period accumulation such as that in the Levaron Cave?As an alternatíve explanation, Sutcliffe suggested thaltwo different zones might have existed within an individ­ual den: an entrance zone with its concentration of food

Fig. 52. Distribution of tbe spotted hyena. Crocuta crocuta. within his­torie times.

debris, and en inner residential zone with abundant hyenaremains. The former zone could have been representedby Levaton Cave, the latter by Kirkdale and Tornewton.

very shortly after Bucklands assertion that spottedhyenas were responsible for bone accurnulations in cavesand that they ate one another in their lairs, the concept ofthe bone-collecting, cannibalistic hyena was chaJlengedby Robert Knox, who traveled extensively in South M­rica between 1812 and 1817. He published his remarksunder the title "Notice Relative to the Habits of theHyaenas in Southern Africa" (Knox 1822), and his Afri­can research has been further remarked upon by Kirby(1940). Concerning hyena behavior he wrote (Knox 1822,p. 383): "yet I do nol wish it lo be understood that I denythat hyaenas ever drag their prey, including the bones, tothe caverns, or wild mountain rracts, they inhabit. Manyinstances occurred, however which indicare that they donot , .. the young of these animal s follow them early intothe field, so that 1 much question if jhey ever carry aportion of their prey, 00 any occasion whatever, to theirdens."

Despite doubts expressed by Knox, the concept of thebone-accumulating, cannibalistic hyena persisted. Daw­kins (1874) interpreted the very extensive bone accumu­lation in the Wookey Hole Cave near Wells in Somersetas representing a spotted hyena den, and hyenas wereprobab1y gene rally regarded as the most important bonecollectors in caves until tbis concept was challenged byR. A. Dart afier the discovery and description of the ñrstaustralopithecine remains from Makapansgat Limeworks(Dart 1948a-<). Dart was attempting lo explain the originof the extensive bone accumulation in the Makapansgatgray breccia, wbich íncluded the australopithecine re­mains, and he reacted strongly both to Buckland's earlyhyena hypothesis and to more recent writers, such asBroom (in Broom and Schepers 1946), Von Koenigswald(1953), and Oakley (1954a,b), who expressed the opinionthat the bones, including the australopithecine ones, inthe well-known Transvaal caves were probably takenthere by camivores.

In searching for a solution to the Makapansgat boneaccumulation problem, Dart enlisted the aid of A. R,Hughes, who undertook a study of modem spotted hyenalairs. He visited the farm Malamala, 26 km north ofSkukuza, close to the Kruger National Park, where hethoroughly investigated two lairs (Hughes 1954a.b). Oneconsisted of a series of nine ant-bear holes, the vicinity ofwhich had been virtually c1eared of vegetation by thetrampling of the hyenas. Hughes excavated one of thetunnel systems and found that it covered a surface area of13 m by 5 m, descending to a depth of aImost 2 m.Although the tunnels had c1early been used as a hyenabreeding lair, they were empty except for a single tortoisecarapace. Outside the entrance to the lair were fourchewed bones and one set of hyena droppings.

The second lair at Malamala consisted of a low shelterunder an outcrop ofexfoliating granite. lt too was empty,although a few broken bones, a tortoise shell, and twofecal evacuations were found outside it,

In his search for additional evídence, Hughes (1958)also investigated four spotted hyena lairs in the KalahariNational Parle These consisted oC two hyena restingplace s beneath ca1crete outcrops along lhe east bank oflhe Allob River, both of which were devoid of bones ordroppings, and two breeding lairs. The tatter took theform oflow recesses beneath calcrete outcrops, one in theeast bank of the Nossob River al Saínt Joho's Dam, the

Food Remains of Carnivores in African Caves 57

other in the bank of the Auob River, 10 km upstream fromTwee Rivieren .. No bones were found insíde either lair.although two chewed springbok homs were found outsidethe first and eighteen bones and horns outside the second.

On the basis ofthis evidence, Dart (1956<1. 1957b) ve­hemently rejected earlier claims that hyenas could be re­sponsible for extensive bone accumulations in caves, Heproposed instead that such accumulations were the workof hominids-either australopithecines or early men-andin his paper "The Myth of the Bone-AccumulatingHyena" Dart (19500) concluded that the abundant spot­ted hyena remains ín caves such as Kirkdale had beentaken there by primítive men; that the whitened droppingsthat others had interpreted as hyena feces could havebeen produced by bone-crunching hominids or, if theywere in fact hyena reces, could have been collected bypeople for therapeutic purposes. Dart pointed out that inmore recent times such whitened droppings have beenwidely used medicinally, being termed atbum graecum,One prescription is as follows:

Take Album Graecum, 1 ounce; Pulp of the Conserveof Roses. 2 cunees; Syrup of white Poppies, as muchas is sufficient. This is to be spread pretty thick, andapply'd to the Throat, from Ear to Ear; and renewedevery 6 or 7 hours, or oftener, if dry. A Poultessagainst Quinseys. [The Pharmacopoeia Offícínolis el

Extemporünea; or, A Complete Engíish Disnensatory(1736), p. 681]

Despite Dart's forceful and eloquent rejection oíhyenas as bone-collecting agents, evtdence has beensteadily building that spotted hyenas do take bones lobreeding laírs and that these may accumulate either insidethe lair or within the catchment area of its mouth, It hasnow been clearly established that breeding lairs of twokinds may be used: either burrows in the earth orshelters-recesses or caves among rock outcrops. Suchlairs will now be considered in greater detail.

Crocuta Lairs in Burrows

I have already mentíoned the breeding den Hughesexamined at Malamala; it took the form of tunnels origi­nally dug into friahle granite subsoil by ant bears (Oryc­teropus ofer), which had then been used by warthogs(Phacochoerus aethiopicus¡ befare the spotted hyenasmoved in (Hughes 1954a).

A rather similar system of burrows was studied in 1935by Harrison Matthews (1939) on the Balbal Plains in Tan­zania, An atea of about 30 m by 15 m was honeycombedwith burrows and was bounded on the west sirle by adonga about 6 m deep, with steep earth banks. The lirstburrows had evidentIy been dug as adits into the side ofdonga, but as they approached the surface the roofs col­lapsed so that the breeding dens opened into steep-sidedcraters. Matthews concluded that the warrens were theproducts of sorne years' work by the hyenas.

After observations 00 a variety of Crocuta breedingdeos in Ngorongoro and 00 the Serengeti plains, Kmukwas ahle to give tbis general description (1972, p. 242):

Hyena dens usuaUy have a numher oC entrances,sometimes a dozen or more. They are, with few ex­ceptions, on flat ground; only rarely do hyenas useholes of an underground-river system or tunnels dngout io soft tuffs in a vertical surface. The tunnels are

58 A Guide to the Interpretation of Bone Accumulations in African Caves

low tunnels, apparently occupied by hyaenas at its inner­most extremity. The openings of these tunnels werenevertheless still within sight of daylight."

In addition to this cave, Sutcliffe recorded two lairsnear Kajiado Hospital in Kenya. They were in naturalhorizontal cracks in lava that apparently extended sornedistance into the rock but were too low to aJlow in­vestigation beyond about 4.6 m. Bones were visible insidethese lairs for a distance of about 7.6 m.

The Kalahari National Park in the northem Cape is anarea of duneland traversed by two riverbeds, those of theAuob River and Nossob Ríver, which are normallywaterless. These are bounded by low ca1crete scarps inwhich many solution cavities occur; one is described inchapter 5 as an exampJe of a porcupioe lair. Many of thesolution cavities, whicb are often low but extensive re­cesses, are used by spotted hyenas. Hughes (1958) re­corded two resting lairs and two breeding Lairs in suchplaces, and Milis and Milis (1977) described bones col­lected from a breeding lair at Urikaruus on the AuobRiver. This lair, which has a100 been used by porcupines(see below), has three entrances, one ofwhich is shown infigure 53.

In the arid regions of the northem Cape and South­West África, I have investigated five reputed spottedhyena lairs: but in each case I found that porcupines hadalso been tenants of the sites. Bones present in the lairscould have been collected by either agent. In fact.(!.!te..tISe I

oí. hyena lairs by more than one species must be takeninto serious COII'9ideratíon "Ren ah'· -hitel ptetstiOW jj ufbones il>lairsiHNuloo;:]

A den that is likely to provide valuable evidence onspotted hyena bone-collectíng has recently been de­scribed by Andrew HilI (l978).lt is on the dried-up ñats ofLake Amboseli in Kenya and consists of tunnels beneatha calcrete crust, opening to the surface through a trench inthe calcrete. According to Hill, porcupines have not beeninvolved in collecting the numerous bones that litter theentrance trench. Hill's full description of the bone as­semblage, and of the damage suffered by individualpieces, is awaited with interest.

Mention has a1ready been made of spotted hyena lairs inBritish caves such as Kirkdale, Tomewton, and Levaton,and similar sites are known from continental Europe. InAftica it has become cIear that\liI habitats where deep soilexists, spotted hyenas favor ant-bear burrows for theirlairs; in rocky or mountainous habitats, however, they Use 01 Lairs by More Than One Speciesmake use of recesses or caves both for resting places, rretreats, and breeding lairs.] (J'he interpretation of bone accumulations in hyena Iaírs

In the course of bis study on the bionomics of the spot- can be very seriously complicated by coHecting by por­ted hyena, Harrison Matthews described lairs of both cupineü 11 appears that the kind of sites spotted hyenaskinds. Concerning the cave lairs he wrote (1939, p. 49): prefer as breeding lairs coincide closely with those that"Near the Serronea river the country is dotted with sma1l porcupines select as diurnal retreats and, presumably,granite boulders. Amongsl the rocks the hyaenas make breeding sites, Both animals select burrows, caves, ortheir dens, which are natural cavities and are not enlarged recesses that are low-roofed, with lighting subdued oror dug out by the animals." virtually absenl.

The use ofcaves as Jairs in rocky country ofEast Africa As is discussed in chapter S, porcupines are extremelyhas been recorded by Percival (1924). This reference is active collectors of bones, and Ibis must be borne in mindquoted by Kruuk (1972), though 1 have nol read the erigí- whenever abone accumulation from a hyena lair isnal account myself. studied unless porcupioe involvement can be definitely

An interest in hyena lairs in Britain led Sulcliffe (1969, excluded. 11 may seem strange that spolted hyenas wiLL1970, 1973a) to investígate caves in East Africa and lo tolerate porcupines in a lair they are currently using, butrecord the bones they contained. Although Crocuta food this does seem to happen, Al the Urikaruus den in theremains and droppings were found in caves on both Auob River, referred lo above, Milis and Milis (1977) sawMount EIgon and Mount Suswa, SutcIiffe found no evi- two porcupines enter the lair while the hyenas were there,dence that these caves had been used as lairs. He wrote but the species ignored each other. Another lair, known(I973a, p. 46): "The nearest approach to a hyaena lair as Wright's den, on the Nossob River was found to becave which we found was on the shore of Lake Langano, used successively by spotted hyenas, brown hyenas. andEth¡opia, where an impressíve cave opening with scat- porcupines. 00 the basis oC other observations in thelered bones and droppings on ils ftoor closed down lo lwo Kalabari Nalional Park. F. C. Eloff (1975) reports peace.

Crocuta Lairs in Caves and Recesses

mostly oval in section. more wide than high, and theynarrow down from an entrance width of \I.!: to I m tosornetimes less than 25 cm; they are I \I.!: to 3 m long.Sometimes large tunnels extend considerably farther;they rnay go on for several meters, usually horizon­tally, about v.z to J m under the surface, Most of themjoin underground.

One of the Ngorongoro hyena breeding dens wassludied by Sutclíffe (l973a) during 1967and 1970when hesucceeded in photographing a juvenile Crocuta, aboutthree months old, inside one of the tunnels. Two similarburrows were also excavated in alluvial sediments in theQueen Elizabeth Park, Uganda (Sutcliffe 1970). The firstwas simple in form. with a single terminal chamber abouta meter below ground level, opening to the surface by atrifurcating tunnel, The second was more complicated,with a central chamber and a series of tunnels leading toat least ten openings situated around it.

Disused ant-bear burrows that have been taken over byspotted hyenas as breeding dens have also been desctibedfrom Zululand by Deane (1962). On an exposed andwindswepl ridge called "Nkwankwa" in the HluhluweGame Reserve, two colonies are situated about 180 mapart, each consisting of about six burrows. Around thewarrens the grass had been trampled out by the hyenas andthe warthogs tbat shared the breeding dens (see below).

U se of ant-bear burrows by spotted hyenas has beenrecorded in the Kalabari Naticnal Park by MiLIs and Milis(1977) and in the Kruger National Park by Stevenson­Hamilton (1934).

It seems that, in areas where ant-bear burrows are avail­able, these are greatly favored by spotted hyenas asbreeding lairs. They are frequently enlarged and modifiedby the cubs thernselves.

fui coexistence of spotted hyenas and porcupines in thehyenas' breeding lair.

Sutcliffe (l973a) has likewise stressed that sorne of thespotled hyena lairs he studied in East Africa showed evi­dence of multiple use. He pointed out that the hyenas maytake over burrows from warthogs that may in tUID havetaken them froro the ant bears that originally excavatedthero. He found that a lair in the Ngorongoro crater thathad been used by spotted hyenas in 1967 was used bybat-eared foxes in 1970.

FrOID the taphDnomic 'paint of view, the.:multiple use oflairs is of the greatest sígnificaIlc.e¿ The lack of aggressionby spotted hyenas toward other tenants of the same lairsis surprising but well documented. A detailed descriptionof interactions between spotted hyenas and warthog§..at aseries of cornmunal breeding burrows in the HluWuweGame Reserve has been given by Deane (1962, pp.34-35):

The warthogs arrive anything up to two hours beforesunset, slowly feeding their way towards the warrensand showing no signs of fear of the hyaenas and, infact, will graze within a few feet of them. On theother hand the hyaenas, particularly the younger ani­mals, ofien display a degree of uncertainty at the arri­val of the warthogs. The very young whelps watchthem intently, often bolting into one ofthe warrens if

"- '.

..

Food Remains of Camivores in African Caves 59

they approach too c1ose. The warthogs arrive singlyor in family parties and as many as seven have beenobserved arriving together and entering one of thewarrens.

On one occasion four mature hyaenas were stand­ing close together when a large warthog female ap­peared and approached to within a distance of twelvefeet from them. They stood watching the warthog in­tently and were obviously uneasy about her attitude.After they had eyed each other for a few seconds, thewarthog suddenly trotted straight at them with the re­sult that the hyaenas scattered. The warthog thenstopped where the four hyaenas had been standingand after having satisfied herself that they had run offto a respectable distance, she approached one of thewarrens and entered without further ado. As soon asshe had entered the warren, the four hyaenas re­turned and lay down in the sun.

On another occasion a noise was inadvertentlymade and two hyaena whelps, who were playing out·side the warrens, took fright and immediately dasheddown the nearest warren, which two warthogs hadentered only a few minutes previously. After aboutfifteen seconds they appeared at the entrance of thehole and after peering around cautiously they carneout again, apparently having met no hostility from thewarthog below.

'.

"~~ .

Pig. 53. A spoued hyena tsldng the skull of a gemsbok into íts den beneath a ca1crete bank in the Auob River bed.Kalahari National Parle. Photo by M. G. L. MíIls, from Milis and Mills 1977.

60 A Guide to the Interpretation of Bone Accumulations in African Caves

Fig. 54. The left patella of a subadult zebra--allthat could be found al theSalara kilI sile after one nighl of spotted hyena activity.

~o

rneters

~S'(eofkilling

I\,,

\

"Righl fropl lag

I)t Mandíble Pjece,

¡¡kS;1Bone chip~

\

Stream oed

been dismembered and carTÍed away, and, although wesearched game paths for severa! kilometers around fromthe site, no parts of the zebra carcass could be found.

Fig. 55. Skelch map showing Ihe disuibulion of remains lWO days afler agiraffe was killed by lions close lo lhe Sa= resl camp, Kruger NalíonalPark. Spolte<l hyenas were responsil>Je for disarticulating and scalleringthe skeletal parts. '

Kili H40. A medium-sized adult giraffe of unknown sexwas killed by several lions in the early hours of 13 July1968, on the bank of a small tributary of the NwanedziRiver, 0.8 km east of Satara rest campo Soon after day­Jight, A. C. Kemp saw three lions feeding on the carcass.At 6:45 the next morning Kemp again visited the kili sitebut found only the thorax (rib cage with articulatedthoracic vertebrae) and right front leg at the kili site.Vultures were feeding, a black-backedjackal was nearby,and tracks of severa! hyenas were discernible. In thecourse of the day the left hind leg was found by gameguards west of the Salara rest camp, having been draggedby hyenas for 1.4 km.

The following day, 15 July 1968, Kemp and 1 againvisited the site. The thorax, seen the previous day, hadcompletely disappeared except for bone fragments, and itwas c1ear that feeding had taken place at two spots, on thenorth and south sides of the streambed (fig. 55). The rightfront leg had been moved about46 m to the northwest andwas intact and articulated except for the proximal end ofthe humerus, which had been chewed away, and theolecranon process of the ulna, which had been removed(fig. 56b). Qne hooded vulture was feeding on the legwhen we aITived.

Besides the leg, the remains consisted of the fol1owing:

,....-. leO hind leg \"

1.4 km "

{:~.~:~ Bone \ChiPs~"::'_'..·I :

1

¡t..-uame path1

Transport olBones by Spotted Hyenas lo SecludedFeeding Places and Lairs

The complication introduced because more than onespecies of animal may use a spotted hyena lair means wecannot assume that al! bones found in such a lair werebrought there by the hyenas. Thus any direct observa­tions of hyenas actually carrying bones to their feedingplace s and lairs are especial1y significanl. Such observa­tions have been made by a number of zoologists inter­ested in hyena behavior.

As Kruuk (1972) has pointed out, the spotted hyena,with its strong, wel1-developed neck and forequarters andrelalively weak hindquarters, is well adapted to carryíngheavy objects in its jaws. He has observed a Crocufarunning at considerable speed while carrying a 15 kg wil­debeest head well clear ofthe ground. Far heavier objectsmay be alternately carried and drag~d for appreciabledistances. The early hunter F. C. Selous (1908) was of theopinion that a spotted hyena may carry a heavy object bythrowing it over its shoulders while galloping off. He re­lates that when he was camping one night in westernMatabeleland near the Gwaai River, he tethered a fat goatthat weighed 23 kg (50 lb) to one of the wheels of hiswagon. In the night, before Selous was asleep, a spottedhyena approached, seized the goat by the back of theneck, broke the tethering thong, and ran off, apparentlywith the goat thrown over its shoulders.

Stevenson-Hamilton (1934) also relates that two spot­ted hyenas in the Transvaal lowveld kilIed more than ahundred goats over a three-year periodo lt was found thatsorne were being carried 10 km to a den consisting ofsevera! oId ant-bear holes at the side of a dry torrent bed.Stevenson-Hamilton wrote that, judging from the bonesand other indications, the den must have been used bysuccessíve generations of hyenas.

The ability of hyenas to carry off large parts of a car­cass became very clear to me during 1967 and 1968 when Imade observations on spotted hyenas in the Kruger Na­tional Park. At that time, A. C. Kemp of the TransvaalMuseum was undertaking ornithological fieldwork in thevicinity of Satara and kindly located forty-one !ion killsthatl was then able to keep under observation for varyingperiods up to fifteen months. The objective was to studythe scattering of bones and the survival of skeletal parts.Two of the kilis were of particular relevanee here becausethey were made very close to the Satara rest camp insituations where the hyenas did not feel at ease. Bothskeletons were broken up and carried away with remark­able rapidity, as the fol1owing descriptions indicate.

Kill H28. A subadult zebra (Equus burchelli), ofunknownsex and weighing an estimated 180 kg, was killed by asingle male líon in the early hours of 10 August 1968within sight of the east fence of the Satara rest campoSoon after dawn the lion was seen feeding on the rib cageof the zebra, but apart from damage to the ribs the skele­ton was virtually intacto Qne spotted hyena and threeblack-backedjackals were waiting about 30 m away. Withlhe daylight, the lion, hyena, and jackals retreated, andthe carcass remained untouched until the evening. Thelion did not return, but five spotted hyenas did; they wereclearly nervous at being so close to the rest camp, andthey made a good deal of noise. The following morning,11 August, the vicinity of the kili site was searched withgreat care, but only one bone could be found-the leftpatella shown in figure 54. The rest of the skeleton had

Food Remains of Carnivores in African Caves 61

north bank area, 128 chewed fragments of rib, 1 piece ofmandible with a full set of incisors; south bank area, 148chewed ríb fragments, I neural spine from a thoracicvertebra, and 4 sesamoid bones (fig. 560). Apart fromlhese fragments, the giraffe skeleton had been dis­articulated and scattered by spotted hyenas in two nights.A search of the sUITounding countryside failed to revealwhere the hyenas had taken the parts, including suchweíghty objects as the head, neck, right hind limb, andleft forelímb.There seems little doubt that the rapidity with which thehyenas dismembered this giraffe carcass was infiuencedby the proximíty of the kill site to the Satara rest campo Asecond giraffe had been killed four months earlier fartheraway from the rest camp, and its skeletal elements werenot scattered to the same degrce. Detaíls fo\low.

Kili H39. An adult giraffe of unknown sex was killed bylions c10se to the Nwanedzi road about 4 km southeast ofSatara rest camp on 11 March 1968. The kíll was visitedon the first day by Kemp, who also observed feeding bylions, spotted hyenas, and vultures over the next threedays. The feeding process was not disturbed by theproximity of humans and was far more leisurely than wíth

the two kills c10se to the rest-camp fence. There was verylittle activity of any kind after the first seven days, and 1fina\ly made a sketch plan of the distribution of the re­mains on 18 July 1968 (fig. 57). Certain skeletal elementshad disappeared, but the greater part of the skeleton waspresent, although scattered as much as 23 m in variousdirections from the original kill site. The distributíon wasrechecked on 1 September 1968 and found to beessential1y the same. Thc pelvis was still attached to thefive lumbar vertebrae and last four thoracic vertebrae.Damage was restricted to the edges of the ilía, the rightischium, and the transverse processes of the vertebrae.The skull was almost complete, except for damage to thenasals and left ossicone.

Hyena activity at this kilI was probably more typícalthan at the first two, in that most of the feeding took placec10se to the kilI site. However the first two kills are ofinterest because they indicate that, when disturbed anduneasy, spotted hyenas are capable of dismeIPbering acarcass with surprising rapidity and carrying off the partsto secluded feeding places.

An instance of transport of bones to a lair was recordedby Sutcliffe (1970) at Kajíado, Kenya, where, amongother bones, three human craniums were found in a low

.'.;t I~J.;

,~

O,--~_..::3__---,,6 .ins.

,.•

••••. ' ...

" .,

b

a

Fig. 56. (a) Remains ofa giraffe skelelon, H40, collected al the kili site (see lig. 55) af!ertwo nights o[ spotted hyenaaction. The hyenas were uneasy al lhe close proximity o[ the rest camp and scattered lhe remains more rapidly lhanusual. (b) The right [ront leg o[ Che giraffe showing damage caused by hyenas chewing the proximal end o[ Ihehumerus.

ebw",-~'~ I.-H~-4~..(. _ ~, ;.,1:....

62 A Guide to the Interpretatíon of Bone Accumulations in African Cav~

Fig. 57. Distribution of skeletal parts four montbs after a giraffe. H39. was killed by u- nens and red upon by sportedhyenas in the Satara area of the Kruger National Park: A, four thoracic v~rlebrae;:; rqht tibia; C. two thoracicvertebrae. left metararsel and tarsal; D, right metatarsal; E, articulated cervical verteeeerae. twc ríbs, and left tibia;F. tborecíc vertebrae, ribo and hoor; G. hoor, right astragalus, and right radius/ulea: 1: skull and right metacarpal;1. atlas vertebra and hyena droppings; J, right tarsal; K. proximal humeros fruKment: L _ . moracíc vertebra; M. boneftake and hyena droppings; N, bone flake; 0, pelvis witb nine articulated vcr1cbrae::- . TWO carpa] bones.

There is also corn:rnetition between hyaenas, so that ina very short lime narts of carcasses, by beíng passedfrom onc hYilena ro another, may be carried a distanceof several miles. Vlie tried placing a head of a wilde­beest which we naad taken from a hyaena, outsideone of the Ngorormgoro lairs. In a short time a halfgrown hyacna carene out ando waIking backwards,pulled the head mnto the entrance of its lair where itjammed by the hocrns. The hyaena repeated this whenwe carried the heead away again. Its motive for doingthis was apparerrtmy to take the head somewhere whereit could eat quietttv . without competiticn from otherhyaenas or líons oor jackals or vulrures. [Sutcliffe(19730), p. 49]

o (jIllOlI

In another study of spotted hyenas in the Ngorongorocráter, van LawicKr.-Goodall and van Lawick-Goodall(1970) descrihcd ho-sw a female carried a large piece ofantelope backhone -eack lo the den where she had left hertwins. Finding that me cubs were no longer there, shemoved from den to den with the bone, calling at each,until she relocnted heer young. She then put the bone downnext to them nnd las- down to rest.

In the course of "nis study oí spotted hyenas in theTransvaal lowveld. :Bearder (1977) observed hyenas car­rying off parts of a ccarcass at dawn, presumably so theycould feed undisturi:'eed in a seC'Iiided place. Likewise , in

KJ'

L~ ~LSITE ., ,

N I

1 G~' HíF.} E"", O

f " e-Oo,o 10

""e,. '-'\meters <-

ROCK

\]OUTCROPS

D~

A B/•

NI

recess used by spotted hyenas. The skulls were found tohave beco dug up by the hyenas at Kajiado hospital cem­etery 4 km away and carried lo the lair.

On the basis of his detailed study of spotted hyenas.Kruuk was able lo remark as follows (1972, p. 244):

Hyenas also do not carry any substantial amount offood to their cubs; they may occasionally take aboneor a head along, but this seems comparable to taking itto their lying-up place, where they can chew in peace.Cubs may chew these bits, but only after the originalowner has finished with them. Older cubs are morelikely lo carry bones lo theír den than adults: theysometimes take them down their holes, but more usu­alIy these are left in front of the den, or in the en­trance. Sorne dens accurnulate quite a coUection ofbones around the entrance, contrary to the belief ex­pressed, for example, by Hughes, but this does nolofien happen; it may be more common among brownhyenas.

Working in collaboration with Kruuk, Sutcliffe (19730)made further observations on bone-carrying. He con­cluded that, in the Ngorongoro crater, where spottedhyenas~ kili their~ prey only to lose the carcass toscavenging lions,

the only hyaena to get a good meal was one that haddetached a leg and had run off before the Iion arrived.

Food Remains of Camivores in African Caves 63

the Kalahari National Park F. C. ElotT (1975) has ob­served a female spotted hyena dragging parts of aspringbok carcass lo a den where there were two sets ofcubs of about three and eight months old, and Milis hasphotographed a spotted hyena taki~ga gemsbok skull ínto'he Urikaruus den on thc Auob River (fig. 53).

Bones and Other Animal Remains Recorded al SpotledHyena Lairs

Although many observations have been made oC foodtransport by spotted hyenas, particularly to lairs, rela­tively few analyses have been undertaken of bone collec­liaos in and around such lairs. [ will now discuss the fewthat are currently available.

During his early observations in the Kalahari NationalPark, Hughes (1958) listed bones from two spotted hyenabreeding denso From the Saint John's lair on the NossobRiver he recovered two pieces of springbok horn, andfrom Site 2, another breeding den in the bank of lhe AuobRiver, he retrieved 18 pieces as follows: 4 springbokhoms, 7 antelope bone fragments, 4 rnedium-sized verte­brae, 1 ulna, 1 innominate, and 1 sacrum from a smallantelope.

More reeently, Mili, and Milis (1977) have studiedthree more deos in the Kalahari National Park: Kas­persdraai, used only by spotted hyenas; Urikaruus, usedby spotted hyenas and poreupines, and Wright's den,used by both spotted and brown hyenas as well as por­cupines. Wrighfs den yielded only 9 bone pieces that willnot be considered further.

Milis recovered 16 booe pieces from the Kaspersdraaiden, as follows:

Antídorcas marsupialís, springbok: 3 adults, based 00 3right mandible pieces, I left mandible píece, I right and Ileft hom sheath.

Alcelaphus buseíaphus, hartebeest: I old adult, basedon 1 mandible piece.

Antelope elass 11I: I ríght libia, I ealcaneus piece, Iright metatarsal, and 1 right metacarpal; 1 indeterminaterib pieee and 4 bone ftakes.

The 71 bone pieces collected at the Urikaruus den wereas folIows:

Oryx gazella, gemsbok: I adult, based on I eranialpíece and 1 horn sheath; 2 juveniles, based on 2 eranialpieces; 1 young juvenile, based on 1 mandible piece.

Antidorcas marsupiaíis, springbok: 1 adulto based on Iarticulated forefoot.

Antelope class 11: 2 scapula pieees.Antelope elass 11I: I metatarsal pieee3 indeterminate pieces and 59 bone flakes. There is no

doubt that many of lhe bone ftakes in the Urikaruus sam­pie were derived from regurgitations-this wiU be dis­cussed further.

From lhe Queen Elizabelh Park in Uganda, Sutcliffe(970) ha. deseribed lhe bone eontent of two spottedhyena burrows, though unfortunately he has not provideddetailed analyses. One burrow had a simple form with asingle terminal ehamber about 1 m below ground. Theaccess tunnel was found to contain a few fragments ofantelope bone and sorne juvenile hippo bones; the tenni·nal ehamber yielded the complete skeleton of a juvenileCrocuta and a single dropping. The second bUlTow con­sisted of a central chamber, opening to the surface by atleast ten tunnels that contained numerous bones includingskulls of a juvenile hippo, warthog, and kob, bones of a

buffalo , a juvenile elephant humerus, and skull fragmentsof a juvenile hyena. Most of these remains had been dam­aged by hyenas.

In the Ngcrongoro crater, a spotted hyena burrowinvestigated by Sutclíffe (1970) contained many bones,rnamly of zebra and wildebeest , together with the skull ofan adult hyena; and two natural horizontal cracks in lava,used by spotted hyenas at Kajiado, Kenya, contained sub­stantial quantities of bones. These included remains ofdomestic donkey, cow, and dog, bones of an ostrich andparts of an ostrich eggshell, and also human remains suchas 3 calvariae and a mandible.

In the course of his study oC Crocuta in the eastemTransvaal, Bearder (1977) eolleeted 409 bone pieees fromsix breeding dens in the Timbavati reserve. His analysesof these pieees is presented in table 35. Of lhe 409pieces, 123 could be assigned to species, as shown in thelast colurnn oftable 36. The animals involved were zebras,giraffes, impalas, wildebeests, and kudus.

Regurgiíations as a Source ofBanes at Lairs

It has been established, it seems, that spotted hyenas donot regurgitate food for their cubs (Kruuk, 1972; Bearder.1977), but it has been equally firmly established that theyregularly regurgitate indigestible residues of recent meals.As Kruuk has pointed out (1972, p. 244):

Adults do regurgitate near tbe den, but what comesout is almost invariably a large slirny ball of hair andbone slivers with nothing edible about it. The reacticnof any hyaena (also a cub) to the sight and sound ofanother one regurgitating is to run up, sníff the resultsof this action, and then immediately roll in it, shoulderfirst, then wilh the whole back, legs bent and headlifted. Often the regurgitating animal joins in.

Regurgitation and rolling were likewise observedarnong Ngorongoro spotted hyenas by van Lawick­Goodall, who also reeorded that the hyenas will piek outpieces of undigested bone from the fresh regurgitationsand eat them again. Despite this habito accumulations ofregurgitations may build up near spotted hyena dens; inthe Timbavati reserve, Bearder (n.d.) found that up to 35regurgitations accumulated at well-used den sites in thecourse of a month, and 200 complete regurgitations werecollected in the vicinity offour such dens during hís studyperiodo Bearder was able to identify the animals that hadcontributed hair and other undigested resldues to the re­gurgitations; his list is given in the first column oftable 36.He showed that, although bone pieces, chips, and sliversoccurred in both regurgitations and scats, they tended lobe larger in the former than in the latter.

Bone pieces that have undergone corrosíon in spottedhyena stomachs before regurgitation show recognizablecharacteristics, Corrosión of bones, almost certainlycaused by spotted hyenas' gastric juices, has been ob­served on very numerous specimeos from !he Pin HoleCave in the Creswell Crags 00 the border betweenDerbyshire and Nottinghamshire in England. 1t was heretha' Magins Mello and Boyd Dawkins worked from 1875onward and demonstrated for the first time that man hadbeeo a contemporary of the mammoth. In his accouot ofthe mammalian fauDal remains from the cave. Dawkins(1877) eoneluded that lhe numerous eorroded bones hadsuffered through the action of carbonic acid in the cavesoil. This expJanation seemed unlikeIy to James Kitching

64 A Guide to the Interpretation of Bone Accumulations in African Caves

(1963), who had samples of the cave earth analyzed andfound that they were highly alkaline. Kitching concludedthat the corroded bones were from spotted hyena re­gurgitations or from the stomachs of hyenas dismemberedin the cave. His ilIustrations show that corroded bonesfrom the Pin Hole Cave compare very closely with bonefragments frorn the stomachs of spotted hyenas shot inthe Kalahari Nalional Park. Kitching concluded that2,946 bone ñakes showed signs of partial digestion, tbesame being true for the following specimens: 32 erodedjuvenile incisors of Equus, 101 deciduous upper and 68lower molars of Rangifer, and 45 deciduous molar platesof Elephas, as well as 15 astragali, 29 phalanges, and 142masloid, carpal, and larsal bones of reindeer.

One of the more striking of the corroded bones fromPin Hole Cave was part of a bison's rib with numerousholes in ít. This specimen had been regarded by A. L. Arm­strong (1936). excavator of the cave. as a human arti­fact; he tenned it a bull-roarer. Kitching, however, hasargued that it is simply a rib fragmenl that had sufferedcorrosión in a hyena stcmach.

Jhe similarity of hyena-corroded bone fragmenls lohuman artifacts is sometimes striking, This should beborne in mind when interpreting bone accumulations towhich hyenas may have contributed.

Droppings o/ Spotted Hyenas

The abundant presence of hyena droppings in certainBritish caves has already been mentioned; such droppingsmay also be found in African caves. such as at Redcliff inRhodesia, where their abundance in an excavated profilehas been plotted (Brain, l969a). It was found that con­centrations of hyena coprolites coincided with layen inwhich quantities of human culturalmaterialwere low, theinference being that hyenas used the cave at times whenthe human inhabitants were elsewhere.

Although spotted hyenas may defecate inside their lairsor in the immediate vicinity, they also tend to havespecific "Iatrine areas" where they indulge in social def­ecation. The existence of such areas has been known fora long lime--under the heading "scatology," HarrisonMatthews wrole as follows (1939, p. 47):

The faecal masses of the Spotted Hyaena resemblethose of a large dog in shape and size, and consistalmost entirely of mineral matter derived from bones,with sometimes a few hairs or feathers, The lower partof the intestine in all specímens exmined contained agreen paste. with sorne lumps, which tumed white ondrying in the airo When fresbly deposited the faecesare green in colour, but as soon as they are thoroughlydry they become quite white, and are then the classi­cal "a1bum graecum." The animals' habit is lo defae­cate at regular latrine areas, spaces of limitedextent covered wilb a plentíful deposit of a1bumgraecum. These places are conspicuous at a distancewhen the sun is shining on the chalky white faeces.Some of the latrine areas are very large: one seen nearthe Serronea river must have covered quite a quarterof an acre. A resident of Arusha was heard to describesuch a deposil as "like a fall of snow"; nor is thesimile greatly exaggerated. It is uncertain whether theanimals consciously repair to these areas to defaecate,but it seems more probable that the sight of the de­posit, when casually encountered, provides the neces­sary stimulus to the individual.

Hyena droppings (fig. 58) can provide valuable in­formation about the diet of the animals that producedthem. In the course of his East African Crocuto study,Kruuk (1972) collected 810 fecal samples, ground themup, and extracted hair for identitication ofthe prey eaten.His analysis provided infonnation about hyena diet dur­ing the wet and dry seasons as well as in various habitats.In an attempt to assess the reliability of fecal analysis asan indicator of actual diet, Kruuk plotted the percentagesoftotal occurrence ofwildebeest, zebra, and Thompson'sgazelle in feces from different ranges in Ngorongoro andfrom Serengeti areas, at different times of the year,against percentages of actual observations of hyenasfeedíng in those areas at the same time. A close correla­tion was found between the two sets of observations,confirming the usefulness of fecal analysis as an indicatorof diet.

Conclusions on spotted hyena diet in the eastemTransvaal have been drawn up by Bearder (1977), largelythrough analysis of 527 droppings and 200 regurgitations,Results are given in table 36. Of particular interest is thedifference in representation of large and small preyspecies in the scats and regurgitations. The percentageoccurrence of giraffe hair in the scats was 38.3 as opposedto 16.3 in the regurgitations. Impala hair, on the otherhand, was presenl in 36.8% of the scats and 60.2% of theregurgitations. Bearder's explanation for this and similardiscrepancies is that a hyena feeding on a small animalsuch as an impala is likely to consume more hair relativeto a given weight of meat than when eating from a largeanimal such as a gíraffe,

In this connection Bearder observed that, in addition tothe usual whitened droppings, brown, organic-rich onescould sometimes be found. These were apparently pro­duced by hyenas that had fed on meal lo the exclusion ofbone, which they periodically do when a large carcass isavailable, Although Kruuk (1972) estimaled that the aver­age food intake for Ngorongoro spotted hyenas in the dryseason was about 2 kg per day, he a1so poínted out that asingle hyena wiU readily eat a third of its body weight alone lime. This has been confirmed by Bearder (n.d.), whoreported that a three-year-old Crocuta ate 18 kg ofelephant meat in one night and then retumed for more thefollowing evening. On another occasion, two adults atemosl of a 50 kg impala in I hr, 20 min.

In conclusíon, the presence ofcharacteristic coprolitesin a cave deposit provides valuahle evidence that hyenashad made use of the site, The chances are good that theycontríbuted lo any bone accumulation that may be pres­enl there,

The Diet 01Spotted Hyenas

Detailed informalion about whal spolted hyenas eal isrelevanl lo Ibis study in that it indicates the range of bonyobjects the hyenas may take back to a cave lair. It hasbeen amply demonstrated, in many parts of the Africancontinent, that spotted hyenas are active and effectivepredators as well as being scavengers,

Kruuk (1972) has collected a weallb of information onCrocuta diet from direct observations in the Serengeti andNgorongoro areas. Results based on 513 carcasses in theformer area and 297 in the lalter are given in lable 37.Wildebeesl forms the mosl importanl dietary Item in bothareas, followed by Thomson's gazelle in the Serengeti andby zebra in Ngorongoro. Surprising as it may seem, the

hyenas were fOllnd to sllccessfully hllnt animals as largeas <ldlllt waterbllck. dand, and buffalo.

Information about spotted hyena kills in the KwgerNatíonal Park oVer a thirty-year period has been compiledby Pienaar (1969). As is shown in table 38, impala forms58.8% of the kills, followed by waterbuck, wildebeest,and kudll.

In the nearby Timbavati reserve, Bearder (1977)showed that impala was also the most frequent prey ítemin regurgitations, although giraffe hair appeared withgreater regularity in scats. These two species were fol­lowed in percentage representation by wildebeest andzebra. No actual kilIs were observed during Bearder'sstudy periodo

In other areas, spotted hyenas clearly feed on whateverthey can gel. In the Kalahari National Park, F. C. Eloff(1964) recorded that three adult gemsbok and twojuveniles were found to have been killed by spottedhyenas between 1958 and 1961, and remains collectedfrom the Kalahari Park dens, referred to earlier, carnefrom gemsbok, springbok, and hartebeesl.

In Zululand, Deane (1962) has recorded Crocuta pred­ation on wildebeest and warthog, as well as on an oldand sick black rhino. Other odd dietary items are fish(Stevenson-Hamilton 1947), tortoises and Iions (Pienaar1969), hippo calves (Cullen 1969), and young elephants(Bere 1966).

Many accounts have been published of spotted hyenasattacking sleeping people and attempting to drag them off.In the Mlanje District of Malawi, severa! attacks havebeen described by Balestra (1962); during September 1955an adult African man, who happened to be the village¡diot, was killed and eaten on a path between two villages.It is estimated that at least four spotted hyenas panici­pated, leaving nothing bllt sorne patches of blood andshreds of clothing. Seven days later an old wornan wasdragged from her hut by a Crocuta and lost an arm beforehelp could reach her. She died the next day. The same

Food Remains of Camivores in African Caves 65

year a six-year-old child was dragged from a verandawhere she was sleeping and eaten; only her head was left.This too wOllld probably have been ea¡en had the hyenasnot been driven off by village dogs.

The habit, which persisted till within historical timesamong variolls African tribes, of plltting aged, feeble, orsick individuals outside villages at night so that they cou\dbe disposed of by scavengers must have meant thatspotted hyenas regularly fed on human flesh. The earlytraveler and hllnter F. C. Selous re\ated an incident inwhích the body of aman was eaten by spotted hyenas.During 1872 Selous was camped near the Jomani River ineastern Matabeleland; at that time one of the residents ofthe area was shot by a Hottentot, who was then clubbedto death by the friends of the rnurdered mano The Hot­tentot's body was left \ying in the bush about 275 m fromSelous's wagon. The first night several hyenas created agood deal of noise around the corpse frorn dark to day­light; on the second night they once again left it alone, buton the third night they devoured it (Selous 1908).

Fig. 58. Characlerislic whilened droppings of a spotled hyena: Salara area, Kroger Nalional Park.

66 A Guide to the Interpretation of Bone Accumulations in African Caves

I am not aware of any accounts providing detailed in­formation about bony remains resulting from hyena ac­tion 00 human bodies, nor is such information easy toobtain experimentally. The closest I have come to it was afeeding experiment I undertook involving a baboon. InOctober 1967 a large mate baboon weighing 27 kg wasshot at Satara in the Kruger National Park while making anuisance of itself in the rest campo The body of the ba­boon was securely tied down to the ground 50 m from therest-camp fence, and several hyenas were heard to ap­proach soon after darkness fell. The following momingthe area was carefulIy searched for rernains, but, apartfrom a patch of blood where the body had been eaten,sorne tufts of hair, and a piece of intestine 20 cm long,nothing was left.

The early observations of Buckland (1822) on the boneaccumulation in Kirkdale Cave, Yorkshire, led to theconclusion that the spotted hyenas that lived there hadbeen cannibalistic, feeding upon one another inside thecavem. The concept of cannibalistic hyenas was vehe­mently rejected by Dart (l956a, 1957a) and Hughes(J954a,b, 1958) in their consideration of the possible roleof hyenas in accumulating bones at Makapansgat. Despitethís. however. evidence has been coming to light thatclearly shows that spotted hyenas do eat one anotherupon occasion, and that Crocuta remains in a cave couldwell be food remains of their conspecifics.

When Hughes (1958) wrole "Sorne Ancienl and RecenlObservations on Hyaenas," he quoted M. Cowie, wardenof the Royal National Park in Kenya, as follows: "withspotted hyaenas I have seen many willing to eat their ownkind, but only after a dead one has become almostputrefied. In other words although 1 have seen spottedhyaenas killing each other, either in fighting for a ladyor over a kili. 1 have never seen them eat each other untilthe dead one has been Iying for something like four days,and has become what to us would be a most offensiveobject. "

The lendency of spotted hyenas lo wait until lhe bodyof one of their fellows has partially ~mposed beforethey are willing to eat it is borne out by two observations Imade in the Salara arca of the Kruger National Park.Details of the incidents are as follows:

Kili H44. An adult Crocuta of undetermined sex wasfound about 90 m from the Salara rest-camp gate on thebank of the Shitsakana stream on 28 May 1968. It hadapparently been killed by lions fíve days before and,although decomposing, was complele. Two days later,that is, seven days after death, it was eaten by otherspotted hyenas, the only remains being the leñ and rightscapulae, shown in figure 59a. These remains were Ieftunti\.27 June 1968when, there being no change, they werecollected.

Kili H45. An adult Crocuto of undetermined sex wasfound lo have been kilIed by lions 0.8 km wesl of Salaraon 20 March 1968. Four days later il was beginning lo rolbut was stíll complete. Within the next two days it wasconsumed by one or more spotted hyenas, except for anumber of ribs and chips of bone. Tbese were left wherethey lay till 30 April 1968 when, there being no furtheractivity, they were collected. As is shown in figure 59b,the remains consisted of 13 complete ribs, 6 sternalbones, 2 slernal cartilages, 2 proximal tibiae, 1 patella and29 bone ftakes.

Another incident of cannibalism in the Kruger NationalPark was recorded by Pienaar (1969). On a night in Au­gust 1966, two spotted hyenas were seen feeding onthe hindquarters of a younger one along the NahpeRoad. Whether this carcass was fresh or putrid was notrecorded.

Spotted hyenas do not appear to have a strong inhibitionagainst feeding on young oftheir own species. F. C. Eloff(1975) reports a case in tbe Kalahari National Park whenan adult was seen feeding on the remains of a very youngcub , "which was probably its own," and Kruuk (1968)records an incident in which his tame Crocuta cub, Sol­omon, was carried off one night by a wild adulto The cubwas outside Kruuks house at Seronera one moonlessnight when, at 2 A.M., it was grabbed by a wild adulto Thecub's cries caused Kruuk to give chase and recover thecub, which had suffered a gashed throat, puncturedwindpipe, and broken jaw. With penicillin treatment thecub recovered.

Two other incidents of spotted hyena cannibalism havebeen reported from the Ngorongoro crater in Tanzania.The first, reported by Kruuk (1972), occurred in Septem­ber 1%7 when a hyena from the Mungi clan killed a wil­debeest within the territory of the adjacent ScratchingRocks clan, A number of Mungi clan hyenas started lofeed on the carcass, but they were challenged by theScratching Rocks residents, and one was severely injuredin a fight. The following day it was partly consumed bysevera! hyenas that, at the time of Kruuk's observation,had already eaten about a third of the body,

The second incident, recorded by van Lawick-Goodalland van Lawick-Goodall (1970), look place in the samearea, on lop of the Scralching Rocks hilL Here one of thehyenas was killed by a lion; the body was left untouchedthroughout the following day but was then consumed bymembers of the clan to which it belonged, a number ofcubs participating in the feeding,

The conclusión to be drawn from all these observationsis that cannibalísm is frequent among spotted hyenas;both adults and juveniles may be involved, and it maytake place within a single clan. It is highly probable thathyena bones would be added to abone accumulation inany long-term Crocuta lair through cannibalism.

Food Storage in Water: A Possíble Source 01 Rones inCaves

In dolomitic countryside it is not unusual for a pool ofwater to exist either inside the entrance lo a cave orwithin íts catehmenl arca. Tbis is particular1y true ofsituations where streams ei!her disappear undergroundthrough a cave entrance Of emerge again from a cavem.Spotted hyenas have a habit of storing surplus food inwater for future consumption, and this can result in anaccumulation of bones al the boltom of a water hole. lfthe water hole happens to be close lo a cave entrance, thisstrange hyena habit could contribute to abone accumula­tion inside the cave.

In connection with the food-storage habit of Crocuta,Kruuk has written (1972, p. 119):

Hyenas leaving a kili often carry a large chunk ofmeat or bone away with them, which is usually eatenquietly sorne distance off. Occasíonally, however,they cache it by dropping it ínto 30-50 cm of standingwater and collect it later. In Ngorongoro, I have seentbem do this three limes, always in the cenlrallake,

Food Remains of Carnivores in African Caves 67

and in the Serengeti both Schaller and 1 once saw ahyena leave part of a carcass in a smaH waterhole. Onseveral occasions 1 fOllnd fresh pieces of carcasspartly or wholly submerged in Serengeü waterholes,without hyenas nearby; but from the way the foodwas partly eaten, it was aH bu! certain that the hyenashad put them there. In the crater lake, hyenas had towade up to 10 m from the shore to reach the proper

depth. They usually retumed to a cache within a day,but the many bones tha! were exposed when par! ofthe Ngorongoro lake dried IIp probably meanl tha!several caches had been lefl unused. A cache canmost likely be recovered only by visual relocalion,and this is no! always effective enough. 1 twice saw ahyena wade ¡nto the lake and repeatedly plunge itshead deep under water in a small area, obviously

f'¿/jO 15ti===:::i__.t:'=:::::::::l-1

cm

b

Fig. 59. (a and bJ Skeletal remains of two spotled hyenas coUected after their decomposing bodíes had been eaten byother spotted hyenas in the Satara area of the Kruger National Parle

68 A Guide to the Interpretation of Bone Accumulanons in African Caves

searching but eventually leaving without finding any­thing. But 1 also twice saw a hyena successfully re­cover a cache in this way. once the head of a wil­debeest. and the second time the remains oC a deadIion: neither of these was in any way visible fromaboye the surface of the extrernely muddy water.

An earlier but rather similar observation was made by aformer warden in tbe Kruger Natlonal Park and publishedby Wolhuter in his Memones 01 a Game Ranger (1948,pp. 178-79):

One day I was resting under a big tree at the'Mbeamede spruit near a small pool of water about 18inches deep. Suddenly 1 saw a hyaena trotting downthe bank, carrying something in his mouth. He ad­vanced to the pool. and dropped what he was carryinginto the water. and then stood back to watch the ef­fects of his action. Apparently he was astoníshed anddissatisfied with the fact that the piece of meat (orwhatever it was) ñoared on the surface for presentlyhe entered the water, seized the meat in his mouth,and pushed it below the surface once more. Itpromptly rose and floated again, and this seemed tocause the hyaena a great deal of concem. I couldalmost imagine that 1 saw the expression of astonish­ment in his Cace. Hís expression, in fact, was soJudicrous that 1 was unable to control a chucklc. andthis startled him so that he glanced nervously in rnydirection before ambling off. 1 waited a while, but ashe did not return 1 went down to the pool to find outwhat the object was that he obviously valued sohighly. It tumed out to be a fragment oflung, of eitherwildebeest or waterbuck, and of course, being lightand bouyant, It naturally ñoated near the surface ofthe water. On scratching round the bottom of the pool1 found sorne decomposed bones, which preved thatthis unusual place was evidently his larder for tbe leantimes.

Food storage of this kind may seem insignificant intransient water holes, but where a cave is involved inwhich bones may accumulate over sorne thousands of

~I ./

years, the water-storage procedure could be of greatimportance.

The Impar/unce of Scüvenging as Opposed lo Hunting

Spotted hyenas have obvious scavenging adaptatíons inthe form ofharnrnerJike premolars mounted in robustjawsthat are powered by very large muscles. They are able tocrack remarkably large bones, thus making a meal of re­mains left by much larger predators such as Iions. But,despite these adaptations, it has been appreciated formany years that sr atE I 11' I 0:&0- tLalitlf&í1lt:is intheir own right. Referring to spotted hyenas in theKalahari NationaI Park, F. C. Eloff(1964) described themas "a curious mixture of hunter and scavenger," but theeffectíveness of their hunting was first demonstrated byKruuk's study in Tanzania, OC 1,052 hyenas that wereobserved feeding, Kruuk (1966) found that 82% wereeating from animals killed by hyenas, whereas only 11%fed on animals killed by other predators, such as jackals,lions. wild dogs, leopards, and cbeetabs; cause of death inthe remaining 7% was doubtful.~t, in the Ngorongorocrater the--~traditioR""re'ztiODship:'bctw.eeD:.--hyenRS 'andlions was reversed: the lions obtained most of their mearby scavenging from hyellailills.

For interpreting abone accumulation in a cave to whichhyenas may have contrfbuted, it would be useful lo knowwbich bones resulted from bunting and which fromscavenging. Active hunting by spotted hyenas is likely tohave been more frequent in environments where few car­casses were available for scavenging. The abundant pres­ence of carnivores larger than hyenas, fossil or modern,may be a reasonable indicator of the avaílability of car­casses upon which hyenas could have scavenged,

Tñe HuntinglScavenging Range 01 Spotted Hyenas

Abone accumulation In a cave, to which carnívores con­tributed, can best be evaluated if we know over whatrange the carnivores transported the bones or carcasses.In spotted hyenas, hunting or scavenging range is likely tobe affected particularly by two factors: availability offood and density of the hyena population. Reference hasalready been made to the considerable distances spottedbyenas in southem Africa travel in search of food:Stevenson-Hamilton (1934) recorded that in the eastemTransvaaI lowveld spotted hyenas carried parts of goatsabout 10 km to their den, and in the Kalahari NationalPark F. C. Eloff (1964) described an incident in whichspotted byenas killed some goats near the warden's houseand were then tracked 40 km back to theír den, fromwhicb, according to their tracks, they had come the pre­vious evening.

The hunting or scavenging ranges of spotted hyenas inthe two East African areas studíed by Kruuk (1966, 1972)were found to differ a good deal. Hyeoas living Inside theNgorongoro crater were found to belong to eight clans,each containing 10 to 100 individuals; each clan had itsown defined area, in which most of the feeding by thoseindividuals took place. On the adjoining Serengeti plainshowever. the situation was different. Here sorne of thehyenas had the same kind of clan system, but others weremigratory, following the wildebeest during their annualmovements. A third group were termed the •'cornrnuters";they retained the same dens on the plains througbout theyear, from which they made long excursions, perhaps

Food Remains of Camivores in African Caves 69

lasting several days, to where the wildebeest happened tobe. Such excursions might involve a distance of 80 kmeach way.

le is difficult to judge what system would have beenfollowed by spotted hyenas living in the Sterkfontein val­ley during lower Pleistocene times. The habitat is knownto have been fairly open, and a Serengeti type of ar­rangement could well have operated.

Sorne Notes on the Mechanics 01 Crocuta ChewingAction and Resulting Darnage fo Bones

Spotted hyenas are remarkable among camivores fortheir exceptional ability to crack· bones with theirteeth-an ability that allows them to benefit from remainsother camivores have discarded. In this bone-crushingadaptation the premolars, especiaJly p3 and P", togetherwith the jaw muscles that power them, are of particularsignificance. The anterior and posterior ~of thesepremolars have been reduced and the central cusps en­larged and widened, so that the teeth have been convertedfrom bladelike stmctures into heavy conical hammers(Ewer 1973).

Powerful closure of the jaws is made possible by threesets of muscles; the action of two of the sets is shown infigure 60. The temporalis muscle arises from the lateralsurface of the braincase and from the ligament behind theeye, whence its fibers run down to insert on the upper partof the coronoid process. lts anterior fibers (TI) pull di­rectly upward, while the more posterior ones (T2 and T3)pul! upward and backward. The fact that the coronoidprocess is situated lateral to the braincase means that thetemporaJis fibers also tend to pull the jaw mediad.

Fibers of the masseter muscle originate along the lowerborder of the zygoma, nmning down to insert on the angleof the mandíble and ín the masseteríc fossa. Those fibersarising from the anterior part of the zygoma (MI) are themost superficial ones and wrap around the angular pro-

Pig. 60. Diagrams of a spotted hyena skull showing sorne of the musclesinvolved in closing the jaw. TI-3 are the anterior, middle, and posteriorfibers of the temporalis muscle, and MI-3 are the superficial, ínter·mediate. and deep fibers of the masseter. The lower jaw is shown at rest(a), half open (b), and fuUy open (e): the percentage changes in length ofeach muscJe component are indicated for the different jaw openings.

cess of the mandible. They tend to pul! the jaw forwardand upward. Fibers of the intermediate masseter layerarise from the middle of the zygomatic arch (M2) and ronstraight down to insert on the outer surface of the angularprocess; their action is a simple c10sing one. The deepestmasseter fibers arise from the posterior part of thezygoma CM3) and nln forward to insert in the massetericfossa. They pull the mandible backward and Ilpward. Inoppositíon to the temporalis musc1e, the masseter exerts alateral pull on the mandible, since the zygoma is lateraUyplaced. The zygomaticomandiblllaris mllscle is partícu­larly effective in this respect but is not easily separatedfrom the deep layers of the masseter (Ewer 1973).

A third set of muscles involved in jaw c1osure, bllt notshown in figure 60, is the pterygoideus, which originate onthe side of the skull below the orbit and nln back to theinner surfaces of the mandible. The pull is anteriad andmediad.

übservatíons I have been able to make on spottedhyenas in the Kruger National Park, and on remains theyleft there, have made it clear to me that the chewingacüon on bones falls into two distinct categories: crackingbones between the premolars, and gnawing Ilncrackablepieces with the incisors and canines.

Bone Cracking wilh Prernolars. As is shown in figure 61,the main teeth involved are pa and P3- 4, the former actingas a hammer, the latter as a cradle or anviL The mosteffective bone-cracking appears to occur when the jawsare abollt half open, that is, when there is a gap of about 4cm between the crowns of pa and Pa, as shown in figure6Ob. With the aid of a model based on the skull of an adultCrocuta. I have been able to measure the percentagechange in length of each of the muscle components whenthe jaws are half open (fig. 6Ob) and fully open (fig. 60c).

Pig. 61. (a) Premolars ofaspolted hyena. which are ofspecial significanceas bone·crackers. (b) Bone tlakes produced from bovid limb bones byspotted hyenas with the help oC their premolars.

70 A Guide to the Interpretation of Bone Accumulations in African Caves

The lengths of the relevant mllsele fibers and the percent­age change in length per component are given in table 39.

1 do not know the relationship between the percentagechange in musele fiber-length and the effectiveness ofthatparticular muscle component. BlIt it is clear that al! mlls­ele components, with the exception of M3, show apprecia­ble changes in length when the jaw is moved from theelosed to the half·open position.

With thejaws fully open, as shown in figure 60c, lengthchanges on all components of the temporalis mllsde areconsiderable; they are appreciably greater than masseterlength changes. It seems that the combination of mllsc!esCrocuta has allows powerflll jaw closllre, to a markeddegree, over a wide range ofjaw openings. A good deal offorce elearly is exerted on the premolars when crackinglimb bone shafts, which may be up to 6 cm in diameter.The typical result of sllch shaft-cracking is the productionof bone flakes, as shown in figure 6tb. These are verysimilar in form to those produced by hominid-wieldedhammerstones. A comparison of hyena- and human-

produced bone flakes is given in chapter 7, together with adiscllssion of [he form of the fractures that bone shafts areJikely to suffer.

The form of other premolar-shattered bone pieces wil!depend on the anatomy of the parts concemed. For in­stance, the blades of scapulae are Iikely to be severelyshattered, as are the ascending rami of mandibles. Dam­age to horizontal rami is consistently characteristic, as isshown in figure 62. The lower margins of the mandiblesare typically broken away by a series of premolarcnlnches, exposing the tooth roots and the marrowaround them. Very similar damage has been describedfor hyena-chewed bones from Ethiopia by Shipman andPhillips (1976).

Gnawing 01 Eones J1!ith (he /ncisors and Canines. Manyof the bones in which spotted hyenas interest themselvesare too large or cumbersome to be cracked between thepremolars. They are therefore worked npon by the in­cisors and canines (fig. 63a), often at wide jaw openings.

Fig. 62. A series of artiodaclyl mandibles from the Satara area of lhe Kruger NationaJ Park showing characteristicdamage caused by lhe premolars of spotted hyenas.

72 A Guide to lhe Interpretation of Bone Accumulations in African Caves

The appearance of such gnawed bones is very differentfrom that of premolar-shattered ones. The bone will typi­cally have been reduced through the systematic removalof small pieces that have been either scooped or chippedout. This effect may be seen on the proximal end of agiraffe hUmeI1ls (fig. 63h) that KI1lger Park spotted hyenaswere unable to crack, but which they systematicallyg'nawed away.

The products of this process are a characteristicallygnawed piece of humera! head (fig. 63e) and a humerosshowing a ragged margin at its proximal end (fig. 63d).

Gnawing damage on ungulate craniums also tends to becharacteristic. Medium-sized antelope skulls, for in­stance, are too large to be taken between the premolars.but horn-cores and nasals are frequently gnawed away(e.g., fig. 64a). A more advanced product of gnawing is a"skull bowl" such as that shown in figure 64h, where onlythe frontals, horn-core bases, and parts of the parietalsare left. The gnawed edge of the bowl is ragged andirregular.

Extensive gnawing of antelope craniums also tends to

free the maxillae entirely from the rest of the skull, asshown in fig. 64<:.

Vario\1s skeletal parts, such as vertebrae of Jargemammals, may show damage caused both by the pre­molars and by the incisors. Neura! spines and transverseprocesses are typically cracked off with the premolars,and the vertebral bodies are extensively gnawed.

In his paper "Spolted Hyaena: Crosher, Gnawer, Di­gester and Collector of Bones," Sutchffe (1970) describedfour types of bone damage he observed to be caused byCroeuta in East Africa and compared them with the dam­age reported by Zapfe (1939) when bones were fed to aspotted hyena at the Schonbrunn Zoo in Vienna.

The types were:1. Splintering of bones by adult animals.2. Gnawing of bones by juvenile animals.3. Scooping out of cancellous bone.4. Damage by partial digestion.

Concerning type 2, Sutcliffe concluded that younghyenas with milk teeth were unable to splinter bones asadults do but simply gnawed them, leaving striations at

e

Fig. 64. Examples el. skulls gnawed by spotted hyenas al a breeding den in the southern Kalahari: (a) Cranium oCajuvenile gemsbok showing charaCleristic damage lo the nasals and horncores. (b) A gemsbok "skuU bow"" (e) Amaxilla isolated by gnawíng.

right angles to the long axes of the bones. Such markswould be indistinguishable from those caused by othersmall camivores but are very different from the effects,classified under types l. 3. and 4, characteristic of adultspotted hyenas.

Sutcliffe's first category would certainly be equivalentto rny "premolar-cracked" class, and his third could beequated with incisor-canine gnawing. Damage to re­gurgitated bone pieces by partial digestion has alreadybeen touched upon.

The Brown Hyena, Hyaena brunnea Thunberg

As will be discussed in chapter 8, the earliest recordedrepresentative of the Hyaena brunnea stock appears lohave been the mid-Pliocene form, H. pyrenaica, fromEurope, which entered Africa from the north and, in thesouthern part of the continent, replaced the stripedhyena, H. hyaena. The latter species has not been re­corded in southem Africa since the accumulation timeof Makapansgat Limeworks. The distribulion of brownhyenas as represented in the last decade is shown in fig­ure 65, although within historie limes brown hyenas stillranged throughout the Cape, to the shores of Table Bay(Skinner 1976).

The brown hyena is typically rather smaller thanCrocuta. though its jaws may be as powerful. It ischaracterized by a sloping back, shaggy hair, andcharacteristic banding on the front legs (fig. 66).

Brown Hyena Lairs in Burrows

In the Kalahari National Park and adjacent areas, Milisand Milis (1977) have shown that brown hyenas regularlyuse ant-bear holes as breeding denso These may be en­larged and modified through digging by the hyenas them­selves. At the "Botswana den," three series of disusedant-bear boles wilbin an arca of I km' were modified andused; at Kaspersdraai, seven ant-bear burrows in an areaof 2 km' were used successively by the same litter ofcubs: the Kannaguass den consisted of a single burrow; atRooíkop. five ant-bear burrows were used within an areaof I km' and at Kwang two ant-bear burrows 200 m apartwere used.

On the banks of the Sabi River, near Skukuza in theKruger Nalional Park, Stevenson-Hamilton (1947) found

Fig. 65. Current distribution of the brown hyena, Hyaena brunnea,

Food Remains of Carnivores in African Caves 73

a brown hyena den consisting of a series of ant-bear boleswith about six entrances, "covering sorne thirty feet ofopen space." The síte was screened by thick bush andreeds and was surrounded by an appreciable bone ac­cumulation that will be discussed shortly.

During his study of brown hyenas in the Transvaal,Skínner (1976) reported 00 two deos in modified ant-bearholes; these were on the farms Leeufontein and Tweeput­koppies. He provided a plan and sections of theLeeufontein den. which had three entrances; the internaldimensions of the tunnels varied in height from 25 to 35cm and in width from 35 to 53 cm. Brown hyenas wereseen entering the tunnels, and Skinner suggests that therather small dimensions helped keep larger predatorsaway from the cubs.

Brown Hyena Lairs in Caves and Recesses

During 1955 a cave was pointed out to me by the lateJ. Templeton on the farm Uitkomst, north of the Sterk­fontein valley; it was used as a breeding lair by brownhyenas and opened into the side ofa steep dolomitíc valley(fig. 67a). Inside, io almost total darkness, were numerousbones among which were scattered 395 whitened hyenadroppings. Sorne of the bones and droppings are shown infigure 67b. According to Templeton. the cave had beenused as a brown hyena breeding site for a number ofsuccessive years before 1955.

Another low cave used as a breeding lair occurs 00 thefarm Boekenhoutkloof, near Cullinan. northeast of Pre­toria. The site and resident hyenas have been describedby Skinner (1976), and I first visited the place on 6 Sep­tember 1975 in Skinner's company. I subsequently re­turned and surveyed the site, drawing up the plan shownin figure 68.

As is shown in figure 69, the site is close to the top of awooded ridge of Waterberg sandstone, facing east. It con­sists of a low cliff, about 5 m high, forming a comer on thesouth side; severa11arge sandstone blocks are scattered infront, as is shown in the plan. The lair consists of a lowrecess, generally about 0.5 m hígh, beneath the sandstonecliff, descending at an angle, with a substrate of fine,loose dust. At the time ofthe first visit, spoorofadult andjuvenile brown hyenas was visible both outside and inside

Fig. 66. A mounted specimen of a brown hyena from the Transvaal. Thesloping back, shaggy coar, and banding on the foreJimbs are dearlyshown.

74 A Guide to the Interpretation of Bone Accumulations in African Caves

the lair. On either side of the entrance to the lair wereconcentrations of hyena droppings: 316 were collected.

It is c1ear that brown hyenas, Iike spotted hyenas, usebreeding lairs of two types: either modified ant-bear bur­rows or natural caves and recesses.

Use 01 Brown Hyena Lairs by O/her Species

It appears that. like CroclIfa. brown hyenas share theirdens with other animals, which they appear to tolerate.The den figllred and excavated by Skinner (see aboye)was found to house a porcupine as well as brown hyenas,and "Wright's den" described by Mills and Milis (1977)from the Kalahari National Park was used, perhaps inrotation, by spotted and brown hyenas, as well as byporcupines.

The Composi/ion 01Bone Accumulalions al BrownHyena Dens

At the time of my first visit to the Boekenhoutkloof lair, apartly eaten Iynx, Felis caraca', was Iying about 12 mfrom the cave entrance, the position being indicated infigure 68. The Iynx is shown in figure 70: its forequarterswere intact, but the abdomen, inc111ding the vertebral col­nmn between the thoracic regíon and the sacrum, wasmissing. The pelvis showed carnivore damage, and thehind legs, except for the head of the right femur. wereabsent.

Fig. 67. (a) Entrance to a dolomilic cave used by brown hyenas as abreeding lair on the farm Uilkomsl. (b) Bones and hyena droppings on thefioor of lhe inner chamber in lolal darkness.

A short while before this observatíon was made, thelocal farmer fOllnd the hindquarters of a large domesticdog slightly closer to the entrance of the lair. He retumeda few days later to find that these remains had beenfurther eaten, so that only fragments of the hind limbsremained. The only other bones fOllnd sllbseqllently atthe lair were the head, shoulders, and thorax of a do­mestic calf, although a number of other species wereidentified by means of hair and feathers (Skinner 1976).

In the course of his brown hyena study in the northernTransvaal. Skinner collected 39 bone pieces from the vi­cinity of five dens on the farm Tweeputkoppies. Thesewere found to come from 15 individual mammals be­longing to seven species as Iisted in table 40A. From thebreeding lair on the farm Leeufontein 6 bone pieces werecollected; they carne from 5 individuals of four species aslisted in table 40B.

The bone aCCllmlllation from the brown hyena den de­scribed by Stevenson-Hamilton from the banks of theSabi River in the eastern Transvaal is of great interest.Concerning it, Stevenson-Hamilton wrote as follows(1947. p. 210):

The vicinity was Iittered with bones. and it was quiteevident that most jf not aH the animals had been seizedalive and killed by these hyaenas. It was to me rathera remarkable discovery, since hyaenas, as a tribe,have never been regarded as primarily hllnters. In thepresent instance there was no doubt about the matter.The heads of fourteen fllll-grown impala rams, aH quiterecently killed, the skulls of several baboons, and oftwo chitas (one of them a full-grown animal), remainsof guinea fowls, and a large tree snake ("boomslang")partly chewed, were among the exhibits. The carcas­ses of the animals had seemingly been dragged down "the holes, inc1uding any homless females, for onlyheads offully grown horned males were fOllnd outside;the prevailing odour was fairly good proof of what hadhappened. These hyaenas must have developed asound hunting technique to be able to catch and kili

. SAMDSIOME BlDCKS

• CAVE

~'\~~ H'iAEHA DROPflNGS

..... SlDPE

Fig. 68. Sketch plan o[ a brown hyena breedíng lair on ,lhe farmBoekenhoulkloof in lhe CuUinan dislricl of Ihe Transvaal.

Food Remains of Carnivores in African Caves 75

" ."- ","':.

.' .,.'. - .. " ~

./} ':~:-~;;;I~~~;~~';Fig. 69. The Boekenhout brown hyena lair; ¡he entnmce is indicaled by the arrow al lhe base of lhe sandstone c1iff.

Fig. 70. The partly consumed body of a caraca! dose to the entrance of lhe Boekenhoulkloof brown hyena Jair.

76 A Guide to the Interpretation of Bone Accumulations in African Caves

so wary and quick an animal as an impala, and one sorelatively fonnidable as a chita, or a baboon.

These observations were made in 1941, butStevenson-Hamilton gave further details of the site in hisannual reports to the National Parks Board for 1942 and1943. These have been made available by Pienaar (1969,p. 139):

1942-The family of brown hyaenas again bred inthe north bank of the Sabi a few miles east ofSkukuza, but changed their breeding holes for othersnearer the Sand River junction, their old home havingbeen to a great extent destroyed by some Iions whichhad dug into it extensively to reach the meat whichthey cOllld smell inside. Established at their new site,these hyaenas continued to kill considerable numbersof animals, and skulls of impala, bushbuck and ba­boons were found scattered olltside, evidently huntedand killed by the occllpants of the den.

1943--The family of brown hyaenas on the northbank ofthe Sabi continued in their new abode, to taketoll of other animals. In addition to the remains offuIl-grown impala, several baboons and one morecheetah skull was fOllnd.

Unfortllnately, Stevenson-Hamilton did not provide adetailed list of the bones found at the Sabi River dens, butthe following is a summary of animals recorded: 1941,14 adult impala rams, severa! baboons, 2 cheetahs (Iadult, 1 subadult), several guinea fowls, water tortoises,and a tree snake; 1942, several impalas, baboons, and abushbllck; 1943, impalas, several baboons, and 1 cheetah.

The Uitkomst cave, IIsed as a brown hyena breeding

lair (fig. 67a,b), which 1 investigated in 1955, was foundto contain, besides 395 hyena droppings, 90 identifiablebones, ofwhich 60 belonged to hares and hyraxes, as wellas 128 bone f1akes. 1 recently reexarnined this collectionand found that, though many of the bones showed c1earcamivore damage, others had been gnawed by porcupinesor charred in fires. The cave has obviously been used byporcupines and Iron Age people in addition to brownhyenas, and aH three agents will have contributed to thebone accumulation. For our present pllrposes, therefore,the composition of this bone accumulation is of limitedrelevance.

Also on the farm Uitkomst, the brown hyenas made useofa spot between sorne large chert boulders in the dry bedof the "python valley" as a resting and feeding place.When I made a routine visit in this place on 6 June 1969, 1found that the hyenas had brought back the right hind legof a donkey, still articulated to the innominate bone. AHmeat had been eaten from the leg, and damage had beendone to the pelvic bone, the greater trochanter, and thecalcaneus. In addition, there was the complete palate of acalf and the ríght mandibular ramus of a blesbok, showingdamage that appears very characteristic of hyena action(see fig. 76). These were the only bones found at the spot,although hair and droppings have been regularly seenthere. The hind leg ofthe donkey had been brought by thehyenas from beyond the boundaries of the Uitkomst re­serve, probably over a distance of 5 to 7 km.

The most detailed and significant information on bonesaround brown hyena dens comes from the Kalahari Na­tional Park, where M. G. L. Milis has been conducting hisstudy of H. brunnea (Milis 1973, 1974, 1976; Milis andMilis 1977). Milis has kindly made his collection of bonesfound around the Kalahari brown hyena dens (fig. 71)

Fig. 71. Bones accumulated by brown hyenas around the enlrance lo lheir breeding lair, Bolswana den, KaJahariNalionaJ Park. Pholo by M. G. L. MiUs. fmm Milis and Milis 1977.

available to me. The sample consists of 235 pieces fromthe following sites: Botswana dens (53 píeces), Kas­persdraai dens (14 pieces). Kannaguass den (66 pieces),Rooikop dens (22 pieces). and Kwang deos (80 pieces}.Panículars of the dens have already been discussed andare given by Milis and Milis (1977).

The 235 bones are found to come from a minimum of 75individual animals, belonging to sixteen taxa as depictedin figure 72. The bones by which each taxon is repre­sented are listed in table 41.

The most remarkable feature of thís bone accumulationis that carnivores are represented in exceptional abun­dance. Of the 75 individual animals whose remains arefound, 35, or 46.7% of the total. are other camivores,black-backed jackals and bat-eared foxes being the moslnurnerous, followed by lynx. ratel, and aardwolf. As willbe discussed shortly, these animals are frequently repre­sented by surprisingly complete skulls. 11 is now wellestablished that brown hyenas, at least when rearingcubs, deJiberately hunt other carnivores (up to the size ofcheetahs, on the Kruger Park evidence) and take them torhe breeding dens as food for Ihe cubs, Although mosl ofthe skeletons of these prey items may be eaten, skulls aretypically left. This camivore-hunting habit of brownhyenas has. 1 suspect, been of the greatest importance inthe accumulation of bones in southem African caves.

Of the 3 animals represented by bony remains at theBoekenhoutkloof lair, discussed aboye, 2 were carni­vores: the proportíon is appreciably lower in the 45bones collected by Skinner at the Tweeputkoppies andLeeufontein lairs in the northem Transvaal. but this mayreftect tbe fact that small carnivores are locally scarcethere, or that brown hyena cubs were not actually beingreared in the dens at the time the collections were made.

An upper Pleistocene bone accumulation containingabundant carnivore remains has been described by Klein(l975a) from the southwestem Cape. It ocurs in an ir­regular fissure hollowed out in calcareous aeolianítesforming a sleep eliff flanking False Bay at Swartklip, 30km east-southeast of Cape Town. The site shows no evi­dence of porcupine or human involvernent, and Klein hassuggested that the bones were collected by carnivores,probably hyenas. Since the collection ineludes remains of2 adult and 2 juvenile H. brunnea, it seems conceivablethat Swartklip was a brown hyena breeding lair. Bonesfrom a total of 32 species have been recorded at the site;of these, 13, or 40.6%, are of other camivores.

Other bone-filled burrows of upper Pleislocene agehave been described from Duinefontein, 50 km northof Cape Town (Klein 1976b). These are stilI being in­vestigated but probably result from hyena action.

Diet 01H. brunnea as a Guide to the Potetuial BoneContent 01Lairs

In addition lo the species represented by the bonesalready listed from the various laírs, further infonnation isavailable from analysis of hair in scats and regurgitations(Skinner 1976) and from direcl observalions in lhe KrugerNational Park. Idenlilied from lhe hair analysis were lbefollowing species besides those already listed: Protelescristatus, aardwoJf; Suricata suricatta, suricate; Her·pestes sanguíneus, slender mongoose; Lepus saxatilis,scmb hare; and Pronolagus crassícaudatus, red rackhare.

For the Kmger National Park. Pienaar (1969) gives the

Food Remains of Camivores in African Caves Tl

following list oC 239animals found killed by brown hyenasduring the periods 1936-46 and 1954--66: kudu 78. impala49. waterbuck 44, zebra 8, baboon 7. guinea fowl Z, watertortoise 6, wildebeest 4, sable 3. tsessebe 3, buffalo 2,reedbuck 2. warthog 1, roan antelope l. eland 1, duiker 1.bushbuck 1, ant bear 1, and tree snake 1. In additionPienaar states that brown hyenas prey 00 lion cubs amiostriches.

From these observations it is clear that brown hyenasare able lo hunt a wide variety of animals, including theJargest antelopes. Wheo rearing cubs they also hunt thefull range of carnivores that they are able to overpower,up lo the size of cheetabs.

Cannibalism appears to occur among brown hyenas, asít does inCrocuta. In his list offood items eaten by brownhyena cubs in the Kalabati National Park, Milis (1974)notes cannibalism 00 two occasions, without specifyingwhether the animals eaten were other cubs or adults.

I have rnade one personal observation 00 brown hyenaseating a dead Crocuta close to the Nossob rest carnp inthe Kalabari Park. On 16 August 1975. in the company ofM. G. L. Milis, I examined the remains of an adult spot­ted hyena (fig. 73) that had died as a result of porcupinequill injuries and had been partly consumed by brownhyenas. On the other hand, Milis (1974) has recorded twoincidents in which spotted hyenas ate brown hyenas. Inthe fírst, a 14-month-old brown hyena was killed near awater hale; in the second, an adult that was recoveringfrom the effects of an irnmobilizing drug was killed by sixspotted hyenas.

Brown Hyena Predation on Primates

1 am not aware of any authenticated reports of brownhyena attacks on people , although Stevenson-Harnilton(1947) quoted Africans in Gazaland as saying tbat brownhyenas sometímes took small children from their huts atn~ht or attacked sleeping adults. -~--.. __,,--Predalion on baboons appears lo mí" fairl,,",common. )

there being evídence of it both in th~-~r National---­Park (7 baboon skulls were among the remains al the SabiRiver breeding den) and al the northem TransvaaI densinvestigated by Skinner. Baboon{Ñe probably one oflbecharacteristic prey species brought back to lairs whencubs are being:f.il~.,

I was able to arrange one feeding experiment involvinga baboon on the farm Uttkomst. An adult male baboonweíghing 25 kg was shol and securely tied lo a tree, eloselo the brown hyena feeding place in Pylbon valley re­ferred lo earlier. It was ealen lwo days later on 30 May1969 by two brown hyenas, and the remains were thencollected. They consisted of 2 pieces of mandible and 3fragmenls of the cranial vault (lig. 74). The maxillarytoothrows were swallowed by one of the hyenas.

Relatíve lmportance 01Hunting and Scavenging

It is well established that brown hyenas both hunt andscavenge, bul little has been published thus far on therelative importance of each food-getting means. The re­search currently being undertaken by M. G. L. Milis willprovide valuable infonnatioo of this kind.

Where the brown hyena's range extends to the coast.the animal has become known as the ··strandwolf." fromíts habit of scavenging from dead marine animals washedIIp 00 the beach. It is worth remembering that bones of

78 A Guide to the Interpretation of Bone Accumulations in African Caves

marine vertebrates in coastal caves could very well havebeen brought there by brown hyenas,

The Hunting or Scavenging Range 01 Brown Hyenas

With ..the"aid"llf"~le~, Milis (e..g .. , 1974) isestablishing the movement pattems oí H. brunnea in theKalahari National Park .. He Fe!lIJt'W--lIllIt'll''ÍOl 111" MM'"lance of 50 km 'pe"'lÚMbt is __eeptiorlilpf", 'íI "'IU'"~-'~:Such trips are generally undertaken singly, andwhen fhe cubs first start to forage they do so on their own.

Distances traveled in the course of a night will certainlybe influenced by the availability offood and will thereforevary greatly from región to región. The home range of

340 to 544 km' estimated by Milis (1976) for the KalahariPark brown hyenas is a good deaJ larger than the rangeSkinner (1976) estimated for six adult and two juvenilebrown hyenas living within an area of 20 km! at Boeken­houtkloof in the Transvaal.

It would probably be reasonable to suggest that brownhyenas might carry bones back to a lair over a distance ofat least 10 km ..

Food Storage hy Brown Hyenas

The storage of surplus food in water practiced by spottedhyenas has not been observed in H. brunnea, lnstead,brown hyenas simply hide food in long grass or take it

BLACKBACKEO JACKAl 13 edult8

1 Juvenlle

13 lndiv¡duea.

11 kKIlvlduals

BATEARED FOX 13 adultsSPAINGBOK 10 8dults

3 juveniles

STEENBOK 9 adulh 2 Juvenil••

5 Indh,ldu8'.

lYNX 4 HU". 1 JuvenI~GEMSBOK 2 edult8 3 juvenil.

2 indlvlduals

WllDEBEEST 2 adults

DUIKEA adult

Juvenlle

AATEl 2 Hult. :rOSTAICH 2 eduits 1 899

1 Indlvldu81

AAADWOLF 1 edult

~DOMESTIC CALF

1 juvenil.

~~SHEEP or GOAT HAR'l'1:BEEST

1 .dult 1 adutt

ANTBEAR 1 Mili"

Fig. 72. AnimaJs represented by remains coUected by brown hyenas at their breedinl dens in tbe KaJahari NationaJPark. Mirumum numbers oC individuals are indicated.

Food Remains of Carnivores in African Caves 79

Hg. 73. ReTruÚns of a sponed hyena partly consumed by brown hyenas nearNossob Camp, Kalahari NlUional Park.

down a hole (MilIs 1973, 1974). One attempt by a brownhyena to bury surpllls food at Boekenhoutkloof has beenreported by Skinner (1976).

Sorne Characteristícs 01Brown Hyena Food Rernaíns

The skull structure and dentition of H. brunnea are ba­sically similar to those of Crocula, although there areadaptatíonal differences. According to Ewer (1973),Hyaena dentitions are characterized by less efficient car­nassial shear, so that P4 and the anterior part of p4 are alsomodified for bone-crushing. In fact, third molars ofHyaena are less dominant than they are in Crocuta.

Bony food remains of the two species of hyena appearremarkably similar, except for those of soft-bodied preyanimals, which brown hyenas take to their dens when

Fig. 74. Remains of an adult male baboon consumed by brown hyenas onthe farm Uitkomst, Transvaal.

rearing cubs. These are highly diagnostico At brownhyena feeding places, chewed bones indicate that H.hrunnea follows the Crocula pattem of cracking manage­able bones wilh its premolars and gnalving larger oneswith ils incisors and canines. Hyaena-cracked limb bonesand mandible pieces are shown in figure 75. The damageis indistinguishable from that inflicted by spotted hyenas,as portrayed in figures 61 and 62. The characteristics ofgnalVed bones are probably also very comparable. Figure76a shows a calf palate, representing the gnawed remnantof a complete craníum, and figure 76h shows a "skullbowl" that resulted from brown hyena gnawing of aspringbok skllll.

An important difference between H. brunnea andCrocUla, from the point of view of this discussion, is thatthe~r actively kills small animals, particularly othercamívores, and brings their bodies to the cub~. The moststrikíng aspect ofthe camivore remaíns found around theKalahari brown hyena breeding lairs is the undamagedstate of many of the craniums. On the basís of skuUs, thefollowíng numbers of individual carnivores are repre­sented by the collection: black-backed jackal, 13 adults,1 juvenile; bat-eared fox, 13 adults; caracal, 4 adults,1 juvenile; ratel, 2 adults; aardwolf, 1 adult. Yet, despítethe abundant craniums of these animals, postcranial bonesare almost unrepresented. For instance, the black-backedjackals are represented by a single proximal humeralpiece; bat-eared foxes by a single articulated forefootand most of a vertebral column; and the only othercamivore remains are 5 isolated vertebrae, a piece ofinnominate, and a tibia. It is obvious that the hyenashave consístently eaten the bodies of the camívores,leaving only the heads, or parts of them. This is sur­prising, since a single crunch of an adult brown hyena's

80 A Guide to the Interpretation of Bone Accumulations in Mrican Caves

Fig. 75. Examples of bones crac:ked by the premolars of brown hyenas. The specimens are from the KalahariNational Park and from Uitkomsl, Transvaal.

jaws would shatter the braincase of a jackal, fox, orcaracal, making the brain írnmediately accessible. Yetthe adult hyenas have clearly not done thís. 1 can onlysuggest that cub-rearing brown hyenas have an inhíbitionagainst feeding on the animal they bring as food for the

Fig. 76. Examples of bones gnawed by brown hyenas: ¡he palate of a calffrom Uilkomst, Transvaal, and a springbok "skuU bowl" from theKalahari Nalional Park.

cubs and that the cubs are unable, or disinclined, to breakup the skulls.

The parts by whích each species of camivore is repre­sented are listed in table 41, but it can also be said that, ofthe black-backed jackal skulls, 6 had undamaged brain­cases (see fig. 77); 11 of the bat-eared fox calvariae werecomplete, and all the broíncases of the caracal and ratelskulls were undamaged. Mandibles were typically presentand reasonably undamaged, except for the bat-earedfoxes, where all but one of the mandibles had dis­appeared. It appears that the brown hyena cubs were ableto chew up and sWallow the rather weak fox mandiblesbut typically rejected those of the other carnivores.

1 suspect that cub-rearing brown hyenas purposelyselect small camivores and, ín sorne areas, primates asfood for young cubs, sínce their skeletons are morecronchable than those of bovids of the same size (seechapo 2). When the cubs are larger they seem to be fedlargely on antelopes that the hyenas hunt or scavenge. Bythis time the cubs are probably able to cope with the morerobust bovid bones. The easily ehewed eamivore andprimate bones would provide the cubs with a readily ac­cessible source of calcium, but why brown hyenas shouldfeed their young in tbis way while spotted hyenas do not,1 am unabIe to explain.

The brown hyenas' habit of bringing the bodies of car­nivores (and baboons, on Kruger Park evidence) to theirlaírs ano of often leaving the heads of these prey animalsundamaged is of great signifi~ance to our understanding ofbone accumulations in caves.Brown hyenas undoubtedlyuse dolomitic caves as breeding lairs and could be re­sponsible for extensive bone accumulations mad~ overmany years.

Food Remains of Camivores in African Caves 81

Fí¡:. 77. SkuL1s of smaIJ camivores coL1ecled around brown hyena breeding lairs in lhe Kalahari NalÍonaJ Park. Thecamivore bodies served as food for lhe hyena cubs: (a) black.backedjackaJ;(b) bal·eared fox;(c) caracal; (d) mIel.

The Striped Hyena, Hyaena hyaena Linnaeus

As will be mentioned in chapter 8, it appears thatHyaenahyaena was derived from the late Pliocene fonn H. ab­ronia. known from Langebaanweg in the sOllthwestemCape. The Iineage was a rather conservative one andlInderwent little change in the COlme of the Pleistocene,the fossil form H. h. makapanensis from MakapansgatLimeworks, more than three million years old, being onlysubspecifically different from the living striped hyena.

As is shown in figure 78, the striped hyena is cUITentlydistributed in north and northeast Africa, overlapping inrange extensiveIy with the spotted hyena, but nowherewith the brown hyena. Outside Africa, striped hyenas arefOllnd from the Gulf of Bengal westward throughoutsouthem Asia and the Middle East and then as far northas southem Siberia and the Caucaslls (Kruuk 1976).

The striped hyena is appreciably smaller than Crocutabut has a large head and long, pointed ears; it has a prom­inent 'erectile mane of long hair, and its general colorvaries from pale brown to white, with sharply definedstripes on the body and transverse bamng on the legs. Itis generally solitary and nocturnal.

1 have made no personal observations on stripedhyenas, but the animals are of significance here becausethey are known to carry bones to cave lairs. 1 will there­fore present a brief review of sorne relevant infonnationfrom the literature.

Slriped Hyena Lairs and Breeding Siles

Referring to the striped hyena in India, Pocock (1941)remarked; "By day it lies up in enlarged porcupine·bUITows, caves, or in crevices ltnder boulders." In East

Africa, Kmuk (1976) has photographed a den under over­hanging boulders on a rocky hin about 10 km northeastof Seronera in Tanzania. It therefore appears that H.hyaena. Iike H. brunnea and Crocula, makes use of eitherbUITOWS or caves and recesses as lairs and breeding sites.

There is nnfortunately not mnch information availabIeabOllt the composition of bone accumulations at stripedhyena lairs. The late L. S. B. Leakey once toId me thathehad seen striped hyena cave lairs in East Africa wellstocked with bones; but, to my knowledge, furtherinformation was not published. In search of such infor-

Fig. 78. Dislribulion of the striped hyena, Hyaeno hyaeno, wilhin hi~loric

times.

82 A Guide to the Interpretation of Bone Accumulations in African Caves

mation, Hughes (1961) paid a visit to East Africa; un­fortunately, in Uganda he was not successful in locanngstriped hyenas or their Iairs, but at Olorgesailie in Kenyahe saw a number of holes and overhangs used by H.hyaena. At the time of his visit, two sites in the area wereoccupied by porcupines, but Hughes decided to excavatea hale from which striped hyenas were seen emerging.From the front of this hole, and in the entrance, Hughesrecovered 61 bones tbat be described as "old, dry andporcupine-gnawed"; he concluded that these had beencollected by porcupines, and thus bis East African obser­vations served to confinn his conviction that hyenas donot transport bones to their lairs.

More recently, Kruuk (1976) bas Iisted 9 items near aH. hyaena den in the Serengeti; these carne from wil­debeest, kongoni, impala, Thomson's gazelle, ostrich,vulture, tortoise, and dung beetle. From regurgitated bairballs near tbe den be was able to add tbe following fooditems: zebra, lion, spotted hyena, hare, hedgehog, fruit,and domestic rubbish.

An interesting eomparison between the diets of stripedand spotted byenas bas been provided by Kruuk (1976),wbo analyzed tbe contents of 50 feces of the formerspecies and 42 of tbe latter from Ngare Nanyukí, in lbeSerengeti. Remains oí large marnmals such as zebra, wil­debeest, kongoni, and topi featured in only 26% of theH. kyaena scats. But in 74% of the Crocuta seats, arnongmedium-sized mammals sucb as gazelle and impala, tbepercentages were equivalent (68 and 71 respectively), andin the small and very small marnmal categories, repre­sentation in the H. hyaena seats far outweighed repre­sentation in the Crocuta seats (28%: 2%). Likewise,reptiles, birds, insects, and vegetable foods were eaten inlarge quantities by striped byenas but were largely ig­nored by spotted byenas. Moreover, Kruuk fcund that 20droppings of striped hyenas had a mean calcium contentof about 12% as opposed to 25% for spotted hyena scatsfrom the same area. These figures indicate that Crocutawas eating more bone, relative to other foods, than wasH. byaena.

To surnmarize Kruuk's results, it appears that the dietof striped hyenas in the Serengeti consists mainly ofsmaller vertebrates, partícularly mammals, either huntedor scavenged, in addition to substantial quantities of fruitand insects.

Striped hyenas are still found in Israel, wbere variousaspects of lbeir natural bistory bave been described bylIani (1975). Of particular relevance bere are lIani's com­mentson lbeuscofcaves byH. hyaena (1975, pp. 12-13):

At times of inactivity and wben the cubs are growingup, hyenas stay in caves. 'lbese usually have narrowentrances and widen out considerahly inside. Hyenasare good at digging and if lbe interior of tbe cave doesnot suit their requirements they do their best lo ím­prove and enlarge ít, In regions where there are nonatural caves, hyenas make their bornes in the lairs ofporcupines or badgers. It has even been claimed lbathyenas occasionally share their lairs with porcupines,but tbis may stem from a rnisunderstanding of the liv­ing babits of tbe byenas. My observations tend to in­dicate that hyenas make use of severa! caves at onetime and do not remain pennanentiy in any one lair.Nevertheless it is worthwhile to examine this theorymore closely, Hyenas do indeed share their lairs withbats and invertebrates common to dark caves and theirskins are consequently infested with ticks.

Let me describe a typicallair located sorne 3 km.south of Arad that 1 investigated in November. 1970.The entrance to the cave was on a c1iffabout 4 metersaboye a dry wadi bed and was accessible by a narrowrack ledge. The diameter ofthe opening was about 40cm. and we had to crawl on our stomachs for fourmeters befare the tunnel widened out and we foundourselves in a roomy chamber 1.40 m. high. From thecave walls five openings led to smaller cave-charnbers.In the center of tbe big room there was a pite of bro­ken bones and animal skulls. We checked sorne ofthismaterial by electric torch-light and found that arnongothers there were hundreds of dog skulls, hundreds ofskulls and other skeletal fragments or domestic ani­mals such as donkeys, camels, goats and sheep as wellas two human skulls that were apparently dug out of aBedouin cemetery in the vicinity. There were alsosevera! gazelle skulls, about ten skulls of ibex , bonesof foxes and badgers, hedgehog skins and porcupinequills. Tbe floor of tbe cave was of dry, powdery clayin which were innumerable hyena spoor. We managedwith great difficulty to crawl into sorne of the adjoiningsmaller caves and, to OUT surprise, discovered in eachone of them additional but smaller accumulations ofbones. We wondered at what seerned to be a habit ofdragging foad to remate corners of the lair and it wasonly two years later when I watched the eating habitsof older hyena eubs that I could advance a plausibleexplanation for this. Food brought by the adults is noteaten together but each cub apparently seizes a pieceof meat and attempts to eat it as far as possible fromhis siblings leaving separate pites of bones in variousplaces. Other caves that I examined were on the wholesimilar in size, layout and contents.

In front of the cave openings, one usually findshyena droppings. These are easy to identify because oftheir white color that is due to the calciurn salts thatremain from the bones eaten by the hyena. Less corn­mon, are wads of undigested hair that are exgurgitatedoutside the lair opening. These wads usually indicatethat the cave has not only been used as a shelter butthat cubs were raised in it.

More recently, J. D. Skinner undertook a study ofstriped hyenas in Israel from December 1977 lo January1978. He inspected five dens and referred to one inparticular-a maternity den near Arad in northem Negev,in the same area as the one deseribed by lIani. Skinnerwrote (n.d., p. 16):

At one particular matemity cave we crawledthrough a passage sorne ten metres long to enter alarge cavem wbere byaenas bad been feeding over avery long period in a special area. The den was oc­cupied and the fresh remains of a goat Capra hírcuswere in evidence on top of a layer of bones coveríngthe floor of the cave for 40 m'. Hundreds of skulls andbone fragments eovered the floor and two squaremetres were sampled ten metres aparto

The sarnple of bones was analyzed by S. Davis at theUniversity of Jerusalem and was found to consist of 267pieees from a minimum of 57 individual animals. Theseincluded 14 camela, 14 sheep or goats, 12 dogs, II don­keys, and single individuals of fox, cow, porcupine,gazelle, pig, and mano If one assumes Iba! tbe layer ofbones was uniformly dense over the 40 m2 of floor area,

•••one may estimate that the cave contained 5,340 precesfrom a mínimum of 1,140 individuals,

Although informalion is nol yet available on theskeletal parts present in the sampie, Skinner (n.d.) di s­counts the acttvtties oí porcupines in the collectingprocess.

H. hyaena is well represented among fossüs fromthe gray breccia al Makapansgat Limeworks, and lbeobservations made in Israel could be. lo my mind, ofthe greatest significance in the interpretation oí the Maka­pansgat fossil accumulation. This matter will be dis­cussed again in chapter 13.

The Possíble Role 01 Extinct Hyenas as Collectors 01Bones in Southern African Caves

Extinct hyenas, whose rernains have beco found in theSterkfontein valley caves, belong to four genera:Crocuta, Hyaena, Hyaeníctis, and Euryboas, Althoughseveral species oí Crocuta have been described on thebasis of Transvaal fossils (see chapo 8), these are cur­rently regarded as subspecies of Crocuta crocuta, thewell-known spotted hyena. Morphologically, Afrícancrocutas of two or tbree million years ago were verysimilar lo those of today, and there is no reason to believethat their behavior, with particular relevance to bone­collecting in caves, was any different.

Remains ofHyaena brunnea virtua1ly identical to thoseof the eontemporary brown hyena are known fromSwartkrans Members 1 and 2, and a less firm identifica­tion has been made from Kromdraai B. Remains ofa verysimilar hyena, H. bellax, have been described fromKromdraai A. Again, 1 am inclined to assume that theseanimals behaved much like their modem brown hyenacounterparts.

The situation with the other two genera, Hyaenictis andEuryboas, is very different. Together they constiture the"hunting hyenas" lo be described in chapter 8. Theywere predacious, fast-runníng animals, and representa­tives of both genera were perhaps derived from a PlioceneHyuenictis stock.

It is not easy to reconstruct the behavior of mammalswhose living representatives no longer exísr, but a fewspeculations may be ventured. Since Hyaenictis andEuryboas were true hyaenids, they very probably rearedtheir cubs in burrows or caves as the three contemporaryforms of hyena do; their dentitions and limb bones in­dicate that they kilIed their own prey, almosl certainlyrunning il down in a cheetahlike fashion, allbougb, likeother hyenas, they probably scavenged as welI; they werevery probably social, bunting in packs like spotted hyenasor wild dogs, and, finally, they very probably broughtback parts of their kilIs as food for their cubs.

Tbat remains of these hunting hyenas are found inabundance in the Sterkfontein va1ley cave brecciasstrongly suggests that Hyaenictis and Euryboas fre­quented caves, at least when they were rearing cubs, Irthis was so, their contribution to the bone accumulationsat such sites could have been considerable. An importantdifference between food remains of spotted and brownhyenas al breeding sites is that the latter inelude manyskulls of small carnivores, from prey that the brownhyena parents fed lo their cubs. 11 is not known whelberhunting hyenas fed their cubs other camivores or not, butthe faet that hunting hyenas appear to have been activepredators suggests that their food remains would consistmore predominantly ofhunted than ofscavenged compon-

Food Remains of Carnivores in African Caves 83

ents. Furthermore, those hunted components are likelyto have included bones fmm medium or large bovids andequids that were pursued and dismembered by huntinghyena packs.

Tbe Leopard, Panthera pardus Linnaeus

Leopards have a remarkably wide distribution both inAfrica and in Asia, and, aJthough they have been exter­minated from many areas, they are usually the last of theIarge predators to disappear under pressure from humanencroachment. Tbey are clearly very successful animals ,able to adapt to a variety of circumstances and enviren­ments. Furthermore, they have been present in theSterkfontein valley since Swartkrans Mernber I times-aprobable span of 1.5 million years, showing negligiblechange in form ayer the periodo

Withín the living species, however, there is a good dealoí variability in size. body conformation, and color pat­temo In his taxonomic review of the leopards of África,R. 1. Pocock (1932) regarded sevenleén subspecies asvalid and considered that these could be placed in fivegroups linked to the environment in which they occur: the"savanna or veldt type," the "desert type," the "rnoun­tain type," the "forest type," and various "dwarfforms," perhaps adversely atTected by their harsh sur­roundings. Only thirteen of the former subspecies wereadmitted by Smithers (l971b) in the most recenttaxonomic evaluation; he remarked that there seems to bea tendency for the unusual specimens to find their wayinto the hands of taxonomists rather than representativesamples of the populations concemed.

Leopards are particularly relevant to this study in thatthey are secretive predators, making use of caves as re­treats, feeding places, and breeding lairs. In addition.they nave the habit of storing food in trees, which. in thecase of the Sterkfontein valley caves. may have had spe­ciallocal significance from the taphonomic point of view.Although frequent passing reference is made in the liter­ature to the use of caves by leopards, there are very fewspecific descriptions of caves that have served as habituallairs or feeding places. In the course of his study of cavesin the Easl Mrican Rift Yalley, Sutcliñe (l973a) found

~.. .'.:K..v)' .. ~¿..Ji7." .Fig. 19. Distribution oC the leopard, Panthera pardus, within historietimes.

84 A Guide to the Interpretation of Bone Accumulations in African Caves

evidence of leopard occupatíon in two caves on MountElgon. The Kitum Cave contained a scatter of brokenbones that appeared to be food remains of a residentleopardo and, a kilometer away , the Makingnen Cave wasof special interest for this. The cave opens at the head of asmall valley, behind a crescent-shaped cliff fall passableonly by way of an elephant path between the rockfall andone of the cave waJls. The path is extremely steep, butdroppings along it and within the cave show thatelephants are able to negotiate it. It may seem strange thatelephants should wish to enter a cave, but they do. in fact,regularly visit sorne of the Mount Elgon cavems in searchof salt, mainly mirabilite, which crystalizes on the cavewalls. Various other animal s are similarLy attracted, and itis on these that the leopards appear to prey,

In the case of the Makingnen Cave, Sutclíffe IocatedSOrne habitual 1eopard feeding places on the right-handside of the cave entrance, while scattered behind the en­trance rockfall in the north charnber were numerous bro­ken bones oC antefopes, monkeys, and forest hogs-­apparently food remains of the leopard. Unfortunately,Sutcliffe did not fully complete his investigatíon: "Thelast intended stage of the investigation of this cave was tohave been a photographic session of the north chamberwhere the leopard eating place was situated. On this oc­casion Una and I were alone at the cave mouth, the lightwas beginning to fail (it gets very dark very quickly inEast África) and an animal of unknown identity made anoise at us from the back oí the cave. We confess totaking fright, abandoning the photographic session andconcluding the investigation prematurely!" (Sutcliffe1973a, p. 56).

Dr. and Mrs. Sutcliffe have my sympathy. I have sev­eral times experienced the apprehension of entering acave knowing a leopard was inside: perhaps ape men feltthe sarne apprehension on nights long ago when leopardsprowled around their sleeping sites.

Use 01 (he Mount Suswa Lava Caves by Leopards

Sorne of the best evidence for the use of caves byleopards in East Africa has come from Mount Suswa orOldoinyo Nyokie, "the Red Mountain," a dormant vol­cano on the floor ofthe Rift Valley ahoul48 km northwestof Nairobi. The central part ofthe mountain consists of awide, shallow caldera, aboul II km across, Ihe floor ofwhich lies al an altilude of 1,830 m. On lhe northeaslslope of the mountaín is a series of aOOut forty-five holes,roughly circular in outline and varying from 1.5 m lo 60 min diameter. These consist of collapses into underlyinglava tunnels or tubes that were fonned when the phono­litic or trachytic lava of which Ihe mountain is composedwas still fluid (Glover el al. 1964).

In the course oC her initial study of bone remains fromIhese lava tunnels, Coryndon (1964) concluded lhal anumber of collecting agenls had been al work (see alsoIsaac 1967b), including leopards, hyenas, owls, and menoIn addition, sorne oC the surface holes concealed by veg­etation had acted as natural traps, eVldenced by theskeletons of 3 rhinos, a giraffe, and various antelopes.Conceming leopard involvement, Coryndon wrote:"Predatory animals prefer certain definite conditions inwhich to take their food-thus the leopard seems to prefera very dark recess where no light can penetrate. One suchcave had a floor strewn with the remains of rnany baboonsand a few small anlelopes, presumably broughl in byleopards (Coryndon 1964, p. 61).

A more specific study of leopard involvement in thebone accumulations of the Mount Suswa caves has beenmade by J. W. Simons (1966), who studied both theskeletal remains of the Leopards themselves and those oftheir prey. Remains of 8 or 9 leopards were found in vari­ous parts of the lava cave system: 3leopard skeletons layon a talus cone beneath the vertical-sided, 21 m-deep ,"caldera blowhole;" together with the dehydrated re­mains of 3 adult and 1 juvenile female baboons. Simonsconcluded that there was little doubt that the leopards fellto their death while trying to catch baboons, which reg­ularly make use of the blowhole as a sleeping place. Thebeboons may also have fallen while being chased byleopards.

Another leopard skeleton, associated with that of a ba­boon. was found beneath collapse hole 6, which has adrop of about 12 m. In Ihe leopard skeleton, the shafts oftbe right humerus and left tibia were broken, damage thatSimons suggests might have resulted from a fall whileattempting to catch the baboon. Scattered remains of 3 or4 other leopards were found in other parts ofthe lava cavesystem.

Concerning the prey of the leopards, Simons estimatedthat there were parts of at least 37 baboons in elevendifferent areas of the cave system. Most of the remainswere concentrated in the tunnels linkíng surface halesdesígnated 6, 36, 37 and holes 18a,b,c,d. Certain verycharacteristic damage appeared to have been done to thebaboon remains: such damage to the skulls may be sum­marized as follows: lateral pterygoid plates on the ventralsurface chewed; tooth holes in the borders of the orbits;zygomatic arches sornetimes chewed away; tooth hotes inthe sides of the muzzle, and brow ridges frequentJychewed. Vertebrae and ribs were found lo be underrep­resented, and a good deal of damage had been done to theother skeletal parts.

The Suswa lava cave situation is fairly clear as regardsthe origin of its bone accumulation: baboons regularly usethe holes in the lava slope as sleeping sites, and they arepreyed on Ihere by leopards that then feed on them in thesubterranean tunneIs. OccasionaJly the leopards and theirintended victims fall to their deaths below the cliff faces.No actual leopard hunts have been observed at Suswa,but Simons has seen a baboon troop corning to sleep onledges around collapse ho1e 18. Mounds of baboon drop­pings below the collapse holes indicate that this has beena long-standing baboon habil.

During June 1970 I was able lo examine aU the bonescoUected up to that time from the Suswa caves andhoused in the Centre for Prehistory in Nairobi. It wassoon apparenl that, as Coryndon (1964)and Isaac (l967b)had observed, the bones owed their presence in the cavesto a variety oCagents. Sorne were charred and showeddamage clearly attributable lo humans; olhers appearedto have been cracked by hyenas; and still olhers had beengnawed by porcupines. Only one area in the cave sys­tem seemed to represent a reasonably uncontaminatedleopard feeding site. It is the tunnel running from collapsebole 36, where live baboons are frequently observed, 10­ward hole 37. Here, in almost total darkness, a steep pas­sage leads upward Cor abaut 6 m from the main tunnel to alow cavem in which baboon remains were scattered ingreat profusion. Apart rrom leopard-inflicted damage, thebones showed a smalJ amollot oC porcupine gnawing, andit is not impossible that porcupines added sorne bones tothe leopard food remains.

The bone collection from this particular spot. desig-

nated "leopard laír 36E," consists of 176 pieces that werecollected in March 1964. Particulars of the individualbones and the damage they have suffered are given intable 42. The remains come from a minimum of 10 anubisbaboons as follows: 1 juvenile male, 1 subadult male,2 adult males, 5 adult females, and 1 juvenile of un­detennined sexo There are also remains of 1adult leopard,1 adult maJe klipspringer, and 2 unidentified birds.

Practically al! the baboon remains show extensive car­nivore damage, incIllding numerolls punctate marks pre­sllmably caused by the leopard canines. Sklllls show theleast damage of al! the skeletal elements, and in no casewas a braincase broken in sllch a way that the leopardscould have gained access to the braln, although the dam­age suggests that they fed on the jaw muscles, tongues,and eyes.

The scarcity of vertebrae, and the almost total absenceofribs. shows that the baboon body from neck to tail wastypically eaten. This is tme of the hands and feet as weIl,and the limbs were clearly disarticulated, each segmentbeing chewed from both ends. It is of interest that manyof the long bones show transverse breaks in their shaftssimilar to the "spiral fractures" described by Dart (1957a)as resulting from hominid activity. In the case of theseSuswa bones there does not seem to be any reasonabledoubt that the breaks were caused by leopards. Someexamples of characterístic damage are shown in figure 80.

A Study 01 Leopard Lairs in South- West Africa

Between 1968 and 19701 was able to investigate a numberof leopard lairs in South-West Africa through the kindcooperation of Attila F. Port, a weIl-known game con­servationist and farmer in the area. At that time Mr. Portwas owner of Valencia Ranch, and the adjoining farmknown as Portsmut, in the Hakos Mountains, about 160km southwest of Wíndhoek, an area where leopards wereformerly extremely abundant and where they stiU occur inreduced numbers. It was in this vicinity that the observa­tions described here were made (see also chapo 2).

The country rock in the Valencia Ranch area is Pre-

Food Remains of Carnivores in African Caves 85

cambrian mica schist, which gives rise to rugged teTrain(lig. 8 J) traversed by watercourses leading westward to­ward the Namib plain. Average annual rainfall is about175 mm, falling in summer, and the vegetation is an aridthorn savanna. In the absence of a well-developed soilcover, the mica schist OlltCropS extensively, giving rise tonumerous overhangs and shelters but seldom to caves ofany size. Mr. Port was able to show me eight overhangsor holes, as well as one cave, where leopards had beenseen; four of these could be regarded as habitualleopardlairs, used eíther as breeding or as feeding sites. Qne wassituated on Valencia Ranch, two on Portsmut, and one onVerloren, as indicated on the map (fig. 82). 1 will givedetails of two breeding and two feeding laírs and thendiscuss some of theír consistent features.

The Portsmut Breeding Lair. The site is in the wall of asteep-sided tributary of the Hakos River about 1 km north­east of the Portsmut homestead and consists of a tunnelless than 1 m in diameter (fig. 83) extending 8 m into themica schist wall. As is shown in the plan (fig. 84), thetunnel opens into a dark chamber about 3 m wide. Bothtunnel and chamber contained a shallow f100r deposit ofloose micaceous dust.

The recognition of this site as a leopard breeding lairoccurred during 1927 in rather dramatic circumstances.Attíla Port's father, first owner of the farro Portsmut,was a military man who believed bis son should grow uptOllgh and fearless. Suspecting that a leopard was rearingcubs in the hole, Mr. Port provided his five-year-old sonwith a thorny branch for protection and induced him tocrawl into the tunne!. Eqllipped with a torch and a .450­caliber revolver, the father followed. As was to be ex­pected. the leopard charged the intnJders, and Mr. Portshot it over his son's head. After they dragged the deadleopard from the lair, the boy crawled back in and broughtout the two young cubs.

The site has apparently been used as a breeding lair onseveral occasions since 1927. and in this interval 8leopards have been trapped at the water hole a fewhundred meters downstream.

Fíg. 80. Baboon bones from a leopard feeding lair in a lava lunneJ on Mounl Suswa, Kenya. Ragged edge damage andpunctale marks are shown.

86 A Guide to the Interpretation of Bone Accumulations in African Caves

Fig. 81. An aerial view oí lhe Raleos Mounlains in Soulh.Wesl Africa, where studies oí Jeopard lairs were under·laken. The arrow indicales lhe Quartzberg lair in an oUlcrop oí white, crystaUine quartz.

YERlOflEN ::12

5<;,1. t. '1000

l) kllom¡j,.-;=J

VERDW¡l.,.,1. 41

HDlJSE

i

Fig. 82. A map oí lhe Raleos Mountain area in South·West Afríca showing the position oí !he various leopard lairsdescríbed in the leX!.

Food Remains of Carnivores in African Caves 87

Fig. 83. The enlrance to ¡he Portsmut leopard breeding (air. Mr. AUíla Port and helpers are shown sieving theoúcaceous dUSl removed from the floor of the lair.

1I

JI! 1I

~ I

My first visit to the lair was on 20 March 1968, whenleopard tracks were visible in the.;9u~ at the entrance andthe stomach of a freshly killed ~ssiyl was found there. 1was anxious to investigate the interior of the lair but ap­prehensive about my reception by its possible occtlpant.Nor was Mr. Port's attitude--doubtless hardened by hisown experiences at the place-particularly reassuring. Inresponse to my query about what to do ifthe ¡eopard wereto attack me, he handed me the tibia of a zebra that hap­pened to be Iying there and suggested I thmst it into thecreature's mouth! After crawling sorne distance into thetunnel, I carne upon the intestines of a dassie and heardominous noises from the dark chamber, suggesting itwould be inadvisable to proceed farther. I therefore re­treated and returned a few days later with more suitableeqnipment. On the second visit the lair was unoccupied; Iundertook a survey and removed and sieved the floordeposit.

flOCK

Ol/TCROPS

Fig. 84. Sketch plan of the PortsmUl leopard breeding lair.

The resulting collection consisted of 192 bone piecesand 17 cultural objects apparently left in the cave by ahllnter-gatherer. The presence of these latter objectsraises the possibility that the cave was used for shelter byprimitive people and that bones found there might repre­sent human food remains. 1 have doubts about this, how­ever, in view of the low roof of the access tunnel and thedarkness of the inner chamber. The grooved stones,grooved pottery piece, pesUes, and hematite and quartzpieces were found together in the inuermost comer of thecave, suggesting that a hunter had stored them there buthad failed to come back for them.

Porcllpine gnawing was discernible on 19 of the bones,in particular on the weathered cranium and mandible of ababoon and the equally weathered zebra limb bones.These and sorne of the other bones were very probablybrought to the cave by porcupines. [, )_

As is detailed in table 43, the remains were found to (/217­have come from an adult male baboon, an adult leopard, 2 ~juvenile klipspringers, 2 c1ass II bovids-probably sheepor goats-l c1ass III antelope, 2 adult mountain zebras, 4dassies, and a tortoise. At least sorne of these animals arelikely to have been leopard prey, though the remains ofthe leopard itself probably came from an animal that diedthere.

There is no way of tellíng from these remains whetherthe prey was brought to feed cubs in the lair or whetherthe síte was used as a feeding retreat between litters.

The Hakos Ríver Breedíng Lair. The site consists of threeintersecting tunnels in the north bank of the Hakos River(figs. 85 and 86) about 9 km west of the Portsmllt home­stead. The tunnels open just aboye the level of theriverbed, which is normally dry, and have formed in a softband of inclined mica schist; they contained a deposit ofmicaceous sand varying in depth from 15 cm to 40 cm,

88 A Guide to the Interpretaüon of Bone Accumulations in African Caves

which was carefully removed and sieved to recover al!bones.

According to Attila Port's records, the site was lastlIsed as a breeding lair by leopards in October 1967, whenthe female was not disturbed. Before this, 104 leopardswere trapped in the riverbed outside the lair between 1910and 1950.

Details of the 339 bones recovered from the lair aregiven in table 44; 65 ofthe pieces showed evidence ofpor­cupine gnawing, and so it is to be expected that some ofthe collection was made by porcupines. Five of the boneshad been partially burned-these may have been pickedup by the porcupines at a human fire place. As at thePortsmut lair, bones of Ieopards were encountered in theassemblage. These carne from a young cub and an adult.It therefore appears not unusual for remains of theleopards themselves to be found in habituallairs.

It was somewhat unexpected to find remains of an adultwild dog among the bones. Whether this animal had beenkilJed by a leopard or whether it had used the lair itse!f Iam unable to sayo Its skeletal parts showed no camivoredamage.

Remains of the various bovids and mountain zebras,many of them from yollng animals, could well representleopard prey. As in the Portsmut lair, however, [t is notpossible to decide whether the prey was taken to the lairat times of cub-rearing.

The indications from these two small caves, which havecertainly been used as leopard breeding places, is thatsl1ch places may be expected to contain remains not onlyof leopards themselves but of their prey.

The Quartzberg Feeding Lair. Although this cave is onthe fann Verloren, which adjoins Portsmut to the north,

the countryside in this vicinity is extremely broken, andno tracks or roads link the two farms. Access to theQuartzberg lair is therefore gained by the Walvis BayRoad at the foot of the Gamsberg pass. The surrollndingsof the cave are easily recognized, since they are purewhite quartz, the only outcrop of its kind in a vast area ofdrab micaceous schist (fig. 81). Here a small stream, fedby a perennial spring, makes its way through a spectacll­lar gorge with glistening walls of white crystalline quartz.As one of the few water sources in these barren moun­tains, the spring attracts a variety of animals, some ofwhich falI prey to the resident leopards.

In the side of the gorge, just below the spring, is a eaveof considerable size, its rounded entrance partiallyobscured by a fig tree and by large fallen blocks of rock(fig. 87). The cave has resuIted from the dissolution of acalcite seam in the quartz and has a complex form, asshown in the sketch plan. Passing between the fallen en­trance blocks, one enters a dimly lit chamber that be­comes progressively darker toward the back (fig. 88),where a low passage Ieads into a second cavem hung withstalactites on the right-hand side. Beyond these a tunnelcontinues, decreasing in height and width.

At the time the plan of the cave was drawn, on 25March -1968, a leopard suillÚng itself on the rocks at thecave entrance was disturbed by our approach and re­treated into the darkness. We followed its tracks acrossthe f100r of the outer cave, through the low passage intothe dark inner cavem, and thence into the tunne! on theright-hand side. Rere we left it, deciding that peacefulcoexistence was preferable to confrontation. My comopanions guarded the entrance to the tunnel with their gunswhile 1 surveyed the cave and collected the bones fromthe inner chamber.

Fíg. 85. Enlrance to ¡he Hakos River leopard breeding lair.

RIVER IIED

LDW WAI,.I,.

Df tl,lIle.SCHIST

Fig. 86. Sketch plan of the Hakos leopard hreeding lair.

The Quartzberg cave is of interest in this investigationbecause it has served both as a habitual leopard feedingplace and as a retreat when resident leopards are dis­turbed. It is likely that it has alsc been used as a breedinglair, but its inaccessability means few people have visitedit, and cub-rearing has not been observed there.

That porcupines also make use of the Quartzberg cavecomplicares any interpretation of the bones found there.Fortunately, porcupines prefer dry and naturally defattedbones (see chapo 5 for discussion on this topic) that havebeen lying in the open for sorne time, and it is these thatthey carry lnto their lairs. They would probably have ig­nored any fresh leopard food remains until these were dryand defatted. AIl the bones collecled in the Quartzbergcave seerned to fall into two groups: unweathered remainswith tissue adhering, derived from animals that were killedor died in the cave, and second, cIean, defatted bonesclearly derived frorn outside the lair, showing abundantsigns of gnawing and almost certainly carried in byporcupines.

The cave was divided into three zones, the light, thetwilight, and the dark, as shown on the plan; bones col­lected from each zone were separately marked. Table 45lists the bones found in the Quartzberg cave. Of the 211pieces, 147 are regarded as probably porcupine-collectedand 64 as probably leopard-collected, As is shown in lable46. it is estirnated that the porcupine component carnefrom a minimum of6 indíviduals, while the mínimum num­ber of individuals in the leopard componenl is 15; 80.2%of the bones in the presumed porcupine component weregnawed by porcupines, and 23.4% in the presumedleopard cornponent showed such marks, suggesting thatthe porcupines may have picked up sorne of the dry re­mains from leopard kills at the cave entrance and gnawedthem in their part of the lair.

!2f F Table 47 gives sorne figures on the distribution ofbones in the Quartzberg cave relative lo the three lightingzones. Only 3% of the bones in lbe presumed porcupinecomponenl were found in the light zone, whereas 36% ofthe leopard componenl carne from this zone. These datasuggest that, in caves. porcupines prefer the twiligbt anddark rones, avoiding the well-lit area, whereas leopardsappear equally al home in lighl, twilighl, and dark zanes.They probably do most of lheir undislurbed feeding al lheentrance to the cave. but they do not hesitate to retreatioto the darkest interior when necessary.L."".... damage lo bones at the Quartzberg lsirshowed _sorne interesting features. Lower limb -bones ofklipspringers were characleristically left um!amaged-andarticulated (fig. 90), while lhe remains of 6 individual das­síes were'almost exclusively cranial. The skulls had beendamaged in a consistently typícal fashion that wiJI be dis­cussed fllrther.

Food Remains of Carnivores in African Caves 89

The Valencia Fecding Lair, This small cave is of interestbecause it was a feeding site of the leopard upon whicbfeeding experiments-c--shortly to be described-wereundertaken. After its escape from a cage at the Valenciahornestead in October 1968, this large rnale leopard madeits headquarters for approximately two weeks at the Val­encia dam, 00 a tributary ofthe Noah Rtver. 0.7 km westof the house. Here it killed an adult karakul sheep and asubadult goal, dragging both of them into a small cave inthe mica schist a short distance aboye the dam wall (fig.91). After eating parts of these, as well as most of a sub­adult dassie, the leopard moved away and was never seenagain.

The cave consists of a triangular recess , 15 m wide , inthe steep rocky hillside. As shown in the plans and sec­tions (fig, 92), the outer overhang leads by way of a lowdivide into an inner chamber with a máximum height of'.5 rn. It was on the ñoor of this inner chamber, litteredwith dassie droppíngs, that the remains of the leopard'sprey were found. AIIthe bones were removed for cleaningand study; the skeleton of the karakul sheep was virtuallycomplete, suggesting that the leopard may have beendisturbed and left the area before finishing its meal. Thedamage to the goat skeleton was more extenslve: the axisvertebra had been chewed through and a1l vertebrae be­tween this and the sacrum consumed, The pelvis showedtooth marks; both scapulae were extensively darnaged,one with a characteristic punctate mark; the proximalends of both humeri had been chewed off', and both kneejoints were chewed through, resulting in the loss of thedistal femurs and proximal tibiae. The skeleton wasotherwise complete.

The dassie remains consisted of a characteristicallydamaged skull with.no postcranial bones. TIte nature ofJeopard damage to ?as'51e prey will be considered further.

Leopard Damage /0 Dassle Prey Skeletons

Remains of 6 dassies were found in the Quartzbergleopard lair and of 1 al the Valencia feeding site. AlQuartzberg, only 3 postcraniaí bones were found as­sociated with the 6 craniums; at Valencia there were noneal all. Furthermore, il appeared in each case that thebraincase and posterior part of the mandible had beensheared off, leaving ragged, toolb-marked edges, Exarn­pies from these lwo feediog sites are shown in figure 93,together with a similarly treated skull from the Suswa lavacaves in Kenya, a leopard feeding situation dlscussedearlier. It seemed bighly likely that when leopards fed ondassies they lypically consumed the whole posteranialskeleton and that, working up toward the head from theback, they sheared off the braincase and ascending man­dibular rami to remove the brain and tengue. The re­maining unchewable maxillary and mandible parts wererejected.

The oniy way lo conlirm Ibis would be by direcl obser­vation of a feediog free-ranging leopard, or by conlrolledfeeding experimenls on a captive one. The possibility ofobserViog feediog behavior of a wild leopard io lbe Valen­cia area seemed remote owing to their extremely secre­live behavior lhere, bul Altila Port kindly offered lo calchone for experimental purposes. The trap was con­sequently selon 20 March 1968 in one oflhe lribularies ofthe Hakos River near the Portsmut homestead. It con­sísted of a subst:.tntial rectangular cage with falling doorsat each end activated by a foot treadle. The cage waspositioned 00 the floor of the steep-sided valley, and a

90 A Guide to the Interpretation of Bone Accumulations in African Caves

Fig. 87. Entnlnce to the Quartzberg leopard lair. The large blocks of rock at the enlnlnCe are use<! by the residentleopards as feedí ng places.

Fig. 88. [nside the Quartzberg leopard lair cave: the 'dimly Jit chamber leading froro the entrance.

••>~barricade ot\ihomyranches was built on each side of it sothat, we hop~ any leopard making its way along thevalley would pass through ir. During the first week thecage was sprung, and we approached it with great anti­cipation but found, to our surprise and disappointment,that we had caught nothing more spectacular than a largemountain tortoíse. But finally, on 26 April, a leopard en­tered the trap and captured itself (fig. 94). This proved tobe the largest male 1 had seen, weighing % kg (211 lb) and

SKETCH PLAN DF THE

QUARTZBERG

LEOPARO LAIR

OARK ZONE

TW' L1GHT ZDNE

>? L1GHT ZONE

20 40, ,

FEET

Fíg. 89. Sketch plan of the Quanzberg leopard lair cave.

Food Remains of Camivores in African Caves 91

was so ill tempered that it could be handled only with thegreatest difficulty. Most authorities quote the maximumweight for adult leopards at about 82 kg (180 lb), so thiswas an unusual specimen; Roland Ward' s Records ofBigGame states that leopards weighing more than 200 lb haveoccasionally been recorded (Bltst 1973), and Tumbull­Kemp (1967) in his book The Leopard admits seven suchrecords.

The captive leopard was taken to the Valencia home­stead and transferred to a specially constructed cage.Handling was difficult, since any entry to the cage pre­cipitated an uncontrollable attack; to overcome thisproblem, the cage was divided into two components witha lockable door between them. Food could be placed inone compartment while the leopard was in the other andthis door then opened.

Records kept by Attila Port over many years showedthat food preferences of wild leopards in the Valenciaarea were, first, domestic ca1f; second, kudu calf; third,young mountain zebra; and fourth, gemsbok calf. Sheepwere taken occasionally, but goats rarely. Typically nomore than 10 kg of meat would be eaten from the prey,and the leopard usually did not return to the carcass, butwould kill again within the next few days. In the absenceof hyenas in the Valencia area, the leopards never usedtrees to store their food, although they would often dragprey to a habitual feeding site where disturbance was notlikely. The relatively easy life of Valencia leopards mayhave permítted the exceptional size of sorne índividuals.

In captivity, the male leopard ate even less from thecarcasses provided for it than had been observed in thewild state. Weights eaten from three kudu calves andthree baboon carcasses were as follows:

Kudu calf 1 : 3 lb 14 OZ or 1.8 kg eatenKudu calf 2: 4 lb 2 OZ or 1.8 kg eatenKudu ca1f 3: 2 lb 3 oz or 1.0 kg eatenBaboon adult female : 4 lb 7 oz or 2.0 kg eatenBaboon adult female : 4 lb 1 oz or 1.8 kg eatenBaboon adult male : 2 lb 8 oz or 1.1 kg eaten

Pig. 90. Remains oC !WO Iclipspringers and a babean, lefl by a leopard tha! used the large blocks oC rack al lheentrance lo the Quartzberg cave as a feeding place.

92 A Guide to the Interpretatíon of Bone Accumulatíons in Afrícan Caves

Fig. 91. The Valencia leopard lair, a small cave in the mica schis[ hillside,used as a feeding retrear.

Fig. 92. Plan and section of [he Valencia leopard lair.

The mean weight of meat eaten from these six carcas­ses was 1.6 kg, and on no occasion did the leopard retumto the remains to eat a second time. The purpose ofcatching the leopard had been to aIlow it to eat dassies incontroIled circumstances so that the remains could becompared with those found in the natural leopard lairs.Unfortunately, only two dassies had been fed to theleopard before it escaped one night, killed a youngmountain zebra in an adjoiníng enclosure, and dís­appeared. Subsequently we found that the leopard hadmade use of the Valencia dam cave as a feeding lair, asdescribed earlier, before moving away altogether.

In September 1%8, when the feeding experiments ondassies were performed, the captive leopard was still ex­tremely aggressive and secretive about its feeding. It re­fused to eat during the day or when an observer was insight. What observations were possible were made atnight from a concealed position. Sorne particulars of thetwo feeding experiments are as follows:

Dassie 1:Adult femate, shot 21 September 1%8, weight 61b, 4

oz, or 2.8 kg; eaten during the rught, a total weight of 5lb, OOZ, or 2.3 kg, being eaten. A good deal offur wasfirst plucked from the ventral surface of the dassie, thebody cavíty was opened, and the stomach and in­testines were removed and placed to one side. Theentire body and ske1eton was then eaten, the only partremairung being the snout. As is shown in figure 95,the damage to the skull is remarkably similar to thatobserved in dassie craniums from the naturalleopardfeeding lairs. The braíncase has been chewed away,

and a punctate mark is present betow the right orbit.Al! of the right and most of the left ascending man­dibular rami have been chewed away, leaving typicalragged edges. No postcranial bones remained.Dassie 2:Subadult male shot 23 September 1968, weight 5 lb, 8oz, or 2.5 kg, eaten during the night, a total of 4 lb, Ooz, or 1.8 kg, beíng eaten. As before, the fur wasplucked out, the alimentary canal was removed andrejected, and the entire body, with the exception ofthe front part of the head and the left hind foot, waseaten. Damage to the cranium was virtually identícalto that of dassie l.

These two feeding experiments confirm that leopardstend to leave the anterior part of a dassie skull, damagedin a characteristic manner, but that Httle or nothing of thepostcranía] skeleton remains. Thus, in a fossil context,leopard food remains might reasonably consist of dassiecraniums showing "sheared-off' characteristics.

It is perhaps worth recording that when cheetahs feedon dassies they employ almost exactly the same proce­dure, except that on no occasion have 1 seen them pluckfur from the body before starting to feed. Working withthe captive cheetahs at Valencia described in chapter 2, Iwas able to make detailed observations on four dassies,each eaten by a single cheetah. Extracts from my fieldnotes conceming one of the experiments read:

19 March 1968,2 P.M. Shot an adult dassie on the cliffbelow the Valencia farmhouse and offered the body toan adult female cheetah, who took it ,immediately. Shefirst chewed off the right front leg, then opened thechest from the side and below, eating the lungs, liver,and heart. She then worked backwards past the día­phragm and removed most of the alimentary canalcomplete. This was dumped under a tree, and the re~tof the body was moved to another patch of shade.

The cheetah next chewed off the head of the dassieat the neck and worked on the back of the head for 10min (fig. 96), eating the brain but leaving the snoutwith the fur on it. The dentition of the dassie appearedundamaged. She then ate away all the flesh and bonesof the body, detaching the remaining legs one by one,chewing and swallowing each.

After 1~ hr a patch of back skin about 8 in by 6 inwas left, together with the snout and the atimentarycanal. The cheetah moved away, and so I enteredthe enclosure to collect the remains. As 1 was leaving,the cheetah attacked me and put on an impressiveintimidation display. She slowly withdrew when 1stood still.

Records of the other three dassie feeding experimentswere very similar, and in each case the snout andalímentary canal remained. It is fairly clear why the das­sie stomach is rejected by both leopards and cheetahs. Inmost specimens I have examined, the stomach, a bean­shaped organ up to 15 cm long, was packed to capacitywith green 1caves and other vegetable matter. This wasso tightly compacted that it is surprising digestion couldproceed.

The reason for these apparently overfilled dassiestomachs has to do with the animals' feeding behavior,which has been described by Sale (1965, 1966). On thebasís ofEast African observations, the average daily foodintake of dassies is not excessive rclative to their body

Food Remains of Carnivores in African Caves 93

b

"..-:)s~~,\

\\j

Fig. 93. Characleris[icalIy damaged skulls of dassies found in three leopard feeding lairs: (a) QUarlzberg cave: (b)

Valencia ¡air; (e) Suswa leopard lair, Kenya.

weight, but the mode of ingestion is unusuaJ. The incísorsare not used and appear unsuitable for biting off shootsand leaves. Instead, the head is tumed sideways, at rightangles to the body, so that the entire premolar-molartoothrow may be used in shearing the vegetation, allow­ing remarkably rapid ingestion. In fact feeding, which isusually performed in groups, usually occupies less thanan hour a day, split into p10rning and aftemoon sessioos.During these sessions the dassies' stomachs are rapidlypacked with vegetation. The briefness of the feeding ses­sioos appears to be an antipredator adaptation, for it iswhen foraging away from the protection of their rockyretreats that dassies are most vulnerable to predation.

The Diet oi Leopards as Reflected in Observed Kills

There are two obvious ways of establishing what leopardseat: direct observations of their kills and aoalysis of theirdroppings. The two methods are likeIy to produce strik­ingly different results, as will be discussed shortly. In­formation on leopard diet is relevant to the present studyof bone assemblages in caves because it will indicate therange of animals that could be consumed in feeding lairs.Five detailed studies of leopard kilIs are currently avail­able, one from the Transvaal, one from Rhodesia, one

from Zambia, and two from Tanzania. Sorne particulars ofthe areas in question are now provided.

Kruger Nalional Park, TransvaaF. Situated along theeastem boundary of the Transvaal, this park had an areaof 19,084 km2 and incIudes a variety of veld types, themaio ones being lowveld, sour bushveld, arid lowveld,mopani veld, and mixed bush veld (Edwards 1974). De­tails of kiHs have been recorded by park staff over manyyears, and Pienaar (1969) collated them for the periods1933-46 and 1954--66. The recorded kills totaled 1,940 forthe first period and 5,525 for the second, a remarkabletotal of 7,465.

Rhodes Matopos Nalional Park, Rhodesia. This nationalpark covers an area of about 43,200 ha situated 48 kmsouthwest of Bulawayo and consists of broken granitehills and gorges surrounded by woodland containingpatches of grassland and savanna woodland. The areawas mentioned in chapter 3 when I discussed thePomongwe Cave. A study of leopards ín the Matopos hasbeen made over a number of years (Wilson 1969; Wilsonand Grobler 1972; Smith 1977), and informatíon is avail­able both from records of kills and from analyses of scats.In the most recent paper, Smith provides data on 38 killsobserved between 1960 and 1974.

94 A Guide to the Interpretation of Bone Accumulations in African Caves

Kafue Na/ianal Park, Zambia. Covering 8,600 miz, thepark líes in the basin of the Kafue River and carrieswoodlands of the Brachystegia-Julbernardia complexwith open drainage lines supportíng grass savannas andaquatíc grasslands in the wetter patches. Kill recordswere kept between June 1960 and May 1963, a total of96being listed by Mitchell, Shenton, and Uys (1965).

Serengeti National Park and Surrounding Areas, Tan­zania. A vast area of plains and hilis Iying betweenlatitudes of lOS and 3°30'S, the national park itselfhas anarea of about 13 ,250 km2, supporting two main vegetationtypes: grassy plains and wooded grasslands with widelyscattered trees, creating a parklike aspect (Schaller 1972).

Tumer observed leopard predation from 1957 till 1964;he was thenjoined by Kruuk until the end of 1965, and 55kills were recorded in this time (Kruuk and Turner 1967).Schaller's observations were mostly made in the Seroneraarea and the edges of the woodlands, in dry seasons be­tween June 1%6 and September 1%9. Details of 164 killswere recorded (Schaller 1972). The full list of prey ob­served in the course of these five studies is given in table48, and a summary ofthe main prey categories is providedin table 49 and depicted graphically in figure 97. Each ofthe five studies reveals that bovids in size cIass 11 forro thebulk of the observed kills, the species varying from areato area, depending on the habitat. In the Kruger Park andthe Matopos, for instance, impala kills are the mostnumerous, whereas reedbuek and puku dominate in theKafue area and Thomson's gazelIes in the Serengetistudies.

Antelopes in other size classes are less prevalent,although duikers are favored prey in the Matopos andKafue. Sorne of the larger antelopes such as wildebeestsand sables are regularly taken, but in such cases youn,ganimals are usually involved. In the very large species,"itis invariably calves that are killed.

Amoog nonbovid ungulates, zebras and warthogs fea­ture fairly promineotly; among primates, kills of baboonsand vervet monkeys have been recorded; and a widerange of camivores is known to have been killed byleopards. The literature contains many references to thestrong taste these cats have for domestic dogs and to theirdisregard for danger in obtaioing them. 1 will cite twoinstances. The first, related by Major C. Graham, oc­curred in the Mongalla Province of the Sudan. It con­

Fig. 95. One oC the dassies fed experimenla1ly lO (he leopard: above. !he cerned the late C. H. Stigand, who was in civil charge ofentire animal: below. discarded remains-lhe snoUl, lufts of fur, and, the region (Graham 1953, pp. 244--45):a1imenlary caDaI.

Fig. 94. The exceplionally large male leopard used in dassie feedingexperimenls described in the lex!.

Fig. 96. A caplive cheetah on Valeacia Ranch eating a dassie. The wholebody, wilh the exception oC lhe slornach and snoul, is crunched up by ¡hecheelah's camassials. Damage lo the skull is similar to thal observed inleopard feeding.

one night he and Iris wife were woken up by a com­motion in the room next to their bedroom. Theythought theír dog was attacking a small duiker kid theyhad inside the house, so they jumped up and rushed tothe rescue. You can imagine the shock they got whena leopard dashed past them out of the room and intotheir bedroom! He had come right into their solidlybuilt stone house, forcing his way through a rough butstout wooden door, and then, passing through a smallentrance hall, had gone for the dog in the room oppo­site. The dog was only badIy mauled and they shot theleopard through a window of their bathroom.

The second incident took place in Angola while theBenguella raiIway line was under construction. Qne ofthe locomotive drivers named Jones had a fox temerdog which stayed with him io the gangers' cottage. Qneevening the driver

Food Remains of Carmvores in African Caves 95

had been sitting at a small table in his lantem lit room.eatíng his supper. His dog had been lying on the ñoorat his side with íts back to the door. Suddenly a hugeleopard sprang into the room, but owing to the srnoothcernent ñoor, it slid across the room knccking Jonesand his table overo JODes, tbe table, the dog and thefurniture, skidded in a circular movement around theroom. Unfortunately the leopard regained its balanceñrst, and ñeeing towards the door, caught the dog initsjaws and disappeared from sight. [Ryan 1961, p. 68]

Certain leopards also have a preference for jackals-one was observed by Estes (1967) in the Ngorongorocrater to kilI eleven in three weeks. Sorne of the largercamivores are also preyed upon: Pienaar (1969) provídesa photograph ofan adult male cheetah weighíng 45 kg thatwas strangled by a leopard and hnisled 3.6 m into the forkof a marola tree in the Kruger Nationa! Park.

Bírds seen lo have been killed by leopards include anumber of European storks, Schaller (1972) suggests thatthese may be rather naive about large fetine predatorsthat are absent from their nesting areas in the northemhemisphere.

The Overaií Diet ofLeopards

It is well known that leopards frequently feed on smallprey that would seldom feature in any liste of observedküls. For this reason, information derived from such kilIobservations is likely to be strongly biased in favor oflarge prey items. The analysis of leopard reces can, how­ever, provide very valuable information about their totaldieto For some years leopard droppings have been reg­ularly collected in the Malopos National Park, and ananalysis of 358 samples as given by Smith (1977) is listedin table SO. It is not usualJy possible lo determine from

fecal analyses how many ofeach kind of animal have beeneaten, but the percentage occurrence of remains from aparticular prey item in the leopard scats must surely beara close relationship to the frequency with which theseiterns were eaten.

In the Matopos study area it is possible to compareresults of kili observations directly with those of fecalanalysis. On the basis of observed kills , the sequence ofimportance of prey iterns, as listed in table 48. was, first,impala (32%); second, reedbuck (13%); thírd, duiker andsable, both wíth 11%; and, fourth, wildebeesl (8%). Dalafrom fecal analysís provides a very different picture: firstis dassie, with 46% occurrence; second, klipspringer(10%); third, hare (8%); fourth, duiker (7%); then rats andmice (5%) (fig. 98). JI is likely that when many of thesmaller prey animals are eaten by a leopard in a feedinglair, there would be no bony food remains.

Leopard Predation on Primates

From the point of view of the interpretation of theSterkfoatein valley bone assemblages, Jeopard predationon primates, particularly baboons and hominids, couldhave been very significant. Neither in the lists ofobservedkills nor in the results of fecal analysis do primates featureas important leopard food sources. Yet in certain circum­stances, such as on Suswa mountain, where baboonscoming to their sleeping sites are preved upon, predationon primates can assume important proportions,

Baboons as Leopard Prey

Although a single baboon cannot contend with a hungryleopard, a baboon troop in an aggressive mood Is a differ­ent proposition. Conceming leopard/baboon interactlons

10050

IUlUGER PARK1PI.rul8r 1

NATOPOSfSmltb)

SERENGETI

l$chelle.}SERENQETI

(Kru".1k and Tun'•• )I

---;5O~ o 50 90 op....--~··

PECENTAGE OF OBSERVEO KILLS

Fig. 97. Broad cetegcries of leopard prey as determíned by five field studies in various parts of Africa. See tables 48

and49. (r 2 9q \

.J.trrELOPEm~ CLASS I

~NTELOPE""~LA8SI1

1M( PR.'_M_AT_E_"-IL --i ... -+ + _Jf'fcA.N,vaR

~:--~-:~-.-IIr------

~ILES'~R~~H ~ -:

o 50 o

Leopard Predatíon on Monkeys

Leopards are known to prey on both species of monkeythat occur in southern Africa-Cercopithecus aethiops,the vervet, and C. milis, the samango, but few directobservations have been published on these interactions.Wilson (pers. comm.) has sent me notes on one observa-

17 January, 1960,5.10 pm. Adult leopard foundfeeding on large male baboon. Leopard spent 45 min­utes at carcass and only part of it was eaten. The car­cass was then carried up into a large Marola treegrowing al the base of a granite rocky outcrop.

18January, 6.20 amo When 1 visited the spol 1 wassurprised lo find two leopards al the carcass whichwas now on the ground, Only the skull in fact re­mained together with pieces of skin and a couple oí legbones. As a result of disturbing them the leopards ranoff and díd not retum to the carcass.

Events at the other two kills were rather similar, exceptthat only one leopard was involved in each.

Several observations suggest that individual leopardsmay develop ataste or preference for a particular kind ofprey, In certain situations they may hunt baboons alrnostexclusively, and in vartous arcas of the Transvaalleopards are regarded as regulating factors in baboonpopulation numbers (Stoltz 1977).

daytight hours, solitary males or stragglers away fromthe troop.

The second incident, which took place in the KrugerNationat Park, was repcrred among readers' letters to thejournal Afrícan Wildlife (Bates 1971, p. 154):

We witnessed a rather exciting epísode during July1966 at Manurge Kop. It was about 5 pm and we weretaking a slow drive home before retuming to Pre­toriuskop rest camp when from the rocks which makeup Manurge Kop there arose a great clamour.

Out of the bush burst a leopard followed by fivelarge mate baboons. The leopard leapt into the tree atthe roadside followed by the baboons, who werebarking and making a great deal of noise. The leopardproceeded upward and outward anta the thinnestbranch-whilst his tummy worked furiously-the re­sult falling onto the roadside in front of us!

The baboons sat still for a few minutes and thenwith a great bark pounced at the leopard, who withgreat dexterity managed to evoid thern and from atleasl 20 fL high leapt out of the tree lo disappear intothe grassland alongside the road.

The baboons , with great chattering, retumed to thesafety of Manurge Kop.

The records of these incidents serve to emphasize theeffectiveness of coordinated action within the baboongroup. Successful hunting of baboons will therefore de­pend 00 stealth and surprise attack, most ofwhich is prob­ably carried out al night.

It appears that leopards, when hungry, will eat most oral! of a baboon body, leaving the skull comparativelyintaet. V. J. Wilson, former curator oí the NationalMuseum al Bulawayo, has kindly sent me extracts fromhis field notes concerning three baboon kills he found nearChipangali in eastem Tanzania. The description of one isas follows:

KI ipspringer 10 %

T«Hare 8 %

T

tttOuiker 7 %

..Dassie 46 %

ir

SCAT ANALYSIS

T~

Rats and mice 5 %

Sable 11 %and

Wildebeest 8 %

r(Impala 32%

OBSERVEO KILlS

Reedbuck 13 %

in the Kruger Natíonal Park, Pienaar has written (1969,p. 123):

Leopards kili many baboons when they are stealthilystalked or pounced upon at theír roosts during thenight. Occasional1y however, a leopard is driven torashness by hunger and atternpts to snatch one out ofa troop in broad daylight. In such cases a number ofbig male baboons will usually come lo the aid of theshrieking victim and in the ensuing free-for-all theleopard is often severely wounded or tom to pieces bythe enraged primates.

Leopards seem fuUy aware of the danger posed by ababoon troop, as evidenced by the following two in­cidents, Tbe first was quoted by Smithers, in his Mam­mals 01 Botswana (197111, p. 116):

While the predation oí leopards on baboons has beenwidely stressed in literature, an adult leopard is nomatch for the co-operative efforts of a troop oí ba­boons during daylight hours. Al Wankie, Rhodesia, aníncident of this lype was observed, the leopard Iyingup near a waterhole and, being spotted by a large ba­boon, took to its heels. When the baboon charged it,the troop including many juveniles then followed untilthe leopard took lo a high tree where ít was harried bymembers of the troop, the remainder sitting around onthe ground at the base of the tree until sundown, whenthey straggled off leaving the leopard still marooned inthe high branches. Baboons do, however, figureamong the prey species but are probably taken frcmthe fringes oC the troop resting after dark or, if during

Fig. 98. On the basis of ñetdwork done in the Matopos of Zimbabwe(Smith 1977), it is possible to compare resulta uf leopard k.iII observationswirh those of fecal anaIysis. The sequence of impcrtence of prey itemsdiffers considerably in the two methods, as is shown here.

Ouiker

96 A Guide to the Interpretation of Bone Accumulations ín African Caves

tion made at Kalichoo, eastem Zambia, when on 6 Octo­ber 1962 he encountered a leopard feeding on a largevervet monkey.

It spent quite some time licking the hair off the ani­mal and then ate the skin and complete carcass, leav­ing only the skull together with the lower jaw. In factthe entire head was chewed off and left intact togetherwith the eyes, skin, etc.

That leopards tend to leave the heads of their monkeyprey intact was confirmed by observations 00 the feedingbehavior of two young captive leopards, caught andreared in eastem Zambia (Wilsoo, pers. comm.; Wilsooand Child 1966). Wilson (pers. comm.) has expressed theopinión that skulls of monkeys and baboons frequentlyfound among rock outcrops in rugged country of eastemZambia are food remains of leopards.

Leopard Predation on Horno

Evidence of leopard predation on people is relevant tothis study for the light it might throw on the possibilitythat leopards may have preyed 00 early hominids. Theancient and modero situations are, however. very díffer­ent. Through their technology, men have become highlydangerous adversaries, and leopards leam early either10 avoid contact with people or to treat tbem with greatcircumspection.

Cases of man-eating by leopards are not uncornrnonand show at least one aspect that is important to us here.When a leopard takes to eating people, it generally con­tinues to do so, perhaps to the virtual exclusion of otherprey sources. It is frequently suggested in the literaturethat man-eaters, once they acquire the taste for humanflesh, become exclusive predators of people. The evi­dence confirms that this is at least partly true, but suchbehavior may not be as unusual as it sounds. Leopardsappear to develop strong food preferences, which varyfrom place to place and perhaps from time to time in anindividualleopard's Iife. The leopards 1 was able to studyin the Satara area of the Kruger National Park seemed tohave a virtually exclusive preference for impala, whereasin other parts of the park certain individuals showed astrong preference for reedbuck or waterbuck. TumbulI­Kemp (1967) described feeding preferences amongleopards marooned on islands that formed in the Karibadam during its filling process. One leopard would go togreat lengths to catch agile duikers 00 an istand, wheocatching baboons, enfeebled by malnutrition, would haverequired hardly any effort, That particular leopard'shabitual diet was duiker, and substitute foods were notfavored as long as a single duiker was available.Likewise, the bone accumulation in the Suswa feedioglair suggests thal the leopards involved there had devel­oped a preference for baboon meat that was almost exclu­sive of other food sources.

So it is, it seerns, with man-eaters. Human flesh mayfirst have been sampled as carrion or through a chanceencounter, The well-known man-eating leopard of Rud­raprayag in India started preying on humans in 1918afteran influenza epidemic c1aimed many lives in that area.The Hindu custom is to cremate the dead, but because theepidemic was killing so many people, full cremation wasimpossible. Hindu tradition thus was satisfied by placing alive coa) into the mouth of the deceased person and tip­ping the body down one of the steep valleys that abound

Food Remains of Carnivores in African Caves 97

there. Leopards are known to be attracted to carrion,and this may have started the man-eating habit of thisparticular lndian leopard (Corbett 1954). Despite con­certed efforts by the authoritie s in the Garhwal region ofIndia, as well as by various sportsmen, the leopard killedal least 125 people between 9 June 1918 and 14 Aprill926,when it was final1y shot by Col. Jim Corbett. During thisperiod it obviously ate other prey, but a number of goatsused as bait were killed but not eaten, the leopard tryinginstead to kili members of the party hunting it. As anindication of the strong preference the Rudraprayagleopard showed for human flesh, the followiog quotationfrom Corbett's book The Man-Eating Leopard of Rud­raprayag is interesting and relevant (Corbett 1954, pp.11-12):

A boy, an orphan aged fourteen, was ernployed tolook after a flock of forty goals. He was of thedepressed-untouchable-class, and each eveningwhen he retumed with his charges he was given hisfood and then shut up in a small room with the goats.The room was on the ground floor of a long row ofdouble-storied buildings and was immediately belowthe room occupied by the boy's master, the owner ofthe goats. To prevent the goats crowding 00 him as heslept , the boy had fenced oITthe far left-hand corner ofthe room.

This room had no windows and only the one door,and when the boy and the goats were safely inside, theboy's master pulled the door to, and fastened it bypassing the hasp, which was attached to a short lengthof chain to the door, over the staple fixed in the lintel.A piece of wood was then inserted in the staple tokeep the hasp in place, and on his side of the door theboy, for his better safety, rolled a stone agaiost it.

/,

".- ­t .

98 A Guide to the Interpretation of Bone Accumulations in African Caves

On the night the orphan was gathered to his fathershis master asserts the door was fastened as usual, andI have no reason to question the truth of his assertion.In support of it, the door showed many deep clawmarks and it is possible that in his attempts to clawopen the door the leopard displaced the piece of woodthat was keeping the hasp in place, after whieh itwould have been easy for him to push the stone asideand enter the room.

Forty goats packed into a small room, one cor-ner of which was fenced off, could not have left theintruder much space to manoeuvre in, and it is left toconjecture whether the leopard covered the distancefrom the door to the hoy's comer of the room over thebacks of the goats or under their bellies, for at thisstage of the proceedings all the goats must have been00 their feet.

11 is besl lo assume that the boy slept through all thenoise the leopard must have made when trying to forceopen the door, and that the goats must have madewhen the leopard had entered the room, and that hedid not cry for help to deaf ears, only screened frcmhim and the danger that menaced him by a thin plank.

After killing tbe boy in the fenced-off comer, theleopard carried him across the empty room-the goatshad escaped inlo the night-c-down a sreep hillside, andthen over sorne terraced fields to a deep boulder­strewn ravine. It was here, after the sun had been up afew hours, that the master fcund all that the leopardhad left of his servant. Incredible as it may seem, notone of the forty goats had received so much as ascratch.

This particular leopard was one of many that, in recenttimes, had become man-eaters in India. Perhaps the mostnotorious was the man-eater of Panar, which was reputedto have killed four hundred people. Incidents continue tobe reported. G. B. Schaller worked in the Kanha NationalPark of central India between 1963and 1965and reported(1967) that three people were killed there by a leopard in1961. Then, between 2 August 1964 and 19 March 1965when the leopard was shot, there were six attacks 00

people, four of which were fatal. The leopard proved lobe a male weighing 97 lb (44 kg) with Its lower rightcanines missing and showing evidence of having receiveda gunshot wound in its Cace some time previously.

In Afriea, incidents of man-eating have been reportedfrom many places. Swayne (1899) related that for severalyears before 1899 a leopard had haunted the Mirso ledgeof the Golis Range, inland from Berbera in Somalia, andwas reputed lo have killed more than a hundred people. It

. had the habít of Iying in wail al a comer of a dark, roughpath overhung by Iarge rocks. The Somalis pointed out aboulder, a meter from the path, from the lop of which theleopard was said to spring on unsuspecting travelers.

Between 1936 and 1937 a leopard was reported lo havekilled sixty-seven people on the Chambezi River in east­em Zambia and 10 have met its end when it jumped on aman who happened lo be carrying a large fishing spear(Brelsford (950). The same author reported that eightpeople were killed by a leopard in the Isoka area of theLuangwa val1ey during 1938, and an African woman'sbaby was snalched from her back by a leopard al Luwinguin 1943.

Finally, the activities of a man-eater at Kanganga sta­tion during the construction of the Benguella railway line

have been described by Ryan (1961). This particularleopard appears to have started preying on humans whenit carne across a two-year-old Afriean child Iying on a skinrug where its mother had placed it while she was workingin a field. The leopard grabbed the child and disappearedinto the undergrowth. Thereafter it attacked five otherpeople, three of them Iatally, before it was shot in a pittrap.

It would be interesting to know whether man-eatingleopards had any particular characteristics in common.For instance, the Rudraprayag animal proved to be arnale, somewhat past its prime, with one eanine toothbroken. Turnbull-Kernp (1967) stales that, of 152 re­corded man-eating leopards, only 9, or 6%, were females.He provides further infonnation about 78 man-eaters,showing that, of these, 80% were mature individualswithout injuries; 12% were aged: and 4% were immatureleopards. It seems safe to assume that the man-eatinghabit is acquired typically by mature male leopards, un­affecled by injuries.

Leopard Hunting Ranges

It would be useful, when interpreting abone accumulationin a leopard feeding laír, to know over what geographicrange the prey had been hunled. Fortunately, sorne infor­mation is available on home ranges and hunting areasof leopards. In the Seronera area of the Serengeti, Schal­ler (1972) found that seven leopards occupied an area of200 km2 and that the minimum ranges of two femalesthere were 40 and 60 km'. He wrole (1972, pp, 284-85):

The ranges of resident leopards overlapped consid­erably, although each animal tended lo focus its activ­ity in an area little used by others at the time. Twoadults were occasionally within .5 km of each other,and during one period three females and a subadultmale hunled along the same 5 km-long stretch of river,yet I only once saw two adults together when theywere not courting. This indicates a strong mutualavoidance probably based both on dírect visual con­tact and on such indirect methods as marking withseent. There was no evidence that these females ac­tívely defended a territory, but the facl that only oneadult male used the area suggesls that he possiblydid so.

Further infonnation has been fortheoming from a re­markably detailed study carried out in the Malopo HilIs ofRhodesia (Smith 1977): 730 sets of leopard tracks werefollowed until they were lost, and when plolted theyshowed deflnite clumping. Judging from differences inspoor size, it was concluded that each of the arcas show­ing spoor clumping (fig. 99) represenled the home range ofan adult male and one or two adult females, there beingseven home ranges involved with an average area of 18km2 and a variation of from 10 to 19 km>, These rangeswere occupied by about twenty residenl leopards as weHas by a number of transitory individuals that wanderedover much wider areas.

As the figure shows, there is a certain amount oí"over­lap between the home ranges, and Smíth estimated thatthe overallleopard density in that par! of the Matopo HilIswas one animal per 4.5 lo 5 km'. Although this soundshigh, it was found in the Wilpattu National Park ofCeylonthat leopard home ranges did nol exceed \0 km' and thatthose of the males were virtually exclusive but over-

j

Japped ranges of femaies (Eisenberg 1970; Eisenberg andLockhart 1972).

There are sorne indications that , where the horne rangeof a leopard is fairly extensive. a regular routíne is dis­cernible in its movements. Astley Maberley (1953) re­ported that, on his farm Narína in the Letaba District ofthe northeastern Transvaal, a leopard made predictable,near monthly, visits to the valley where the farmhousewas situated. The following dates índicate the regularityof the process:

Arríval Departure28 June 4 July

12 August 20 August5 Seplember 12 Septernber

Evidence collected by Corbett (1954) suggesls that theman-eating leopard of Rudraprayag operated ever an ex­ceptionally large area. If all the human kills are included,an area of 500 mi? (1,296 km-) is involved. However, asis shown in figure 100. almost aJl the kills occurred in asmaller area, namely 150 rni> (388 km-).

In its search for human victirns, this man-eater clearlytrespassed on other males' home ranges, as evidenced bya vicious fight that took place on 14 Apri11926 between itand a territorial male leopard. The man-eater sufferedminar injuries and was forced to retreat, covering 29 kmin the course of that níght.

Transport o/ Prey by Leopards

When interpretíng bones in a leopard feeding lair, it wouldalso be useful to know over what distances they are likelyto have been transported. In completely undisturbed cir-

Fig. 99. Land utilization by leopards in tbe Matopos area of Zimbabwe(from Smith 1977). Abol'f·. the clumped distribution of leopard tracks,giving an indication of the borne range pattern of individual animals shownbelow. From Smith 1977.

Food Remains of Carnivores in Mrican Caves 99

curnstances a leopard may feed where it kills, but thisseems rare. and the prey is usuaJly canied or draggedsorne distance to a feeding site . My own observations onthis are scanty, In the Satara area ofthe Kruger NationalPark, where 1 examined twelve irnpala kills, all but onewere placed in trees, probably very clase to where theactual killings took place. The abundance ofsuitable treesin this area meant that it was probably never necessary todrag the carcass more than J00 m.

1 have been fortunate in having access lo a motion pie­ture sequence of a leopard kili taken at a dam in the cen­tral area of the Kruger Natiooal Park during the winter of1966. The incídent was observed by B. Stapelberg ofWhite River, who filmed it in 8 mm color film, from which1 have made black and white prints and selected singleframes. As is shown in the prinls (fig. 101). a leopardapprcached three juvenile impalas drinking at the edge ofthe dam and managed to overpower one of them while theothers escaped. After kiIJing the impala, it dragged it forabout 100 m to a rocky kopje, where it fed 00 it. The latterpart of the film is indistinct but shows that the leopardlicked the hair from the ventral body surface before .starting to feed on the abdomen. During the dragging, theimpala was gripped by the neck and straddted by the fronlIegs of tbe leopard.

Evidence from the Valencia feeding lair, discussed ear­lier, indicated that a mature karakul sheep weighing 30 kgand an immature goat weighing rather less were killedwhen drinkíng at the dam and were dragged about 300 mup a steep and rocky slope to the cave where they wereeaten.

Valuable information on transport distances for leop­ards in the Malopo HiIIs has been provided by Smith(1977); dala for 28 kills is given in table 51. 11 is of par­ticular ínterest that the prey was moved a good dealfarther in the wet season than in the dry season. Fifteen

•

1t,

,. ,

.,r,,.

3. 10 l. l.

Fig. 100. The area in northem India where the "man-eating íeopard orRudraprayag" killed üs victims between 1918 and 1926. Fi¡UTes on themap indicare the number of pecple kiUed at eecb locaJity. Data fromJ. Corbett (1954).

102 A Guide to the Interpretation of Bone Accumulations in African Caves

kills in the dry season were moved an average of 120 m,the range being 0-400 m, whereas during the wet seasoneleven kills were moved an average of 260 m, the rangebeing 50--1 ,000 m. It is clear tbat leopards prefer to feed indry circurnstances, so that duríng the rains they draggedtheir prey up inro the granite kopjes where they foundsbelter and better drainage. The longest drag distance wasfor a sable calf that was transported 1,000 m. Accordingto Smith, large prey was dragged straddled between theleopard's front legs, and it was often eviscerated alongthe way, the viscera being covered by brush or plantlitter.

Smith has recorded a single observation of a dassiebeing killed by a leopard; it was carried 1,600 m beforebeing eaten,

Corbett (1954) has mentioned the distances over whichsix of the Rudraprayag man-eaters victims were trans­ported. AIl but one were moved no more than a fewhundred meters, the exception being an adult man whowas dragged 2 mi (3.2 km) up the slope of a well-woodedhill and then the same distance down tbe other sirlethrough dense scrub jungle, a total of about 6.4 km. It isof interest that on two occasions the leopard took itshuman prey down 3.6 m high walls, suggesting that in thepast Ieopards would not have had difficulty in taking theirkills down into somewhat inaccessible caves.

In conclusión, what evídence we have suggests thatprey would normally not be transported more than a fewhundred meten to a feeding lair, although distances ofupto 6 km are not impossible.

Food Storage by Leopards

In many parts of their geographic range, leopards storesurpIus food. The habit seems best developed in areaswhere hyenas and other scavengers are troublesome, andmy own observations, as well as those of others (e.g.,Pienaar 1969), suggest that leopards are frequently sub­ordinate to spotted hyenas in competition for a carcass.This being so, unless a leopard can promptly take its preyinto a tree or other inaccessible spot, it risks Iosing it tohyenas. In the Satara area of the Kruger NationaJ Park,all but one of the twelve impala kills I studied were storedin trees, while one or more spotted hyenas waited hope­fully below (tig. 102). In the one exception. the impalawas dragged into a lbicket but was stolen by hyenas be­fore the leopard could eat it.

A variation in this pattem is found in the ValenciaRanch area, where hyenas are absent and leopard fcod isabundant. Here, at the time of my observations, leopardsnever used trees for food storage, although they did usesecluded feeding laírs. Frequently they failed lo retumto prey carcasses, simply killing again when the werehungry.

The habit of storing food in prolected places and re­turning to it 00 severa! occasions is clearIy economicaI.The weight of leopard prey frequently exceeds what canbe consumed al a single sitting, Tumbull-Kemp (1967)quotes fifteen cases in which the meat consumed by asingle leopard in a 12-hr period has been measured. Meatweights eaten were found to range from 18 to 39 lb (8.2­17.7 kg), with a mean of 28 lb (12.7 kg). Depending on thesize of the prey, therefore, feeding duration will vary. Inlhe Matopos, Srnith (1977) found that an adult leopardwould consume a full-grown impala in fOUT days, and ayearling wildebeest lasted six days.

A well-documented case of a leopard retuming to itskili 00 successive nights was provided by Schaller (1%7)from observauons in the Kanha National Park of India.The prey was a 72 lb (32.7 kg) goat, killed on the night of9 July 1964 and eaten during four consecutive nights, 00the tirst night the leopard fed for 60 min and ale I7.0 lb(7.7 kg); 00 the second night 9.5 lb (4.3 kg) was eaten in 40min; on the third night, 10.5 lb (4.7 kg) in 65 min; and onthe fourth night 2.0 lb (0.9 kg) in about 20 mino Thismeans thal a total of 39.0 lb (17_7 kg) was eaten in about185 min by a leopard estimated to weigh 100 lb (45 kg).

Food storage in trees aboye a cave, 01" within the caveitself, may have been of special significance in the case ofSwartkrans, to be discussed in chapter 10.

Leopard Damage 10 lmpala Prey Skeletons

Observations on eleven impala carcasses, stored and fedon by leopards in the Satara area of the Kruger NationalPark, suggested a pattem of skeletal damage that maycharacterize antelope prey of tbis size subjected to re­peated leopard feeding sessions. The following extraetfrom my field notes describes the feeding process at oneof the kills:

16Ju/y 1968. A freshly killed female impala was seenin a tree on the east bank oftbe Timbavati River, 14km upstream from the Timbavati rest spot. A largoadult leopard hadjust started lo feed when dis­covered during the late aftemoon.

17 Ju/y, 8:50 A.M.: The kili has been moved lo anothertree, 30 m south ofthe first ene, and is being caten bya second leopard, considerably smaller than the tirst.The leopard is eating from the back of the head. Novertebral column or body remains, only a large area ofskin linking the head with one front leg (the other hasfallen) and both hind legs, hanging down on strips ofskin. In the front Iegs the shafts of the radius/ulnahave been chewed through, as have both tibia shafts inthe hind Iimbs.

9:30 A.M.: The leopard hasjust eaten the ears and isbusy on the throat, but the skull and mandible are ap­parently undamaged.

/0:30 A.M.: The head hasjust fallen and been taken bya waíting hyena. The leopard has climbed down thelree and departed, leaving onIy !bree legs of the impalahanging from strips of skin over a branch about 10mup. These were collected and photographed.

These remains are shown in tigure 103. The feedingprocess described here is atypical in that two leopards,inslead of the usual single individual, fed on the car­cass, thereby reducing the typical consumption time con­siderably.

In undisturbed circurnstances the leopard's feeding inall the observed cases was neat and orderly, resulting inthe "eaten-out" appearance oí the carcass that variousauthors (e.g., Tumbull-Kemp 1967) have remarked upon.The body, neck, and upper Iimbs are eaten, so to speak,out of the skin, which is left surprisingly Intact, linking therejected skeletal parts----the head and lower leg segments.ldeaUy, it should be possible to collect these rernains as aunit at the end of the third or fourth day, when the leopardfinally leaves the carcass. In practice, the head and sorneofthe limbs typically fall, partly as a result ofputrefaction

104 A Guide to the lnterpretation of Bone Accumulations in African Caves

11! 'i

¡, I

of the linking skin, and are snatched up by waitinghyenas. In one case we decided to remove and examinethe carcass befare the head fell and was lost to hyenas.The kili was a juvenile impala that, in September 1968,had beco stored 8 m high in the fork of a marola tree nearSalara. At the end of the second day we climbed the tree ,put the leopard to flight, and recovered the remainsshown in figure 104. The linking skin was artificially re­moved befare the photograph was taken to show thatthe vertebral column had disappeared between the axisand the lumbar vertebrae, while both humeri had becochewed through just aboye their distal extremities. Hadwe left the carcass for another night, the rest of the ver­tebral column and upper limb bones would almost cer­tainly have been eaten by the leopard, but the head wouldvery probably have fallen to a waiting hyena.

The Possible Role o/Extinct Cats as Collectors o/Ronesin Southern Afrícan Caves

Cats that may be significant from this point of víew. andwhose remains are preserved in the Sterkfontein valleycave breccias, are these: the leopard; a related extinctfeline, Panthera crassídens: and the false and truesaber-toothed cats.

Remains of leopards frorn Swartkrans have been de­scribed as a separate subspecies, Panthera pardus incurva(see ehap. 9), slightly smaller than typical living Ieopards.However, there is a good deal of size variation in P. par­dus over its wide geographic range. There is no reasonto suspect that (he leopards that lived in the Sterkfonteinvalley two million years ago behaved differently fromtheir contemporary descendants.

Remains of P. crassídens, a cat that seemed to havehad characterístics of both a leopard and cheetah (seechap, 9) are known from Kromdraai A and also from vari­ous localíties in East Africa. What little infonnation thereis suggests that we may be dealing with a leopardlikefeline in which cursoriaJ adaptations were emphasized.Whether the "crassidens cat' was a regular frequenter ofcaves is not known. If it was, then it was presumably farless common than leopards were. Food remains of thetwo eats would probably have been similar, although thecrassidens cat may have actively pursued its prey ratherthan taking it by stealth as is usual with leopards.

False saber-toothed cats of the genus Dlnofeíís wereprobably of singular irnportance as accumulators of bonesin the Sterkfontein valley caves. As is described inChapter 9, remains from three species are known fromsouthern Africa: D. diastemata from Langebaanweg, D.barlowí from Sterkfontein, Makapansgat, and Bolt'sFarm, and D. píveteaut from Kromdraaí A. The threeappear to have formed an evolutionary sequence, with D.diastemata showing the most primitive dental charactersand D. piveteaui the most advanced. The indícations arethat Dlnofeíts was a heavy-bodied feline with a short tail;as is shown in figure 105, the eraniums of both D. bar­lowi and D. piveteaui are appreciably larger than those oflarge male leopards.

The frequency with which Dinofelis remaíos are foundin the TransvaaJ cave breccias suggests that these werereasonably common. cave-frequenting cats. Their bodilyconformation suggests that they would have relied onstealth. rather than sustained speed, to capture their prey;their behavior was probably similar to that of leopards.However, their appreciably larger size means they would

have been able to overpower larger prey--dass 111 an­telopes would have been more readily accessible to themthan they are to leopards.

For what purpose did Dinofelís use its enlargedcanines? They were clearty highJy effective killing weap­ons and, in combination with the powerful fore­limbs, would have enabled Dínofeíis to hunt not onlylarge, but also dangerous prey. 1 am tempted to suggestthat Dinofelís would have been a very efficient hunter ofprimates such as baboons and early hominids. Individu­ally, a baboon or australopithecine may not have been aparticularly dangerous adversary, but an attack on an in­dividual would probably have precipitated retaliation bythe entire group. In sech circumstances, sílent, efficientkilJing would have been a great advantage to the predator.One may visualize Dínofelis biding íts time in a concealedposition while a troop of baboons or hominids passed:selecting a straggler, the cat would then overpower itsprey, holding it down with its powerfuJ forelirnbs whilekilling it silently and quickly by a throat bite.

The suggestion that Dínofelis frequented caves gainssorne support from fossils discovered in Pit 23 of Bolt'sWorkings (see chapo 9). Here rernains of three Dinofelisindividuals were preserved alongside the remains of sev­eral baboons. It seems that the cave represented a naturaltrap into whieh the baboons had fallen. The eats followedand probably fed upon the baboons, only to díscover thatthey could not escape either.

Remains of true saber-toothed cats fall naturally intotwo groups-those with crenulate sabers and highly spe­cialized carnassials (Machoírodus and Homotherium) andthose with smooth-edged canines and less specializedcheek teeth (Mcgantereon). These animals are furtherdiscussed in chapter 9. From the regularity with whichtheir remains are found in Transvaal cave deposits,saber-toothed cats appear to have been both fairly com­mon and prone to frequent caves. Their size and formida­ble eanines suggest that they probably preyed on largeanimals, though the structure of their cheek teeth makes itclear that they would have done minimal damage tobones. It is quite conceivable that Transvaal saber­toothed cats used caves as feeding and breeding lairs andthat their food remaíns should have accumulated in suchplaces. Such rernains would be characterized, 1 suspect,by the large size ofthe animals involved and by the lack ofdamage to the bones themselves.

The Blaek Eagle, Aqui/a FerretJUXÍ Lesson

Black eagles probably never enter caves, but they do con­sume very large numbers of dassies, whose remains ac­eumulate below their nests and habitual perches. If eitherthe nest or the perch is within the catchment area of acave mouth, black-eagle food remains wiII form part ofthe bone accumulation in the cave.

The black eagle is a most spectacuJar bird, coa! blackwith a white back that fonns a white Vwhen the wings areclosed. Aeeording to Leslie Brown (1970), its graee inflight is surpassed only by thal of the Lammergeier. Inmany ways it is very Jike the golden eagle. Aquilachrysaetos, which it replaces in Africa south oC the Sa­hara. Its usual habitat ineludes rocky gorges and kopjeswhere dassies abound, and in Kenya and Ethiopia itbreeds at altitudes ofup to 13,500 ft (4,117 m).

The biology of black eagles has been studied in greatdetail in various parts of the birds' geographic range.

106 A Guide to the lnterpretation of Bone Accumulations in African Caves

After sorne early breeding observations in South-WestAfrica by Hoesch (1936), Rowe (1947) located a conven­iently placed nest on a rocky hill near Mbulu in north­central Tanzania. A crevice on the nesting cliff allowedhim to observe at a range of 21 m, and he watched theeagles for a total of 1,013 hr. The first egg, laid on 13 June1944, hatched on 27 July; the second. laid on 16 June,hatched on 29-30 July, and the surviving eaglet flew on1 November, which meant a laying-fledging period of 140

Fig. 105. Slculls of two species of the exlinct cal Dinofelis compared withthe skull of a large male leopard. (a) D. barlowi rrom Boll's Farm; (b) D.pjveleaui from Kromdraai A; (e) a modero Panrhera pardus.

Pig. 106. Distribut¡on of the black eagle. Aquila verreauxi. Data kindlyprovíded by D. Snow, Brilish Museum of Natural History.

days, Rowe described the process whereby the secondchick to hatch was pecked and harried by the first until itdisappeared from the nest on 6 August; he ascribed itsdeath to maltreatment, starvation, and exposure. Rowespeculated about the biological significance of this "Cainand Abel" process and suggested that, since the firstchick is stimulated by the presence of the second to takemore food than it needs, a more robust eaglet results.

During Rowe's observation of the nest, 41 carcasseswere brought-24 mammals and 17 birds. Among themammals were 22 dassies, 1 hare, and 1 mongoose, andthe birds were francolins, guinea fowl, and poultry. Abouthalf the mammals had been decapitated and evisceratedbefore being brought to the nest, a process that may berelated to the adults degree of hunger. Rowe observedthat the eagles ate most of the bones of the prey, so thatonly skulls and larger skeletaI elements accumulatedbelow the nest.

Further observations on black eagle nests were made inthe Embu District on the eastem slopes of Mount Kenyaand elsewhere in East Africa (Brown 1952, 1955, 1966), atJansenville in the Orange Free State (Visser 1963), in theMatopo Hills of Rhodesia (to be discussed shortly), andelsewhere, allowing Siegfried (I968) to collate 99 breedingrecords' for sourthem Africa. These showed that, over aconsiderable geographic range, most eggs were laid inMay and JUne, the usual clutch being two, thoughclutches of one and three were occasionally encountered.

Fig. 107. A black eagle, photographed in the Hakos Mountains of South-West Africa by A. C. Kemp. '

Almost invariably the ñrst-hatched chick killed the sec­ond or subsequent ones, and about this Síegfried wrote(1968, p. 144): "the apparent biologicai waste involved inthe two-egg clutch, followed by the killing of thc onechick by its sibling partner, is 001 understood."

Sorne experiments designed to throw light on the "Caínand Abe!" struggle were carried out at black eagle nestsin the Matopos by V. Gargett (1967, 1970a). She con­cjuded that it was not shortage of food that led to thesurvival of one rather than two chícks; it appeared thatthe proxirnity of a second chick elicited an aggressiveresponse as marked al six, seven, and eight weeks as itwas in the first few days after hatching. The experimentsdemonstrated that there was no reason the second chickshould 001 grow to maturity if shielded from the aggres­sion of the first.

The Matopo HiIIs of Rhodesia, where the experimentswere carried out, are remarkable for black eagles, prob­ably supporting an eagle population more dense than anyother in the world. For many years it has been known thatblack eagíes are particularly abundant in the Matopos(e.g.,Mouritz 1915), a spectacular area of granite hiDs southof Bulawayo (see also chapo 3). By 1960 lhe positions of35 black-eagle aeries had been detennined in the MatoposNational Park, and this informatíon was the basis for asurvey of black eagles in the park (Vernon 1965). Fromthis survey, conducted in 1964, Vernon was able to con­elude that there were then 41 nest sites in 160 mil of theMatopos, that one breeding pair occurred in every 5.3mil, although the actual population density was evengreater, and that this density was limited by the need foreacb pair of eagles to have a territory large enough todisplay in, The suitability of nest sites was determinedmore by security considerations than by aspect, althoughit was found that adjacent nests that overlooked oneanother had to be more than a mile apart or one pairwould inhibit the breeding of the other.

This 1964 survey has been continued by an ornithologi­cal team under the leadership of V. Gargett and has re­sulted in an impressive body ofinformation on the biologyand population dynamics of the eagles (Gargett 1965,1967, I97Oa.b. 1971, 1972a,b, 1975,1977). Witbin the 620km2 study arca, and over the thirteen-year study period,442 breeding attempts were recorded and 339 youngeagles seen to be f1edged, a reproductive success of 0.52young per pair per year, Tenitory size vatied from 585.4ha to 1,437,6 ha, depending nn the nature of the terrainand the availability of food, and tenitory shape in theairspace appeared lo be an inverted truncated cone,which resulted in airspace overlapping aboye ground­level boundaties.

lt is of particular relevance to the present stndy of boneaccumu1ations that about nine months of a black eagle'syear would nonnally be taken up with some part of thebreeding cycle, centered on tbe nest, Apart from this,habitual perches are used, whose positíons, relative to thenests in eighteen adjacent territories, are shown in figure108, Food remains could be expected to accumulatemainly below nests, but also beneath these perches. Preywas almost exclusively dassies, many individuals havingbeen decapitated by the eagles at a perch before theirbodies were taken to the nests. Gargett (1971) quotedseveral observations of actual kills. It appears that theeagles oñen fly low around the rocky outcrops, exploitingsurprise, and knock the prey off the rocks with their out­stretched talons. They then circle and pick up the bodies.

Food Remains of Carnivores in African Caves 107

Through the cooperation of J. H. Grobler of the De­partment of National Parks and Wildlife Management inZimbabwe, 1 have been able lo study a sample of blackeagle food remains collected below nests and perches inthe Matopos. Details of the 111 bones are given in table52; they carne frorn a minimum of 8 adult common das­sies, Procavta capensis, 40 yellow-spotted dassies,Heterohyrax brucei (39 adult and subadult, l juvenile),and 1 hinged tortoise, Kinyxis belltana. The three plastronpíeces of this tortoise were in a regurgitation, surroundedby dassie fur, and it is remarkable that the largest of thesepieces was 7.5 by 8.5 cm-Iarger, 1 would have thought,than an eagle could swallow and regurgitate. Yet onecertainly did so.

I have been able to examine two other collections madebelow black eagle perches, The ftrst was from theMagaJiesberg at Wonderboom, within the municípal lirnítsof Pretoria, The collection was made in 1969by O. P. M.Prozesky and W. Spofford; it consisted of 56 píeces, aslisted in table 53. Common dassies, Procavia capensis,dominated the prey and ineluded 8 adults, 5 subadults,and 4 juveniles. Parts of 2 bares. a mountain tortoise, and2 francolin-sized birds were also present.

The second collection, consisting of 131 bone pieces,listed in table 54, was made by A. C. Kemp beneath ablack eagle perch close to its nest on the Portsmut Farmin South-West Aftica----a locality already described inconnection with leopard lairs----in June 1969. Representedby these remains were 9 adult, I subadult, and 2 juvenileProcavia capensis, as well as 1 hare,

The three collections of food remaíns clearly confirrnearlier observations that the main prey of bJack eagles isdassies, either Procavía or Heterohyrax. Other marnrnals,

o Nuh

o

N

t

.,~

\ lO.,

Fig. lOS.Tenitories of eighteen pairs of black eagles in the Matapos areaoCZimbabwe. In each case, habitual perches are linked lo nest site5. FromGargett (1917).

108 A Guide to the Interpretation of Bone Accumulations in African Caves

birds, and tortoises are of Jimited importance. The foodremains of dassies are mainly cranial parts, together withpelvises and a few of the larger Iimb bones. The skullsshow very characterístic damage: in most instances thebraincases have been opened from the back or the side toremove the brain. In sorne cases the en tire calvaría hasbeen eaten away. Damage is also common on the anglesof the mandibles, probably inflicted when the tongue wasremoved from behind or below.

The damage to skulls, as shown in figure 109, has been

inflicted by the poínts of the black eagles' sharp recurvedbilis. Frequently a round opening ís made in thebraincase-an ínterestíng difference from the "sheared­off" damage to dassie skulls caused by leopard or cheetahfeedíng.

If dassíe skulls that had been discarded by black eagleswere found in a fossil bone accumulation, it should bepossible to separate them, by the nature of the damagethey have suffered, from remains [eft both by felid pred­ators and by human hUnters.

Fig. 109. Dassie skuUs coUecled benealh black eagle perches and nesl sites, showing characlerislíc damage 10 lhebr:ilncases.

5 Porcupines as Bone Collectors in African Caves

It is rather surprising that a vegetarían rodent like Hystrixafrícaeaustralis should prove to be an important collectorof bones; yet 1 suspect porcupines carry more bones toAfrican caves than does any other species. As long ago as1934, Shortridge wrote in bis aceount nf tbe mammals ofSouth-West Africa: "The porcupine has a well-knownhabit oí gnawing bones, accumulations of which afienbetray the whereabouts oí a warren." He went on toquote earlier accounts by Lydekker and Pitman aboutthe gnawing abilities and collecting habits oí Africanporcupines.

Attention was specifically focused on this strange be­havior of porcupines during the 1950s when R. A. Dartand A. R. Hughes were trying to make sense oí the re­markable bone accumulation in the gray breccia alMakapansgat Limeworks (Hughes 19540; Dar! 1957a).About that time Hughes started bis observalions of con­temporary animallairs, including those of porcupines. InSeptember 1954 he visited the Kalahari Gemsbok Na­tional Park in tbe nortbern Cape and examined severallairs in the banks of tbe dry Auob and Nossob riversthere. Two of tbese were clearly inhabited by porcupines,and from tbem Hughes (1958) took 147 bones and otberobjecls that the porcupines had apparently hoarded. Bothsites were in lhe east bank of tbe Auob River, 17 and26 km upstream from tbe main camp at Twee Rivieren.

Sorne direct informanon on porcupine hoarding be­havior came to light about the same time. In January 1956a young male porcupine was caught near Grahamstownand taken to lhe Zoology department of Rhodes Univer­sity, where it was observed ayer a considerable perlad(A. J. A1exander 1956). The porcupine, named Aristotle(fig. 111), made its lair in the photographic darkroom andwandered out freely at night. Before long it started lostock its lair witb objects collected during its nocturnalwanderings-bones, a tortoise carapace, "a wicker bas­ket, an enamel dish, a duster, a piece offlex, a drain pipe,a buck horn and a piece of softboard." Manageable ob­jeets it simply canied in its mouth, and larger ones itdragged, walking backward all tbe way to its lair.

Of particular interest was Anne Alexander's observa­tion that the porcupine showed no interest in fresh boneswith meat on them, though dry bones were immediatelytaken lo tbe lair and gnawed there. Tbe purpose of thegnawing was soon apparent: it had lo do with wearingdown the incisors rather than with nutrition. Like otherrodents, the African porcupine has open-rooted lncisorstbat grow throughout its life (fig. 11 Z); they require regularattrition lo keep them al a usable length. So it seems that

porcupínes have developed a behavior pattem that re­quires them lo collect dry bones and other hard objectsand hoard thern in their lairs. When resting during theday, the porcupines select sorne of their favored objectsand gnaw them. The coUecting behavior appears to havebecome a compulsion-tbey will bring back far more ob­jects than they can possibly use and do not get around tognawing anywhere near all the treasures they collect.

Marks 00 bones caused by porcupine incisors arehigbly characterislic (fig. 113), as are the end products ofprolonged gnawing 00 various parts ofbovid skeletons. Inher dissertation on various aspects of fossil and livingporcupines, Judy Maguire (I976) has considered the na­ture and distribution of gnawing marks on differentskeletal elements. Her conclusions may be summarizedthus:

Craniums:Gnawing tends to be random and widespread.Mandibles:Tbe symphyseal región and ascending ramus tend lobe gnawed first; thereafter the horizontal ramus isgnawed, which rnay cause the teeth to fall out. Theangle of the rnandible tends to remain till last.Vertebrae:Neural spines and transverse processes are gnawed

Fig. t 10. Distribution of the African pcrcupine , Hvstrix africaeaustralls,

109

¡lO A Guide 10 the Interpretation of Bone Accumulations in African Caves

Fig. 11 I. A lame Africanporcupine, Aristolle. photographed jn [he Zoology Departmenl ofRhodcs UIÚversilY during1955. This animal provided the first clea¡ evidence of bone-hoarding and bone-gnawing.

Fig. 1(2. The skull of an African porcupine showing jts prominent incisors, which grow throughout Jife and requíreregular atlrition.

first, thereafter the zygapophyses and centro. An ir­regular and unrecognizable fragment may result.Ribs:The ends and body are prone to gnawing, Damagealong the anterior and posterior edges causes ribs tosplit 10ngitudinally, producing thin slivers of bone inwhich the porcupines show little further interest.Scapulae:The compact head tends to disappear first, and trunpieces of blade, of little interest to the porcupines, areleft.Pe/vises:Gnawing proceeds frorn all sides, usually leaving theacetabular regíon tilllast.

Porcupines as Bone ColIectors in African Caves 111

Long bones:Proximal and distal ends of the various limb-bone ele­ments tend to be destroyed first, and the gnawing thenproceeds down the shafts from each end, to result incharacteristic pieces shaped like napkin rings, withgnawing scars around each circumference.Asrragali, calcanei, and phalanges:Gnawing tends to occur on all surfaces of these bones.

An interesting comparison be'tween a porcupine collec­tion and abone assemblage from a human occupation sitehas been made by Hendey and Singer (1965). Construc­tion of a dam and canals on the Kougha River in theGamtoos valley near Port Elizabeth necessitated the de-

Fig, 113. Examples ofporcupine-gnawed bones, The marks caused by the chisellike íncísors are highly characteristic.

SKETCH PllN Of THE

NOSSOS PORCUPINE lAIR1

clean ground well consolidated by the porcupines thathad lain there sleepfng.

1 have been able to visir the site twice-s-in March 1968when all bones and other objects were again removedfrom lb.e lair, and in June 1969, wlten the skelch planshown In figure 114 was drawn. The site itself and theabundance of bones inside its entrence are shown infigures liS and 116. Besides tbe main entranee al A on theplan, the lair has three openings to the surface, The roof islow, varying from 4S cm lo 60 cm, a situation that makes~etrieva1 of bones diffícult. Two porcupines were visiblel~ .the farther recesses of the lair at the time of myvisits-s-they shock their quills in a threatening manner butwere not otherwise troublesome. More discomfort wascaused by the voracious tampan ticks that emerged inlarge numbers from the dusty substrate of the lair whilewe were creeping around inside it.

The 1956 collecíion, made by Alun Hughes and kindlyput at my disposal, consísled of 1,328 pieces as listed inlable 56. By 1968 380 more objects were available forcoDection, making a total of 1,708. Tbese were alI bone,born, or tortoise sbell, except 71 píeces of wood and 17metal objects (iron bars, rusty cans, enamel mugs, and thelike). As is reñected in table 57, the remains in the wholecol1ection carne from a minimum of 106 individuals:bovids such as springbok, gemsbok, hartebeesl, and wil­debeest were particularly well represented on the basís oftheir homs.

The study of the bone assemblage from the Nossob lairhas províded ínformatíon on a number of questtons:

What ls the Rate o/ Accumulation of Objects in a Por­cupine Lair? The Nossob lair was cleared by Hughes in

Fi,. I ~4. Plan oC. and section tbroush. the NO$sob porcupine lair. ThisextenSlve .hut low-roofed cave ha, formed in lhe cakrete bank of theNossob River. Kalabari National Park.

The lair was situated in one of a number of inter­communicating solution eavities in the calcrele at thetop o( the east bank of lbe Nossob R. 8 miles from thebase camp at Twee Rivieren in lbe soulb of theKalabari Gemsbok National Park. In the &Ontoftheentrance was a 72 sq. ft. area ofloose graund 6 ins.-Ift. in depth whiclt contained 411 porcupine quiDs andmany very small hairlike quiDs, gnnwed bones, skulls~d homs of largo and smaU animals, 9 tins, pieces oftron, bottles. 105pie<:Cs ofwood up to 3 ft. in 1enstJ>and other articles coDecled by poreupines. Tbe insideo( !be lair was nbout 425 sq. ft. in extent and laperedaway from the entrance for 36 ft. to a smalI opening inthe calcrele suñace superadjacent to the river bank.00 the inside oCtbc lair I especially the sides near theentrance, were scattered the bones, skuDsand homsof animals whieh had died during the Iast 25 years aswas indicaled by the presenee of the homs of domesticrams and cows; domeslic catlle bad not Iived in lbeneighbourhood since 1932 when the park was pro­claimed. Inside lbe lair was a circular raised piece of

I12 A Guide to (he Interpretation of Bone Accumulatíons in African Caves

molition of three rack shelters in 1961 and 1962. Two oftbese, relevant here, were on the west bank of the Gam­toos valley al Andrieskraal. The ñrst. designated An­drieskraal 1, coeststed ofa smaJJ rack shelter containing ashalJow deposit rich in stone artifacts and bone fragments(J. Deacon 1965). The second, Andrieskraal 2, a fewhundred meters downstream but al the same height aboyethe river t consisted of a crevice in the Enon con­glomerate, too low to have been occupied by man butcontaining a remarkable collectíon of gnawed bones. Thesignificance of these two bone accumulations Crom adja­ceot sites is clear: a direct comparison can be made be­lween abone assemblage resulting from Later Stone Agehuman feedíng and one collected by porcupines, The"h~man site;" AK 1, yielded 11,056 bone pieces, ofwhich 8,887 (80.4%) were too frngmentary lo beidentified; the "porcupine lair," AK 2, produced 1,105pieces, of whicb only 465 (42.1%) were unidentifiable.Tbere was therefore a striking difference in frngmentationof the bones-tbis is also reflected in the weigbt of eacbsample. The 11,056 bone pieces from AK I weigbed 20.9kg, givíng an average weigbt per piece of 1.9 g; the 1, 105pieces from AK 2 weigbed 8S.3 kg, an average weíght perpiece of 77.2 g. Ukewise, only 0.3% of the identifiedbones from AK I sbowed signs of poreupine gnawíng,,,:hereas 60.0% of those from AK 2 did so. The strikingdifferences between these two bone accumulations areparticularly significant becanse the rango of animalspecíes represented in the two collections is very similaras is reñected in table 55. '

A porcupine lair 1 have been able lo study in sornedetnil is in the easl bank of tbe Nossob River, 12.9 kmupstream from Twee Rivieren in the Kalahari NatíonalPark. Analysis of íts bone accumulation bas provided in­sights into a number of aspects of colleetions made byporcupines.

The Nossob Poreoplne Lair and lts Evidence

The sile was lirst visiled by A. R. Hughes in 1956; on 21May of thnt year be removed from it 1,420 bones andother ~hjects thnt appeared to have becn coUecled byporeupmes. At the time, Ite made tbe foUowing notes(Hugbes. pers. comm.):

Il,[

ji

Porcupines as Bone Collectors in African Caves 113

Fig. 115. The main enlrance lO the Nossob porcupine lair.

Fig. 116. The interior of Ihe Nossob lair showing abundanl bones Ihat ha!! been coUecled by the residenI porcupines.

114 A Guide to the Interpretation of Bone Accumulenons in African Caves,¡

t i

!!., 1.., i¡i 'l,!,

¡1

May J956 and again by me in March 1%8, giving a timelapse of alrnost twelve years. During this period 380 ob­jects were taken to the Iair, arate of about 32 per year. 1was able to observe two porcupines in the lair on twooccasions, but there rnay wel1 have been others in thedeeper recesses that wcre out of sight. The rate oC collec­tion of bones would depend on their availability: 1 suspectthat the Nossob loeality would generally províde anabundance of bones for porcupines to collect,

The only other estímate of accumulation Tate of which 1am aware was made by Maguire (1976) at the Hartebeest­hcek porcupine lair west of Pretoria. The Iair, carefulIycleared and reexamined after precisely one year, wasfound lo eontain 44 gnawed bones, 5 gnawed branehes, 6gnawed mealie cobs, and 2 gnawed stones-a total nf 57objects, whieh is slightly higher than the figure for theNossob lair.

Do (he Bones in. the Porcupine Coílection Reftect theNatural Abundanee ofAnimal. from Which They Come?It is fortunate that the Nosscb lair is situated in theKalahari National Park-an area for which census dala onthe contemporary bovid populations are available. Theseunpublished data have been kindly provided by G. DeGraaff (De Graaff, Bothma, and Muolman, unpublished)of the National Parks Board. The eensus figures forspringbok, gemsbok, hartebeest, wildebeest, eland,steenbok, and duiker are given in table 58 and may becompared with the minimum numbers of individuals re­flected in the porcupine collection. The relevant numbersfor tbese species are springbok, 40; gemsbok, 15; harte­beest, 9; wildebeest, 5; steenbok, 5; duiker, 2. Elandhas not been identified from cranial remains, but atleast 3 individuals of an eland-sized bovid are representedin the posteranial parts.

Percentage abundanee of the various species is plottedin figure 117, from which it is evident that the minimumnumbers of individual animals represented by theporcupine-collected remains do indeed rnirror the actualabundance of the antelope species,

An interesting point here is that, although morespringbok are represented by their homs !han are lacgerantelopes, the posteranial remains of the larger bovids aremore numerous iban are those of their smalJer counter­parts. For convenience, southem African bovids havebeen divided into four size elasses (table 1) according toweight: elass 1, up to about 23 kg (50 lb); 11, 23-90 kg(50-200 lb); I1I, 90-295 kg (200-650 lb); and IV, aboye295 kg (650 lb). Springbok fall into elass 11; gemsbok,hartebeesl, and wildebeest, into elass III.

Table 59 lists the bovid skeletal parts in the 1968 col­leetion from the Nossob lair. On the basis of hnms, the1968 eollection contaíns parta oí 15 clase 11 antelopes(springbok) and 9 from class III (gemsbok, hartebeest,and wildebeest); nevertheless, elass Il antelopes are rep­resented by 132 pieces and class 111 by 183. The reasonfor this discrepaney is probably that the homs of almoslevery antelope kilIed survive and may be eolleeted byporcupines, but that much lees is left of the skeleton oCa smaller antelope afier predator and scavengers arefinished.

What Parts 01 Bovid Skeletons Are Represented, and Arelhese the Most Resistam Elements in (he Skeleton? Table60 lists the parts of bovid skeletons represented in the twoNossob colleetions. There is wide representation of

almost all skeletal parís. On the basis of horns, at least 81individual antelopes appear to have contributed to thesample. Working with this total. it is possible to calculatehow many individual parts of the skeletons there shouldhave been if none was lost and al1 were coUected. Thesefigures are given in the third column of table 61. Surpris­ingly enough, fol1owing pelvis pieces, atlas and axisvertebrae have the highest survival figures, whereascaudal vertebrae have the lowest. Following through thevertebral colurnn, the sequence oí percentage survivalfrom highest to lowest is as follows: atlas (30.9%); axis(28.4%); other cervical (18.8%); lumbar (15.6%); sacra!(8.6%); thnracie (5.0%); caudal (0.1%).

These survival figures very probably reñect the robus­ticity of the different vertebra! types but may also havebeen inñuenced by the porcupines' selective preference.Perhaps porcupines simply prefer a substantial-looking,ehunky cervical vertebra lo a scrawny-looking thoracieone with íts long neural spine.

The individual parts of a bovid skeleton vary consid­erably in strength and in resistance to destructive treat­mento Under a given destructive regime, the individualparts will survive in proportíonto their robusticity. Studyof the survival of goat bones subjeeted to Hottentot anddog feeding (chap. 2) has shown which parts of the skele­ton are best able to withstand destructive treatment.Table 5 Iists the various parts of rhe goat skeletons andprovides figures 00 their percentage survival. Horns hadthe highest survival, followed by mandibles, maxillae,and distal humeri. Proximal humeri and caudal vertebraewere found to have a ni! survival value. Survival may beeorrelated with the compactness of tbe bone, expressedas specific gravity, and witb fusion times of the epiphysesof the long bones. In fact, the survival of parts will followan entirely predictable pattem if the destruetíve in­fluences are known.

o 1020 ICI <1050 o lO20IO.cJloeo

"" ü ....... _ '% l""'" ~~oll_l..

Fig. 117.The histogntm 00 tlle left aboWI tlle perceota&e abundlUlCC ofvarious marnrnals in tile Kalahari National Park. according to NationatParksBoard census data; that00 thc rigbt reflects tbe percentage abun~

dance of me same species as representedby bones collected by the por·cupines.

Porcupínes as Bone Collectors in African Caves 115

In figure liS the percentage survival of bovid skeletalparts in tbe Nossob sample is plotted and contrasted withthat for the same parts in the Hottentot goat-bone sample.The Nossob survival sequence does not folJow the Hot­tentot pattern; to my mind it does not follow the patternthat could be predicted for skeletons subjected to a gooddeal of destructive treatment. For instance, the ratio ofproximal to distal humeri in the Hottentot sample is O: 82;in the Nossob sample it is 9: 20. It is perfectly clear thatthe bovid skeletons in this par! of lbe KaJahari Park havenot been subjected to extreme scavenging pressure;complete bones frequently survive, as do delicate ele­ments. The residue left for porcupines to collect thereforedoes not necessarily eonsist oí only the most resistantelements.

Do Porcupines Preferentiaítv Collect Objects 01 a Par­ticular Stze and Weigh/? To answer tbis question, eachindividual object in the Nossob collection was weighedand measured-lbe greatest length per piece being re­corded in inches. The results given in tables 62 and 63 areplotted in hislogram form in figure 119. As far as weight isconcemed, the greal bu1k of the sample fa1ls in the (}-50g category, though sorne objecls up lo 750 g are in­cluded in the sample. In length, the bulk of the objectshave máximum dimensions of between 1and 6 in, thoughsorne up to 36 in (mostly gemsbok horns and pieces ofwood) do occur.

Do Porcupines Preferemíaiy Gnaw Objects 01 a Particu­lar Size? Tab1es 64 and 65 give the percentages of gnawed

bones in each weight and length class. It is immediatelyapparent (fig. 119) that the percentages of gnawed bonesin the smallest weight and length classes are IOIVa thanthose in the larger categories. The conclusion, therefore.is that, although porcupines colleet large quantities ofsmall bone píeces, they prefer to gnaw the larger ones. Itis probably more difficult for a porcupine to hold a smallobject between its forepaws while gnawing than to hold alarger one.

Are the Coílected Bones Generally Fresh or Weathered?Any curator of an osteological collection will have foundthat, unless bones are defatted before being stored, theywiU continue to exude grease indefinitely. An interestingpoint about the Nossob porcupine collection is that only avery small number ofbones (no more than 15out ofa totalof 1,620) showed appreciable traces of fattiness. 1 havelittle doubt that if the bones had been brought lo the lair ina fatty condition, this fattiness would have remaíned in­definitely in the shaded conditions of the lair. 1 formedthis conviction as a result of a bone-weathering experi­ment 1 set up more than ten years ago at the Namib DesertResearch Station in South-West Afriea. Two series oífresh bones were placed in a wooden frame; one serieswas fully exposed, protected only by wire netting. Theotber series was shielded from direct sun by surroundingscreens of \iI-in pegboard with holes I in aparto The framewas set up in a fully exposed situation on 1 December1965; effects were observed at yearly íntervals, and sornespecimens were removed on 21 March 1973 for photog­raphy, after a time Iapse of seven years. Two halves oí apig mandible showed striking differences (fig. 120). The

..'".. ...•os

....~"""T"""T",~~ 3 4 5 I J a • lO 11 Il 1) 14 15 • i"~hn

§ 10 I§ 20 25 » » ~ cms

"

.."

80 20 40 60 80'4 SUflYlyAl HOTTENTOT GOATS

02040110

NOSSOB lAIR

CAUDAL nrtatlrn

,toxillal

llj.lal

SAtRAl ~"tlb,..

RAOIUS & ULNA diftal

PH~lAIII6ES

HUMUluS proliln.l

THOflACIC nrtabr••

RIBS

Fig. 118. Hístograms showing the percentage survival of various bovidskeletal parts in the porcuptne-collecred sample from the ND5SOb lair. Thefigures are contrasted with those for the Hottentol goat bones described inchapter 2.

rll. 119. Histograms showing tbe percentaBe abundanee of bcees of dif­ferent weight and length ciasses in tlle porcuptne-conected sample fromthe Nossob lair. The proportioos of gnawed and ungnawed bones in eacbclass are also shown.

116 A Guide to the Intrpretation of Bone Accumulations in African Caves

half that had been exposed to full sun was bleached andcracked, while the other, which had been shaded from thesun in a well-ventilated situation, retained enough fatafter seven years' storage for dust to be adhering to it. Thedefatting process of the fully exposed bone was completewithin one year, and I would suspect within three months.In a shaded situation, defatting may take decades.

On the basis of these observations, I have no doubt thatthe bones generaUy brought to the Nossob lair by por­cupines are ones that have been naturally defattedthrough surface exposure. In fact, the porcupines show amarked preference for bleached, defatted bones.

Maguire (1976) reached a similar conclusion after anexperiment in which I3 ungnawed bones, varying In

freshness, were placed 5 m from the entrance of aporcupine-occupied ant-bear hole near Klaserie in theeastem Transvaal. Gradually the porcupines removed aHthese bones, but they left the undefatted, greasy ones til1last. After three months these were found to have beennaturally degreased, and they were then attractive to theporcupines.

Conclusion

The most reliable indicatíon that abone accumulation in acave has been built up by porcupines is the presence oftypical gnawing marks on defatted and frequently wea­thered bones falling within the size and weight range

Fig. 120. The long-term persistence of faL in bones is demonsLrated by Lhese Lwo halves of a pig mandible, both ofwhich were removed from a bone-weathering experiment after seven years. The half on the lefi was subjected to fullsun and was fully defatted wiLhin one year; that on Lhe right was stored in a shaded, well-ventílated siLuation andretaíned its faL after the full seven years.

suggested in this chapter. The incidence of gnawed bonesin the whole collection from any site will depend on theabundance of bones available 10 the porcupines at thetime. Where bones were readily available collections willshow less gnawing than where borres were scarce. Severalbone accumulations attributable to porcupines have nowbeen studied. Percentage abundances of gnawed bones inthese collections varied as follows:

22.0% gnawed. Winkelhoek lair, northem Natal,463 objects (Magnire 1976).

60.0% gnawed. Andrieskraal 21air, eastem Cape,640 identified bones (Hendey andSinger 1965).

61.l% gnawed. Nossob lair, 1,708 objects (Brain,this study).

69.0% gnawed. Auob lair, Kalabari National Park,287 objects (Brain, unpublished),

Porcupines as Bone Collectors in African Caves t J7

76.0% gnawed. Rand van Tweespruir lair, westernTransvaal, 106 objects (Maguire,1976).

80.0% gnawed. Kimberley lair, northem Cape, 22objects (Maguire, 1976).

96.0% gnawed. Hartebeesthoek lair, central Trans­vaal, 479 objects (Maguire, 1976).

100.0% gnawed. Hanglip lair, northem Transvaal,51 objects (Maguire, 1976).

It is perhaps of interest that abone accumulation re­marked upon many years ago by Peringuey (1911) in his"The Stone Ages of South Africa as Represented in theCollection of the South African Museum," from a lowcave al Hawston on the Cape coast, shows clear signs ofbeing a porcupíne collection. Of the 115 bones 1 havebeen able to examine from the site, 44 showed the imprintof porcupine incisors.

6 The Contribution of OwIs

Though I have devoted comparatively little attentionto systematic zoology, 1 have frorn time te time madepretty extensíve collections oí the bones of smallmammals, and the best hunting grounds 1 have in­variably found to be the haunts of owls. In the dis­gorged pellets afien found in great abundance in rackclefts the small mammal skulls are usually preserveduninjured, and the owJs frequently obtain specimenswhich the collector of skins will not readily comeacross. [Broom 1907, p. 262]

The habit of certain southem African owls of roostingand nesting close to the entrances oí caves is of consider­able taphonomic signiñcance. For it is during daytirne restperiods that the owls may regurgitate pelIets containingthe indigestible remains of their prey. These pellets ac­cumulate 00 the cave floor in considerable numbers and,in certain circumstances, become fossilized as a brecciarich in microvertebrate remaíns (ligs. 121 and 122).

In southern Africa, the owl species most importantfrom this point of view is undoubtedly the barn owl, Tytoalba (Scopoli), although certain species of eagle owl mayalso play a role in the accumulation oí microvertebrateremains within caves.

The Mechanism of P.lIet Production

As early as 1927 an American study was made of pelletformation in the great homed owl, Bubo virginianus, byReed and Reed (1928). Seven nestlings were ftuoroscopedafter ingesting barium paste or food mixed with paste. Itwas Cound that pellets were ejected from 12 to 20 hr afierfeeding and that the Cormation of the pellets was madepossible by structural peculiarities of the stomach. Thepyloric opening preved lo be only about 1 mm in diameterando in the normal position, was found 10 lie on a levelwith the opening of the esophagus so that practically thewhole stomach, even when greatly distended, lay belowthe level of the pylorus. The small size of tbe pyloricopening provided a mechanical bar to anything but finelydivided or liquefied food.

Very Iittle Cree acid was detectable in the gastríccontents at any stage oí digestión, or in extracts from thepellets. Total acidity of a sample taken 5 hr after eatingwas 0.16%; 2.5 hr after a meal it was 0.44%; and at 4 hr itwas 0.28%. Despite the lack of acidity, peptic activity wasCound to be gene rally about three times as potent as thatof dog gastric juice at comparable stages of digestión.

118

Summing up the situation, Reed and Reed concludedthat pellet Cormation involved three factors: first the highplacement and small size of the pyloric opening; secandofeeble gastric motility tbat would preelude stirring up orfreeing of hair or feathers enmeshed in the whole rnass;and, third, potent peptic activíty tbat would digest ñeshfree from hair, feathers, and bones and liquefy it so thatpyloric passage would be possible. The actual ejection ofa pellet was similar to vomiting-s-the owls showedevidence of nausea for 15 to 20 min before ejection.

It seems likely that other species of owl may showvariations in the action of their gastric juíces. A pellet ofthe wood owl, Cíccaba woodfordi, from Zimbabwe ...contained fur and feathers but no bones at aH (Brooke1967), suggesting that digestión ofbones may occur in thisspecies, as it appareotIy does in birds such as theblack-headed herén, Ardea melanocepñala, whichcharacteristically ejects bone-free pellets (Crass 1944). Adigestive oddity in the barn owl, Tyto alba. recorded InCalifornia is that tbese birds appear unable to digesteggshell or yolk (Banks 1965).

A number of studies (e.g., Clark 1972) have suggestedthat hawks digest bone more extensively than do owls.This has, in fact, been demonstrated in a series ofexperiments 00 gastric digestion in raptors (Duke et al.1975), in which pH and proteolytic activity of gastricjuices in several hawks and eagles (Calconiformes) wascompared witb tbat in owls (strigiformes). The basal, orpreprandial, pH in Calconiform gastric juice was found torange from 1.3 to 1.8, witb a mean of 1.6, while that of tbestrigiform equivalent ranged from 2.2 to 2.5 witb a meanoC2.35. TIte greater acidity oC hawk gastric juice couldhe expected to damage bones more extensively thanbappens in owls, and tbis has now been experimental1ydemonstrated (Cummings, Duke, and Jegers 1976).Mouse bones were incubated in solutions simulatinggastric juices witb pH levels of \.6 and 2.35, and it wasshown tbat tbe former corroded bones signilicantiy more,at all pepsin concentrations, tban the latter. Proteolyticaetivity proved slight1y greater in solutions with a lowerpH.

In an important paper 00 the role of avían predators asaccumulators offossil mammal material, Mayhew (1977)emphasized that mammal remains in peHets of diurnalpredators showed ccnsistent differences from those inpellets of nocturnal predators. He concluded that diurnalbirds oC prey were able to digest bones more completelythan were nocturnal ones, but that the fonner sornetirnes

120 A Guide to the Interpretation of Bone Accumulations in African Caves

built up assemblages of mammalian teeth, as in theCromerian Upper Freshwater Bed al West Runton inEngland.

It is also to be expected that sorne parta of preyskeletons should suffer more from digestión than others,and this has been shown to be true for artificially fed preyof three specíes of European owls (Raczynski andRuprecht 1974). Determinations were made, by analyzingpellets, of elements missing from skeletons of birds andmammals fed to the owls. It was found that the pelvicgirdle was most often missing (up to 80%), followed bycraniums (up to 35%) and mandibles (25%). The largestnumber of missing skeletal elements occurred in thetawny owl, Strix aluco (51%), with fewer in thelong-eared owl, Asia otus (46%), and bam owl, Tvto alba(34%). Furthermore, the quantity of missing skeletalelements showed an inverse relationship 10 tbe age of theowls and of their prey.

Furtber information on pellet formation has resultedfrom a laboratory study on the short-eared owl, Asiaflammeus I a diumal species common in Scotland andWales (Chitty 1938). An experimental owl was housed ina specially constructed cage that allowed a record lo bekept of the time of feeding and ejection of pellets. Itwas found that the interval between eating and pelletproduction varied from 1.5 to 13 hr, depending on theweight of the meal. A food ítem weighing 4.8 g led toejectíon ofa pellet after 1.5 hr, whereas one weighing 37.8g tock 13 hr. Likewise, the size of a pellet depended onthe nature of the meal; pellet lengths varied from lOto 65mm, the largest recorded pellet resulting from digestion ofa 54.2 g voleo Tbis 65 mm-long pellet had a volume of 13.5ce and a wet weight of 10.9 g. Typically, the weight of adry pellet is one-half lo one-third that of the fresh pellet.

The timing of pellet production is clearJy importantfrom the point oC view of microfaunal accumulations incaves. Only pellets ejected during rest periods in the caveroost will contribute to the deposito It seems likely thatpellet-ejection times will vary a great deal in naturalcircumstances. A study of twenty pairs of tawny owls,Strix aluco, ayer a period of eight years in Britain(Southem 1954) revealed that tbis species typically castsall its pellets before going to roost.

After a study of bam owls in France, Guérin (1928)concluded that there are characteristically two pelletejections in each 24-hr periodo The first results from preycaught early in the evening and takes place away from theroost at the "station noctume." Tbereafter, the owl tendslo feed again before retiring lo its roost, or "stationdiume. " Here, in the course of tbe day, the second pelletis produced, JI tbis procedure ís typical of barn owls ingeneral-e-and it is confirmed by WalIace's study (1948) inMichigan-it is clear that only part of the diet will bereftected in the pellets that aceumulate under a day roost.In addition, the consumption may be unexpected\y high;Guérin showed that a wild bam owl captured a total of1,197 animals in a 9O-day observation perlod; of these, itate 837.

Various attempts have been made lo assess the foodrequirements of owls by collecting and analyzing theirpellets. Montgomery (1899) studied the short-eared owl,Asio fíammeus, frorn this poínt of view but concluded thatmany pellets would have been disgorged at casual feedingperches that could not be located. The same is thought tobe true of the bam owl in the Netherlands (Honer 1963),which was found to hunt over fixed routes, using highpoints as lookouts and pellet-casting posts.

The rate of pellet-casting by barn owls al a diurnal roostin the Transvaal has been established by Dean (l973a).He found tbat the rate varied from 0.6 to 1.5 pellets perday, with a mean of 1.10 per owl per day. This involved amean daily ccnsurnption of vertebrate biomass rangingfrom 41.7 to 81.9 g per day, with a mean of 55.3 g per dayper owL Consumption of food was highest in late autumnand winter, the months when young are reared. Thefigures given by Dean were based only on pellets cast atthe roost. They therefore represent only part ofthe owl'stotal díet.

Soutbern Afrlcan Owls Responsible for MicrovertebraleColIections

As 1 mentioned earlier, the species of importance here arethe eagle owls and bam owl. Tbe former are perhaps oflesser significance than the latter, but they certainly war­rant rnention. Three species of eagle owl oceur in south­em Africa: the spotted eagle owl, Buba africanus; theCape eagle owl, B. capensis, and the giant eagle owl,B. lacteus.

Spotted Eagle OwJs. Bubo tifrú:anus (TemmJnck)

The spotted eagle owl is the cornmonest large owl insouthern Africaand has a wide distribution farther to thenorth, It tends to rest during the day amcng rocks or inlarge trees, emerging at dusk, when it may be seen reg­ularly on vantage points in the hunting area. It is fre­quently killed by cars on country roads at night.

Though less important than bam owls as accumuletorsof bones in caves, spotted eagle owls cannot be ignored,since they certainly roost and nest in caves from time totime. While prospecting for cave deposits in South-WestAfrica, 1 carne across the nest of a spotted eagle owlinside a dolomitic cave on the farm Cbicagc, in the OtaviDistrict, on 23 May 1954. The cave consisted of a verticalshaft about 9 m deep, overhung by a large fig tree whoseroots descended the shaft and aided entry. Tbe nest wason the bare floor of the cave, in the twiligbt zone, andcontained three eggs and two downy owlets. Tbe vicinityofthe nest was Iittered with pellets, which I unfortunatelyfailed to collect, but a freshly kiIled rodent close to thenest was identified as an African bush rat, Aethomyschrysophilus.

Fig. 123.Distribution of thespotted eaglecwt, Bubo a!ricQllus.

¡

Spotted eagle owls, like many other predators, are op­portunists, taking whatever prey is to be had at the time,whether rodents or invertebrates. The stomachs of7 adultowls killed on roads between 1956 and 1961, examined byBenson (1962), contained nothíng but insects and onesnail; similarly, an owl killed nearTrelawney in Zimbabwe(Camegie 1961) had eaten 11 beetles and a locust.

On the other hand, analyses of pellet accumulationsbelow spotted eagle owl roosts have revealed a variety ofvertebrate prey species. Nel (1969) identified a total of 565individual animals from a collection of 252 pellets that hadaccumulated on the floor of an old building at Sossus Vleiin the Namib of South-West Africa. The prey sample wasdominated by desert gerbils, Gerhillurus vallinus (62%),followed by golden moles, Eremita/po granti, various ro­dents (10%), and geckos and birds (7%). A rather similarprey composition was found in pcllets collected below ac1iff roost at Homeb, on the Kuiseb River in the NamibDesert Park (Brain 1974b). Other rodents feature prom­inentIy in the prey of spotted eagle owls from other

The Contribution of Owls 121

southem African locatities. In the sandveld area of thesouthwestem Cape, the gerbil, Talero afra. was found tobe the commonest prey (Siegfried 1965), whereas in aJohannesburg sample the vlei rat, Otomys irroratus,dominated (Welboume 197~).

A few years ago 1 had an opportunity to compare thefaunal content of spotted eagle owl pellets with those ofbam owls hunting simultaneously over the same area.During September 1972 a pair of eagle owls roosted insome large Ce/lis trees against a dolomitic c1iff 300 msouthwest of Swartkrans, while a bam owl was using theInner Cave at Swartkrans as its day roost. In the course ofthe month 1 collected pellets beneath both roosts; thcre­after the eagle owls left, and the barn owl continued tomake intennittent use of the Swartkrans cave roost untilJuly 1975. Prey animals identified froro the two pelletcollections are Usted in table 66, and the major preygroupings are depicted in figure 125. The contributionsmade by the various prey animals to the diets of the twospecies of owl were surprisingly similar.

Fig. 124. A spotted eagle owl, Bubo africanus, ptlotograptled by A. C. Kemp.

well as quite large birds.·· The northernmost subspecies,B. c. ditlonlí, has alsc received scant attention from or­nithologists , although Brown (1970) has observed it in theBale Mountains of Ethiopia. He states that in the cedarforests there ít nests at the bases of low cliffs, and thenumerous pellets he collected contained the remains offruit bats and mole rats in about equal proportions.

Most of the information available about Bubo capensisIs based on the subspecies mackinderí, named after SirHalford Mackinder, the first man to climb Mount Kenyawhere, incidentally, the owl occurs up to an altitude of13,500ft (4,117 m). The ñrst detailed study of B. c. mac­kinderi was made by Sessions (1966, 1972) at LengetiaFarm on the Mau Narok plateau clase to the west wall ofthe Central Rift Valley, 56 km wesl of Gilgil in Kenya.Here ten nest sites were found, each breeding pair havinga territory of about 2 km", The nests were 00 bare ground,beneath suitable cover, and invariably close to water.Conceming feeding behavior, Sessions wrote (1972, p. 4):

Typical of many eagle owls, Mackinder's Owl is nolentirely nocturnal and is often seen hunting beforesunset and again in the early moming; sometimes onemay see it sunning itself by day on an exposed rack ocbranch.

The usual routine is, at sundown, the owl gives apreparatory hoot from its daytime roost, sometimesgiving a few answering calls to its mate. It then flies upthe valley in short stages, pausing en route on suitableposts, until it finally reaches its hunting ground on theridge topo Here it will utilise a favourite perch such asa taU tree or telephone pole, or even a chimney top,from which to spot its prey, which may be eaten onthe ground or on the percho Afier a couple of hours'hunting, the birds retum to the roost site in the vaIley,where they appear to spend a considerable part of theníght settled on the ground near the water. The pur­pose ofthis habit is not clear, bUIit may be in order tocatch frogs and crabs. Before dawn they are backhunting again on top of the ridge, retuming to roost asthe sun comes up.

From an analysis of the contents of pellets, Sessionsconcluded that the main prey of Ibe owls was mole rats,Tachyoryctes splendens, which weigb up to 250 g, and asingle owl may eat three each night. Second in importance

Fil. 126. Distribution of lhe cape eagle owl, Subo capensis,

o 20 40 60 o 20 40 60

E.I,Fig. 125. Histograms showing the percentage contributions made by vari­ous animals to the diet of spctted eagle owls and bam owls. hunted simuJ­taneously over tbe same range al Swartkrans. Prey selection by lhe twoowl species preves lo be remarkably similar.

The tentative conclusion is that microvertebrate boneaccumulations built up in any particular cave by spottedeagle and by barn owls would probably not differ much inspecies compositíon. In fact , in these terms it would notbe possibLe to separate one set of food remains from theother.

Cape EagJe Owls, Bubo capensis Smilb

In South Africa this is a rare species, not easy to distin­guish from the spotted eagle owl, but with very muchlarger feet that allow it to overpower larger prey. Threesubspecies are currently recognized (Benson and Irwin1967):

B. c. capensis, confined to the Republíc of South Af­rica, in the mountainous arca from Cape Town eastward10 Natal and ZUluland.

B. c. makinderl, found al scattered localities fromZimbabwe 10 the highlands of Kenya and wesl 10 MountEIgon. Lasger than capensis, with less distinct barring onthe lower abdomen and thighs,

B. c. dillonii, an owl intennediate in size between thelast two, from the higblands of Ethíopia and Eritrea.

As figure 126 shows, the presenl distribution of the owlis higbly discontinuous, bul the currently isolaled relictpopulations were probably linked during cold periods inthe Pleistocene.

Very little is known about rhe South African sub­species, although Clancey (1964) states thal these birdscan overpower "quite large marnmals such as thespringhare, Pedetes caffer, the fawns of antelopes, as

122 A Guide lo the Interpretation of Bone Accumulations in African Caves

are freshwater crabs; it is not unusual to find a wholecarapace up to 25 mm across in a pelle!, or a claw 39 mmlong. Crabs are aJso brought to the nestlings, and on oneoccasion Sessions found apile of their carapaces on afeeding block about 5 m from the nest. The crabs hadnot been eaten whole as they would have been by anadult owl; the meat had been picked out and the sheHdiscarded.

At the top of the Teleki valley on Mount Kenya, Ses­sions also found numerous roosts of Mackinder's owland, from an examination of the food remains beneaththem, concluded that dassies had been the main prey. Theheads of the dassies had not been swaJlowed, and theskulls were simply discarded aiongside the pellets.

Bubo capensis was not known from Zimbabwe untilspecimens unexpectedly came to light from Inyanga caveand Bambata cave in the Matopos (Benson and Irwin1967). The first information about the owls' diet there waspublished by Brooke (1973), who listed the contents of acoUection of pellets from a small cave on the ShotsheKopje in the Matopos. Prey included three hares, con­firming a chance observation by Jackson (1973a,b), whofound that a female owl collected in Mozambique near theZimbabwean border had the leg of a red rock hare in itsesophagus. A number of breeding and roosting sites ofB. c. mackinderi have now beenlocated in the Matopo HiIls(Gargett and Grobler 1976; Gargett 1976; Gargett, n.d.).These show that the owls are sparsely distributed in com­parison with, say, black eagles. Seven breeding sites havebeen located, indicating that the owls occur at a density ofone pair to 88.6 km 2 • Since 1972, regular visits have beenmade to the nest and roost sites in the Matopos, and allpellets and other food remains have been collected and

The Contribulion of Owls 123

analyzed. A remarkable list of 925 prey items has nowbeen compiled (Gargett and Grobler 1976), a simplifiedversion of which is given in labJe 67 and figure 128.Working with the liveweights of the various prey species,it has become clear that hares, principally the red rockhare, make up 7Cffo of the owls' dieto Second in im­portance are the two forms of dassie occurring in theMatopos, followed by the two species of cane ral and avarjety of mamma(ian prey including springhare,hedgehog, squirreI, mongoose, genet, and juvenile eivet.Birds, Iizards, and invertebrates also contributed, albeitinsignificantly, to the diet.

Concerning treatment of the prey skeletons, Gargettand Grobler reported that the hare craniums showedcharacteristic damage to the calvariae and nasals. Sku\lsof prey of this size had not been swaUowed but occurredside by side with the smaller skeletal elements re­gurgitated in the pellets.

1 have been able to observe a pair of B. capensis roost­ing in a small cave 15 m up on a cliff face adjacent to theSkeerpoort River on the farm Uitkomst, about JO kmnortheast of Swartkrans. Remains of the owls' prey in thecave, collected on 29 September 1972, carne from 2 adultand 2 juvenile hares, probably Pronolagus randensis andfrom 1 hedgehog, Erinaceusfrontalis. The hares had beendismembered and eaten piecemeal, but the remains ofthe hedgehog were contained in a single pellet shown infigure 129a. The hedgehog had obviously been swallowedwhole, despite its many spines, and at the time this struckme as very odd. Subsequently 1 found that the eating ofhedgehogs by eagle owls is not as uncommon or as un­likely as it seemed. Hedgehogs were recorded arnong thefood remains of B. capensis from the Matopos (Gargett

Fig. 127. Young of the Cape eagle ow1, pho[ographed in the Matopo HíUs of Zimbabwe by Mrs. V. Gargett. Thecomparaüvely large fee[ of [he species are well shown,

Barn Owls, Tyto alba Sc:opoU

TItese familiar birds are of great significance as ac­cumulators of microvertebrate bones in southem Africancaves, a fael emphasized by Davis (1959) in bis paper"TIte Barn-Owl's Contribution lo Ecology and Palaeo­ecology.'

As figure 132 shows, the bam owl has a remarkablywide distribution throughout Africa, Europe, America,the Far East, and Australia, feeding on whatever preyhappens lo be available in the particular habítat, Evenwithin the southern African region, there is a good deal ofvariation in principal prey species, as has become appar­ent from no fewer than eighteen published analyses ofbarn owl pellets from seventy-seven separale localities insouthern Africa. References lo !hese sludies are given intable 68 along with particu1ars of the species occurringmost abundantly in each pellel collection. Insectivoresdominated in three collections, rodents in sixty-eight,birds in five, and reptiles in one. Among the rodents, themultimarnmate mouse, Praomys natalensis, proved lo bethe dominanl species in by far the Iargest number of col­lectíons: Srnithers (1971a) gave the mean weight of 14adult male P. natalensis from Bolswana as 64.5 g, andthat of 16 adult females as 56.8 g, TItese figures suggest

distribution from the Cape to Ethiopia and Senegal,wherever its preferred habitat of riverine forest, broad­leaved woodlands, or forests interspersed wíth grassyglades oecurs. It is the only eagle owl in Africa in whichthe eyes are dark brown rather than briHiant yellow ororange as in the other species. les eyelids, however, arebright orange or pink and are doubtless used in socialsignaling. According lo McLaehlan and Liversidge (1970)and Brown (1970), gianl eagle owls lend lo breed in dis­used nests of eagles and vultures, incomplete nests ofhamerkops, or occasionally in hollow trees. They prob­ably make líule regular eontribution to bone accumula­tions in caves, but could well do so if their roost or nestwas situated within the catchment area of a cave en­trance. Like the Cape eagle owI, this species is known tofeed on hedgehogs. Benson (1962) recorded an inslance ofone of these owls swallowing a complete adult hedgehogal Nyamandhlovu near Bulawayo, while Vernon (1971)found hedgehog remains below a B. lacteus roost 00

Tsumis Estate in South-west Africa. Other prey itemsrecorded here were ground squirrel, Xerus inauris, suri­cate, Suricata suricatta, yellow mongoose, Cynictispenicillata, a gerbil, Tatera sp., a raptorial bird, and akorhaan.

Hedgebog prey of B. tacteus has also been recorded byBrown (1965) al a nesl 21 km from Nairobi. Herehedgehogs proved lo be the most cornmon prey species; 4skulls and at least 11 skins were collected below the nest,and many other skins were not picked up. The adult owlshad skinned the hedgehogs before offering them lo theyoung, Other prey remains carne from rodents, fruit bats,a bushbaby, various birds inc1uding a bam owl, a housesnake, and numerous unidentified frogs.

Of possible interest here is Broom's (l937b) record of afossil hedgehog, Atelerix major, from Bolt's Farro. Thespecimen carne from the same small cave deposit whichyielded abundaat remains of a game bird, possibly a fran­colin, as well as bones of a bird of prey. It is conceivablethat the fossils represenled the food remains ofa large owlsuch as an ancestral B. lacteus,

I

ro ~ ~ ~ ~ D ~ ~ ~ ~ ~

% R__allon 01 pnlY

B"Aft OWl

,-

Fig. 128. (a) Percentage contributions to tbe diet ofCape eaaJe owls in theMatopo HiIls of Zimbabwe by animals of VariOU5 k.inds (data from Gargettand Grobler 1976). (b) The weights of prey items in Cape eegle owl dietcompared with those in the diet of bam owts.

OTHEA

PAEY

CANE

AATS

DASSIE5

HAAES

--

Gianl Eagle Owls, Bubo tacteus (Temminek)

By far the largesl of the southem African owls, with awingspan of up to 1.6 m, this specíes has a very wide

% CONTRIBl/TION TO

OIET

and Grobler 1976), and gianl eagle owls also prey uponthem, as discussed below.

A second visit was made to the Uitkomst cave roost on18 April 1978; the owls were not there but had left a largenumber of bony food remains on the cave ftoor. Thesewere found to come from 18 hares, of which at least 10were immature, from 4 birds, 1 crab, 1 dassie, 1 elephantshrew, and 26 rodents, of which 17 were vIei rats of thegenus Otomys. A striking feature of the hare skulls wasthe characteristic damage they had suffered to both nasalsand calvariae, as shown in figure 129b. This was the sameas that described and figured from a Cape eagle owI roostin the Malopos by Gargett and Grobler (1976).

In conclusion, it appears that Cape eagle owls are likelyto contribute significantIy to bone accumulations incaves. They have the habit of roosting and nesting insuitable caves, and they leave uneaten food remains andregurgitated pellets there. In contrast to the bam andspotted eagle owls, their mammalian prey is oñen com­paratively 1arge, and their diet also ineludes freshwatercrabs.

01020 30 40 5060 70

124 A Guide to the lnterpretation of Bone Accumulatíons in African Caves

that barn owls, over much of their southem Africanrange, typical1y feed on prey weighing approximately 60 gper item.

Sorne interesting information is available on chick­rearing by bam owls in circumstances where food issuperabundant. The observations were made by Wilson(1970) at the eland research station (21°2WS, 29°42'E) inZimbabwe, between January and December 1967, at thetime of a Praomys nalalensis population explosion that,in that area, peaked during luly and August and thengradually subsided. Bam owls were nesting in a hollowbaobab tree and, during the eleven months of observa­tions, four clutches varying from seven to nine eggswere laid and the young raised. Young owls were foundto fiy for the first time at about forty-five days of age and,since the eggs were laid and hatched in sequence, youngwere still in the nest when the eggs of the next clutch werelaid. In fact, the young birds probably served to incubate

The Contribution of Owls 125

the new batch of eggs. The rapidity of the breedingsequence observed here was doubtless a response to theabnormally abundant food supply, but it does showthe reproductive potential of barn owls in favorablecircumstances.

Dean (l973a) made a study of pellets collected atmonthly intervals throughout 1971 at a roost nearWarmbaths, Transvaal. The prey was also strongly domi­nated by Praomys natalensis: there were at least 431Praomys individuals out of a total of 647 animals. EachPraomys specimen was placed in one of six age classes,depending on the degree of wear on the first upper molars(Dean 1973h), and it was established that a substantialpart of the sample consisted of subadult individuals. Deanestimated that the mean liveweight for the Praomys indi­viduals in the pellet sample would have been 38.8 g. Par­ticulars of the species present in the Warmbaths sample,together wíth mean liveweight estimates, are given in

Fig. 129. (a) A regurgitated pellet of a Cape eagle owl containing the spines of a hedgehog. (h) Skulls of haresdiscarded by Cape eagle owls in a cave roost, Uitkomst. Transvaal. Damage to lhe nasals and calvariae is charac­teristic.

ticular area with the changing seasons. A monthly analy­sis of bam owl diet based on owl peIlets col1ected inBryanston, Johannesburg, during 1955 and 1956, enabledde Graaff (196Gb) to conclude that the owls concentratedmainly on birds in summer and rodents in winter. Asimilar situation had previollsly been observed in theTransvaal by Kolbe (1946) and Davis (1959). Generally,however, in favorable habitats a bam owl's diet is verydiverse, so that clear-cut altemations between prey taxamay be difficlllt to isolate.

A particularly uncomplicated example of prey alterna­tion came to 1ight recentIy in the Namib ofSouth-West M­rica. The Mirabib Rock Shelter in the Namib Park containsledges along its back walJ that are cllrrently used as bamowl roosts (fig. 135) and apparently have been for at 1eastsix thousand years (Brain 1974b). Bones from the result­ing pellets have been incorporated in the stratified floordeposit, which also contains vegetable matter and humanoccupation debris. The floor deposit has been extensivelyexcavated by Sandelowsky (1974).

Recently an analysis of microvertebrate remains in thestratified floor deposit has been undertaken (Brain andBrain, 1977), providing a record of ow1 diet over a six­thousand-year period. It was found in alJ layers that avery large proportion of this diet was made up of two preycategories: desert gerbils (Gerbillurus vallinus and G.paeba) and geckos (probably largely Pachydactylus !Jih­roni). Together, individuals of these two categories oftenmade up 90% of the total prey items taken. Inevitably,therefore, when the percentage abundance of one cate­gory's individuals, for example, gerbils, increased, thatof the other, the geckos, declined. This reciprocal re­lationship is shown graphically in figure 136.

The key to this striking relationship lies in the fact thatthe gerbil population on the plains surrounding theMirabib Hill is prone to remarkable fluctuations. Mterrain, which in this part of the Namib is sporadic andsparse, annual grasses appear in great abundance. In re­sponse to the grass cover, the gerbil population builds upwithin a few months to a relatively high .density. If no rainfalls in the following year, the population again declines toa residual leve!.

When gerbils are abundant, the barn owls prey on themto the virtual exclusion of al1 other prey sources. Whenthey become rare, however, as in poor rain years, thebam owls fall back on the standing gecko resollrce,mainly on the rocky hillsides, which is hUle utilized whenthe gerbils are readily available. It seems reasonable to

table 69. Gn the basis of these data, it has been possible tocount the number of individual animals in each weightcIass; figures are given in table 70. As is apparent in figure134, the great majority of prey items in this sample faH inthe 20-40 g weight range.

It is abundantly clear that, although bam owls mayhave some dietary preferences, they will feed on the mostreadily available food source as long as it is palatable andof manageable size. This means that, over the owIs' widegeographic range, dominant prey will vary from one areato another depending on the species composition of thepotential prey fauna. Apart fram this, fluctuations fromone dominant prey taxon to another may occur in a par-

Fig. 130. nistribuLion of the giant eagIe owl, Bubo lacreus.

126 A Guide to the Interpretation of Bone Accumulations in African Caves

Fig. 131. A giant eagle owl, Bubo lacleus. photographed at Lake Mc-11waine in Zimbabwe. Fig. 132. Distribution of the bam owl, Ty10 alba.

correlate high gerbil abundance in the owl's diet withsummers in which adequate rain feH; gecko abundance,on the other hand, tends lo suggest times of drought.

This Namib example provides an indication of howbam owl diet may reftect environmental changes in thehunting area.

Species Composition 01 Barn Owl Prey as a Reflection 01tite Habitat

It would be very valuable from the paleoecological pointof view if microvertebrate bone accumulations derivedfrom owl pellets could be used as indicators of the habitatover which the owls hunted. For any such attempt it isclearly necessary to know what the hunting range of an owlfrom its roost might be. The range is Jikely to be greatlyaffected by the availability of food; when prey is abun­dant, the owl's range will be smaller than in times offoodscarcity.

Reliable infonnation on hunting ranges wiU not be forth­coming until radío telemetric tracking of owls has beencarned out. Meanwhile we have little more than tentativeindications. As a result of a study of bam owls in Israel,

The Contribution of Owls 127

Bodenheimer (I949) concluded that ··it can be assumedthat their hunting grounds do not exceed an area of25 sq.kJm. and are probably much more restricted." Anotherindication has come from a study ofbarn owls on the westcoast of North America. Owls roosting in a small cave onthe northernmost of of the Islas Los Coronados, BajaCalifornia, inc1uded in their pellets prey that had beenobtained on the mainland. Forays to the mainland forfood would have involved ftights of la mi, or 16 km, eachway (Banks 1965).

Evidence from the Mirabib owl roost in the Namib ofSouth-West Africa (Brain 1974b) c1early indicates thatbarn owls in those circumstances do not fty the 25 km tothe nearest dune fields where they may obtain dUne­adapted prey, in the form of Namib golden moles. Themeager evidence currently available suggests that barnowls may radiate as much as 16 km from their roost onnightly forays, but not as much as 25 km. Thus, from thepoint of view of environmental reconstructions, owl preyshould be regarded as coming from within a few kilomet­ers of the roost. When prey is nonnally abundant, sornerecent evidence suggests that owl hunting in the Trans­vaal may be extremely localized. 1 have been able to

Fig. 133. Two young bam owls, from Bríts in the Transvaal, al differenl stages oC deveJopmenl. Barn owl eggsarelypically laido and hatch, in sequence, so that young of diJIerenl sizes are fonnd in one nest.

128 A Guide to the Interpretation of Bone Accumulations in African Caves

ble. Davis (1962) has divided southern ACrica into fourbiotic zones, the main features of each being these:

The southwestem Cape. A distinct area c1imaticalIyand biotically, The southwestern portion has a Mediter­ranean climate and winter rainfall, but along the coasttoward the east the rainfall becomes more regularly dis­tributed throughout the year. The zone corresponds to theCape Macchia vegetation type.

The southwest Arid. The area consists of true desert,the Namib, and semidesert, the Kalabari and Karroo, alIof it receiving, on average, less than 50 cm (20 in) of rain ayear. It is divided by the Orange River and contains thewestem and southwestem portions of the great escarp­meot. It extends across the Kunene River to aboutlatitude 12°S.

Southern savanna. This zone occupies the eastem partof southern Africa. It includes montane and high.veldgrasslands, bush and lowland woodlands, and the tropicalMozambique plain.

Forest, Isolated patches of montane and subtropicalevergreen forest are distributed in the savanna zone andsoulbwestern Cape, mainly below the escarpment, En­demic mammals are extremely rare, and no rodents oftheCamily Muridae can be regarded as true forest forms,

Davis (1962) has listed murid rodent species that areendemic 10 each of the biotic zones. The two zones ofparticular relevance 10 us here are the southem savannaand southwest arid; species occurring in these zonesthat Davis regards as endemic or nearly endemic are asCollows:

Southwest arid loone: endemic: Zelotomys woosnami,Aethomys granti, Petromyscus coííínus, Parotomyslittledalei, Gerbillurus vallinus; near endemic: Malaco­thrix typica, Otomys unisulcatus, Parotomys brantsi,Desmodilíus auricularis, and Gerbillurus paeba.

Southern savanna loone: endemic: ThamnomysdoUchurus, Pelomys fallas, Acomys spp., Cricetomysgambianus, Dendromys mystacalis, D. nyikae, Otomysangoniensis, Tatera inclusa; near endemic: Aethomyschrysophilus, Thallomys paedulcus, Praomys natalensis,Aethomys namaquensis, Rhabdomys pumilio, Musminutoídes, Dasymys incomtus, Lemniscomys griselda,Saccostomus campestris, Steatomys spp., Dendromysmesomelas, D. melanotis, Otomys laminatus, O. ir­roratus, O. saundersiae, O. sloggetti, Mystromysalbicaudatus, Tatera brantsi, and T. Ieucogaster,

Abundant remains of any of these species in micro­vertebrate accumulations would indicate a habitat typicalof one or other oCthe two biotic zones, Owl pellet collec­tions Crom the Kalabari National Park (e.g., Davis 1958;Nel and Nolte 1965)or Namib (e.g., Brain 1974b) that aredominated by Gerbillurus valllnuslpaeba remains are, infact, indicative of true arid zone habitats, whereas thosefrom the very numerous Tmnsvaallocalities. afien domi­oated by Praomys natalensís bones, clearly indicatesavanna zone habitats.

The differences between arid and savanna zonehabitats are often extreme, whereas paleoecologicalreconstructions generally aim at more subtle habitatcharacterization. Within the southem savanna zooe arehabitats as diverse as opeo grassland and woodland; itwould be interesting to know ifthe species composition ofmicrovertebrate accumulations could be used as in­dicators oC such vegetation types. To lest such a possibil­ity, comparable collections oCbarn owl pellets have beenanalyzed Crom an open grassland and a savanna woodlandhabitat.

'20 140gm

80

make pellet collections at two pairs of adjacent barn owlroosts, covering the same time periods, and the resultsshow interesting local variation.

The first pair of roosts were the Swartkrans Cave itselfand the Bolt's Farm roost, 1,470m to the soulbwest. Bothroosts are on the north side oC the Bloubank River valley,Swartkrans being 250 m Crom the streambed and Bol!' s350 m away (see fig. 137). The habitat tooks very similar.Analyses of pellet collections from each site made in 1975are given in tables 66 and 71. Although the same preyspecies occur in botb samples, owls operating traroSwartkrans were exploiting a local concentration ofpygmy mice, Mus mlnutoides, which constituted 27.2% ofthe Swartkrans prey animals but only 1.8% of those atBolt's Farm,

The second pair of roosts were in two quarries close 10the víllage of Irene, south of Pretoria. The grasslandhabitat 00 dolomite country rock is very similar to that inthe Sterkfontein valley. The new quarry is 1,500 m east­southeast ofthe old one; both are in similar surroundíngs,except that the former is 1,300 m from the Hennops Riverand the latter is oniy 120 m away. The Hennops River is aperennial stream wilb fringing trees and a Cnirly extensivealluvial ftoodplain. Pellet collections were made at eachsite in July 1975, and resuíts are given in table 71. As is lobe expected, the same species occur in both collecñons,although there are interesting differences. The NewQuarry collection, as a result of its proximity to thealluvial plain, contains larger oumbers of vlei rats,Otomys írroratuslangoníensis: SO.6% as opposed to32.0% at the Old Quarry. On the olber hand, lbe owlsbased at lbe Old Quarry were exploiting a Tatera colonynearby, so lbat these gerbils constituted 26.0% oC lbe preythere in contrast to S. t% at the other roost.

Taking it into account that extremely localized varia­tioos in prey composition are to be expected at any par­ticular roost, it would still be usefuJ if the composition ofprey could be used as an indicator of habitat type. On agross basis across southern Africa this is certainly possi-

Fig. 134. Histogram showing the percentege representation of variousweight groups in the prey of bam owls roosting al Wannbaths in theTransvaaJ. Data reworked from Dean (I973a).

open grasslands in the Sterkfontein valley hábitat, but it iscurrently so rare there that the barn owls often fail tocatch it.

There is cJearly a need, and a good deal of scope, forfurther study on the hábitat preferences of small southernAfrican mamma1s and other animals cornmonly found inmicrovertebrate accumulations. Severa! small marnmalshave c1ear requirements---knowledge of such re­quirements cou1d be very useful in reconstructing pasthabítats , An example ts the shrew genus Myosorex. inwhich both extant southem African species, M. varíusand M. coffer, require moist and dense1y vegetatedmicrohabitats (Brain and Meester 1964).

The Presence o/ Barn Owl Roosts as Indicators 01 PostCave Form

A large colleetion of owl-pellet-derived microfauna in adolomite cave c1early implies that a suitable roost siteonce existed withio the catchment area of the cave en­trance. If any consistent features of roost sites could be

!L-=-~~'-~"-",,~""'Ill o 10 JO iIO

Pf:1tC11T"OE .......0....1:1

Fig. 141. Pcrcentaae representation ofvariobs animals in the preyofbarnowls huntill8 from tite 8olf, Fann and Makapansgat Llmeworks owlroosts. Despite thc difl'crencc in habita1 ofthe two areas, prev-selectíon isremarkably similar.

The Contribution of Owls 133

found. these would be of value in reconstrncting cavefonn where the relevant sturctures of ancient caves havebeen destroyed by erosiono

Roost sites selected by barn owls are extrernely vari­able. They may be church spires, loñs, disused nests ofthe hamerkop (Scopus umbrettcú, caves, or even gutterpipes. And yet, despite this diversity of chosen sites, thefavored locations of rocsts in natural dolomite caves areoften rather similar. The roost is very often on a ledgebeoeath an overhang in lite twilight zone of the cave.Subdued Iight is delinitely favored, provided the owlshave reasonably direct access to the cave entrance.

In lhe course of this study, peUet accurnutenons be­neath three roosts in natural dolomite caves have beeninvestigated. Theee were at Swartkrans, Bolt's Farm, andMakapansgat Limeworks. Sectional proliles have beendrawn through each cave al the position of the relevantroosts; these are shown in figure 142. Despite differencesin the size of the caves, selected roost positions have abasic similarity.

o 10, , " ,rnetr..

Fis. 142. Scctional profilcs through three caves wnere bato owl roostshave been studied. Arrows indicate the roosts: (a) Swartkrans Inner Cave;(b) Bolfs Film; and (e) Makapanspt Umeworts.

Although the preservation of bones in cave breccias maybe extremely good, fossils are often found that havesuffered the most remarkable distortions, fractures, ordislocations as a result of pressure and movements in thesediment. It may not be easy to separate such effectsfrom injuries the borre sustained before fossilization,particularly when the fossil has been freed from itsmatrix and can no longer be placed in its stratigraphiccontext.

7 Sorne Compressional Effects on Bones Preserved inCave Breccia

-é)3----- ----~~.~~.:~..o...oo,\o- - __@o~:o~::~.-=-ó:.o.- --'o "_ - -. /:F.3-\ f" 'J ~

..': ....00./ t···,·,,"- .•:o./' ~

~......_.~.~..:..,.;.= .•..:¡;;•....., . '::'.~ '_..". .,.:.' 'c " .

TheEffectsofPressure .~."~ ~b:"J!'§<r ..~;-<,..,' '.A fossil buried close to the bottom of a deep cave filJing ; ' ..... oo.,: - ~- ,: ·... .':.00..·0. ," ~...such as that at Sterkfontein may have as much as 30 m of io:iíii¡¡¡¡¡;;,------...", ~--_¡,,¡,¡¡, ~Wiii;·_..·""'·_

sediment aboye it, the weight of which is very consider- d f

able. But, in addition to this weight, the overburden load Fig. 143. Compressional effects on skuUs buried in different positions inon a fossil buried within a dolomitic cave may be vastly cave earth: (a) foramen magnum uppermosl; (h) foramen magnum to one

side; (e) foramen magnum beJow. When overtJurden pressure is applied,increased by subsíderrce of the cave roof. Denudation of the filled skulJ (d) retains its shape; the panly filled cranium (e) collapsesthe hillside aboye a cavern results in the breakup of the partially; and the empty one (f) is completely flattened.

a bFig. 144. Two examples of austra10pithecine craniums from the Swartkrans Member I breccia. (a) The fiUed caJvariaof SK 48 has suffered very little dísrortion. (h) The a1m03r empry cranial vault of SK 79 has collapsed completelyunder the weight oC the overburden.

134

Sorne Compressional Effects on Bones Preserved in Cave Breccia 135

dolomite spanning such a cavern. Roof blocks collapse orsettle onto the fossiliferous sediment, introducing com­pressional stresses of enormous magnitude.

The ability of buríed bones to withstand such pressureswill depend on the structural characteristics of the bonesand the degree to which their cavities are infiltrated bysedimento 1 will restnct my comments here to craniums,although effects on other skeletal parts may readily bevisualized.

From the taphonomic point of view, a skulJ consistsessentially of a hollow, subspherical capsule whose wallswill collapse inward if sufficient pressure is applied tothem. The capsule, or calvaría, is breached by a numberof orífices, of which the foramen magnum is the Iargestand, for us, most significant.

Skulls embedded in cave breccía have generally beencovered by sediment brought into the cavern through thecombined action of stormwater and gravity. Whether thehollow capsule of the calvaría will fill with sediment de-

1-

pends on the position in which the skull comes lO reSl onthe cave floor. rfit is upside down (fig. 143(/), the foramenmagnum is uppermost and total filling of the endocranialspace is Iikely; if it is on its side, the space will fill to thelevel ofthe foramen magnum's upper edge (fig. 143b), andif it lies right way up, with the foramen magnum beneath,the endocranial space typically remains empty (fig. 143c),except for a probable frosting of calcite crystals. Whenoverburden pressure is applied to these skulls, the degreeof compressional collapse will be related to the amountof sediment fiUing. The totalJy filled calvaria will distortbut little under pressure (fig. 143d); the half-filled skullmay suffer collapse of the unfilled portion of its vault(fig. 143e), and the empty calvaría will probably collapsecompletely.

Figure 144 shows two examples of australopithecineskulls that demonstrate the extremes of the process. Theendocranial space of the well-known cranium SK 48 fromSwartkrans (fig. 144a) had been almost completely filled

e

I'1

\ .~

Fig. 145. Examples ofIocaliud damage lo fossils in cave breccia caused by stones in the matrix to which overburdenpressure has been applied. (a) Auslralopilhecus mandible (STS 7) showing localized damage lo right corpus only. (b)Hyaenjclis skuU (SK 314) wilh damage reslricled [O ilS snout. (c) Panlhera skuU (SK 349) with part of a slone (slripes)

still embedded in Íls crushed orbital regían.

136 A Guide to the Interpretatíon of Bone Accumulations in African Caves

....

Fig. 146. An example of a fossil SKUU lhat has been laterally sheared bymovemenlS in the enclosing matrix: a baboon, Difropilheeus (SK 603)from Swartkrans Member l.

the new situation. A common result of such shape ad­justrnents is shearing deformation offossils that happen tobe enclosed in the matrix. Figure 146 shows ao exampleof a baboon skull, Dinopithecus ingens, SK 603, fromSwartkrans Member 1, that has been sheared by forcesacting in the direction of the arrows.

Sorne Effects on Rotted Bone

Fossil bones are found in the Sterkfontein valley cavesin various states of weathering, indicating that they layaround for various leogths of time before they were io­tened and calcmed. But, in addition to such weatheringeffects, it is not unusual to find fossils in which the struc­ture of the bone has broken up into discrete spicules thathave lost their original orientation (fig. 147). For sorne time1 was at a 10ss to explain the circumstances in which suchdrastic effects might occur until, by chance, 1happened toencounter a modem babooo skulllying in a perennial pooldeep in the subterranean portion of Peppercom's Cave at

Fig. 147. Three examples of Auslralopithecus bone from Swartk.ransMember 1 thal had apparently been extensively rolled in waler beforefossilizalion. (a) Mandible (SK 876) showing cracking of !he teelh anddisintegralion of lhe bonc. (b) Similar effects on a palale (SK 57).(e) Disinlegration of (he bone slructure shown by a mandibular ramus(SK 1514).

with sediment and, although the specimen was clearlysubjected to pressure, retajned its shape except for sornecollapse in the occipital area. By contrast, the endocranialspace of SK 79 (fig. 144b) remained empty and collapsedcompletely under pressure from above.

It occasionally happens that a craníum survives intactand undistorted throughout the burial and fossj/izationprocess. Such an example is the well-known skull of"Mrs. PIes" (STS 5) from Sterkfontein Member 4, whoseendocranial cavity remained empty except for a calciticencrustation. It happened that this skull had originallycome to rest in a part of the Sterkfontein cave c10se to theback wall, where the sediment surface was almost incontact with the underside of the roof (see fig. 157). Theresult was that the specimen was not subjected to over­burden pressure 01' the effects of roof collapse. Fossilspreserved in the same matrix but a few meters away fromthe back wall show very marked effects of pressure, prob­ably induced by subsidence of roof blocks.

The effects I have mentioned are to be expected in areasonably homogeneous matrix. They have typically ledto overall flattening, rather as if the skull had been runover by a steamroller. But a cave breccia seldom repre­sents a homogeneous matrix-rather, it is a fine-grainedsediment in which large objects such as stones occur.When one of these stones is in contact with part of a skulland pressure is then applied, localized darnage is thecharacteristic resul1. Part of the skulI may be fractured,depressed, or distorted. Three examples wíll be illus­trated. The first is an Australopithecus mandible, STS 7(fig. 145a), in which the right ramus has been distortedinward by a stone in the adjacent matrix, while the left isrelatively unaffected. The second example is a cranium ofthe extinct hyena Hyaenictis forfex from Member 1 atSwartkrans (fig. 145b), in which the muzzle has been ex­tensively broken and depressed by overburden pressureapplied to a stone that lay against i1. An example of evenmore localízed damage may be seen on a leopard cranium(SK 349) from the same deposi1. Here a piece of chert hascrushed the skull's orbital region (fig. 145c) while the restof the specimeo has remained comparatively uodistorted.In this particular case part of the chert stone has been leftin the depressed fracture. Unfortuoately, many other fos­sils have been freed of their enclosing matrix in thepaleontologicallaboratory, and no record has been keptof the nature of that matrix adjaceot to the bones. In suchcases, when localized damage is observed, it is oot alwayspossible to establish with certainty whether the damageoccurred before interment 01' duriog fossilization. Thisproblem has occurred relative to depressed fractures ont¡aboon and homioid craruums that Dart speculated hadresulted from hominid bludgeoning (see chapo 10). Untilone is familiar with the variety of 10calízed damage thatcan result from pressure in a heterogeneous matrix, suchspeculations are very tempting. Having observed pres­sure effects on many hundreds of fo'Ssils, 1 am now ex­tremely cautious when attempting to isolate instances ofprefossilízation bone injury.

The Effects of Shearing

Dolomite caves are notoriously unstable structures----notonly do theír roofs tend to collapse, but their ftoors fre­quently drop out as other cavities develop below them.When the latter happeos with a filled cave, the filliog itselfbecomes unstable and changes its shape to accommodate

Sorne Compressional Effects 00 Bones Preserved in Cave Breccla 137

Makapansgat. As ] atternpted to lift the skull out of thewater it simply collapsed in my hands, and 1was left witha white malleable sludge that a few moments befare hadconstituted a perfectly recognizable baboon skull. Theteeth had split up into sharp-edged segments, and the en­tire structure of the bone had disintegrated.

It is not difficult to imagine the eífects of overburdenpressure on specimens in this condition. Resultant fossiisare forrnless sheets of bonc whose anatomical featureshave largely disappeared. They are not uncommon inparts of the Swartkrans Member 1 deposn, and 1 havealso encountered them in Sterkfontein Member 4.

8 SummaryBone Accumulations in Southem African Caves:A Search for Interpretative Criteria

The airo of this investigation, which 1 have discussed inthe last five chapters, was 10 provide insights into theinterpretation of bone accumulations in caves-to decidehow the bones found their way to these pIaces and toreconstruct something of the behavior oí the animals in­volved in their accumulation. Certain guidelines haveemerged.

Separation oC an Accumulation ioto Micro- andMacrovertebrate Components

The difference between the bones classified in these twocategories is clearly one oí size-body size oí the animalsthat contríbuted their remains. It is conceivable thatabundant bones oí rnicrovertebrates, such as in­sectivores, small rodents, birds, and reptiles should entercaves in a variety oí ways; yet experience oí contempo­rary southem African situations has shown that, almostinvariably, such bones represent the prey of owls. Incomparison with other agents responsible for collectingbones of thís síze, owls are supremely significant. Por thisreason it is perhaps justifiable, when speaking of themicrovertebrate component of abone accumulation. loimply the owl-collected component.

The Owl Specles Involved

The various specíes of southem African owls likely tocontribute to bone accumulations in contemporary south­em African caves are discussed in chapter 6. They are thebarn owl and three species of eagle owl; presumably theantecedents of these species behaved like their modemcounterparts, though 1 have no good evidence for thisassumption except that fossilized food remains verysimilar to modem equivalents have been found.

RecognItIon oC Owl Species on the Basis oC TIlelr FoodRemains

Some indications have emerged whereby the species ofowl responsible for a particular bone accumulation maybe reeognízed. No fewer than eighteen published studieson the eomposition of bam-owl pellet colleetions fromseventy-seven localities in southem Afriea are available(table 68). A wide range ofanimals is eaten, but prey withbody weights of 2(}..4() g appears to be preferred (table 70;fig. 134).

What little evidenee there is suggests that it wouId not

138

be easy to distinguish food remains of barn owls fromthose of spotted eagle owls if the birds had been huntingover the same range. Prey selection by the two species inthe vicinity of Swartkrans has proved remarkably similar(table 66; fig. 125).

ALthough superficíally similar to the spolted eagle owl,the Cape eagle owl has far more poweñul feet and takesmueh larger prey. As is doeumented in table 67 and figure128, food remains studied in' Zimbabwe carne principallyfrom hares, dassies, cane rats, and other mammals suchas springhare, mongoose, genet, and juvenile civet.StrictIy speaking, such prey wouLd fall into the macro­rether than the microvertebrate component. In East Af­rica, B. capensis showed a strong preference for nestingand roosting close to water and for feeding on freshwatercrabs, whose remains accumulated around such sites.The Uitkomst roost in the Transvaal was similarLy placedclose to water, and freshwater crabs featured in the diet ofthe Cape eagle owls there as well.

Comparatively large mammalian prey also charac­terizes the food rernains of giant eagle owls. althoughthese birds have not yet been observed to roost insidecaves. It appears that hedgehogs are much favored preyof this species, and their remains have been found in allcollections of giant eagle owl food debris studied to date.1 would go so far as 10 infer that abundant hedgehog re­mains in a cave are suggestive ofB. lacteus or B. capensisínvolvement,

Although barn owls roost in a variety of sítuations, itseems that places in caves selected as long-term roostsand nest sites have sorne consistent characteristics (fig.142). The site is usually on a ledge beneath an overhang­ing roof in subdued light. This being so, large concentra­tions oí barn-owl-like food remains, fossilízed in a cavebreccia, suggest that the cave originally had ledges be­neath an overhanging roof and that that particular part ofthe cave was within the twilight zone.

Evaluation of the Macrovertebrate Component

This part of abone accumulation typically consists ofremains from larger animals that found their way to thecave in a variety of ways. Elisabeth Vrba (l976h) hasproposed a valuable scheme for c1assifying agents re­sponsible for accurnulating bones in caves of theSterlcfontein valley type. According to the scheme,biological bone-transport agents may be divided into twogroups: autopod and al/opod. Autopod agents involve

Bone Accumulations in Southern African Caves: A Search for lnterpretative Criter¡a 139

transport on the animals' own feet; allopod agents, on thefeet of others.

Autopod Transport Mechanisms

Animals come to the cave to seek shelter and die there,or they fall accidentally into a precipitous cave entranceand kill themselves. Their articulated, or disarticulated,skeLetons may become buried and fossilized, but charae­teristically the skeletal parts may not have been affectedby carnivore action, lo the Sterkfontein valley contexr.such remains are rare and insignifieant.

AUopodTranspor! Agents

With these agents, bones are transported to the cave or itsvicinity through the activity of other animals: there aretwo distinct pattems involving primary agents-thosepredators that bring their own kills for consumption in thecave, and secondary agents-those animals that transportbones of animals they did not kili themselves-thescavengers and collectors. Prirnary and secondarybone-transport agents are likely to be represented mainlyby meat-eaters, though vegetarían porcupines are un­doubtedly the most important "coliectors" of bones insouthem African caves. Within the meat-eater category,separation between predators and scavengers delineatesthe primary from the secondary agents, Regrettably, sucha delíneation proves fuzzy in sorne cases, sinee carni­vores are invariably opportunists. On arre day a hyena ora hominid is a hunter-a primary predator-but on thenext it scavenges and may then function as a secondarybone-accumulating agent.

Separating remaíns derived from primary predationfrom those of scavenging is not likely to be easy. Vrba(1975, 1976a,b) has poinled lo the two eriteria that, in lhebovíd context, are very significant. Primary predatorstend to prefer prey within a restricted weight bracket (seefig, 97); second, juvenile prey tends to be consumed morecompletely than are adults of the same species; for thisreason, juvenile representation in scavenged samplescharacteristically lends to be low. The applieation ofthese eriteria to the various Sterkfontein valley site unitswill be considered in subsequent chapters.

The ñrst ehapters of this book have been devoted loa consideration oí the various agents, primary and sec­ondary, responsible for bone accumulations in Afrieancaves. I will now summarize sorne of the criteria wherebythe agents may be recognized.

Survival and Disappearance or Skeletal Parts

As was discussed in chapter 2, the individual componentsof a vertebrate skeleton vary a great deal in their ability towithstand destructive treatment of whatever nature. It isnow possible to sor! skeletal parts, at leasl those ofbovids, into groups according to their robusticity andability to resist destruction. In limb bones, ability to with­stand carnivore action is closely related to the compact­ness of the bony part and may be refteeted in the bone'sspecific gravity; it is also affected by the time at which theepiphysis of a particular extremity fuses lo íts shaft (table7; fig. 16).

Another result of the investigations described in chap­ter 2 is that each par! of a bovid skeleton may be alloeateda potential-survival rat;ng, which will indicate how Iikely

Flg. 148. Diagram ofa bovid skeleton in which (he varíous pans have beenallocated "potential survival ratings." Parta shown in black nave highratings: stippled elements have intermediare ratings; and unshaded bonesbeve low survival potential.

ít is to survive a particular destructive regime. As isshown in figure 148, each part of an antelope skeletonmay be placed in one of three categories, having a high,medium, or low potential survival rating. Abone as­semblage composed essentially of elemenrs with highsurvival ratings clearly has been subjected to a good dealof destructive treatment. On the other hand, the skeletonof an animal that has been preserved without carnivoreintervention will also retain fragile elements with low sur­vival ratings.

It is not only individual skeletal parts that vary inrobusticiry and survival potential. As was indicated inchapter 2, skeletons of primates, and probably also ofcarnivores, are far more susceptible to destruction thanare antelope skeletons derived from anirnals of equivalentlive body weight.

Rerognizing CoUeetlonsMade by Poreupines

As was discussed in chapter 5, African porcupines areextremely important colJectors of bones in caves. Beingvegetarians, porcupines do not have the same interest inbones as carnivores do, yet they colleet large numbers ofbones and other objects in their lairs and gnaw thern attheir leisure. As was shown in figure 112, porcupine in­cisors in the upper and lower jaws occlude against oneanother, ensuring that sharp chisellike edges are con­stantly maintained, Nevertheless, unless the porcupineshave hard objects to gnaw, the incisors appear to growunmanageably long (as they frequently do in zoo ani­mals), preventing proper occlusion and sharpening of theedges, Gnawing bones may also pro vide the porcupineswith mineral salts they require. It is unlikely that ealeiumis the objective of bone-gnawing, stnce porcupines in theKalahari, for example, collect and gnaw large numbers ofbones despite the highly ealeareous nature of the soil andcountry rack. It is more Jikely that they are locking forphosphorus, which is generally the motive for bone­chewing by ungulares. It is known that high calcium con­tent in the soil can reduce the availabilíty of phosphoruslo plants, and thus lo herbivores. The phosphorus de­ficiency may then be made up by osteophagia-chewingbones (Sutcliffe 1973b).

Whatever motive porcupines have for collecting andgnawing bones, that they do these things cannot be dís­puted. Although porcupine collections have a number of

140 A Guide to the Interpretation of Bone Accumulations in African Caves

.'

tures of their vaults. He attributed the fractures to pur­poseful hominid bludgeoning with clubs made of antetopehumeri.

If it could be demonstrated beyond doubt that the in­juries had been in!licted before the skulls were buried inthe cave earth, then 1 would have no hesitation in agree­ing that deliberate hominid activity was indicated. But itis extremely difficult to separate the effects of localizedpressure within the enclosing matrix from prefossilizationinjury. Stones or other hard objects in the matrix cancause depressed fractures most suggestive of hominidviolence (Brain 1972). Regrettably, most skulls availableat present showing this particular kind of darnage havebeen freed from their matrix, and so the controversy con­cerning them cannot be resolved (chap. 7).

~\k:! .. ~

\ . a

Bone Damage: Shalt Fractures

As was discussed in chapter 1, a good deal has been writ­ten about the "crack and twist" technique for breakingantelope limb bones. Typically the technique results inspiral fractures of the bone shafts. If spiral fractures re­sulted only from the cracking and twisting of bones, auniquely human procedure, then the presence of spiralIyfractured shafts would be a good indicator of humaninvolvement. But, as is shown in figure 149, spirallyfractured shafts can result not only from human bone­breaking, but a1so from the bone-cracking of spottedhyenas, brown hyenas, and leopards. In his paper "SorneOperational Aspects of Human and Animal Bone Altera­tion," Bonnichsen (n.d.) has shown that a glass tube orcylinder will frequently break with a spiral fracture whenstruck a vertical blow. This observation implies that atube without internal structure of any kind in its wallsmay produce spirally fractured pieces as a result of im­pact. Long-bone shafts, on the other hand, are not devoidof internal strocture. Tappen (1%9) has shown that

consistent characteristics, as outlined in chapter 5, thesurest indicator of porcupine involvement in abone ac­cumulation is the presence of gnawmarks, caused by theanima!s' incisors (fig. 113).

Porcupines frequently collect more bones than theygnaw, and the percentage of gnawed bones depends, itseems, on availability. When suitably defatted and wea­thered bones are abundant in the vicinity of a lair, thepercentage of gnawed bones in a porcupine collectionmay be as low as 20%. When bones are scarce, 100% ofthem may be gnawed. Gnawing on 60-70% of the piecesmay be regarded as typical of bones from lairs in normalcircumstances.

Recognizing Primitive Human Food Remains

The Nature 01 Primitive Human Diets

As I emphasized in chapter 3, the animal remains thatconcem us here represent only part of the primitivehuman diet; the more stable, and frequently larger, com­ponent was of plant origino At this time, thirteen analysesof bone accumulations from southem African caves, rep­resenting human food remains, are available (table 33; fig.48). These show that hunter-gatherers are opportunists,making use of whatever food resources happen to be athand and often participating in a seasonal round to exploitthem. The same was probably true for early horninids,and 1 seriously doubt that prirnitive human food remainscould be positively identified as such on the basis of thekinds of animals that had been hunted. One point that hasemerged from the analyses is that remains of carnivoresare not abundant in the human food refuse. In fact, whencarnivore-ungulate ratios are calculated for the thirteenaccumulations, the mean value proves to be ten. By con­trast, bone accumulations built up by certain carnivoresthemselves-as, for example, brown hyenas-tend tohave carnivore-ungulate ratios far aboye this.

Associalion with Artifacts

Human food remains are frequently associated with ar­tifacts, and the relative abundance of each kind of tracewil1 detennine how a particular site is c1assified (tig. 47).By the scheme outlined in chapter 3, caves containinghuman food refuse would usual1y be c1assified "camp oroccupation sites." In such situations, increased numbersof bone fragments in a particular level are typically as­sociated with more abundant stone artifacts (figs. 26 and50).

Association with Traces 01Fire

Indisputable traces of fire associated with bones in a cavegenerally point to human involvement. Natural tirescaused by lightning in caves have been recorded, but theymust surely be rare.

Apart from the presence of ash, charring of bones maybe taken as an indicator of fire. As shown in figure 51,light charring blackens the bone; more prolonged in­cineration causes the bone to revert to a chalky color andconsistency. (.

Bone Damage: Depressed Fractures

Dart (l949a,b) described a Iarge series of baboon andaustralopithecine craniums that showed depressed frac-

Fig. 149. Examples of spiraJ fractures on bovid limb bones produced bydifferent agents: (a) human action: Namib Houentols; (h) spotted hyena:Kruger NalionaJ Park; (e) brown hyena: KaJahari Nalional Park; (d)leopard: Kruger National Park.

Bone Accumulations in Southern African Caves: A Search for Interpretative Criteria 14l

superficial cracks that develop on weathered bones haveorientations identical with those that may be produced bythe artificial split-line technique. Such cracks and linesfollow the main structural orientations in bones (Tappen1970) and, when bones are subjected to impact, the frac­tures also tend to preferentially follow such orientations.The weathered giraffe humerus in figure 150, showssuperficial cracking that follows the structural orientationof the bone in a spiral fashion. When such a humeral shaftfractures, through either stone impact or camivore toothpressure, the fracture line is very likely to follow the spi­ral path dictated by the structural orientation of the bone.

The appearance of spiral fractures on humeral andother limb bone shafts therefore tends to reflect more ofthe intemal structure of the bone than the nature of thetrauma the shaft has suffered.

Bone Flakes: Human-Made or Hyena-Made?

The shattering of long-bone shafts through either stoneimpact or camivore tooth pressure typically producesbone ftakes, as defined in chapter 1. For interpreting boneaccumulations, it would be useful if bone ftakes producedby the two agencies could be separated. As figure 151demonstrates, ftakes of the two kinds frequently bothshow impact points, with negative f1ake-scars on theirundersides where the hammerstone blow, or tooth pres­sure, was applied. I am not confident in my ability toseparate human-made from hyena-made bone ftakes byshape or by structure.

It has not proved easy to make a large colIection ofbone ftakes from hyena feeding sites or denso This is notbecause spotted and brown hyenas do not produce abun­dant flakes, but because such ftakes are rapidly gulpeddown by the hyenas or carried away by vultures that feedthem to their young (Mundy and Ledger 1976; Plug 1978).To date I have been able to examine only 220 bone flakes

Fig. 150. A giraffe humeros showing wealhering cracks thal follow themain slroctural orienlalion in the bone in a spiraJ manner. Fracture oftheshaft of such abone is likely lO follow a similar spirnJ course.

produced by spotted hyenas and 34 made by brownhyenas in natural circumstances. The lengths of theseflakes are given in tables 73 and 74. The impression givenby these small samples of bone f1akes resulting fromhyena action is that many pieces are longer than thosetypically produced by late human bone-breakage. Muchlarger hyena-produced samples are required te allowcomparison with the extensive human-activity samplesdescribed in chapter 3. For what the exercise is worth,length distributions of the 220 f1akes produced by spottedhyenas are compared with those of the vastly morenumerous f1akes from Pomongwe Cave (fig. 152).Whereas most of the flakes from the human refuse havelengths of less than 5 cm, those resulting from hyenafeeding have a fairly even length spread between 2 and 12cm. These results suggest that the variability in hyena­made flake lengths is greater than that in the humanlyproduced f1akes. It wiJI be interesting to see if larger sam­pIes collected in the future will support this indication.

Problems of Wear and Polish on Bones

It is tempting to interpret wear and polish on bones asresulting from .deliberate human use. Yet, as was ex­plained in chapter 2, such an interpretation may be in­conecl. Wear, both localized and general, as well aspolish, can result from the abrasion of sand regularly dis­turbed by the feet of animals. The positive identificationof wear or polish resulting from deliberate human usewould depend on microscopic examination of scratchesor other marks and the exclusion of natural agencies.

Recogniúng Hyena Food Remains

The teeth of hyenas are specialized for the destructionof bones, and although the damage they cause may bebasically similar to that inflicted by other camivores, thedegree of the damage is greater.

As was outlined in chapter 4, hyena action on bonesfalls into two clear categories: premolar cracking of man­ageable pieces and incisor/canine gnawing of larger, un­crackable pieces. The results of the two types of chewingare easily recognizable, as is shown in figures 61-64.

Specific Attributes ofBrown Hyena Food Remains

It is probably not possible to separate food remains ofspotted and brown hyenas on the basis of damage suf­fered by individual specimens. However, as was de­scribed in chapter 4, brown hyenas actively hunt largenumbers of smaIl mammals, particularly other carnivores,and feed the bodies to their cubs. The craniums of thesesmall camivores frequently remain intacl. Bone as­semblages in caves with high camivore-ungulate ratioscould very easily result from brown hyena cub rearing.

The Possible Role ofExtinct Hyaenids

As was emphasized in chapter 4, the extinct huntinghyenas of the genera Hyaenictis and Euryboas are wellrepresented among the fossils from the Transvaal caves.They may wel! have reared their young in these caves. Itis likely that they were social, cursorial hunters, like wilddogs, and that they preyed on antelopes of al! size classes.They may have contributed very significantly to the fossilbone accumulations.

142 A Guide to the lnterpretation of Bone Accumulations in African Caves

, \ ,

~ J''l¡ .,".\

" ,~

~ io.'::f.,

Leopards are prone to develop strong preferences forcertain types of prey and to seek these preferred animalswith considerable perseverence. A preference for pri­mates, either babooos or people, has been shown todeveJop in various areas throughout the leopard's geo­graphic raoge.

Leopard Damage Lo Eones

Sorne information has come to Iight on characteristicdamage caused by leopards to the skeletons of their prey.In dassies this is highly characteristic (fig. 93). Damage toimpala-sized antelopes is equally diagnostic (figs. 103 and

a

Recognizing Leopat"d Food Remains

Evidence presented in chapter 4 makes it clear that con­temporary leopards use caves both as breeding lairs andas feeding retreats. Food remaios tend to accumulate inboth types of site.

As was índicated in table 48 aod figure 97, leopardsprey 00 a very wide spectrum of animals but show astrong preference io a11 study areas for antclopes in sizeclass n. Observations on kills alooe, however, are io­clined to provide a picture strongly biased toward largerprey (table 50; fig. 98). Analysis of scats provides data onthe numerous smaller prey items not oormally observedas kills.

Fig, 151. Examples of bone flakes produced from bovíd limb bones by (a) a stone chopper on an anvil; (v) lhepremolars of spolted hyenas in the Kruger Naliona] Park, Arrows indicate thc poi nls al which impact or pressure wasapplied to cause the fractures.

Bone AccumuJations in Southern Afrícan Caves: A Search for Interpretative Criteria 143

The Possible Role of Extinct Cats

;\'. ."

a

.....\. A._

b

¡ndicators of the carnivores that left them. As was de­scribed in chapter 4, leopards and cheetahs characteristi­cally ingest almost the whole body, shearing off the backof the skull and mandible in an unmistakable manner (figs.93 and 95). Black eagles typically damage the dassiecraniums with their sharp beaks, creating equally un­mistakable openings (fig. 109), and human dassie-eatersleave a scatter of unchewable skeletal parts (table 15) thatconform to a predictable and consistent pattem.

The Incídence of Camivore Tooth Marks on Bones

In collections of bones accumulated by camivores it isusually possible to find certain pieces bearing c1ear evi­dence of the meat-eater's chewing action. Damage thatcharacteristicalIy results when a particular species of car­nivore, such as a leopard, feeds on a specific prey animal,such as a dassie, has already been discussed. Typical re­sults of hyena bone-cracking and bone-gnawing have alsobeen mentioned.

The recognition of camivore tooth marks on bones is arather subjective procedure, and not al! investigatorswould agree on the criteria to be considered. Por thisdiscussion I will restrict my observations lo "toothmarks" in the' form of (a) punctate marks and gougescaused by pointed teeth and (b) ragged-edge damagewhen a bone has been "worried" by the teeth of a camí­vore and an irregular margin has resulted (see figs. 80 and93).

HUMAN ACTlON

8

4

As is discussed in chapters 4 and 9, severa1 extinct cats,particularly the false sabertooths, Dinofelis, and the truesabertooths, Homolherium, Machairodus, andMeganter­eon, appear to have frequented the Transvaal caves andmay have contributed lo bone accumulations there. Di­nofelis is thought to have hunted and behaved as a verylarge leopard may have done, while the true saber-toothedcats would have been well suited to killing very largeprey, but quite unsuited to utilizing their bones. Truesabertooth food remains, therefore, should be charac­terized by mínimal bone damage.

104), leading to the "eaten-ouC' appearance of the car­cass in which only the head and lower limb segmentssurvive. With primate prey lhe whole body, except thehead, is prone to disappear.

Matching Observed Damage with Specific CarnivoreAction

%Abundance

It should be possible to match the kind of damage ob­served on bones in an assemblage with that characteristi­cally done by a particular camivore. Figure 153 showsthree humeri whose proximal ends have been removed bycamivores. In each case the distal end of the bone provedto be beyond the chewing ability of the speeific camí­vore's dental battery, unless unnatural effort were ex­pended. In example e, a spotted hyena has gnawed off theproximal end of a gíraffe humerus but has abandoned thedistal end; in lJ a leopard has done the same to an impalahumeros, and example a shows a dassíe humerus---Qne ofmany from the Pomongwe Cave-in whteh the proximalend was apparently chewed away by a Stone Age hunter,while the distal end was simply discarded into the ashof his fire. In fact, dassie remains prove to be excel!ent

~Abundance

Length of bone lIakes

cm

Fig. 152. Histograms showing lhe percentll8e abundance oCbone f1akes ofvarious lengths produced by Stone Age human aclivity in Pomongwecave, and spolled hyena activilY in the Kroger National Park. The samplesizes arc nol equivalent: 9,549 bone f1akes from Pomongwe are comparedwith 220 produced by hyenas.

Fig. 153. The damage observed on bones may be matched lO that charac­lerislically done by particular carnjvores. The three humeri shown he re aUh.ad lheir proximal ends removed by carnivorc action, and the scale ofdamage may be equated wilh the chewing abilily of lhe particular carní­vore concerned: (a) a dassie humeros damaged by human teelh; (b) animpala humeros chewed by a leopard; (e) a giraffe humeros gnawed bysponed hyenas.

144 A Guide to the lnterpretation of Bone AccurnuJations in African Caves

It would be useful to know the incidence of bonesshowing these recognizable tooth marks in accurnutationswhere camivores are known to have been involved.Further studies are urgently needed, but sorne informa­tion may be derived from those presented in this book.Bones from leopard lairs will be considered ñrst. Of the176 pieces collected in the Suswa leopard feeding cave,50. or 28.4%, showed clear tooth marks (table 42); 64bones thought to represent leopard food remains werecollected from the Quartzberg cave, and of these 15, or23.4%, bore tooth marks (table 45). Percentages oftooth-marked bones from the two leopard breeding lairswere much lower, as is perhaps to be expected sinceleopards rarely bring food back to the places where cubsare being reared. In fact, most of the bones in thePortsmut and Hakos River breeding laírs (tables 43 and44) were probably introduced by agents other than

Jeopards. The Portsmut lair produced 192 bone pieces, ofwhich 7. or 3.6%, showed tooth marks. Of tbe 333 bonesfrom the Hakos River lair, only 7, or 2.1%, of the totalwere so marked.

The collection of 235 bones from brown hyena breedingdens in the Kalabari National Park (Iable 41) was foundlo contain 96 tooth-marked bones. implying that 40.8%of the total showed recognizable evidence of camivoreaction.

A conclusion to be drawn from these observations isthat the frequency of tooth-rnarked bones in abone ac­cumulation of known camivore crigin is likely to be sur­prisingly low. Fewer tban 30% of bones left by leopards infeeding lairs may show tooth rnarks, and fewer than halfof the bones transported, red upon, and discarded bybrown hyenas at their dens may bear such traces.

I

rI,

Part 2Fossil Assemblages from the Sterkfontein Val1ey Caves:Analysis and Interpretation

9 The Fossil Animals

The remains oí animals preserved in the SterkfonteinvaUey caves fall naturally into two groups according tohow they found their way to the fossilization sites. Onegroup of typically larger animals, referred to here as lbe.' macrovertebrate component," contributed their skeletalremains in a variety of ways; the second, consisting ofcharacteristicaUy small animals termed the "microverte­brate component,' almost certainly fonned the prey oíowls that roosted in the caves. These remains becamefossilized from pellets that owIs regurgítated 00 the cavefloors. The composition of the macro- and mícroverte­brate components will be considered separately.

The Macrovertebrate Component

The fossils by which each species oí animal is representedal the various cave sites and in the different site units willbe considered in subsequent chapters devoted lo specificlocalities. The purpose of this ehapter is to provide someinformation about the animals whose remains are repre­sented by the fossils. Sorne of the species have livingrepresentatives and are therefore familiar; others are ex­tinet ando to a greater or lesser extent, unusual or bizarreo

As table 75 shows, the animals are manunals, birds,reptiles, or mol1usks. Of the 67 genera represented, 24 areextinct, and 54 of the 91 speeies involved have no livingrepresentatives. The order ofpresentation in the table andtext does not fol1ow a phylogenetic scheme but retlectsthe taphonomic significance of the laxa involved. Theprovenanee quoted refers to the Sterkfontein valley siteunits only.

Order Primates

Family HominidaeAustralopithecus africanus DartProvenance: Sterkfontein Member 4

A definition:

A species of the genus Australopithecus characterisedby the fol1owing features: more gracile, líghter con­structíon of the cranium; calvaria hafted lo facialskeleton at a high level, giving a distinct though notmarked forehead and a high supraorbital height index;ectocranial superstructures and pneumatisation not asmarked as in other speeies; sagittal erest commonlyabsent though probably present in sorne individuals;nuchal crest not present, but slight to moderate occip­ital toros commonly present; hony face of moderate

height and varying from moderately tlat and orthog­nathous to markedly prognathous; nasal region slightlyelevated from facial plane; ramus of mandible of mod­erate height and sloping somewhat backward; jawsmoderate in size with lesser development of zygornaticarch, lateral pterygoid plate, temporal crest and tem­poral fossa: palate of more or less even depth, shelv­ing steeply in front of the incisive foramen; premolarsand molars of moderate size and not so markedly ex­panded buccolingually; M3 smaller than M2 inrnesiodistal diameters, but equal in buccolingual diam­eters; mandibular canine larger than in other species,and hence more in harmony with the postcanine teeth;degree of molarisation of lower first deciduous molarless complete; cingulum remnans or derivatives pres­ent on all maxillary molars, weak on buccal surfaces,pronounced on lingual. representing an earlier or moreprimítíve stage in the trend towards reduction oí thecingulum; sockets oí anterior teeth arranged in a mod­erate to marked curve. [Tobias 1968a, pp. 293-94]

The first gracile australopithecine specimen lo be de­scribed was the child skull from Taung that, in his Naturepaper of 7 February 1925, R. A. Dart designated A. af­ricanus in commemoration "first, ofthe extreme southemand unexpected horizon of its discovery, and secondly, ofthe continent in which so many new and important dis­eoveries connected with the early history of man haverecently been made, thus vindicating the Darwinian elaimthat Mrica would prove to be the cradle of mankind." Inhis now classic paper, Dart continued (I925a, pp.19S-99):

It will appear to many a remarkable fact that an ultra­simian and prehuman stock should be discovered, inthe fírst place in Ibis extreme southern poiot of Mrica,and, secondly, in Bechuanaland, for one does not as­sociate with the present climatie conditions obtainingon the eastern fringe of the Kalahari desert an envi­ronment favourahle to higher primate life.... Inanticipating the discovery of the true links between theapes and man in tropical countríes, there has been atendency to overlook the fact that, in the luxuriantforests of the tropical belts, Nature was supplying withprotligate and 1avish hand an easy and sluggish solu­tion, by adaptive specialisation, of the problem ofexistence in ereatures so well equipped mentally asliving anthropoids are. For the production of man adilIerent apprenticeship was needed lo sharpen the

147

148 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

wits aod quicken the higher manifestations ofinteHect-a more open veldt country where competí­tion was keener between swiftness and stealth andwhere adroitness of thinking and movement played apreponderating role in the preservation of the species.Darwin has said, "no country in the world abounds ina greater degree with dangerous beasts than SouthemAfríca" and, in my opinion, Southem Africa, by pro­viding a vast open country with occasional wooded

belts and a relative scarcity ofwater, together with afíerce and bitter marnmalian competition, fumished alaboratory such as was essential to this penultimatephase of human evolution.

In Southern Africa, where climatic conditions ap­pear to have fluctuated little since Cretaeeous timesand where ample dolomitic formations have providedinnumerable refuges during life, and buríal-places afterdeath, for our troglodytic forefathers, we may con­fidently anticípate many complementary discoveriesconceming this period in our evolution.

And, in fact, many complementary discoveries wereindeed waiting 10 be made. The next gracile au­stralopithecine specimen carne to light eleven years lateron 17 August 1936 in tbe Sterkfontein lime quarry, laterknown as tbe Type Site, It consisted of an endocranialcast and part of tbe skull, TM 1511 (fig. 163), of an adultthat Broom tentatively assigned lo the same genus as tbeTaung specirnen but tbat appeared lo belong to a newspecies (Broom 1936, p. 488). After the furtber discoveryof a lower molar, Broom (1937a) proposed tbe name Au­stralopithecus transvaaíensis, but he created the newgenus Plesianthropus after the discovery of the man­dibular symphysis ofa child, TM 1516, which appeared lodiffer in structure from that of the Taung fossil (Broom1938b). Particulars oftbe first postcranial skeletal e1ementof P. transvaalensis were published in the same year(Broom 1938e); the specimen consisted of a 1eft distalfemur piece designated TM 1513. Details of an os mag­num foUowed (Broom J941a).

Descríptions of a11 specimens available at the end of tbewar were pubJished in two Transvaal Museum memoirs(Broorn and Schepers 1946; Broom, Robinson, andSchepers 1950), and renewed Sterkfontein excavations in19471ed to the finding ofSku1l5, "Mrs. PIes," on 18 April(fig. 155). This specímen proved to be tbe most complete

adult australopithecine cranium known. It was pictured intheIllustrated London News of 17 May 1947, and the firstdescription appeared in an issue of Nature of the samedate (Broom 1947). Further anatomical descnprícns werepublished immediately thereafter (Broom and Roblnscn1947a). A remarkably complete pelvis with articulatedvertebral column, found on 1 August 1947 (Broom andRobinson 1947b), provided the first informatíon about thisregión of the skeleton (fig, 165).

After paleontologícal reconnaissance work in theMakapansgat valley led by P. V. Tobias, the first gracileaustralopithecine specimen from the MakapansgatLimeworks was found by l. W. Kitching and his brothersin September 1947. It consisted ofthe occipital region ofan adult and was followed the next year by A. R.Hughes's discovery oí an adolescent mandible (Dart1948a). Pelvic and other bones subsequently carne tolight.

The Makapansgat australopithecíne was described byDart (l948b,e) as A. prometheus, the specific nameprompted by Dart's suspicion that blackened bones as­socíated with the hominid remains had been deliberatelybumed. Dart also concluded that the Makapansgat aus­tralopithecines had been responsible for building up theremarkabJe bone accumulation in the Limeworks graybreccia, or Member 3 of the Makapansgat Formation(Brock, McFadden, and Partrídge 1977), as well as for the"osteodontokeratic culture" (Dart 1957b).This issue willbe touched on again in chapter 13.

Two hominid fossils found al Garusi in Tanzania in1939 by Kohl-Larsen were described by Weinert (1950,1951) as Meganthropus afrieanus. Robinson (l953a) re­examined the specimens and concluded that they hadclose affinities with Plesianthropus from Sterkfontein.They therefore represented the first indications of agracile austraJopithecine outside South Africa.

The following year Robinson (1954) discussed tbe gen­era and specíes of tbe Australopithecinae in detail andconcluded tbat a11 known gracile australopithecine speci­mens should be referred to a single species Australo­pithecus africanus, which would contain two subspecies:A. a. africanas, for the Taung fossil, and A. a. trans­vaalensis, for the material from Sterkfontein,Makapansgat, and Garusi. This scheme would involvesinking the genus Plesianthropus, the taxon Megan­thropus africanus, and the species prometheus. Sub­sequentIy, Dart (I964a) agreed tbat the MakapansgataustraJopitbecine shou1d not be separated specificallyfrom the Sterkfontein one. In tbe interim he had describedremains from higher levels in the Makapansgat sequence:from Member 4 (Dart 1959d) and Member 5 (Dart 1955).

Renewed fieldwork at tbe Sterkfontein site (Tobias andHughes 1969)has recovered furtber gracile australopitbe­cine remains, including the fine cranium desígnated StW12113/17 and mandible StW 14 (Tobias 1973b). Fartbernortb on tbe continent new finds attributed to A. cf. af­ricanus are steadily accumulating.They include the man­dible piece from Lotbagam HiU in northem Kenya datingto the Pliocene but witb clear gracile australopithecineaffinities (Patterson, Behrensmeyer, and SiU1970;Tobías1975; Smart 1976); the distal humeral fragment fromKanapoi in the same geographic area (Patterson andHowells 1967); part of a ríght temporal bone from tbeChemeron Beds west of Lake Baringo (Martyn and To­bias 1967); some of the specimens from Olduvai Gorge,such as OH 13 and OH 24 and the cranium KMN-ER

1813 from the Koobi Fora Formation (R. E. F. Leakey1976a,b).

Recently sorne exceptionally fine specimens, appar­ently representing aD early fonn of gracile australopitbe­cine, have been found in Hadar, Ethiopia. Much of askeletcn (AL 288-1) carne lo light on 24 November 1974(Johanson and Taieb 1976), and the follcwing year re­markably complete skeletons of severaljuvenile and ádultindivíduals were discovered in clase approximation­perhaps the remains of an australopithecine band buriedby a flash flood (Johanson 1976).

To accornmodate these finds, as well as others rnade byM. D. Leakey al Laetoli, a new taxon, Australopithecusaforensís. has been proposed (Johanson, White, andCoppens 1978; Johanson and White 1979). It may wellprove to be ancestral to A. afríconus.

A new development conceming the antiquíty of theoriginal Taung skull is oC sorne interest. Two independentstudies by Partridge (1973) and Butzer (1974b) havesuggested that the type specimen of A. africanus is likeLylo be far younger than previously thought. Partridge'sgeornorphological analysis indicates an age of less thanone miJlion years, and Butzer (1974b, p. 382) suggeststhat "Taung is contemporary with or even younger thanSwartkrans and Kromdraai, rather than broadly coevalwith Sterkfontein and Makapansgat." The implications oíthis new ege assessment have been considered by Tobias(l973a). He poínts out that the lime gap between theTaung skull and the next youngest gracile australopithe­cine specimen may be as much as two million years andthat during this gap robust rather than gracile australo­pithecine populations appear to have existed. Tobias goesso far as lo suggest that the Taung child rnight in faclrepresent a juvenile robust australopithecine rather thanan A. afrícanus of the kind known from Sterkfontein andelsewhere. The kind of taxonomic muddle that would re­sult if the type specimen of A. africanus proved 10 repre­sent the robust rather than the gracile lineage has beenconsidered by Olson (1974).

In a short but significanl paper, Washbum and Patter­son (1951) made the following staternent: "the great evo­lutionary importance of the .man-apes' is that they showthe ínitial division between ape and man to be alocomotor adaptation." The skeletal remains of Aus­tralopithecus showed that, contrary to expectation, thesehominids walked erect before the major expansion of thebrain charaeterislic of Horno took place. As with otheranimals, various parts of lbe body evolved al slrikinglydifferenl rales.

This "mosaic nature" of human evolution has beenconsidered in detail by McHenry (l975b). He poinls outthal lhe idea lhal bipedalism mighl precede lbe expansionof lbe .brain goes back lo Lamarck, who expressed bisideas on lbe subjeel in 1809; lhereafter lhe concept wasfurther developed by bolb Haeckel and Darwin beforefalling inlo temporary disrepute arnong paleoan­thropologists. McHenry mal<es lbe poinl lbal lbe aus­traIopithecine pelvis was basically human in structure.though certain changes were to come about in Horno,where fetuses with enlarged brains necessitated the evo­lution of widened birth canals.

Attempls have been made by McHenry (l975a) loestimate body weight of A. africanus by detennining thecross-sectional area of vertebral centra. A figure of27.6:!: 10.5 kg (60.8 :!: 23.0 lb) for lhe specimen STS 14(which ineludes the pelvic bones referred to aboye) has

The Fossil Animal 149

been obtained, though, as McHenry (l975b) points out,this is probably a low value for the population as a whole.On the basis of a larger vertebra (STS 73), McHenry'sestimate is 43.0 kg (95 lb).

Various estimares have been made of the cranial ca­pacíty of South African gracile australopithecines. Robín­son (1966) gives a figure of 430 cm) for the mean cranialcapacíty of six specimens of A. ofrícanus, and Holloway(l970b) arrives at a figure of 442 cm'. Holloway's (l970a)estimate of the cranial capacity of the Taung specimenwas 405 cm). which would have led 10 an estimated adultvolume of 440 cm)-appreciably less than Darts orig­inal estimate of 525 cm'. Tobias's (1971) estimates areappreciably higher than HoHoway's with a mean for sixSouth African gracile australopithecines of 494 cm). Hisestímate of the population range, based on the samplemean :!: 3 SD, gives values of 370-618 cm'.

Austraíopithecus robustus (Broom)Provenance: Swartkrans Member 1 and Kromdraai B

A definition:

A species of the genus Australopithecus characterisedby the following features: more robust, heavier con­struction of the cranium; calvaria hafted 10 facialskeleton at a low level, giving a low or absent foreheadand a low supra-orbital height index ; well-developedectocranial superstructures and degree of pneumatisa­tion (more marked than in A. afrtcanus, though not aspronounced as inA. boisei); moderate to markedsupra-orbital toros with no "twist" between the me­dial and lateral cornponents; sagittal crest normallypresent; small nuchal crest commonly present; bonyface oflow lo moderale height, and flat or orthognath­ous; nose set in a central facial hollow; ramus of man­dible very high and vertical; jaws large and robusl withstrong development of zygomatic arch, lateralpterygoid plale, temporal crest and temporal fcssa:palate deeper posteríorly than anleriorly; shelvinggradually from the molar región forwards; prernolarsand molars of very Iarge size; M3 cornmonly largerthan M' in both buccolingual and mesiodistal diame­ters; mandibular canine absolutely and relatively smalland hence not in harmony with the postcanine teeth:degree oí molarisation of lower first deciduous molarmore complete; cingulum remnants on1y weakly repre­senled on lingual face and absenl on buccal face ofmaxillary molars, representing a more advanced stagein reduction of!he cingulum: sockets of anterior teetharranged in a low lo moderate curve. [fobias 1968a,pp. 294-95J

The first specimen of a robust australopithecine carneinto Broom's hands on 8 lUDe 1938: the circumstancessurrounding this discovery are discussed in chapter 12.The speeimen Broom (l938b) desiguated as lhe lype ofParanthropus robus/us (Iig. 201) consisted of lbe left sideof a partial cranium and a nearly complete righl man­dibular ramus. Broom estimated that the cranium had acapacity of 600 cm) and observed that the face was re­markably flat and much shorter lhan in a gorilla. Althoughthe second premolar proved lo be half as large agajn as inPlesianthropus, Broom found that the canines and in­cisors were relatively smalL

Laler lhe same year Broom (193&) described lbe firslof the poslcranial parts associated wilh lbe type skull.They consisted of a right dislal humerus, part of a righl

150 Fossi1 Assemblages from the Sterkfontein ValleyCaves: Analysis and Interpretation

¡í

AWIt,."lopi th"wIPGr"~t~ropu.

son explained the differences between the two forros interms of different dietary specializations-he sawParanthropus as a vegetarían, continuing the conserva­tive tradition of the early hominid stock, and Australo­píthecus as an omnivore, taking meat as well as vegetablefood. Robinson visualized two critical points orthresholds where adaptive shifts in the evolutíon ofhominids occurred: the first was the change troro quad­rupedal to bipedallocomotion; the second was the inclu­sion of meat in the diet. He saw the second threshold asresulting from increasing aridity of the environment inwhich Australopithecus lived-from the need to supple­ment dwindling plant food sources with animal protein.Robinson (1963, p. 413) visualized Paranthropus as rep­resenting the basic stock from which Australopithecusand Hamo arose and as retaining certain apelike charac­ters, such as a relatively long ischium that suggested(Robinson 1972)that Parantbropus may have been betteradapted lo tree-elimbing than was Australopíthecus orHorno.

Tobías (1967) eould not agree that the anatomical dif­ferences between robust and gracile australopithecineswere of great significance in tenns of diet or ecologicalseparation. Formerly, L. S. B. Leakey, Tobias, andNapier (1964) had proposed that Australopítñecus andParanthropus should be regarded as subgenera of Aus­tralopithecus; subsequentiy Tobias (1967) proposed that

Later he continuec (1943, p. 690):

Paranthropus, in having a brain of about 650 ce. and avery lightJy built body of probably not more than 80lb. or 90 lb., and in being most probably bipedal, withdelicate slender hands, must have been an animalmuch more like a human being than either the chím­panzee or gorilla. Like the other known Au­stralopithecines, it was certainly not a forest-livingPrimate. but a being who lived among the rocks andon the plains and hills as the baboons do today.

Fieldwork started by Broom and Robinson atSwartkrans in November 1948 resulted in their finding awealth of robust australopithecine remains. Part of amandíble, now known as SK 6 (fig, 192), was desígnatedthe type of a new species, Paranthropus crasstdens, byBroom (1949). Clearly related to the Kromdraaí australo­pithecine, the new form appeared to have more massiveteeth. In the same year Broom and Robinson (l949b)reported the discovery of a thumb metacarpal ofP. eras­sídens, and the abundant remains from Swartkrans, in­cluding a pelvie bone, SK 50, available at the time ofBroom's death, were described in a Transvaal Museummemoir (Broom and Robinson 1952).

In his aceount ofthe genera and species ofthe australo­pithecines, Robinson (1954) listed the main differences, ashe saw them, between Paranthropus and Australopithe­cus. He remarked on the structure of the deciduous mo­lars and of the anterior par! of the nasal eavity lloor; heobserved that the Paranthropus P' tended to have multi­ple roots more commonly than was the case in Aus­tralopithecus; that the canines tended to be smaller thanin Australopithecus; and that the skull shape a1so dif­fered. Robinson proposed that a single species of robustaustralcpíthecine should be reeognized in South Africa,P. robustus with two subspecies, P. r. robustus fromKromdraai, and P. r, crassidens from Swartkrans.

Robinson (1963) developed bis concept that structuraldifferences in the craniums of Paranthropus and Aus­tralopithecus resulted primarily from differences in themastícatory apparatus of the two forms. He pointed outthat proportions in the anterior-posterior parts of theAustralopithecus dentítion followed a human patternwhile those in Paramhropus were strongly aberrant. An­terior teeth of Paranthropus appeared lo be much re­duced in size relative to the postcanine dentition, whichwas specialízed for grinding tough, fíbrous food. Robin-

proximal ulna, and a toe phalanx. These were followed bydescriptions of bones from a left hand (Broom 1942) andpar! of a right talus from the same block (Broom 1943).

Renewed fieldwork at Kromdraai in 1941 led to the dis­covery of a young Paranthropus child mandibie thatBroom (l941b) described as differing from tbat of Aus­traíopithecus from Taung in that the second incisors andcanines were much smaller in Paranthropus, while thefirst and second milk molars also differed in shape andarrangement of cusps. After a restoration of the Krom­draai skull, Broom wrote (1939b, p. 328):

In my opinion the Kromdraai skull differs from theSterkfontein in the unusually advanced position of theeheeks. They are so advaneed that if a ruler is placedon the two cheeks, the rest of the face is behind theruler. Also there is a remarkable degree of flattening ofthe lower part of the face aboye the incisors andcanines.

these two subgenera be sunk and that the two South Afri­can forms be designated A. africanus Dart and A. robus­tus (Broom). This terminology appears to have been fol­lowed by most anthropologists and paleontologists since.However, in his recent extensive review of thehorninidae, Howell (1978) retained the taxaA. crassídensfor robust australopithecines from Swartkrans Member 1,A. robustus for specímens from Kromdraai B. and A.boisei from a variety of East African localities.

Likewise, it is generally agreed that A. robustus was alarger animal than A. africanas. McHenry's (l975a) re­cent estimate of the body weight of a robust australopith­eclne, SK 3981, based on vertebral cross-sectional area,gave a figure of 36.1 :!: 10.7 kg (79.3 ± 23.5 lb), but this isregarded as a minimum adult value. On the other hand,McHenry regards Robinson's estimate (1972) of therobust australopithecine adult weight range of 150-200 lbas excessive. Despite the larger body size, the craníalcapacities of robust australopithecines may not have beengreater than those of their gracile counterparts. Holloway(l970b, 1972) dedueed capaeities of 530 cm' for both theSwartkrans endocast, SK 1585, and Olduvai Hominid 5(Zinjanthropus)-figures within the sample range forgracile australopithecines (Tobias 1971).

Apart from the two originallocalities of Kromdraai andSwartkrans, robust austraJopithecines are now knownfrom a variety of sites. These inelude Olduvai Bed 1,source of the skull of Australopithecus bolset, a craniumthal has been treated to more detailed description (Tobias1967) than any other; Peninj, near Lake Natrón (L. S. B.Leakey and M. D. Leakey 1964); East Rudolf (R. E. F.Leakey 1973, 1974, 1976b); the Omo valley (Howell andCoppens 1976); and the central Mar (Johanson and Taieb1976). In addition, Robinson (l953a) has claimed that"Meganthropus palaeojavanicus:' from Java should beregarded as a robust australopithecine. Ifhe is right, thesecreatures must have ranged widely through Africa and theFar East, a concept that has been challenged by severalauthors (e.g., Le Gros Clark 1978; Tobias and vonKoenigswald 1964; R. E. F. Leakey and Walker 1976)who believe there is no incontestable evidence for theformer presence of Australopithecus outside of Africa.

Horno sp,Provenance: Slerkfontein Member 5, Swartkrans Mem­bers 1 and 2

A definition of the geaus Horno Linnaeus:

A genus of the Hominidae with the following eharac­ters: the structure of the pelvic girdle and of the hind­limb skeleton is adapted lo habitual ereet posture andbípedal gait; the fore-limb is shorter lban the hind­limb; the pollex is well developed and fuUy opposableand the hand is eapable not only of a power grip butof, at the Ieast, a simple and usuaíly well developedprecision grip; the cranial capacity is very variable butis, on the average. larger than the range of capacitiesof members of the genus Australopithecus, ahhoughthe lower part oC the range of capacities in the genusHorno overlaps with the upper part cf the range inAustralopuhecus; the capacity is (00 the average)large relative to body-size and ranges from about 600c.c. in earlier forms lo more than 1,600 c.c.: the mus­cular ridges on the cranium range from very stronglymarked to virtuaUy imperceptible, but the temporalcrests or lines never reach che rnidline; the frontal re-

The Fossil Animals 151

gion of' the cranium is without undue post-orbital con­striction (such as is common in members ofthe genusAustra/opithecus); the supra-orbital región of thefrontal bone is very variable, ranging from a massiveand very salient supra-orbital toros to a complete lackof any supra-orbital projection and a smooth brow re­gion; the facial skeleton varies from moderately prog­nathous to orthognathcus, but it is not concave (ordished) as is common in members of the Australo­pithecinae; the anterior symphyseal contour varíesfrom a marked retreat lo a forward slope, while thebony chín may be entirely lacking, or may vary from aslight to a very strongly developed mental trigone; chedental arcade is evenly rounded with no diastema inmost members of the genus; the first lower premolar isclearly bieuspid wilb a variably developed lingualcusp; the molar teeth are variable in size, but in gen­eral are srnall relative to the síze of these teeth in thegenusAustralopithecus; the size ofthe last uppermolar is highly variable, but it is generally smaller thanthe second upper molar and eornmonly a1so smallerthan the first upper molar; the lower tbird molar issometimes appreciably larger than the second; in rela­tion to che position seen in the Hominoidea as a whole,the canines are small, with little or no overlapping afterthe initíal stages of wear, but when compared withthose of members of the genus Australopíthecus, theincisors and canines are not very small relative to themolars and premolars; the teeth in general, and par­ticularly the rnolars and premolars, are not enlargedbucco-lingually as they are in the genus Australopithe­cus; the first deciduous lower molar shows a variabledegree of molarization. [Leakey, Tobías, and Napier1964, pp. 5--6]

Excavations conducted by Robioson at the Sterkion­tein Extension Site (Member 5) during 1957 and 1958 pro­dueed isolated hominid teeth and a fragmentary juvenilemaxilla in direct association with stone artifacts. Robin­son and Mason (1957) and Robinson (1958, 1962) con­cluded that these remains represented Australopithecussimilar in structure to the more abundant material fromthe Sterkfontein Type Site (Member 4). On the otherhand, Robinson attributed the artifacts to a more ad­vanced hominine, similar to Telanthropus (Horno sp.)from Swartkrans, that moved into the arca in the timespan separating Sterkfontein Members 4 and 5 and there-

152 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

after displaced Australopíthecus, Tobías (1965) was in­clined to regard sorne of the teeth as belonging to Hamo,perhaps H. habilis, a form presumably responsible for thestone culture found in association.

New evidence 00 this controversial matter carne tolight 00 9 August 1976 when Hughes discovered a re­markably complete hominid cranium, StW 53. weather­ing from the side oí a solution pocket in SterkfomeinMember 5. In the opinion of Hughes and Tobias (1977),the new specimen should be referred to Horno habilis orto a forro closely related; tbis hominine was presumablythe Sterkfontein toolmaker.

On 29 April 1949 a mandible was discovered in theouter cave al Swartkrans that appeared strikingly differ­ent from the remains already found there. It was de­scribed by Broom and Robinson (1949a) as belongíng loTelanthropus capensis, a fonn "somewhat allied toHeidelberg man and íntermediate between one of theape-rnen and true man." The mandible, now designatedSK 15 (fig. 186), was closely associated with parts of twopremolars (SK 43, ISa) and a proximal radius piece (SK18b), all apparently from the same individual.

A right mandibular corpus of an old adult, SK 45, wasfound in September 1949 and described by Broom andRobinson (1950) in a paper entitled "Man Contemporane­ous with the Swartkrans Ape-Man." Al about the samelime a maxillary fragment (SK 80) was discovered, andthis, together with the mandible pieces, was referred toTelanthropus capensis by Robinson (l953b) in his paper"Telanthropus and Its Phylogenetic Signíñcance." Thegenus Telanthropus was subsequently sunk by Robinson(l96l), and the Swartkrans specimens were lransferred loHorno erectus,

In July 1969 R. J. Clarke assembled hominid cranialfragmenls collectívely designated SK 846h and 847,which had previously beco referred to Paranthropusrobustus. After assembly it was found that there was aperfect join across the left side between the posteriorpalatal fragmenl of SK 847 and the maxillary fragrnent SKSO, previously c1assified as Horno. The compositecranium SK 846b1847/80 (fig. 191) was reallocated loHorno sp. in a paper "More Evidence of an AdvancedHominid al Swartkrans," by Clarke, Howell, and Brain(1970).

Wilh the separation of Swartkrans faunal remains bytheir stratigraphic provenance (Brain 19700), lbe compos­íte cranium SK 846b!847180 was allocated lo Member 1,while the mandible SK 15and associated remains (SK 18)were referred to Member 2. In the interirn, Clarke (1977)has concluded that remains of two more Horno índividu­als are present in the hominid fossil samplefromMember1. These are represented by a flattened juvenile cranium,SK 27, formerly lhoughl lo be an aberranl Paranthrapus,and lwo upper premolars. SK 2635.

On lhe basis of lbe Swartkrans evidence lbere can beno reasonable doubl lbal early representalives of lbegenus Horno coexisted witb robust australopithecinepopulations in lbe Sterkfontein valley, This was pointedoul for lhe firsl time by Broom and Robinson (1950).Since then. corroborative evidence oC such coexistencehas becn forthcomíng from a number of East African 10­calilies, including Olduvai Bcd I (L. S. B. Leakey, To­bias, and Napier 1964). Member G of lbe Shungura For­malion (Boaz and Howelll977), lbe upper member of lbeKoobi Fora Formation (R. E. F. Leakey and Walker1976), and Hadar, Elbiopia (Joh¡mson and Taieb 1976).

In concjusíon. there is no universal consensus amonganthropologists on which fossil forros should be includedin the genus Horno. While most authorities separare thegracile australopithecines generically from Horno. Robin­son (1972) does not; instead. he propases the taxonHomoafrícanus, which he separates from the genus Paran­thropus,

Family CercopithecidaeThe fossil baboons and monkeys frcm the Sterkfontein

valley caves were initially classified by Freedman (1957,1970) as foUows:Family: Cercopithecidae Gray, 1821

Subfamily: Cercopilbecinae Blanford, 1888Genera: Parapapia Jones, 1937

Papio Erxleben, 1mSimoplthecus Andrews, 1916Dinopithecus Broom, 1936Gargopithecus Broom and Robinson, 1949

Subfamily: UncertainGenus: Cereooithecoides Mollet!, 1947

More recently Eisenhart (1974) has reviewed the sys­tematic classification of the Cercopithecoidea and drawnup a scheme based on the writings of Delson (1973), Jolly(1972), Kuhn (1967), Maier (l970a), Napier and Napier(1967), and Simons (1972). In lhis scheme the genus Cer­copithecoides is plaeed in the subfamily Colobinae Blyth,1875, while the genus Simopithecus is reduced in status tothat of a subgenus of Tberopithecus I. Geoffroy, 1843.

The Genus ParapapioThe generic name Parapapio was first used' by Jones

(1937) in describing a collectíon of fossil baboon skullsfrom Sterkfontein. AlI the specimens in this collection,except one, were thought to belong lo a single newspecies, P. brooml; the exception was later found 10 be aspecimen of Cercopithecoides williamsi. As more mate­rial from Sterkfonteín became avaílable, Broom (1940)concluded thal there were in fael three species ofParapapio present. They differed in size-s-the largestspecies he namedP. whitei and the smallestP.jonesi. andhe retained the name P. broomi for baboons with inter­mediate dimensions. Representatives of these threespecies are now known from other southern African cavesites of Taung, Makapansgal, Swartkrans, Bolt's Farm,and Kromdraai A, and lbe genus also appears lo be repre­sented al various East African fossil sites (Howell, Fich­ter, and Eck 1969: Patterson 1968; Simons and Delson1978).

It does seem remarkable lbal three closely relaledspecies ofParapapío, differing onIy in síze, should havelived synchronously in the immediate vicinity oíSlerkfontein and Makapansgat and presumably elsewhereas well. Yel lhis scems lo have becn the case, and lbere is

..no evidence al present that the specimens from Member 4al Sterkfontein are separated in time or that the threespecies form a chronocline. They occur in approximatelythe same relative abundance at Sterkfontein andMakapansgat, with P. broomí specimens being the mostnumerous, followed by P. jonesi, then P. whitei, Thisstudy indicates that the collection studied from Member 4at Sterkfontein contained remains from a mínimum of 91P. broomí individuals, 37 P. jonesi, and 10 P. whitei (fig.167).

A recent reexamination in statistical terms oí theParapapio material from Sterkfontein by Freedman andStenhouse (1972) has confirmed the validity of the threespecies apparently occurring there.

MorphoJogically the dentitions of Parapapio and Papioare probably indistinguishable. The main anatomical dif­ference between the genera lles in the profile of the muz­zle when vlewed from the side, In Parapapio the nasalsand frontal s fcrm a straíght line or slightly concave curvefrom the glabella to the posterior margin of the nasalaperture, whereas in Papio the profile is sharply concaveimrnediately anterior lo the glabella.

In his recent detatled study of the skull of Papio ur­sinus, Jones (1978) concluded Íbat a distinction between

Paplo and Parapapio based un muzzle profile may not beas reliable as was previously thought, In a single troop ofbaboons from Zimbabwe, Jones found individuals withboth the typical Papio and the typical Parapapio muzzleprofiles.

Parapapio jonesí Brocm.Provenance: Sterkfontein Member 4, SwartkransMember 1 and Kromdraai A.

Parapapio broomi JonesProvenance: Sterkfontein Member 4

The Fossil Animals 153

The Genus Papio ErxlebenLiving baboons from Africa and fossil forms from both

Africa and India are included in this genus, the typespecies of which is Papío papio (Desmarest, 1820).Characteristic of the genus is the Long, doglike snout,particularly well developed in males. Viewed from theside, the profile of the muzzle is strikingly different fromthat of Parapapio (see iIlustration), in that there is amarked drop in the interorbital region and then a fairlygradual slope to the alveolar margin of the incisors. Sex­ual dimorphísm is marked, males being appreciahly largerthan females and possessing much more robust canines.

Five species of living Papio-P. anubís, P. cynoce­phalus, P. ñamadryas. P. papio, and P. ursinus-arecurrently recognized by many primatologists. These ap­pear to be allopatric species with eontaet at the marginsoftheir ranges and even a certain amount ofhybridizationthere. Severa! authors have therefore questíoned thestatus of the five species, and Buettner-Janusch (1966) hasproposed that a single po1ytypic species exists throughoutthe African range, Papio cynocephalus, which in­corporates the other four fonns, previously consideredspeciñcally distinct. For the purpose of this discussion,the only baboon currently oecuning in southem Africa,P. ursinus, is still regarded as a separate and validspecies,

Papio ursínus (Kerr, 1792). Chacma baboonProvenance: Swartkrans Member 2 (cf.)

The welI-known living Chacma babeen of southem Af­rica. widely distributed through a great diversity of

habitats-in fact, the success of the species can probablybe attributed to its adaptability and environmental toler­anee. Although several species oí baboons coexisted inthe Sterkfontein valley previously, P. ursinus is the onlyliving baboon in the subcontinent today.

Papia robtnsoni FreedmanProvenance: Swartkrans Member 1, Kromdraai A and B

A species of extinct baboon, very similar lo P. urstnusand almost certainly ancestral to it, origínally describedby Freedman (1957) on the basis of a large sample fromSwartkrans. The main difference between P. robínsoniand Papio ursinus Hes in the shape and structure of themuzzle: that of P. robinsoni is ñatter, with the maxillaeoverlapping the nasals to a greater extent than in the liv­ing chacma baboon.

According lo Freedman (1970) the taxon is also kncwnfrom Cooper's site, Swartkrans H, Gladysvale, Bolt'sFarrn, and Skurveberg.

154 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpreraticn

been marked. Adults were large, with fernales about thesize of an adult male chacma baboon ."

Dinopíthecus ingens BroomProvenance: Swartkrans Member 1

In appearance, D. íngens was probabJy like an excep­tionally large chacma baboon, with adult males weighingmore than 45 kg (100 lb).

The Genus Dínopithecus BroomVery large extlnct baboons with marked sexual di­

morpbism, robust skulls. and long muzzles. The teeth arelarge and bread, and the dental arcade is horseshoe­shaped, somewhat narrowed posteriorly. Maier (1973)has suggested that Dinopíthecus may have relationshipswith Gorgopíthecus and Papio.

The genus was originally described by Broom (l937b)on the basis of a mate jnandibie from Skurveberg. Freed­man (1957) subsequent1y assigned a large number ofspecimens from Swartkrans to the same taxon.

Sorne teeth from Omo were described by Arambourg(1947)as D. brumpíi, but accordíng lo Simons and Delson(1978) rnost of these can be referred lo Theroptthecus,The latter authors consider, however, that sorne un­described specimens from Leba in AngoJa may representDínopttñecus.

Theropíthecus iSímopitñecus¡ danieli (Freedman)Provenance: Swartkrans Member 1

A typical theropithecine baboon, known only fromSwartkrans, in which sexual dimorphism appears to have

become very upright , According to Simons and Delson(1978) the fcrelimb was long but the phalanges were shortand stout. These features reftecl lhe terrestrial life-style ofTheropithecus in open grasslands where resistant grassseeds formed the bulk of the díet.

Th"ropif.hI10U8 da.,.ti

Papto angustíceps (Broom)Provenance: Krorndraai A and B

The species was orígínally descríbed by Broom (1940)as Parapapio angusticeps 00 the basís of a female skullfrom Kromdraai A. The specimen was damaged in theinterorbital area, and Broom assumed that the muzzleprofile would have been that of a rypical Parapapio.However, when more complete material became avail­able, Freedman (1957) was able lo conclude that thespecies angusticeps definitely belonged in the genusPapio. characterized by the steep dip in muzzle profile inthe interorbital región.

The skull of P. angusticeps is similar lo that of P. ur­sínus in almost all respects except size. An adult male ofP. angusticeps appears to have been typically smallerthan a female ofP. ursinus,

The species has also been recorded from Cooper's siteand Minnaar's Cave.

The Genus Theropithecus GeoffroyThese are large baboons with hígh, ñat faces that form a

striking contrast lo the dog-faced baboons of the genusPapio. The cheek teeth show an unmistakeable pattem ofhigh cusps, separated by deep clefts, foveae. and fossae.Lingual cusps of upper molars and buccal cusps of lowerones are connected by high ridges of enamel.

The first specimen from a South African cave depositwas described by BlOOm and Jensen (1946) fromMakapansgat Limeworks as Papio darti but was sub­sequently transferred lo the genus Gorgopithecus byKitehing (1953). In his 1957 study, Freedman placed allthe Makapansgat specimens into a single species,Símopíthecus darti, and also described a second, largerfonn from Swartkrans as S. daníeli.

Jol1y (1966, 1970, 1972) argues that Simopithecusshould be regarded as a subgenus of Theropithecus, thegenus lo which the living gelada of Ethiopia belongs. Thísscheme is fol1owed here, However, Maier (l970a.b)favors retention of the genus Símopithecus for the fossilfonns and placea Theroplthecus 00 a related lineage thatdeviated fromSimopithecus in the course ofthe Pliocene.

Sorne of the cranial adaptations found in Theropíthecusappear lo mirror those of the robust australopithecines.There is a reduction in the size of the anterior teeth rela­tive to that of the molars, while the ascending ramus has

The Genus Gorgopíthecus Broom and RobinsonThe type specimen from the Kromdraai Faunal Site

was described originally by Broom (1940) as Parapüpío

majar and consisted of two considerably worn upper roo­lars. Later, when more complete material became avail­able from Kromdraai, Broom and Robinson (l949r)created the genus Gorgopithecus to accommodate it. Asecond species was subsequently described by Kitching(1953) from Makapansgat Limeworks as G. wellsi, butthis was sunk into Slmopithecus darti by Freedman(1957). Thus Gorgopíthecus rernains a monotypic genus.

On the grounds of size in particular, Delson (1975)suggested placing Gorgopithecus as a subgenus of Di­nopishecus, but this was not upheld in a subsequent re­view (Sirnons and Delson 1978).

Gorgopithecas majar (Broom)Provenance: Kromdraai A

An exceptionally well " preserved male skull fromKromdraai A (fig. 207) shows G. majar to be a large ba-

. ,

boon with a high, narrow, and relatively short muzzle.Other specimens indicate that, unlike Dinopithecus, Gor­gopithecus had very slight sexual dimorphism. Adultswere probably intermediate in size between Dinapithecusand Papío ursinus.

The Genus Cercopithecoides MollettIt is remarkable that, although all living rnonkeys in

southern Africa belong to the genus Cercopithecus, allfossil fonns from the austraIopithecine caves are col­obines, having no direct phylogenetic link with the livingrepresentatives.

Tbe genus Cercopíthecoides was ñrst described byMollett (1947) on the basis of a single male cranium andmandible from Makapansgat Umeworks, attributed lo C.wil/iamsi. Freedman (1957) subsequently described a sec­ond, larger species from Swartkrans that he called C.molletti. However, three years later he sank C. mollettiinto C. williamsi, which is currentIy the only recognizedspecies.

Cercopitñecoides monkeys were fairly large-s-certaínlyheavier than a living vervet monkey-with exceptionaUyshort muzzles and big brains. The large circular orbitswere widely spaced and lay beneath pronounced supraor­bital torio The teeth were characteristically colobine, mo­lars showing high cusps separated by deep foveae. Ac­cording to Simons and Delson (1978), Cercopitñecoídes

The Fossil Animals 155

was most cJosely related to Paracotobas on dental andfacial morphology.

Living colobine monkeys are typically arboreal animalsrestricted to evergreen forest areas. It is not known whatthe habitat requirements of Cercopitñecoídes monkeyswere-wooded ccuntry was doubtless required, butperhaps not as tblckly forested as is preferred by livingrepresentatives such as Cofobus. In fact, it seems that theCercoplthecoídes monkeys were using a niche later to betaken over by southern African vervet and samango mon­keys of the genus Cercopithecus,

Cercopíthecoides willíamsi Mollett (fig. 167d)Provenance: Sterkfontein Member 4. SwartkransMember 2, and Kromdraai B

Al one time this species was apparently common andwídespread. In southern Africa, remains have also been

found at Taung, Makapansgat Limeworks, Cooper's site,Swartkrans 11,Sterkfontein Graveyard, and Bolt's Farm,while in East Africa very similar remains have been re­covered from the Omo Group deposits and the Awashvalley (Eck 1976) and the East Rudolf succession (M. G.Leakey 1976).

Order Carnívora

Family FelidaeRemains of cats from the Sterkfontein valley caves

come from two familiar species, the leopard and lion, oneextinct and poorly known creature witb characteristicsboth ofthe leopard and the cheetah, and from a range ofextinct camivores showing sabertooth adaptations to agreater or lesser extent. The latter are c1assified in twosubfamilies: the Felinae or true cats and theMachairodontinae or sabertoothed cats.

Subfamily FelinaePanthera pardus Linnaeus, 1758. LeopardProvenance: Sterkfontein Member 4, Swartkrans mero­bers I and 2, and channel fill, Kromdraai B

On the basis of !he fossils from Member 1 atSwartkrans, Ewer described a new subspecies of theleopard, P. pardus tncurva, which differed from the livingsouthern African leopards on the following points: "an­gular process of the mandible scooped out rather after themanner ofa hyaena; anterior cusp oC p3 more reduced; MIlarger and milk carnassial with a distinct talonid; probablya trifle smaller than the living leopard" (Ewer 19560, p.84).

156 Fossil Assemblages from the Sterkfontein Valley Caves: Analysís and Interpretatíon

Smil.odon

Saber-toothed Cats: False and TrueThe sabertooth adaptation-the tendency to greatly in­

crease the size of the upper canines as lethal killingweapons, has been experimented with by various carni­vores over a long span of geological time. Saber-toothedforros existed in four subfamiHes of the Felidae: in theFelinae, Hoplophoneinae, Nimravinae. and Machairo­dontinae. The only living cat that shows a tendency 10­ward exceptional enlargement of the canines Is theclouded leopard of Asia, Neofelis nebulosa. In this case,however, there is enlargement of both upper and lowercanines-a situation not typical of the extinct saber­toothed cats, where the upper canine is greatly enlargedat the expense of the lower.

Ewer has discussed the characteristic sabertooth ad­aptation (l954a, 1973), pointing out that the killingmethod used by a normal feline is different from thatof a sabertooth, for example, a machairodont. Typicalanatomical differences between a feline and a machairo­dont are shown in the iIlustration. Although sabertoothsare now regrettabJy extinct, they were the first successful

Panthel"a

have been found in southem Africa thus far; the closest toit is perhaps F. obscura Hendey, from "E" Quarry atLangebaanweg. A single specimen is known that Hendey(1974a) considered to come from a eat slightly smallerthan a leopard with a p3 reminiscent of a cheetahAcinonyx jubatus (Sehreber). However, the fact that theLangebaanweg specimen lacked a p2 indicates that itsaffinities do not tie with Acinonyx or Panthera. Hendeypoints to the similarity between F. obscura andcats ofthegenus Sivafelis Pilgrim, known as fossíls from India andChina.

Remains from East Aftica are also coming to Iight.Ewer (1%5) tentatively assigned a mandible from OlduvaiBed II to F. crassidens, and three postcranial bones fromOlduvai have been referred to Panthera crassidens(Broom) by G. Petter (1973). More recently, two postcra­nial bones from the Koobi Fora Formation of the EastRudolf succession have been described underP. cf. eras­sídens by M. G. Leakey (1976). From the Shungura For­mation in the Omo valley Howell and G. Petter (1976)have also assigned a mandible and postcranial pieces to P.crassidens. They tentatively equate this taxon withSívofeíis potens Pilgrim from the Siwaliks of India.

Remains from the site units other than SwartkransMember 1 are too incomplete to allow allocation to aparticular subspecies--all that can be said is that theycarne from leopards. It appears that leopards were partof the Sterkfontein valley fauna throughout the timespanned by the various cave deposits.

have been examined by Hendey and assigned (pers.eomm.) to ef. Panthera leo.

Clearly the taxonomie status of the fossil lions from theSterkfontein valley caves is somewhat uncertain. Whatevidence we have suggests that the Swartkrans andKromd.raai Iions were more robust than their modemcounterparts.

Felis crassídens Broom. The "crassidens cat"Provenance: Kromd.raai A

This remarkable eat was described by BlOOm (1948a)on the basis of two maxillary pieces and one mandibularpíece, The teeth have features in common with both lheleopard and the eheetah but are typieal of neither; theyclearly carne from a short-faced feline of leopardlike pro­portions. No further remains of this tantalizing creature

Panthera leo (Línnaeus, 1758). LionProvenance: Sterkfontein Member 5, Swartkrans Mem­ber 2, and Kromdraai A

The isolated teeth from Krorndraai were described byEwer (19560) as corning from a lionlike felid but beinglarger and heavier than the corresponding teeth in a livinglion. She tentatively equated them with the large eat pre­viously described by Broom (l948a) as Felís shawi fromBolt's Farm. Hendey has reexamined the specimens andeoncluded that they ean best be assigned to P. leo (pers.comm.).

Ewer (19560) pointed out that, although the Swartkransfossils belonged to a tionlike cat, the P-1 is exceptionallybread in proportion to Its length and also shows unusualfeatures. She concluded: "From the cbaracters of P-I itseems clear that specimen SK 359 is not identical with theliving South African Lion, but whether the differenceshould be accorded specifíc or subspecific recognitioncannot be decided until further material becomes avail­able" (Ewer 19560).

The specimens from Member 5 at Sterkfontein consistof two isolated teelh and a right distal humerus. These

The Fossil Animals 157

fonns a single blade, and in advanced species theloolh is further lengthened by Ihe developmenl of anexlra anlerior lobe (a) in line wilh Ihe resl of Ihe blade(see Figures). In Ihese forms Ihe very long upper ear­nassial usually cuts against two teeth in the lower jaw;but the bone-crushing teeth in front of the camassialsare so reduced ás to be almost functionless, andc1early the sabre-tooths cannot have been able to cope

large felid predators, dominating the scene from theOligocene to the Pliocene.

A normal cat kiJIs its prey with its upper and lowercanines, forcefuIly driven in by the closing oí the jaws,through tbe action oí the large temporalis muscles thatron down 00 each side of the skuU, Irom the upper part oíthe braincase to the coronoid processes of the Jower jaw.Where the coronoid processes are fairly high, as in typicalfelínes, considerable force can be transmitted even whenthe jaws are wide open. A low coronoid results in a lessfavorable mechanical advantage and a weaker bite. Hav­ing killed its prey, the cal uses its bladelike eamassialteeth, P" aboye and MI below, in a scissor aetion to slicethe meat. Finally, manageable bones may be cracked bythe premolars, anterior to the carnassials.

The anatomy of a saber-toothed skull makes it abun­dantly clear that these carnivores kílled with their uppercanines only, driving them into the prey not by the actionof the temporalis muscles closing the jaws, but by adownward thrust of tbe head, powered by the neck mus­eles. These neck musc1es, the c1eidomastoids and ster­nomastoids, were attached to the great1y enlarged mas­toid process of the skull, whieh lay farther below the jointof the skull with the neek than in normal cats, Thechanged position improved the mechanical advantage ofthe system. adding power to the downward jab of thehead.

Other structural changes allowed the lower jaw to beswung down, out of the way of the upper canines, Thecoronoid processes were reduced in height and breadth,allowing the earnassiaI teeth to be brought cIoser to thehinges of the jaw and thus adding power to the scissoraction. In fact, considerable force eould still be exertedby the temporalis muscles, but now only when the jawswere a1most closed, As rneat-shearing mechanisms, thesabertooth camassials were even more efficient thanthose oí true cats. Ewer has deseribed the situation asfollows (I954a, p. 34):

In cats, the upper camassial consists oí three lobesIying along the length of the jaw and an inner lobeprojecling on lo the palale almosl al right angles lo themain axis ofthe tooth (see illustration). The two post­erior lobes (2, 3) are narrow and form the cuttingblade, working against the edge of the lower camassialwhile!he stouter anterior (1) and (4) lobes form part ofthe crushing mechanism. In sabre-tooths the wholelength of the upper carnassial becomes incorporated inthe cutting blade. This invoves the reduction andevenlualloss of!he inner lobe, so Ihal !he whole loolh

Bomotheriu/IlMegantlill"eOn

Machairodus. had upper canines in the form of wide, lat­erally compressed blades, not particularly elongated butwith edges serrated like a bread knife. Upper camassialstend to be very specialized, with the inner lobe reduced orabsent and an accessory cusp adding to the Iength of theblade. The other adaplive type, characterized by Ihegenus Megantereon, had very elongated but much lessñattened, smooth-edged canines; the carnassials wereless specialized, having small inner lobes and no acces­sory cusps.

It appears that the two kinds ofcaníne would have beenused in slíghtly different ways. The serrated canines ofthe Machairodus-Homotherium group would have beenmost effective in slicing, while the sharp, smooth-edgedsabers of Megantereon were stabbing daggers. Ewer(1973) has suggested that Megantereon specialízed inkilling large, heavy prey with a lough protective hide,against whieh the stabbing daggers would have been ef­fective. On the other hand, Ihe bladelike slicing eanines ofMachairodus or Homotherium would have been bettersuited for use on thinner-skinned prey: they would havebeen driven in and then, with a backward pulloused to ripopen the victim's throat.

In addition to these two maehairodont adaptive types, agroup of troe cats with sabertooth tendencies must beconsidered. These are plaeed in the genus Dinofelis andshow characters intermediate between those of normalfelines and maehairodonts; the upper canines were notexcessively enlarged. nor were the lowers much reduced.The carnassials and other teeth do not show the extremespeciaJizations found in the machairodonts. Cats of thegenus Dinofelis are referred to as "false sabertooths."

It is interesting to speculate on possible reasons for theextinction of the sabertooths. We do not know exaetlywhen the extinctions oceurred in southem Africa. Themost reeent machairodont known from the subcontineotcomes from Elandsfontein. It is a Megantereon and maybe associated with the "Final Acheulean" stage of theEarly Slone Age (Hendey 1974a). It may be as mueh ashalf a miUion years old. Elsewhere in the world, saber­tooths survjved much longer; for instanee, the well­known American genus Smilodon persisted until the

with bones of any size. This is hardly surprising as theenlarged upper canines would surely have made themanipulation of bones very diffícult, and might havebeen in danger oí getting broken themselves in theprocess. The large canines must also have been ratherin the way when getting started on a meal, and theenlargement of the incisor teeth helps to get over thisdifficulty, and would also facilitate the process ofgrooming.

Among the machairodonts. two main adaptive typeshad beeome dominanl by the Plioeene, and both are rep­resented in the fossils from the Sterkfontein valley caves.One, typified by the genera Homotherium and

lb)2

(a)~

4~

158 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Holocene. Ewer (1973) has pointed out that themachairodonts were smaller brained and less agite thanthe felines that became progressively more dominant afterthe end of the Pliocene. She suggests that they may nothave beco abJe to compete with the felínes in hunting themódem, swift ungulate fauna that was evolving al thattime. Another factor was almost certainly the rtse ofhuman intelligence and technology. Sabertooths, like alllarge carnivores, must surely have posed a threat lo earlymano One may be SUTe that human hunters would havetaken steps lo minimize this threat.

Dínofeíis, though a feline, was rather specialized andheavy bodied. It went the way oí machairodonts, prob­ably for the same reasons.

There are sorne indications that the sabertooths suf­fered from bone diseases that may have weakened themand perhaps contributed lo their demise. A complete rightradius (SE 680) attributed to cf. Megantereon sp. fromSterkfontein Member 5 shows pathology of the distal end,perhaps attributable to osteoarthritis. Among machairo­dont postcranial material from Langebaanweg, Hendey(l974a) found similar evidence of pathology.

A femur and tibia, apparently belonging to the sameindividual Megantereon sp. from Elandsfontein in theCape, showed osteítis of the distal fémur and proximaltibia. Four bones, possibly from a single animal referredto ?Machairodus sp .• show the fol1owing effect (Hendey1974a, p. 184):

Right tibia. Osteitis, mainly near the proximal end,Osteoarthritis at the proximal end, with severe ebur­nation of the lateral articular facet and lipping of thebone,

Left and right ca1canei. Osteitis on lateral surfacesanterior lo distal extremity.

Metapodial. Osteitis on dorsal surface. Osteo­arthritis (eburnation) of distal articular faeet.

In addition, the foUowing effects were noted on bonesfrom other ?Machairodus sp. individuals:

Proximal end of left ulna and distal end of lefthumeros. Severe osteítis 00 lateraJ and medial sides ofulna and a less extensive but similar condition on thehumeros, particularly the arch enclosing the en­tepicondylar foramen.

Distal end of right tibia. Severa osteitis on the shaft.ProximaJ end of left humeros. Extensive osteitis.Left radius laeking distal epiphysis (immature). Se-

vere osteitis of the shaft.

The term .,osteoarthritis' is used for any condition af­fecting the joints; "osteítis" refers to any other bone in­flammation (see Brolhwell 1963). Although the postcra­nial bones of sabertooths are not common in the southemAfrican deposits, the observed incidence of pathology onthem is remarkably high. Could it be lhat sabertoothswere susceptible to chronic disorders of the bones andjoints? It is tempting to speculate that lhey sulfered froma calcium deficiency as a result of their inability to crunchup any but the most delicate of bones. This may havebeen a price they paid for the remarkable dental spe­cializations that improved their hunting performance inother ways.

In an interwoven ecological situation, the extinction ofone component of the fauna cannot accur without re­percussions elsewhere in the system. In this regard. thefortunes of hyenas appear to have beeo linked with those

of the sabertooths. It is obvious that sabertooths were noteffective crackers of bones and that the remains of theirprey would have provided a regular source offood for anyscavengers able to utilize them. In fact, Ewer (1967) hasargued that the rise and decline of the hyaenids was re­gulated by their relationship with the machairodonts. Notonly did the sabertooths províde the food, but the natureof this food made it expedient for the hyenas to perfect aseries of bone-crushing adaptations.

Toward the end of the sabertooth's period of dorni­nance, hyenas showed a diversity of fono unkoowntoday. During the accumulation of Member 1, at Swart­krans, for instance, six recognizably different hyenataxa had their remains preserved as fossils. Today onlytwo different hyenas, the spotted, Crocuta crocula, andthe brown, Hyaena brunnea, occur in southernAfrica-they represent the rernnants of a formerly richerhyaenid fauna, a residue that preves viable in the current,postsabertooth era.

The Genus Dínofeiis ZdanskyThe genus Dinofelís ccntains a group of cats that show

a range of sabertooth tendencies but have not specializedas far as the true machairodont sabertooths have. Theyare thus retained in the subfamily Felinae and are referredto as false sabertooths.

The first southern African representarive was describedinitiaJly by Broom (l937b) as Meganlereon borlowi on thebasis of a damaged skuH and canine from SterkfonteinType Site. It was named after G. W. Barlow, manager ofthe Sterkfonteín Limeworks, who had also discovered theñrst australopithecine specirnen there.

Ewer (1955c) reexamined Broom's specimens and con­c1uded that they did not rightly belong in the genusMeg antereon, She found that the canine was not a typicalslender, curved blade as a Megantereon saber should be,but was an almost straight-edged dagger; Iikewise, theskull profile was not a typical machairodont one. Shetherefore transferred the specimens to Therailurus bar­lowi (Broom).

In the same paper Ewer described a new species.Tñeraiturus pivereaui, on tbe basis of a reasonably com­plete skull and two mandible fragments from KromdraaiA. She pointed out that the specimens showed featuresintermediate between those of the true cats (Felinae) andthe true sabertooths (Machairodontinae). One species hadpreviously been described in the genus Therailurus: T.diastemata Astre, 1929, and Ewer found close similaritiesberween this and the Kromdraai cat, a1though the saber­tooth specializations had been carried further in T.ptveteau! than in T. diastemata.

Meanwhile, sorne saber-toothed cat remains were dis­covered at lhe Makapansgat Limeworks by J. W. Kitch­ing and his two brothers. These consisted of a well­preserved mandible píece and articulated maxillary andmandible fragments from a second individual; they weredescribed by Toerien (1955) as belonging to a new speciesof the genus Machairodus: M. darti. The following yearEwer 0956<1) reexamined the Makapansgat material andshowed that M. darti should. in fact. be equated withTheraiJurus barlowi from Sterkfontein.

Hemmer (1965) reviewed the nomenclature and dis­tribution of the genus Vino!elis Zdansky. 1924, and con­c1uded that the known species of Therailurus rightly be­longed in the genus Dinofe/is. which would thus containfour species:

V. abeíi Zdansky, 1924 (formerly Machairodus hor­ribilis), from the Pontian stage of the PLiocene and theearly Pleistocene.

D. diastemata (Astre), 1929, from the Astían stage ofthe Upper Pliocene of France.

V. bartowí (Broorn), 1937, from the Sterkfontein TypeSite , Makapansgat Limeworks, and Bolt's Farm.

D. piveteaui (Ewer), 1955, from Kromdraai A.He concluded that, of these four, D. diastemata was

the mas! primitive in the specialization of its teeth whileD. píveteaui was the most advanced. He suggested thatD. diastemata-D, barlowi-D. piveteaui formed an evolu­tionary sequence, while D. abeli occupied a more isolatedposítion. This phylogenetic concept has been supportedby Hendey (19740).

Sorne exceptionally fine fossils of Dtnofelis barlowiwere recovered from Bolt's Farm in the course oí theUniversity of Calífornia's excavatlons there in 1947-48. Atearn under the direction oí Camp and Peabody found

tm 10

remains oí three Dlnofelis individuals in Pit 23 oí theBolt's Farm complex. These, together with the other fos­sils collected on the expedition, were taken back to theUniversity of California in Berkeley, wbere tbey werestudied hy H, B. S. Cooke in 1957-58. Cooke (n.d.) con­cluded that the three skulls, together witb their postcra­nial skeletons, belonged to an adu1t male, adult female,and young male of D. baríowi, The presence of threeweíl-preserved Dinofelis skeletons, fossílízed in closeproximity to one another, is remarkable and suggestssorne unusual circumstances at the time of death. Cookehas pointed out that the remaíns oí the three cats occurredin association with those oí at least eight baboons but thatbovid or equine remains, characteristic ofthe other Bolt'sFarm localities, were absent. He suggests that Pit 23 rep­resented sorne kind of natural trap into whicb the baboonshad fallen. They were followed by tbe three Dinofelíscats, perhaps members oí a single family, who then foundthemselves unable to escape from the trap. The presenceoí both baboon and cat coprolítes supports the suggestionthat the animals were alive for a while in the fossilizationsite before dying.

The large mate Dinofeíis skull from Pit 23 is nowhoused in the Transvaal Museum, through the courtesy ofthe University of California and H. B. S. Cooke,

Both cranial and postcraoial remains of Dínofelis di­astemata have been found at Langebaanweg by Hendey.They come from the Quartzose Sand Member and Pelletal

The Fossil Animals 159

Phosphorite Member in E Quarry (Hendey 19740). Thepostcranial remains are of particular interest because theyprovíde sorne insight into the build of the animal. Forinstance, the five caudal vertebrae are about the size of

adult cheetah vertebrae in diameter but are less than halfthe length, The tail of Dinofelís apparently was apprecia­bly shorter than that of an adult cheetah or leopard. Con­cerning body build, Hendey wrote (19740, p. 176):

The Langebaanweg Dinofeíís was evidently a heavilybuilt animal. with the fore- and hindfeet, and perhapsthe limbs in general, being more equally proportionedthan in either the leopard or cheetah, and possessing arelatively short tail. Its locomotion is likely to havebeen ambulatory. The indícations are thus that it hadparalleled the developments in the Machairodontinaein its postcranial skeleton as well as the skuIl.

More recently. remains oí D. barlowí and D. er.piveteauí have been reported from the Lower and Ileretmernbers respectively oí the Koobi Fora Formation atLake Turkana (M. G. Leakey 1976). There is also a reportof Dinofeíis sp. from the Usno and Shungura Formationsin the Omo valley {Howell and Petter 1976). The in­dications are that the genus had a wide range in bothspace and time.

The True Sabertooths: Subfamily MachairodontinaeAs mentioned earliet, the tfUe sabertooths from the

Sterkfontein valley caves fall into two natural groups: theMachairodus-Homotherium group with crenulated sabersand highly specialized carnassials, and tbe Megantereongroup with srnooth-edged sabers and less specializedcheek teeth. In his díscussíon of European sabertooths,Kurtén (1968) equates the Machairodus-Homotheriumgroup with the tribe Homotheriini and place s Meganle­reon in the tribe Smilodontini.

The genus Macbairodus is not represented among thefossils from the cave sites specifically studied here but hasbeen reponed from Bolt's Farm, wbere Broom (19390)described the crown of an upper canine and an isolatedupper carnassial (whicb may or may not be associated) asM. transvaalensís. Further remains from the Transvaalhave not come to light in the interirn, but sorne have beendescríbed from the Quartzose Sand Member in E Quarryat Langebaanweg by Hendey (19740), He points out thatif the carnassial from Bclt's Farro is, in fact, referrable toM. transvaaíensís, then the Langebaanweg specimensmust represent a different species, but in the absence ofmore complete material he refers them 10 Machairodussp.

160 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

The Genus HomotheriurnThe genus is particularly well represented as H. sain­

retti, the greater scimitar cat, among fossils from the Vil­lafranchian and early Middle Pleistocene of Europe. It is

also well known from North America, China, andelsewhere. Sorne of its anatomical characteristics havealready been discussed.

In her paper on machairodont remains from the Trans­vaal caves, Ewer (l955c) deseribed two lower camassialteeth (KA 66) from Kromdraai A. She poinled out that thespecimens were cemarkably similar to Epimachairoduscrenatidens (Fabrini) of Europe and less so toMachairodus ultimus Teilhard, from China. She desig­nated the Kromdraai specimens Epimachairodus? sp.More recently they have been reexamined by Hendey,who concluded that they can best be accornmodated inHomotherium sp. (pers. comm.).

No other traces of Homotherium have been found inthe Sterkfontein valley caves, but a fine skull with articu­Iated mandible was discovered by Kitching in the bandedtravertines close to the so-called ancient entrance of theMakapansgat Limeworks. The specimen was describedby Collings (1972) as Megantereon problematícus, bUIHendey (19740) poinled OUl that il fíts best in the genusHomotherium and refers to it as Homotherium sp. indet,(Hendey 1974a, p. 159). More recently Collings el al.(1976) have reassigned the Makapansgal specimen lo H.cf. nestianus.

A very spectacular though incornplete left upper caninehas been found in the Pelletal Phosphorite Member of EQuarry al Langebaanweg and referred by Hendey (l974a)to ef. Homotherium sp. Remains have also been reportedfrom various horizons throughout the Koobi Fora andKoobi Algi formalions near Lake Turkana (M. G. Leakey1976) and from the Usno and Shungura formations in theOmo valley (Howell and Petter 1976).

The Genus MegantereonThe genus is well known on the basis of the European

species M. megasuereon Croizet and Joubert, a VilIa­franchian guide fossil, referred to as the Dirk tooth. Other

species have been descrihed from the Siwalik deposits ofIndia, and from China and elsewhere.

In 1937 Broom described par! of a felid Ieft mandiblefrom Schurveberg as .. Feíis whitei, ., stressing that he wasuncertain about the most appropriate genus for thespecimen. Later (1939a, 19480) he suggested that the fos­sil should probably be referred to Megantereon, where ithas remained pending the discovery of more completespecimens.

A second specimen, consisting of the right mandibularramus with P4 and Mio was described by Broom (l948a)from the Sterkfontein Type Site as Megantereon gracíle,It is similar in many respects to M. whiteí but differs in theproportions of the camassial.

The first reasonably complete specírnen of a Megante­reon carne from Kromdraai A and was described by Ewer(1955c) as M. eurynodon-a new species, larger than M.whitei but of about the same dimensions as M. gracíle andM. megantereon, The uppercanines were long, recurved,much flattened blades with sharp edges, devoíd oí crenu­lations.

A single specimen from Swartkrans, consisting of a leftmandible fragment with Pa- t , was tentatively assigned toMegantereon by Ewer (I955c). This assignation was sub­sequently confirmed by Hendey (1974b).

Fossils from the lIerel Member of the Koobi ForaFonnation have recently been referred to M. eurynodon(M. G. Leakey 1976),and representalives of the genus arealso known from the Usno and Shungura fonnations inthe Omo valley (Howell and Petter 1976) and from Bed 1al Olduvai, site FLK NNI (M. G. Leakey 1976). Thegenus Megantereon has also beco reported fromElandsfontein (Hendey 1974a)-a record that apparentlyrepresents the most recent appearance of a sabertooth insouthem Africa,

Postcranial remains are not common at the local sites,but what indications we have support Kurtén's view(1968, p. 75) that the European Megantereon typícallyhad very powerful neck muscles, used in the stabbingmovement, "short but massive front legs and reJativelyfeeble hind quarters. The animal was obviously not a fastrunner but relied on its tusks and the great strength of itsfront paws, so the prey may be visualised as a relativelylarge, slow-moving animal. In the Villafranchian faunathe rhinoceros may be a possibility, or perhaps youngmastodonts oc elephants."

As mentioned earlier, two fossil forms, originally de­scribed as species of Megantereon, have come to rest inother genera. M. barlowi, described by Broom in 1937, isnow regarded as Dinofelis barlowi (Broom), while M.problematicus Collings, 1972, has become Homotheriumcf. nestianus.

"Nirnravidae Indet.'When Ewer (J 955e) discussed the fossil rnachairodonts

from the Transvaal caves she considered that a singlespecimen from Swartkrans may have represented a nim­ravid. The suhfamily Nimravinae contains saber-toothedfelids, lypically Oligocene in age; the discovery of a pos­sible representative at Swartkrans was therefore unex­pected.

The specimen, SK 336, consists of a left mandiblefragment and MI' In his reassessment of the Swartkranscarnivores, Hendey (t974b) concluded that the fossil ac­tually carne from a hunting hyena, Euryboas nitidula typeA (primítive), so that what was regarded as the most

problematieal identiñcation among the Swartkrans carni­vores has been resolved.

Family HyaenidaeHyaenids had their origin in the Miocene but did n~t

become abundant until the Pliocene, when theír mamradiation oecurred. They appear to have arisen from vi­verrid ancestors, perhaps not unlike the modem civets,and then to have undergone dental adaptations that madethe efficient crushing ofbones possible. Although hyenashave very effective carnassials for shcing meat, their an­terior premolars have become extremely robust to ser:eas bone crushers. As aJready mentioned, the hyena radia­tion was probably rnade possible by the ectivities ofsaber-toothed cats that were unable to eat the bones oftheir prey themselves. A valuable food source thus be­carne available to any seavenger able to use it. During thePliocene, the hyenas appear to have made full use of thisbony resource provided by the sabertooths.

Fossil hyenas from the Sterkfontein valley caves arecurrently c1assified in four genera: Hyaenictis, Euryboas,Hyaena, and Crocuta. The first two are extinet and ap­pear to have beeo secondarily predacious camivores,while the latter two have living representatives almostidentical to their fossil antecedents.

The Hunting Hyenas: Hyaenictts and EuryboasThe Genus Hyaenictís

This genus was created when Gaudry described an im­mature hyaenid mandible from Pikermi and named it H.graeca. It was characterized by a low, flat mandibleprofile, not deepened below M" and by the retention ofM2 • The genus is now known to have had a míd-Pliocenetime range, occurring in Eurasia and Afriea.

During the early phase of work at Swartkrans, ahyaenid cranium and mandible (SK 314) was found andconsidered by Broom and Robinson (1952) to belong In

the genus Hyaenictis. No speeific name was suggested.Ewer (l955a) reexamined the specimen and concludedthat, although Mil was índeed present, other charactersdiagnostic of the genus Hyaenictis were not. She there­fore described SK 314 as a new species of the genusLeecyaena Young and Liu, 1948, naming it L. forfexEwer. Later, Ewer (1967) expressed the opinion that L.forfex was not closely related to the only other knownspecíes of Leecyaena, L. lycyaenoides from the latePliocene of China. Hendey (l974b) tended to agree withher and because he did not think the two formed parts ofthe same lineage, he concluded that the Swartkrans fossilshould be named Hyaenictis forfex. There is, however,still sorne doubt about its generic identity,

From Langebaanweg, Hendey (l974a) describedhyaenid remains rather similar lo the Swartkrans speci­men as Hyaenictis preforfex, a form he regarded as an­cestral ío Hs forfex, In addition lo cranial rernains, excel­lent postcranial specímens were obtained al Langebaan­weg that showed that the limbs ofpreforfex were long andslender, indicating a cursorial habit. Hendey remarks(l974a):

This is of ínterest because of the suggestion (Thenius,1966)that the long-legged .. hunting hyaena"Euryboas, was descended from Hyaeníctis graeca,Plioeene Hyaenictis may thus have been a group oflong-limbed fonns whieh evolved in two direetions,one brnnch becoming inereasingly cursorial and ac-

The Fossil Animals 161

tively predacious (Euryboas], and the other parallelingHyaena and Crocuta (H. boseí, H. [orfex. H. pre­forfex).

Hyaenictís forfex (Ewer, ]955)Provenanee: Swartkrans Member 1

This animal had a head about the size of a brown hyena(H. brunnea) but had longer limbs and a more slender

body. Dentally it is characterized by anterior cheek ~eeth

resembling those of an advaneed Hvaena, a relativelylong MI' and the presence of a small M':!.

The Genus EuryboasThe genus is best known by the European hunting

hyena, E. lunensis Del Campana, found as fossils in vari­ous sites of Villafranchian age. Although deflnitely ahyena, Euryboas does not show dental specializations forbone crushing earried to an extreme degree. Its teeth arethose of a predator, while its limb bones are almost aslong and slender as those of a cheetah. One may assurnethat Euryboas was a fast eursorial hunrer, running downits prey after the fashion of a cheetah. Kurtén (1968) con­siders the prey ofthe European huntíng hyena to have beenlargely the Bourbon gazelle, Gazella borbonica Depéret,or the chamois antelope, Procamptoceras brivatenseSchaub.

The southem Afriean representatives of the genus areknown only from Sterkfontein and Swartkrans and wereinitiaJly placed in the genus Lycyaena. The ftrst specimenfrom Sterkfontein carne to light in a rather unusual man­ner. The Abbé Breuil was in Johannesburg in January1945and happened lo visir an art gallery owned by H. K.Silberberg. There he saw some fossils Silberberg hadpicked up at Sterkfontein a few years earlier, and amongthem was the muzzle of an interestíng-looking carnivore.The abbé asked for the specirnen and took ít to Broom inPretoria. who immediately reeognized that its affinitieslay with the Pliocene genus Lycyaena, known fromEurope and Asia. He described it (Broom 19480) asLycyaena silberbergi,

In her aceount of the fossil carnivores from the Trans­vaal caves, Ewer (19SSb), assigned two further specimensfrom Sterkfontein and one from Swartkrans to L. silber­bergi and amplifted Broom's description of the species.She also deseribed a new species ofLycyaena, named L.nítíduía on the basis of a series of specimens fromSwartkrans and a single one from Sterkfontein. When re­viewing the fossil hyaenids of Africa, Ewer (1967) ex­pressed the opinión that L. nítidula should be regarded asa subspecies oC L. silberbergi. More recently, Hendey(I974b) has proposed that the Lycyaena specimens fromSterkfontein and Swartkrans should be referred to the

162 Fossil AssembLages from the Sterkfontein Va11ey Caves: Analysis and Interpretation

genusEuryboas. This proposal appears to have been gen­erally accepted. Hendey's reassessment of the specimenshas nol yet appeared, bUI he is inclined (pers. comm.) loplace all the specimens from Sterkfontein inE. silberbergiand all those from Swartkrans in E. nitidula, this taxon,however, being split into a type A category, showingsorne primitive characteristics, and a type B group withmore advanced attributes.

Two species of hunting hyena are thus currently rec­ognized in the Sterkfontein valley assemblages:

Euryboas silberbergi (Broom)Provenance: Sterkfontein Member 2 or Member 3 (seebelow) and Member 4

The type specimen was pícked up underground, in whatis now called the Daylight Cave. It probably carne from

Member 2 or Member 3 of the Sterkfontein Formationwhile subsequent specirnens were found in Member 4.

Euryboas nitldula (Ewer)Provenance: Swartkrans Member 1. Both primitive andadvanced forms (types A and B) appear to occur.

E. nitidula may be separated frcm E. silberbergi by theoverlapping and oblique setting of the prernolars and bythe arrangement of accessory cusps. A tentativeidentification of Euryboas has recently been made byHowell and Petter (1976) from Member F of the ShunguraFormation, Omo Group, of southem Ethiopia.

The Genus HyaenaThe genus has two living representatives: H. hyaena

Linné, the striped hyena of tropical and north Africa,Asia, and formerly of Europe, and H. brunnea Thunberg,the brown hyena, now restricted to southem Afriea.

The first fossil Hvaena specimens from southern Africawere described by Toerien (1952) from the MakapansgatLimeworks and named H. makapani. Ewer (1967) did nolregard the differences between this species and H.hyaena as sufficient to warrant speclñc separation, andshe renamed the Makapansgat material H. hyaena maka­pani,. Hendey (l974a) has traced a phylogeny for lhe genus

Hyaena from Pliocene ancestors lctitherium orPalhyaena. He visualizes the late Pliocene H. abroniafrom Langebaanweg and the mid-Pliocene H. pyrenaicafrom Europe as the earliest recorded representatíves ofthe H. hyaena and H. brunnea lineages respectively. Hewrites Hendey 1974a, p. 146):

The H. hyaena lineage, which ineludes H. h. maka­pan; from the Transvaal, was a conservative onewhich underwent comparatively little change from thelate Pliocene onwards. The differentiation of thislineage might well have taken place in Africa. An an­cestor of H. brunnea probably entered Africa from thenorth during the Pliocene. It is first recorded in South

Africa during the Makapanian (Swartkrans) and it ap­parently replaced H. hyaella, which is last recordedearlier in the age (Makapansgat). H. bellas, anotherMakapanian species, is probably an off-shoot of the H.brunnea lineage.

The two species involved in the Sterkfontein Valleycaves are H. brunnea and H. bellax.

Hyaena brunnea Thunberg, 1820. Brown hyenaProvenance: Swartkrans members 1 and 2 and KromdraaiB (tentative identification)

The remains from Swartkrans Member I have been de­scribed by Ewer (19550) as a new subspecies H. brunnea

dispar, distinguished from the extant race by "the slightlymore primitive character of the upper premolars, the rel­atively short P, and the narrowness ofthe anterior palatalforamen" (Ewer 1955a, p. 824).

Hyaena bellax Ewer, 1954Provenance: Kromdraai A

This large extinct hyena was described on the basis of abeauliful cranium with articulated mandible, KA 55 (fig.208). In appearance the animal must have been similar tothe brown hyena, to which it was certainly related. Ewer(I954b) states that the species shows a combination ofprimitive and advanced characters, the upper molars rel­atively as large as those of H. hyaena but the camassialteeth more advanced than those of H. brunnea.

More recently Ewer (1967) has suggested lbalH. belíaxmay have evolved írom Hyaenictís (Lvcyaenaí forfex, butthis view does not appear to have the support of Hendey(l974a).

H. bellax is very probably also represented among fes­sils from Baard's Quarry, Langebaanweg (Hendey1974a).

The Genus CrocutaThe spotted hyena, Crocuta crocuta (Erxleben, 1777),

is a familiar member of the genus Kurtén (1956) hassuggested arose in India, south of the Himalayan range,then spread into Europe and Africa during the MiddlePleistocene: the possibilíty of an earlier rnigration lo Af­rica is considered (Kurtén 1957a,b), Later Ewer (1967)made the countersuggestion that Crocuta may have arisenin Africa, thence spreading to Europe and Asia.

That Crocuta was present in Africa before the MiddlePleistocene is well established. Howell and Petter (1976)record it from Member G of the Shungura Formation; M.G. Leakey (1976) records it from the Lower, Upper, andIleret members of the Koobi Fora Formation, and it isalso known from Laetolil (Díetrich 1942), Bed 1 al 01­duvai (Ewer 1965), and elsewhere.

ji

Howell and Petter have expressed the opimon thatCrocuta did, in fact, migrate into Africa as Kurténsuggests, but considerably earlier than he has visualized.

The following forms have been described from cavedeposits in the Sterkfontein valley at various times:

Crocuta ultra ultra Ewer, 1954Based on several cranial and postcranial specimens fromKromdraai A, distinguished inter alia by a short snout.,compressed incisor region, and the faet that the frontalsand premaxillae do not make contaet.

C. ultra latídens Ewer, 1954Based 00 three specimens from the Clyde Trading Com­pany site in the Sterkfontein valley and differing from C.u. ultra particularly in the structure of the cusps.

C. spelaea capensis Broom, 1939Based 00 a cranium from Kromdraaí that both Broom(1939a) and Ewer (1954b) considered lo resemble theEuropean c. spelaeo more closely than it does the livingC. crocuta. There has been a good deal of controversyover whether C. spelaea should be separated from C.crocuta. The matter remains unresolved. C. cf. spelaeahas also been recorded from the Sterkfontein Type Siteon the basis of a left maxilla (Ewer 1955a).

C. venus/ula Ewer, 1955Based 00 two rnandible pieces from Member 1 atSwartk.rans-a small Crocuta with a slender, shallowmandible.

C. crocuta angelJa Ewer, 1955Based on severa! specimens from Member 1 atSwartkrans and differing from the living spolted hyena incertain minor features of the dentition.

In her review of African fossil hyaenids, Ewer (1967)relegated the prevíous species C. ultra and C. venustu!ato subspecies of C. crocuta. Subsequently Hendey re­marked (1974b): "There has probably been more confu­sion over the taxonomy ofCrocuta than that of any othercarnivore in Africa and there is still a need for a com­prehensíve revíew of the Crocuta material now avail­able .' •·

He agreed with Ewer that C. venustuía should not beregarded as a sepárate species and placed all theSwartkrans material in the taxonC. crocuta. Llkewise, a11other Crocuta materia) from the cave deposits is referredto C. crocuta, pending a comprehensive revision of thegenus (Hendey, pers, comm.).

Family HyaenidaeSubfamily ProlelinaeGenus Proteles

Until recently, the genus contained a single livingspecies, P. cristotus Sparrman, 1783, the aardwolf.Although unquestíonably a hyaenid, the aardwolf is smalland has undergone very marked degeneration of lbe post-

The Fossil AnimaLs 163

canine teeth in response to its diet of tennites and othersmaH invertebrates. Premclars and molars have been re­duced lo ineffectual pegs. Ewer (1973) has suggested thatpredacious Phocene hyenas of the genus Lycyaena mayhave given rise ro Proteles. She argues that they probablyfailed in competition both with the specialized scavengerssuch as Crocuta and with the advanced felines, but thatthey found their salvation in turning to a diet of insects.

Prote1.es

An interesting fossil ancestor of the aardwolf has comelo light from Member 1 al Swartkrans and is probably alsorepresented at the Kromdraai australopithecine site(Hendey 1973, 1974b). Described as Pro/eles trans­vaaíensís Hendey (I974b), it preves lo be bigger and lessdegenerate dentally than is the living P. cristatus. Theremains suggest that sorne of the dental degenerationseen in the living aardwolf has taken place during lbePleístocene.

Sparse remains of the extant P. cristatus are found inSwartkrans Member 2. These do nol appear lo differ fromequivalent parts of the living formo Recent fossils havealso been described from the Black Earth Cave al Taungby Gingerich (1974).

Family Canidaelo striking contrast to the felids, the canids have re­

mained generalized feeders, with dentitions capable ofcoping with either meat or vegetable food. The molarshave not been reduced or lost as they have in cats, andthey provide a grinding battery in additioo to the slicingcarnassials. Canid muzzles are thus typicaJly longer thanfelid rnuzzles when animals of equivalent body weight arecompared.

The smallest members of the family in southem Africaare fcxes, which díffer from jackals and larger canids inthat they do not have frontal sinuses. According to Ewer(l956h). the function of the frontal sinuses is lo increasethe altachmenl area for the anterior fibres of the tem-

164 Fossil Assemblages from the Sterkfontein VaJley Caves: Analysis and Interpretation

C. brevirostris Ewer, 1956. Ewer's short-faced jackalThe species is known on the basis of a single cranium

from Member 4 at Sterkfontein. According to Ewer's

Genus Canis. JackalsTwo species are currently found in southem Africa, the

black-backed jackal, C. mesomelas Schreber, and theside-striped form, C. adustus Sundevall. Fossil remains,particularly of C. mesomelas, occur abundantJy at a varí­ety of sites.

Three species of jackal have been described or identiliedamong the fossils from the Sterkfontein valley caves. Inhís early descriptions, Broom (1937c, 1939a, 1948a) usedthe generic name Thos Oken, but Ewer (1956b) has ar­gued that Thos should be regarded as a subgenus ofCanis.The three species are as follows:

A fossil form intermediate in size between V. chamaand the European fox, V. vulpes Linné, was described byBroom 00 the basis of a mandible from Kromdraai A. Henamed it V. pulcher 8room. More recent1y two specimensfrom Member J at Swartkrans have been referred lo thespecies (Ewer 1956b; and Hendey 1974b). V. pulcher wasc1early a fox slightly more robust than the living Cape fox.It is known only from Kromdraai A and SwartkransMember 1.

C. terblanchei (Broom), Terblanche's jackalA jackal ahout the size of the living C. adustus but

differing from it in certain minar aspects of its eranial

anatomy and dentition. The type skull was described byBroom (1948a) as coming from Kromdraai A, but Ewer(1956b) pointed out that the matrix was unlike thal ofother KA specimens, and she suggested that the type had,in faet, come from elsewhere. Two specimens referred toC. terblanchei by Ewer (1956b) had been found at theCooper's site, but the provenance of the type rernainsuncertain. Apart from these, the only other specímenknown is a left maxillary piece from Member 5 atSterkfontein. It is tentatively identified as C. cf. ter­blanchei.

Genus Vulpes. FoxesThe Cape fox or "silver jackal," V. chama (A. Smith,

1833), is the only southern African representative of thegenus and occurs in the more arid areas of the region,Fossil remains of the species have been found at the sitesof Elandsfontein, Swartklip, and Sea Harvest in the west­ero Cape (Hendey 19740), while much earlier fossils re­ferrable to Vulpes are known from Langebaanweg (Hen­dey 1976). Broom (1937c) described an edentulous man­dible from Taung as V. pauisoní Broom. This may prob­ably be referred to V. chama.

and has an extensive range farther north, feeding on avariety of smal1 prey including insects, arachnids, androdents (Smithers 197Ia). Hendey (1974b, p. 31) haspointed out that "Otocyon dentitions are an exception tothe general rule arnong marnmals in that the more ad­vanced the species, the greater the number of molars andthe greater the complexíty of their occlusal surfaces."

Two specimens that may be referred to Otocyon havebeen found in Member 2 at Swartkrans, an incompletepalate and an isolated lower molar. Hendey concludedthat these were dentally more primitive than O. megalotisand referred them to O. recki, a species known from BedI at Olduvai. He pointed out that the teeth of theSwartkrans specírnens were not quite as primitive incharacter as those from Olduvai.

Genus Otocyon Müller, 1836. Bat-eared foxesThe bat-eared fax, O. megulotis Desmarest, is a famil­

iar camivore of the more arid regions of southem África

poralis muscles without unduly increasing skull weight.Foxes, with their relatively weak jaws, have not foundit necessary to increase the temporalís attachmentareas. Canids are well represented among fossils fromthe Sterkfontein valley caves, specimens having beenallocated to four genera: Otocyon, the bat-eared foxes:vulpes. the foxes: Cante. the jackals; and Lycaon, thehunting dogs.

(1956h) description, it was a small, short-faced jackaí withwell-developed upper canines, extremely shortened pre­molar series, and rather rectangular upper molars. Thedentition suggests that C. brevírostris was a less spe­cialized feeder than are the living jackals.

C. mesomelas Schreber. Black-backed jackalProvenance: Sterkfontein Member 4, Swartkrans mem­bers I and 2 and channel fill, Kromdraai A, and probablyKromdraai B

Remains of this familiar jackal have been found al allthe maín fossil-bearing Iocalíties in the Sterkfonteinvalley.

Broom (1937h) described tbe mandible of an immalurejackal as representing a new species, Thos antiquus

Broom. He stated that it carne from "the same cave asthat in which the skull ofAustralopithecus tronsvaalensiswas found"; however, the catalog entry records thespecimen as coming from Minnaar's Cave, and there isconsequently sorne uncertainty. Ewer (I956h) preparedand examined the specimen further and concluded that itshould be placed in the taxon C. mesomelas. At the sametime she examined all the availabJe specimens frorn theSterkfontein vaIJey caves and decided that, as a group,the fossils were sufficiently differenl from living black­backed jackals to warrant creating a new subspecies, C.mesomelas pappos Ewer. The points distinguishing thefossil subspecíes were summarized as follows (Ewer1956h. p. 113): "díffering from the living C. mesomelas inhaving the lower premolars longer in comparison with thelength of the carnassiai; nasals frequently not extendingbehind the posterior end of the maxilla."

More recently Hendey (1974h) expressed lbe opinionthat C. mesomelas pappos constitutes a heterogeneousand unsatisfactory taxon, and the subspecies name istherefore not used in this work.

Genus Lycaon, Hunting dogsThe hunting or wild dog, L. pictus Temminck, is a

famillar Mrican predator, hunting in packs and relying onits speed and endurance to ron down its medium-sizedungulate prey. A European species was formerIy desig­nated L. lycaonoides Kretzoi but has now been trans­ferred lo tbe genus Xenocyon by Schütt (1974). The formappeared brielly during the early Middle Pleislocene ofEurope but appears lo have failed in competition with theDhole, Cuon sp. (Kurtén 1968). 111e only other species inthe genus is L. atrox, although the Olduvai form Cantsafricanus should possibly also be included (Q. B. Hen­dey, pers. comm.).

Lycaon atrox (Broom). Kromdraai hUDtiog dogProvenance: Swartkran. Member 2 (possibly) and Krom­draai A

The Fossil Animals 165

In 1939 Broom mentioned the lower camassial of alarge canid from Kromdraai A and subsequently de­scribed this tooth (J948a; Broom and Schepers 1946) to­gether with an ísolated upper molar. referring them toCanis atrox Broom. In the course of her reexarnination ofthe carnivore remains from Kromdraai, Ewer (1956b)concluded that C. atrox had affinities with C. afrtcanusfrom Bed 11 at Olduvai, though she concluded that theywere specifical1y distinct.

More recently, Hendey (1974a) has shown that Lycaonremetns from Elandsfontein in the westem Cape exhibitcharacters intennediate between those of C. atrax and L.pie tus . He concluded that hunting dogs of the genusLyeaon acose from a wolflike Canis ancestor during latePltocene or earJy Plelstocene times and that c. atrox wasan early member of the lineage that led lo the living L.pictus, On these grounds he placed C. atrox in the genusLyeaon.

Regrettably, we know nothing of the skeleton of theKromdraaí hunting dog, L. atrox. We may speculate that,

as a primitive member of the Lyeaon lineage, Its cursoriaJadaplations may have been less well developed than theyare in the living huntiog dogs. Its limbs may have beensomewhat shorter and less slender.

Apart from the two Kromdraai teethjust discussed, theonly other hunting dog fossil known from the Sterkfonteinvalley caves is an Isolated upper canine from Member 2 atSwartkrans. This has been referred to Lycaon sp. indet.(Hendey 1974h).

Family MuslelidaeThree subfamilies are represented in the living southem

African fauna: Mustelinae, which incfudes polecats (lc­tonyx Kaup, 1835) and striped weasels (PoecilogaleThomas 1883); Mellivorinae, the honey badgers (Melli­vora Starr, 1780); and Lutrínae, which includes otters ofthe genera Aonyx Lesson, 1827, and Lutra Brisson, 1762.These carnivores present a wide spectrurn of bodily fonn,size, and adaptation, the polecats and weasels beingsmall, active, and highly efficient predators, while thehooey badger and others are mueh larger animals in whichthe crushing function of the dentition is more developedto suit their particular diets.

Mustelid remains are not common among fossils fromthe Sterkfontein valley caves. In his repon 00 fossilmammals collected by the Uoiversity of California ex­pedilion in 1947-48, H. B. S. Cooke (n.d.) menlioned lbepresenee ofthe otter, Aonyx ef. eapensis in Pi13 of Bolt'sFarm and described a new species of polecal from PillOof the same site. He suggested that it represented an off­shoot from lhe slock lhal gave rise lo both IClOnyx and

166 Fossil Assernblages from the Sterkfontein valley Caves: Analysis and Interpretation

Poecilogale, but that it was less specialized than any ofthe living species of these genera.

The only mustelid from the sites that particularly con­eem us here is the following:

Mellivora ef. sívalensis Falconer and Cautley, 1868.Honey badgerProvenance: Swartkrans Member 2

Two isolated teeth and a mandíble piece have been de­scribed by Hendey (l974h). He found that these teeth

were less primitive than those of M. punjabiensisLydekker, 1884, provisionally ídentified from E Quarry alLangebaanweg (Hendey 1974a) but not as advaneed asteeth of the living honey badger M. capensis (Schreber,1776). He therefore referred the Swartkrans specimen loM. cf. sivalensis, a species frorn the Pinjor stage of theSiwaliks in India that shows the same set of characters.

The Swartkrans honey badger was probably very simi­lar to the living one in appearance and habits.

Family ViverridaeCamivores with a very long history, representing the

basic stock from which both the hyenas and the catsarose, the viverrids are an old-world family that retainsmany primitive features, They tend to be mustelinelike intheir adaptation. Two subfamilies occur in southem Af­rica: the Viverrinae, comprising the civets and genets,and the Herpestinae, lhe mongooses. Skulls of lhe lattercan be separated from thcse of the formerby the presenceof well-developed postorbital processes.

Subfamily ViverrinaeGenus víverra Linnaeus, 1758cf. Viverra sp. CivetProvenance: Kromdraai B

The distal end of a right humeros from Kromdraai Bproves lo be morphologícally indislinguishable from theequivalent part of the living civeto V. civetla Schreber orfrom that of the Pliocene form, V./eakey; Petter, found alLangebaanweg and in East Afriea (Hendey 1973). Untílmore complete specímens come to light, little more can besaid than that remains of a civet are represented in the

Kromdraai B assemblage. They havo not been recordedin the other Sterkfontcin valley caves.

Subfamily Hespestinae, MongoosesFossil mongooses from the sues of Sterkfontein,

Swartkrans, and Kromdraai are currently alIocated tothree genera: Herpestes, Cyníctís, and Crossarchus.

Genus Herpestes Illiger, 1811Herpestes mesotes Ewer, 1956Provenance: Kromdraai A

Based on a single skuIl and mandible, this speciesclosely resembles the living Iarge gray mongoose, H.

,-- --~-~

'*- -.=--

ichneumon Grill, 1858. which has an extremely wide Afri­can distribution. The fossil differs from the living form inhaving more robust cheek teeth, a less reduced M2

• and aless prolonged palate behind the last molars (EwerJ956c). In her description, Ewer pointed out that H. me­sotes had close affinities with both H. ichneumon and thewater mongooseAtilax poludinosus (G. Cuvier). She con­sidered that the fossil form had characters that might beexpected in an ancestor of either of the living animals,though she was ínclined to place it low down on thelineage leading lo Atílax, lf this is so, and Hendey (l974a)is inclined to believe it ís, then the speciesmesotes shouldbe placed in the genus Atilax rather than Herpestes.Further material is required to settle the question. Fossttsof H. sp. aff, ichneumon Linné have been recorded fromBed 1 al Olduvai [G. Petter 1973), while Herpestes sp. isknown from Member C of tite Shungura Forrnation, Omovalley (Howell and Petter 1976). More recent fossils ofH.íchneumon come from the Elandsfontein, Swartklip, andSea Harvest sites in the westem Cape (Hendey 1974a).

Herpestes cf. sanguineus Rüppel, 1836. SLender mon­gooseProvenance: Swartkrans Member 2 and possibly Krcm­draai B

Part of a mandible has been deseribed by Hendey(l974h). He points out that it could be referred either lOH. cf. sanguineus, the common slender mongoose, or toH. pulverulentus Wagner, the Cape gray moogoose.However, since these two forms are apparently very

¡

¡

The Fossil Animals 167

Tribe BoviniThree genera are inciuded in Ansell's (1971) African

Iist: Bos Linnaeus, the wild OX; Buba/us H. Smith, thewater buffalo; Syncerus Hodgson, the African buffalo.

Order Artiodactyla

Artiodactyls whose fossils have been preserved in theSterkfontein valJey caves are classiñed in three families:Bovidae, the antelopes; Suidae, the pigs; and Giraffidae ,the giratTes.

•• Makapania'

BuffaloEland, kudu, bushbuck, nyala

Springbok, gazelleSteenbok, klipspringer

Wildebeest, hartebeest,blesbok

Sable, TOanWaterbuck, reedbuckRhebuck

Tribe HippotraginiTribe RedunciniTribe Peleini

Subfami/y AntilopinaeTribe AntilopíniTribe Neotragini

Subfamily CaprinaeTribe Ovibovini

Subfamily BovinaeAnsell (1971) inciudes three tribes in this subfamily:

Bovini, Tragelaphini, and Boselaphini, the last withoutliving African representatives but present in India in theform ofthe nilgai, Boselaphus tragocamelus (Pallas), andthe four-homed antelope Tetracerus quadricornis (DeBlainville). Recently Gentry (1974) has described aboselaphine from Pliocene deposits at Langebaanweg asMesembriportax acrae Gentry, the first representative ofthe tribe to have been found in southern Africa.

Family BovidaeFossils from the Sterkfontein valley caves may be

ciassified, following Ansell (1971), in four subfamiliesthat, in tumo are divided into nine tribes as follows:

Subfamily BovtnaeTribe BoviniTribe Tragelaphini

Subfamily HippotraginaeTribe Alcelaphini

Syncerus sp. BuffaloProvenance: Swartkrans members 1 and 2 and Krom­draai A

The taxon is best represented at Swartkrans, where 13cranial pieces come from at least 3 individuals; a singlejuvenile is recorded from Sterkfontein, and 4 teeth havebeen found at Kromdraai.

Vrba (19700) has pointed out that the dentitions fromthe Sterkfontein valley sites have clear affinities with thatof the extant African buffalo S. caffer (Sparrman) but

signed a fossil palate from Kromdraai A to the sametaxon, albeit somewhat tentatively. Remains have notbeen recorded from other sites, and the relationship ofC.transvaatensís is somewhat obscure. The genus is repre­sented by three species of living long-nosed rnongooseseJsewhere in Africa: C. ansorgei Thomas from Angola; C.obscurus F. Cuvier and C. alexandri Thomas andwroughton from equatorial forest regions, which twospecies may be conspeciñc (Coetzee 1971). These are di­urnal mongooses, hunting in small parties. They aremainly terrestrial, though they may take refuge in trees.

c. penici/lata in the wide spacing of the cheek teeth andthe relatively small size of M,. Although she felt that thefossil was subspecifically distinct, she declined to namethe new subspecies in the absence of more completespecimens. She has, however, expressed the opinión thatthe Swartkrans subspecies could be the same as c. pen­iclllaía brachyodon Ewer that she described from theMakapansgat Limeworks 00 the basis of a reasonablycomplete cranium.

Crossarchus transvaalensís Broom, 1937Provenance: Kromdraai A

Described originally 00 the basis of a right mandibularramus from Bolt's Farm. Broom (1939a) subsequently as-

Cynicris peniciliata (G. Cuvier). Yellow mongooseProvenance: Swartkrans Member 2

The familiar and widely distríbuted yellow rnongoose ischaracterized by a bushy tail with a white tipo

Ewer (1956c) described a mandible from Swartkrans,pointíng out that it differed from mandibles of the living

closely related, the Swartkrans fossil could be ancestralto either.

Two specimens from Kromdraai B have also been re­ferred to Herpestes sp. by Hendey (1973), who expressedthe opinión that they might come either from "a largevariety of a species such as H. sanguineus or a smallvariety of a species such as H. ichneumon:"

168 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

differ in the following respects: the fossil teeth are con­sistently larger; they have simpler enamel patterns 00

their occlusal surfaees; and the paraconid and metaconidof the P~ remain unfused.

The possibility that tbe fossils should be referred to thegenus of the extinct long-homed buffalo, Pelorovis Reck,has been thorcughly discussed by Vrba. They could, forinstanee, have affinities to the widespread P. antíquus(Duvernoy), which is known from many southem Africanlocalities and which appears to have become extinctabout 10,000 years ago (Klein 1974a); other possible rela­tives are P. oldowayensis Reck from Olduvai and a thirdunnamed species knawn from Elandsfontein (Hendey1974a).

For severa! reesons Vrba discounts a close relationshipbetween the Swartkrans fossils andPelorovis. lnstead sbepoints to similarities between these specirnens and fossilsfrom Bed II al Olduvai that appear lo be on the Synceruslineage. The Olduvai species remains lo be named.

It rnay be eoncluded that the bovtne remains fromSterkfontein, Swartkrans, and Kromdraai appear to havecome from buffalo somewhat more robust than, but prob­ably close to, the ancestry of the living African bulfalo S.caffer (Sparrman), which shows a great deal ofvariabilityin its wide African range and which appears to compriseat least two subspecies, S. caffer caffer in southem Africaand S. c. nanus in forests of west and central Africa.

Tribe TragelaphiniOpinions on the classitication of this group of antelope

vary considerably, but Ansell (1971) places a1lliving rep­resentatives in two genera: Tragelaphus De Blainville(bongo, mountain nyala, sitatunga, nyala, bushbuck, andkudu) and Taurotragus Wagner, the eland. AII aremedium to large antelope with spiral borne that may ormay not be present in the females; they are browserspreferring a bushy or wooded environment.

Tragelaphus ef. sctiptus Pallas. BushbuekProvenance: Swartkrans Member 2 and Kromdraai A

Sorne rather fragmentary dentitions have been tenta­tively referred lo this taxon by Yrba (1976a). In lheirstudy of fossil Bovidae from Makapansgal, Wells andCooke (1956) described whal they considered a newspecies of duiker, Cephalophus príceí, perhaps related to

~Á<

the red duiker, C. natalensis A. Srnith. Gentry (1976) hasreclassitied the Makapansgat dentitions (though not thehom-eore) as Tragelaphus pricei (Wells and Cooke),suggesting that tbey have strong affinities with thebushbuck, T. scriptus, Gentry also identifies T. cf. ortcetfrom Member C of the Shungura Formaticn in the Omovalley, pointíng oul that Ibis animal differed from the liv­ing bushbuck in having homs more vertically inserted andless anteroposteriorly compressed.

Apart from the evidence oí these early remains andsorne much later ones from Klasies River Mouth caves(Klein 19760), !be fossil hístory of the busbbuck isobscure.

TrageJaphus cf. strepsiceros (Pallas). Greater kuduProvenance: Swartkrans members 1and 2 and KromdraaiA

Yrba (19760) has shown that the remains-a singlehorn-core piece and various adult and [uvenile

dentitions-appear consistently larger than equivalentparts of the living greater kudu. A similarly large fonn hasbeen described from Makapansgal by Wells and Cooke(1956).

Anolber large fonn of kudu was described by L. S. B.Leakey (1965) as Strepsiceros grandis from upper Bed IIat Okíuvai, Yrba (19760) states that the Slerkfontein val­ley remains are probably similar lo lbese Olduvai fossils,which Gentry (19700) reclassities as Tragelaphus strep­siceros grandis (Leakey).

Fossils that appear to be at least subspecifically differ­enl from the living kudu are known from Melkbos andElandsfonlein (Hendey 1968), while others, modem inappearance, are known from many sites such as theKlasies River Mouth caves (Klein 19760) and Cave ofHearths (H. B. S. Cooke 1962).

The living greater kudu has an exlremely wide Africanrange and, according to Ansell (1971), four subspecies arecurrently recognized.

Tragelaphus sp. a1f. angasi Gray. NyalaProvenance: Sterkfontein Member 4 and SwartkransMember 2

A few jnvenile dentitions are tentalively referred lo thistaxon by Yrba (19760), a11hough they are somewballarger lhan living counterparts. Fossils lhal a1so appear locome from rather large nyalas have been described fromMakapansgal by Wells and Cooke (1956), who comparedlhem lo the mounlain nyala, T. buxtoni Lydekker, aspecies endemic lo lhe Ethiopian highlands.

T. angasi has a limited distribution in southeastem M­rica and is typieally restricted to thick rivenne cover. It is

unusual in that the male is nol only larger than the femalebut strikingly different in appearance,

Taurotragus cf. oryx (Pallas). ElandProvenance: Sterkfontein Member 5, SwartkransMember 2, and Kromdraai A

The onIy remains consisl of isolated teeth that Vrba(19760) considered rnorphologícally similar lo those of

living eland. Similar fossils have been reponed fromMakapansgal Limeworks (Wel\s and Cooke 1956),Melkbos (Hendey 1968), Elandsfontein, Swartldip, andSaldanha (Hendey 19740), and many other southern Afri­can shes.

A more primitive species of eland, T. arkellí, was de­scribed by L. S. B. Leakey (1965) from Bed IV at Olduvaíbut is not known from South Africa.

Ansell (1971) recognizes three subspecies of the livingeland, which has a wide, but frequently discontinuous,distribution, rnaínly south of the equator,

Subfamily HippotragínaeTríbe Alcelaphíni

The tribe contains four living genera confined lo Africa:Connochae/es Lichtenstein (wildebeest), Alcelaphus DeBlainville (hartebeest), Beatragus HeDer (Hunter's har­tebeest), and Damaliscus Sc1aLer and Thomas (blesbok).Gentry (1978) suggests thatAepyceros, the impala, shouldalso be included in the Alcelaphíni. Three additional andextincl genera are represented among the Sterkfonteinvalley fossils: Parmularíus Hopwood, Rabaticeras En­nouchí, and Megaíotragus van Hoepen, All members ofthe tríbe are medium lo large antelopes, and both sexescarry horns. The frontals and hom pedicles are charac­terized by well-developed, smooth-walled sinuses, while

The Fossil Animals 169

the braincase is steeply angled to the axis of the face,which tends to be long and narrow. The molars are hyp­sodont and premolar rows are short.

Genus Connochaetes LichtensteinTwo living species are currently placed in this genus:

the brindled gnu or blue wildebeest C. taurinus (Bur-

chell), with a fairly wide distribution in southem and east­em Africa, and the white-tailed gnu or black wildebeest,C. gnou (Zimmerman), which formerly ranged over theopen central platean of South Africa but is now muchmore restricted. Fossil remains of Connochaetes areknown from a number of sites in the Orange Free State,from Elandsfontein and elsewhere. The genus is alsoabundantly represented in assemblages from east Africansites, and it or a related genus ís present in north Africansites.

Cranial remains, mainly dentitions, of wildebeestlíkealcelaphines are abundant in the coJlections from theSterkfontein valley caves. In her study of fossil Bovidaefrom Sterkfontein and Kromdraai, Vrba (19760) record.the taxon cf. Connochaetes sp. from the site units ofSterkfontein members 4 and 5, Swartkrans members 1and 2, and Krorndraai A. Connochaetes sp. is recordedfrom Kromdraaí B and from Swartkrans channel fill. C.cf. taurinus ls recorded from durnp D16 at Sterkfontein.Vrba expresses the opinión that all these remains prob­ably carne from anirnals 00, or close too the C. taurinuslineage, though sorneof the teeth from Kromdraai A showunusual morphology suggestive of a specialization awayfrom the lineage.

Dentitíons from Member 1 at Swartkrans showed cleardifferences from those oC the living C. taurinas. To con­vert the Swartkrans dental morphology inlo that oC theblll" wildebeesl as known today, Vrba (19760. p. 10)

speculated lhat the simple tooth shape of the fossils wouIdbecóme more complicated and specialized; the anteriorpart of the toothrow would assume greater functionalsigniñcance, despite the gradual loss of p.; a wideningof P3' P4. and M., together with further molarization oíthe premolars, would occur, and increasing hypsodontyof the premolars would result in deepening of the man­dibular ramus below them,

It is clear that encesrors of the blue wildebeest, knownin the area today, were a well represented component ofthe fauna from the Sterkfontein valley sites,

Genus Damallscus Sclater and ThomasAccording lo Ansell (1971), the genus contains three

living species: D. dorcas (Pallas, 1766), the blesbok andbontebok; D. lunütus (Burchell. 1823), the tsessebe, topi,and so forth; and D. hunteri (P. L. Se/ater, 1889), the

170 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Hunters hartebeest. The last species, D. hunteri, has anextremely restricted range between the Tana River,Kenya, and the Juba River, Somalta: it is considered bysorne mammalogists to tit best in a different genus, Be­atragus Hellet, and in this account of the fossil animalsBeatragus is maintained as a genus separate fromDamalíscus.

Small alcelaphine antelope are well represented in theSterkfontein valley fossil assemblages. vrba (19760, p.24) has tabulated differences in the dentitions that allowedher to divide the collections between three taxa:Damaliscus sp. 1. or Parmularius sp.; Damalíscus sp. 2(?D. niro); and D. dorcas, Subsequently her furtherstudies have indicated (Vrba 1978)that all the Damallscussp. 1 or Parmularius sp. material belongs to the genusParmularius. A new species, Parmularius parvus, wasdescribed on the basis of the Kromdraai A smallalcelaphine specimens. In addítion, a specimen fromSwartkrans Membee 1, SK 14104, Is now regarded asParmularius angustícomls (Schwarz) (Vrba, pers.comm.).

Damaliscus sp. 1 or Parmularius sp.Provenanee: This is the dominant antelope species atKromdraai A and is also represented at Swartkrans

Representatives oí the genus Parmularius are knownfrom several levels in the Olduvai sequenee (L. S. B.

Leakey (965), from the Shungura Formatíon of the OmoGroup (Gentry (976), and from the Koobi Fora Formatíonof the Easl Rudolf succession (Harris (976). RecentlyVrba (1977) has described a new species of Parmu­larius-the first from South Africa-from members 2, 3,and possibly 4 of the Makapansgal Formation. Pre­viously, Vrba (19760) had expressed the opinion that P.parvas from Kromdraaí A may be closely related lo P.rugosas L. S. B. Leakey, as known from Beds lll-IV alOlduvai.

Parmularius sp.Malerial from members 4 and 5 al Sterkfontein may

belong nol lo P. parvus, but lo a different species.

Damaliscus sp. 2 (?niro)Provenance: Swartkrans Membee 2, abundant (57 speci­mens from a minimum oí 9 individuals); SterkfonteinMember 4, a single specimen that may have come fromelsewhere; Sterkfontein Member S (7 specimens from aminimum oí 4 individuals)

Vrba (19760) created this taxon lo accommodate smallalcelaphine dentitions similar to those of D. dorcas butsomewhat larger and witb less complex enamel pattems

on the occlusal surfaces of the teeth. A single hom-corepiece from Member 2 at Swartkrans is strongly rem­iniscent of D. niro (Hopwood),

The species D. niro was first deseribed by Hopwood(1936)on the basis of a right born-core fcund by L. S. B.Leakev in Bed IV at Olduvai and sent to the BritishMuseum in the early 1930s. Hopwood described it asHippotragus niro, pointing out the similarity of the horn­core to that of the sable antelope, H. niger. He derivedthe trivial Dameniro from a Masai word meaning brown,the color of the holotype. .

L. S. B. Leakey (1965) pointed out that specimens ofH. niro had similarities to the genus Damaííscus, Gentry(1965) reexamined the material, including a new specimenfrom the Peninj Beds, west ofLake Natron and concludedthat Damaliscus was in fact the most appropriate genus.The taxon thus became D. níro (Hopwood). It representsan alcelaphine antelope of médium size with homs rern­iniseent of those of a sable but differing by prominentwidely spaced transverse ridges on their anterior sur­faces. The horn-cores are typically shorter, however, andthe frontals are less sharply raised between the hornbases.

D. niro is now known from severaJ sites in the OrangeFree State, including Cornelia and Florisbad (H. B. S.Cooke 1974), Driefontein neae Cradock, WonderwerkCave, and Cave of Hearths (H. B. S. Cooke 1962; Wells1970).

Damaliscus cf. dorcas (Pallas, (766). Blesbok or bon­lebokProvenance: Sterkfontein members 5 and 6, SwartkransMembee 2 and channel ñll

Remains indistinguishable from equivalent parts ofmodern blesbok are fairly abundanl in the later Sterkfon­lein valley site uníts, Vrba stales (19760, p. 24):

,

D. dorcas dentitions can be distinguished fromall other extant or extinct alcelaphine species bythe unique comblnaríon of

1. being among the smalIest a1celaphine denti­tions known,

2. having long premolar rows with P2and P''more often present than absent and P3and p3 welldeveloped, and

3. having a complicated molar occiusal surfaceenamel configuration.Extinct a1celaphine species almost always seem tohave a distinctly less complicated dental enamel con­figuration.

Extant representatíves of the species are divided be­tween two subspecies, according lo Ansel! (1971): D. d.dorcas (Pallas, 1766), the bontebok, restricted lo thesouthwestern Cape Province, D. d. phillipsi Harper, 1939,the blesbok, with a much wider distribution in open habi­tats of the interior.

Rabaticerus porrocornutus (Vrba, 1971)Provenence: Swartkrans Member 1

The type specimen was described by Vrba (1971) asDamaliscus porrocornutus on the basis of a frontlet with

hcm-cores, SK 3211a-b, from Swartkrans Member 1.The speciñc name is based 00 the Latin adverb porromeaning "forward" and adjective cornutus, "homed"; itrefers to the tendency of the hora-corea to sweep fartherforward, relative lo the face, than lhey do in related anotelopes (Vrba 1971, p. 60). Ayear afier the initial descrip­tion appeared, the parameters of the type specimen weresubjected to multivariate analysis (Laubscher, Steffens,and Vrba 1972), and the conclusion was reached that lhetype was unlikely lo belong correctly in either of lhe gen­era Damaliscus or Alcelaphus. Subsequently, Vrba(19760) rec1assified it as Rabaticeras porrocornutus.

The genus Rabaticeras was first created by Ennouchi(1953) when he described a fossil skul! from Rabat asR. arambourgi. he placed it in tbe Ovicaprinae, but itsoon became apparent that this was in faet an alcelaphine.

The Swartkrans type specimen differed from R. aram­bourgi in a number of respects, notably that the horn­eares are oriented less forward with respect lo the skull asa whole, and they also show more pronounced mediolat­eral compression (Vrba 19760).

Rabaticeras remains are also known from Olduvai RedII and the III/1V junction and from Elandsfontein. 1t hasbecome cIear that representatives of the genus were pro­bably related lo the ancestors of lhe living hartebeestspecies Alcelaphus lichtensteíní (Peters) and A.buselaphus (Pallas).

The Fossil Animals 171

ce. Beatragus sp.Provenance: Member 2 at Swartkrans

As mentioned earlier, the genus has a single living rep­resentative, Hunter's hartebeest , classified by Ansell

(1971) as Damaliscus (Beatragus) hunterl (P. L. Sclater)and restricted to a small area of Kenya and Somalia. Itappears, however, that aleelaphines of the genus Be­atragus were formerly more widespread. L. S. B. Leakey(1965) described B. antíquus from Olduvai, where it hadbeen recorded from beds 1 and II as well as from high upin the Shungura Formation of the Omo valley (Gentry1976). B. antiquus differed from living species in severa!respects, partieularly its larger síze and more upright hominsertions.

Remains of Beatragus occur at Elandsfontein, andVrba (l976a) quotes Gentry's opinion that B. antiquusmay have been replaeed geographically by a southemspecies whose remains are found at Swartkrans andElandsfontein.

a. Megalotragus sp. Giant hartebeestProvenanee: Sterkfontein Member 4, Swartkrans mern­bers 1 and 2, and Kromdraai A

The genus Megalotragus was ereated by van Hoepen(1932b) when he described a paír of horn-eores from Cor­nelia as M. eucamutus, The earliest description of a gianthartebeest from southem Afriea was that of Broom(1909a) when he named the taxon Bubalis priscus on thebasis of a cranial fragment, with part of the left horn-core,found on lhe bank of the Modder River between Kimber­ley and Bloemfontein.

Subsequently, a variely of giant hartebeest specímenswere described from the Orange Free State under the

172 Fossil Assemblages from the Sterkfonteín Valley Caves: Analysis and Interpretation

genera Peíorocerus van Hoepen, Megalotragus vanHoepen, and Lunatoceras Hoffman. As H. B. S. Cooke(1974) pointed out, the genus Megalotragus has priorityover both Pelorocerus and Lunatoceras; furthermore,Cooke consldered that the type specimen of Broom' sBubalis príscus was inadequate and hence a nomenvanum.

Genlry (1976,1978), on the other hand, eonsiderspris­cus the valid type species of Megolotragus, According toevidence accumulated by Klein (1974a) from sites in thesouthem Cape, M. príscus made its last appearance therein terminal Pleistocene times, l2,OQO-lO,OOO years O.P. Itappears to have been one of a series of mammals thatbecame extinct about that time.

M. priscus was appreciably larger than the blue wil­debeest but less massive than the eland. Farther north asecond species, M. kattwínkeli (Schwarz), is known frornOlduvai Bed II and the Shungura Formation of the Omovalley (Gentry 1976),

Tribe HippotraglniA group of medium to large antelopes with horns pres­

ent in both sexes. The horns are either straight, scimítar­like, or twisted and generally have hollow pedicles 10their cores. The molars are moderately hypsodont withcharaeteristic basal pillars. Aeeording lo Ansell (1971),living hippotragines Call ínto three genera: Addax Lauril­lard, containing the addax found in north African deserts;Hippotragus Sundevall, the extinct blue buck, the sable,and the roan; and Oryx De Blainville, with two species,the scimitar oryx and the gemsbok.

Fossil remains from the Sterkfontein valley caves ap­pear to be restricted to the genus Hippotragus.

Hippotragus cf. equinus (Desrnarest). Roan antelopeProvenance: Sterkfontein Member 4 (but see below) andKromdraai A

The material is very sparse, consisting of two uppermolars from Sterkfontein and a single lower molar fromKromdraai. Previously H. B, S. Cooke (1947) described amandible from tbe "upper quarry" at Sterkfontein asHippotragoides broomí Cooke, but both Mohr (1967) andVrba (19760) agree that the specimen belongs in H.equtnus.

Vrba (19760) speculates whelher Cooke's specirnen aswell as the two Sterkfontein molars might not have comefrom a deposit younger than Member 4.

H. B, S, Cooke (n.d.) has described further materialamong the University of California collecticn fromGladysvale as Hippotragus broomi (Cooke), concludingthat it showed characters similar to those of his "Hippo­tragotdes" from Sterkfontein.

AH tbat can be said, then, is that remalns of roan an­telopes come doubtfully from Sterkfontein Member 4 andcertainly from Krorndraai A.

Hippotragus cf. niger (Harris). Sable antelopeProvenance: Swartkrans Member 2

The sample from Swartkraos carne from a minimum ofsix juveniles and three adults. MorphologicaJly the fossils

cannot be separated from the extant sable antelope,which has a wide distribution in the southern savannazone of África.

?HippotraginiProvenance: Sterkfontein Member 4 and, very tenta­tively, Member 5, Swartkrans Member 1, and KrorndraaiA

This tentative taxon is best represented in SterkfonteinMember 4 and by single specimens only from the othersite units. Vrba (19760) expresses the opinion that most ofthe specimens probably belong 10 a single species that islikely to include material from the MakapansgatLtmeworks, designated Taurotragus cf. oryx by wellsand Cooke (1956), She gives reasons for believing that theSterkfontein valley teeth eould well be híppotraginerather than tragelaphine and concludes that they moslclosely resemble those of the extinct form Hippotragusgigas L. S, B. Leakey, although lhe possíbility of othertribal affinities eannol be ruled out (Vrba 1976a, p. 46),

Hippotragus gigas was descríbed by L. S. B, Leakey(1965) as "a Hippotragus of gigantic proportions, witheharacters reealling both H, equtnus and H. niger;" Theholotype comes from Bed II and the paratype from thetop of Bed 1 al Olduvai, A single tooth from lhe Omovalley could belong lo H, gigas (Gentry 1976), and thetaxon is also recorded ín the Koobi Fora Formation (Har­ris 1976) and Crom Elandsfontein and possibly Florisbad.

Tribe ReduneiniA group of medium lo Iarge antelopes in whieh only

males carry horns; lhese may be bowed, Iyrale, orhooked. Basal pillars are present in both upper and lowermolars, and the latter often show goal folds.

There are two genera of living reduncines, Redunca H.Smith, which contains three species of reedbuck, andKobus A. Smíth, encompassing tbe waterbuck, lechwes,kob, and puku, Togelher lhese animals have a very widesavanna-zone distribution in Mrica but are seldom foundfar from water.

Two fossil representatives of the tribe are found in theSterkfonteín valley assernblages.

Redunca cf. arundinum (Boddaert). ReedbuckProvenance: Sterkfontein Member 4, Swartkrans Mem­ber 1, and Kromdraai A

The material is sparse, consisting of single juveniledentitions from Sterkfontein and Swartkrans and remains

/

",('(

\\

of one adult from Kromdraai A, Vrba (19700) observesthat the fossils are morphologically indistinguishable fromthe extanl reedbuck R. arundinum, bUI the possíbilityalso exists that they had affinities with an extinct fonn R.darti Wells and Cooke. In the course of Iheir study ofMakapansgat fossil Bovidae, Wells and Cooke (1956)found that the rnost cornmonly occuning species in theassemblage was a reedbuck about the size of R. arun­dínum, "but possessing more massive hom cores with adistinctly sigmoid proñle" (Wells and Cooke 1956, p. 17),The f2 was found to be larger than in R. arundmum, andthe cheek teeth were relatively and absolutely broader. Inthese respecta R. darti was reminiscent of the puku,Kobus vardoni (Livingstone), and appeared in sorne re­spects to be intermediate in fonn between the puku andthe reedbuck.

Dr. Vrba's suspicíon that R. dartí could be representedin the Sterkfontein valley assemblages is strengthened bythe fact that a Redunca frontlet, found al Sterkfonleinabout 1935 and preserved in the Analomical Museum ofwitwatersrand Universily, has been identiñed by Wells(l969a) as belonging lo R. darti.

er. Kobus ellipsiprymnus (Ogilby). WalerbuckProvenance: Swartkrans Member 2

Vrba (19700, p. 27) notes that her idenlification of tbetwo Swartkrans specimens is extremely tentative.

Waterbuck have a very wide African dístributíon, andIhe species is divided by Ansell (1971) into an eltipsi­prymnus and a defassa group. The former contains foursubspecies, the latter nine.

Wells (1967) notes that K. ellipsiprymnus is known onlyfrom late fossil contexts in southern Africa. However, in1913 Broom described a fronllel and a right horn-corefrom Haagenstad (now known as Florisbad) as Kobusventerae, which he thought 10 be intermediate in forrobetween Ihe walerbuck and the lechwe, K. leche, Similarremains have since been found at various other OrangeFree State sites, but there does not seem to be justifica­tion for separatingK. venterae from K. leche (Vrba, pers.comm.).

The Fossil Animals 173

; ~

{. '\\~-

, ~... "t 'd/IJ.<'. ':1' • '. i)':~{~~"'-~' , t ~

~- --e, -.."----..

A second species of Kobus has been described byH, B. S. Cooke (1949a) on the basis of an isolated P'from the Vaal River Gravels in the Riverview Estatesarca. The tooth is described as being similar in form 10 thatofK. ellipsiprymnus but twice as large. Nofurther remainsof a giant waterbuck appear to have come 10 light.

According lo Gentry (1976), Ihe ancestor of K. ellipsi­prymnus is K. sígmoídaiis Arambourg, a form originallydescribed from the Omo valley. Rernains are abundant inthe Shungura Forrnation, and Gentry believes that thetransition from K. sigmoidalis 10 K. eilipsiprymnus is de­tectable in Member G. Waterbuck remains are extremelynumerous in the East Rudolf bovid fauna, but the separa­tion of K. sigmoidaíis from K. eilipsiprymnus does notseem to be as readily definable in the Koobi Fora as it is inthe Shungura Formatlon (Harris 1976).

Tribe PeleiniThe tribe contains a single representative, Pelea ca­

preolus (Forsterj, the gray rhebuck or vaal ribbok, con­ñned to South Africa and southeastern Botswana. It is arather long-necked, graceful anlelope weighing 23-27 kg(5O-W lb), with straight, vertical horns carried only by themales. Its preferred habitat is grassy hills and rnountains,where it lives in small groups, each with a dominant male.lt is exclusívely a grazer,

There is sorne uncertainty about whether Pelea shouldbe plaeed in the subfamily Hippotraginae. Gentry (1970a)has suggesled that it might possibly be plaeed in the Ca­prinae, with affinities to the goats or the chamois,Rupicapra rupicapra (Linnaeus).

Peleo cf. capreo/us (Porster, 1790)Provenance: Members 1 and 2 at Swartkrans and Krom­draai A

The most complete remains are found in Member 2 at

174 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Swartkrans and include one almost intact skull of a sub­adult, SK 2735a-e. According to Vrba (l976a), thisshows sorne interesting differences from extant Pelea.' theteeth appear to be larger, and the angle between braincaseand face is perhaps less acute. Possible subspecific varia­tion is suggested. It is perhaps surprising that the earliestfossils, from Swartkrans Member 1, are Indistinguishablein size from the living form, whereas those from the laterMemher 2 and from Kromdraai A are larger.

Remains oí P. capreolus are known from the Cave ofHearths (H. B. S. Cooke 1962), Klasies River Mouthcaves (Klein 19700), and Wilton Shelter (Brain, this vol­ume).

Subfamily AntilopinaeIn his review of African artiodactyls, Ansell (1971)

c1assifies living representatives oí the Antilopinae as fol­lows:Tribe: Antilopini

Genera: Gazella De Blainville, gazelles; AntídorcasSundevall, springbok; Litocranius Kohl, gerenuk.

Tribe: Ammodorcadini (Or Ammodorcini)Genus Ammodorcas Thomas, dibatag.

Tribe: NeotraginiGenera: Oreotragus A. Smith, klipspringer; MadoquaOgilby, dik-diks; Dorcatrogus Noack. Beira antelope;Ourebia Laurillard, oribi; Raphicerus H. Smith, steen­bok and grysbok; Neotragus H. Smith, suni, and soforth.

Tribe AntilopiniGenus Antídorcas Sundevall, 1847

The genus contains a single living representative, A.marsupialis (Zirnmerrnan), the springbok, of which threesubspecies are recognized by AnseU (1971): A. m. mar­supíaíis (Zimmerman), widely distributed in South Africa,A. m. hofmeyri Thomas, from southem South-West Af­rica and Botswana, and A. m. angolensls, from northemSouth-West Africa and Angola.

Antidorcas australis Hendey and Hendey andlor A. mar­supialis (Zimmerman)Provenance: Swartkrans Member 2 and channel fiU

In 1968 Q. B. Hendey and H. Hendey described a newsubspecies of springbok asA. marsupialis australis 00 thebasis of cranial remains from the Swartklip site 00 theFalse Bay coast near Cape Town. The new form difTeredfrom the extant springbok particularly in the shape andsize of the hom-cores: these were intermediate in size

between those of the male and female of the modero forrnand, although lyrate in fonn, showed no inward CUrvetoward the tips. The teeth of the two forms appearedmorphologically indistinguishable, though those of thefossü tended to be rather narrow (Hendey and Hendey1968, p. 56). Although the Swartklip springbok was re­garded as a subspecitic geographic variant of A. mar­supíalis, the authors pointed out that, if A. m. australisshould prove to have been more widely distributed in thepast, speciñc status would be warranted.

When describing Antidorcas fossil material fromSwartkrans, Vrba (1973) a1located several dentitions andhom-cores toA. australis, pointing out that in view ofthetemporal and geographic separation of Swartkrans fromthe westem Cape sites, it seemed highly likely tbat au­stralis should be regarded as a species separate from mllr­supíalis, More recently it has become apparent that theSwartkrans specimens referred to here come from theMember 2 accumulation phase that, when calciñed chan­nel fills are included, certainly spanned a considerableperiod of time. Vrba (19700) now has queried the concJu­sion that A. austraíis is definitely present at Swartkransand suggests that the A. australislike specimens may rep­resent premarsupialis evolutionary stage.

Vrba (1973) provided a diagram of possible phyloge­netic relationship in the genus Antidorcas, suggesting thatin Pliocene times the Antídorcas lineage separated fromthe Gazella; thereafter the A. bondí stream separatedfrom that of A. reckl, and al a subsequent time A. reckigave rise to both A. australis and A. marsupíalis,

Antidorcas bond; (Cooke and WeJls, 1951). Bond'sspringbokProvenance: tentatively recorded from SterkfonteinMember 4, Kromdraai A and B; positively fromSterkfontein Member 6, Swartkrans Member 2 and chan­nelfiJl

In 1942. when describing faunal remains from the Vlak­kraal thermal springs in the Orange Free State, Wells andCooke noted that they were unable to assign three veryhypsodont lower teeth and a corresponding upper tooth toany known bovid species. The teeth were similar to thoseof the impala but transversely narrower; they were listedas Antílope gen. et sp, indet.

Sorne years later Cooke and Wells (1951)described fos­silízed remains from alluvial terraces flanking a tributaryof the Umgusa River on Chelmer Farm, 13 mi west­northwest of Bulawayo. Bond and Summers (1951)showed that the fossils carne from the top of the oIdera1luvium that rests on Karroo (Forest) sandstone. Theassemblage included several teeth similar to those ob-

11,

tained from Vlakkraal and were named "Gazella" bond;Cooke and Wells.

In her study of Antidorcas fossils from Swartkrans.Vrba (1973) noted that in Gemrys opinion "Gazella"bondi should become Aruídorcas bond; (Cooke andWells); Vrba concurred with this suggestioo and provideda detailed expanded diagnosis of the species 00 the basisof abundant material from Swartkrans. She showed thatthe teeth were consistently very ta11 and that the "hypso­donty index" for Mz was higher at all wear stages than inother species of Antidorcas-that is, recki, austraJis, ormarsupiaíis, The hom-cores of the Swartkrans A. bondíspecimens were found to be less mediolaterally com­pressed than in other Antidorcas material, and they typi­cally diverged slightly basally but converged toward thetips, giving the homs the appearance of parentheses whenviewed from the front.

Remains ofA. bondi are now known from many south­ern African sites, and the species appears to have beencommon during later Pleistocene times. Irs precise ex­tinction point has not been established.

Antidorcas recki (Schwarz)Provenance: positively identified from Kromdraai A,tentatively from Sterkfontein members 4 and 5,Swartkrans Member 1, and Kromdraai B

This springbok was appreciably smaller than the livingA. marsupialis and had proportionally shorter Iegs. Itshorn-cores were bent sharply backward toward their tipsand showed more mediolateraJ compression than does theliving formo

The type was originally described by Schwarz asAdenota recki on the basis of a skull and right horn-corefrom Olduvai, but the specimen was destroyed in Munichduring the Second World War. More recent material hasbeen referred to Phenacotragus recki (Schwarz) but isnow regarded as fitting better in the genusAntidorcas (seeVrba 1976a),

When describing fossil marnmals from the Vaal Rivergravels, H, B. S. Cooke (l949a) created the name Gazellawellsi Cooke for two mandibular fragments and an iso­lated left M3 from Power's site near Kimberley. He re­marked that the teeth were remarkably similar to thcse ofthe Mongolian Gazella subgutterosa,

Gentry (1966) pointed out lhat the most common gazel­line species al Olduvai appeared lo be conspecíñc with G,wellsi, particularly 00 the basis of excellent material fromPit 3 at Bolt's Farm (Cooke, n.d.), More recently Gentry(1978) considered that all the G. wellsi material should bereclassified as Antidorcas reckí,

A remarkable coUection of A. recta remains was exca­valed al site SHK II ofOlduvai Gorge by L. S. B, Leakeyin 1935. These remains appear to have come from a herd

The Fossil Animals 175

of nine or ten individuals, all but two of which werejuvenile or young (Gentry 1966, p. 78). The herd mayhave been driven into a swamp by hominid hunters, sincethe bones were preserved in clay about 90 m from a livingfloor in Bed H.

With the recording of A. recki in the Koobi Fora For­mation (Harris I976a) and possibly in the Shungura For­mation (Gentry 1976), ít appears that this probable an­cestor of the living springbok had a wide African distribu­tion.

Gateíía sp.Provenance: Sterkfontein Memher 4, Kromdraai B. andpossibly Swartkrans Member 1

No living species of Gazella occur in southem Africatoday, although ten species are recognízed by Gentry(l971a) from farther north. Antilopines of the genusGazeíío De Blainville differ from those oCAnttdorcas par­ticutarLy in the structure of their frontal bones. In Anti­dorcas the frontal sinus is particularly well developed,causing inñation of the frontal bones and a hollowiog ofthe horn pedicles. For this reason the frontal bones be­tween the hom pedicles are raised aboye the level of theorbit.al rims and the openings of the supraorbital canalsare sunk into bony tubes, None of these developmentsmay be observed in Gazella.

The ñrst fossfl assigned to the genus Gazella fromsouthern Africa was found at Comelia in the Orange FreeState and described as G. helmoedi by van Hoepen(1932b), 1t consisred of a long slender horn-core with agentle spiral twíst: it was perhaps as much as 40 cm longin undamaged form. Initially van Hoepen thought tbehom-eore was from the left side, but more recentlyH. B. S. Cooke (1974) has argued that it is from the rightand that the basal hollowing Is more suggestive of Anti­dorcas tban GazeJla, if in fact, the specimen is from anantilepine. Altematively, it could as well be of alcela­phine affinity, as suggested by Vrba (1973), who, in faet,plaeed it in the Alcelaphini.

From the Limeworks Cave at Makapansgat, Wells andCooke (1956) described two new species of gazelline an­telope: Gaielto gracilior and Phenacotragus vanhoepeni.The former was reminiscent ofthe red-fronted gazelle, G.rufifrons, from west and north Africa but with more deli­cate hom-cores and a relatively Iarger Mt- The secondspecies, P. vanhoepeni, has subsequentIy been removedfrom the genus Phenacolragus by Wells (l969b) andplaced in Gazella on account of the structure of its frontalbones and sinuses, Wells and Cooke (1956) pointed outthat G. vanhoepení resembled Antidorcas recki but hadhorn-cores that were comparatively massive and laterallycornpressed, rising vertically from the frontals.

176 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

cording lo Ansel1 (1971), a single species split into thirteensubspecies occurs over an extremely wide African range.The short black tail, prominenlly displayed, and the"rockíng-horse" gait, interspersed with high leaps, arecharacteristic. The fossil history of the genus appears tobe unknown.

and more restricted east African distribution. Its habitattolerance is considerable, ranging from arid plains to openwoodland, where it browses on low bushes and shrubs.Steenbok occur either alone or in pairs; females are fre­quently larger than males but lack homs.

Gentry (1978) is ofthe opinion that one of the paratypesof Cephalophus prieei described by Wel1s and Cooke(1956) from Makapansgat should be referred loRaphicerus.

Fossils of Ruphicerus are known from Langebaanweg,Elandsfontein, and many other sites (Hendey 1974a).Klein (19760)has reviewed the fossil hislory of the genusin the Cape biotic zone and has suggested that individualswere larger during cold climatic intervals.

Ourebía cf. ourebi (Zimmerman), OribiProvenance: Swartkrans Member 2

This is a graceful small antelope of the open plains,always close to water, and almost entirely a grazer. Ac-

which have the consistency of hard rubber and are trun­cated and blunt, allowing the animal to gain a foothold inthe most precarious places. The coat is thick and bristlyand appears to constitute a cushion against bumps, towhích klipspringers are prone (Dorst and DandeloI1972).

A single living species is recognized with a very widedistríbution in southem and eastern Africa. In all but afew mees, the females are hornless. The Swartkrans re­mains do oot appear to differ from the living formo

Vrba (1976a) has expressed the opinion that theGazcl!a sp. specimens from Sterkfontein, Swartkrans.and Kromdraai could have affinities to G. vanhoepeni.The species has apparently not been recognized at anyother sites.

Oreotragus cf. oreotragus (Zirnmerman). KlipspringerProvenance: Swartkrans Member 2

This is a small compact antelope adapted for life inrocky habitats, particularly in the structure of the hooves,

Tribe NeotraginiThis is a rather heterogeneous group of small antejopes

with short spikelike homs and large preorbital fossae. Sixgenera of living neotragines are recognized, as follows:Oreotragus A. Smith; klipspringerMadoquo Ogilby; dik-dikDorcatragus Ncack; Beira antelopeOurebía Laurillard; oribiRaphícerus H. Smith; steenbok and grysbokNeotragus H. Smith; suni, and so forth

Oreotragus cf. major WellsProvenance: Sterkfonteio Member 5, Swartkrans mem­bers 1 and 2

Accordiog to the available evidence, this antelope wassimilar to the living klipspringer but was 15-20% larger incranial dimensions (WeUs and Cooke 1956). It was origi­nally described by Wel1s (1951) on the basis of a rea­sonably complete skuU from a breccia deposit 00 a farmadjacenl lo Makapansgat. No other fossils have been de­scribed from this locality, and its age relationships areunknown. Further remains have been described from theMakapansgal Limeworks (Wel1s and Cooke 1956).

Raphícerus cf. campestris (Thunberg), SteenbokProvenance: Swartkrans Member 2; Raphicerus sp.:Swartkrans Member 2. Kromdraai A

This is a familiar small antelope with a wide southem

Subfamily CaprinaeTribe Ovibovini

There are two living representatives: the muskox.Ovíbos moschatus (Zimmerman) of high latitudes inNorth America, and the Chinese takin, Budorcastaxicolor Hodgson,

Gentry (l97Ih) has poinled out that the Ovibovini arewel1 represented as fossils, almosl al1 of which are Eura­sian. The only African forms thus far described arespecimens of the genus Makapania Wells and Cooke fromthe Transvaal and a tentatively identified horn-core fromthe Omo valley (Gentry 1970h).

Makapania broomi Wells and CookeProvenance: M. ef. broomi: Sterkfontein members 4 and5: cf Makapania sp.: Swartkrans Member 1

Wells and Cooke (1956) created this taxon for sorneremarkable fossils from Makapansgat Limeworks thatthey coosidered to belong in the tribe AJcelaphini. Themost striking feature oí these antelope was their horns,which were directed strongly lateraUy.

In a subsequent reexamination oí the Makapansgatmaterial, Gentry (l970b) concluded that Makapania wasin fact an ovibovine, Strong affinities to the French Villa­franchian ovibovine Mega/avis latífrons Schaub weresuggested by tbe structure of the basioccipitals as well as"fhe wide insertions oí the horn cores just behind theorbits, their transverse emergence. the gentle aseent to­wards the tips, absence oí transverse ridges 00 the homcores, the short braincase well angled 00 the facial axis,the extent oí the projection of the dorsal parts of theorbital rims, the small localised preorbital fossae, fairlyhypsodont teeth" (Gentry 1970b, p. 64).

Makapania therefore appears to have been an anteJopesimilar in appearance to the rnuskox, with laterally pro­jecting horns. Vrba (l976a) equates the dentitions fromMember 4 at Sterkfontein, and the single tooth fromMember 5, tentatively with M. broomi. The teeth fromMember 1 at Swartkrans, however, are appreciablysmaller, though in most respects Makapanialike. Vrbaconcludes that the Swartkrans material could representeither a smaller, and later, species of Makapania or anaberrant mediurn-sized alcelaphine.

Family SuidaeThree species of indigenous living pig are currently

recognized in sub-Saharan Africa (Anseli 1971): Pota­mochoerus porcus (Linnaeus), the bush pig or redríver hog; Phacochoerus aethiopicus (pallas), the wart­hog: and Hylochoerus meinertzhageni Thomas, the giantforest hog, The wild boar, Sus seroja Linnaeus, still oc­eurs in Tunisia, although it has disappeared from most ofits former north African range,

Three extinct Suid taxa are currently recognízed in theSterkfontein valley assemblages under consideration.

Phacochoerus modestus (van Hoepen and van Hoepen)Provenance: Swartkrans Member 2, Krorndraai A and B

This warthog was c1early very similar in appearance tothe living form and apparently close to the ancestry of it.

The Fossil Animals 177

The species P. antíquus was originally described byBroom (1948a) 00 the basis of an irnrnature skull firstthought to have come from Sterkfontein but , in fact ,almost certainly frorn Kromdraai A. Ewer (1956(', p. 528)redescribed the specímen, placing it in a new subgenusPotamochoerops. She considered it to differ from theliving form by the following characters:

auditory bulla conical and pointed; frontals not con­cave aboye the orbits, at least in the immature animal'jugal not forming a wel1 defined post orbital process; ,side walIs of the snout aboye the canines slightly con­cave; canines more forwardly directed than in P. af­rtcanus; cheek teeth very similar to those of the livingspecies, but differing in the foilowing points, P4 withposterior accessory cusp single, first and second mo­lars distinctly bilobed and with the talon or talonidcontaining only a single cusp. Root formation 00 theanterior columns of M 3 starts before the posterior col­umns come into wear.

In her consideration of adaptive features in the skulls ofAfrican pigs, Ewer (1958b) pointed out that features in theskuU of the bushpig, Potamochoerus koiropotamus, wererelated to this animal's digging habit, while those of thewarthog skull were related 10 a grass-eating habit. Sheconcluded that the adaptations in the skull of P. antiquusthat allowed for the grinding action of the cheek teethwere far less well developed than in the living warthog.

An exceptionally complete skull with articulated atlasand axis vertebrae was found by the University ofCalifornia expedition at Pit 3 of Bol!' s Farm. It has beendescribed in detall by Cooke (n.d.), who states that itsfeatures confirm the conclusions reached by Ewer. Cookeand Wilkinson (1978) use the name P. modestus in prefer­ence to P. antiquus.

Metridiochoerus andrewsi HopwoodProvenance: Swartkrans Member 1 and Kromdraai A

Shaw (1938) reported the discovery of three third mo­lars from Sterkfontein, presumably from Member 4. Theyare not included in this anaIysis.

These large extinct pigs had molars that are mor­phologically similar to those of Phacochoerus but arestrikingly different in anterior skuU structure. In particu­lar. the canines were comparatively straígbt, reminiscentof small elephant tusks; the uppers pmjected laterallyfrom the sides of the muzzle.

The specific name meadowsí was created by Broom(1928) when he named an isolated third molar fmm día­mond gravels on the Vaal River near Kimberley asNotochoerus meadowsi: a few years eartier he hadcreated the genus (Broom 1925)for an exceptionally largelbird molar, a1so from the VaaI River gravels, that henamed N. capensis.

The genus Tapinochoerus was introduced by van

178 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation,

Hoepen and van Hoepen (1932) when they described T.modestos frorn the Comelia Beds. They expressed theopinión that Notochoerus meadowsi should be placed inTapinochoerus. an opinión supported by H. B. S. Cooke(l949h) and subsequent authcrs.

Cooke and Maglio (1972) have expressed the opinionthat two forros found al Olduvai, Afrochoerus nicoli andT. meadowsi, could be regarded as a single speciesshowing considerable sexual dimorphism. The almoststraight, elephantlike tusks of A. nicoli would find theirextreme development in the males. More recently Cooke(1976) abandoned thispossibility, sinking Tapinochoerusinto Metridiochoerus .andAfrochoerus into Stylochoerus.

In their recent review 1 White and Harris (1977) agreethat Tapinochoerus should be regarded as a synonym ofMetridíochoerus, This policy is supported by Cooke andWilkinson (1978).

Cf. Metridiochoerus sp.Provenance: Sterkfontein Member 4

Sívatheríurn maurusiurn (Pomel). SivathereProvenance: Swartkrans Member 2

The single specimen, an isolated and unerupted uppermolar, carne from a subadult sivathere of a form that,

ossicones of this species varied a good deal in form. Herecognized three basic ossicone shapes, as indicated inthe iIIustration.

although now extinct, had a wide African range both inspace and in time. These were heavy-bodied, short­necked giraffids that must have browsed in the same waythat modern giraffes do. According lo Harris (l976c), the

Order Perissodactyla

Two families of odd-toed ungulates are represented inAfrica: Rhinocerotidae, containing two species of rhino,and Bquidae, containing one species of ass and four ofzebra. No rhino remains have thus far been found in theSterkfontein valley caves, bUI equid remains are fairlycommon.

10 cmSivatherium

From Harris 1976c

Genus Sivatheríum Falconer and Cautley, 1835Harris (1976h, p. 317) has provided the fol1owing di­

agnosis of this genus, based on an earlier one by Colbert(1935):

A gigantic Pleistocene giraffid, with four ossicones inthe male, an anterior pair arising from the frontals anda posterior pair situated on the parietals. As in theother gigantic sivatheriines there are deep pits in thetemporal fossae for the temporal muscles, and on thesupraoccipital for the neck muscles. The face is veryshort, the nasals being retracted and strongly curved.The teeth are large with rugose ename!. Body andlimbs heavy, limbs not elongated.

Several forms of sivathere have been described fromSouth Africa, includíng Griquatherium cingulatum,which Haughton (1922) named on the basis of an uppermolar from the VaaJ River gravels, G. haughtoni H. B. S.Cooke, 1949, and Orangiatherium vanrhyni von Hoepen,1932. Singer and Boné (1960) synonymized sorne of theearlier names, recognizing in southem AfricaSivatheriumcingulatum as well as three subspecies ofS. oíduvaiense.

In his description of the only sivathere specimen cur­rently known from the Sterkfonlein valley caves, a singletooth from Swankrans, Churcher (1974) concluded lhatal1southem Mrican material known at that time could beclassiñed as S. maurusium. This view was supported byHarris (l976c), who also described a new species, S. hen­deyi, from the Varswater Forrnation of Langebaanweg.

Until more material becomes avaílable, the validity ofthis new species is questioned by Churcher (1978).

Family GiraflidaeThere are two living representatives of the famíly:

the giraffe, Giraffa cameloparda/is (Linnaeus), and theokapi, Okapiajohnstoni Sclater, placed in the subfamilíesGiraffinae and Palaeotraginae respectively. The thirdsubfamily, the Sivatheriinae, contains a variety of largefossil forms, inc1uding the one recorded from Swartkrans.Giraflids are known from tbe Sterkfontein valley caves bythis single specimen, though Broom (l948a) refers lo amandible of "a very large Giraffid,' without saying fromwhich site it carne. I have not been able to relocate hisspecimen. though it may be the ene described by Cookeand Wells (1947) from the Makapan valley.

Family EquidaeAllliving African equids are plaeed in the genus Equus

Linnaeus by Ansell (1971), though the subgenera Asínns,

Dotichohippus, and Hippotigris are retained. Fossilequids to be considered here fall into two genera, Hip­parían and Equus,

The Genus Hipparion de Christol, 1832Horses of the genus Hipparion were divided from a

Merychippus stock in North Ameríca and appeared firstin Europe 12.5 million years ago (Hooijer 1975). Theywere generally lightly built animals, about the size of apony, though some species were somewhat larger, anexarnple being H. crusafonti Villalta, the terminal speciesin Europe that died out during the Yillafranchian, thoughAfrican forms survived much longer.

In Hípparion there are three functionaJ toes on each ofthe four feet, whereas in Equus the weight of the horse Iscarried through the enlarged third metapodials while thesecond and fourth toes are reduced to insígniñcant"splint" bones. This means that Equus is a single-toedanimal, whereas ín Hipparion the sirle toes presumablylouched the ground during trotting and galloping,

p

Another olear differenee between represeotatives of thetwo genera is found in tbe enamel patterns 00 the uppereheek teeth. As shown in the iIIustration, the protoeonefonns an island in Hipparíon but no! in Equus.

The following hipparionid taxa have beeo deseribedfrom South Mrica:

Hipparion stevtlerl van Hoepen, 1930, from Uitzoek,oear Cornelia, based on four isolated teeth.

Eurygnathohippus cornelianus van Hoepen, 1930, fromCornelia, based on a mandibular symphysis with inclsors,now shown to have belonged to a Hipporion (L. S. B.Leakey 1965).

Stylohipparion hipkini Van Hoepen, 1932, from Uit­zoek, near Camelia, based on an isolated right M2 •

Notohipparíon namaquense Haughton, 1932, from alocality 49 mi easl ofSpringbok, Narnaqualand, and basedon a series of mandibular teeth from a single individual.

Hípparíon (Hipparion) albertense baardi Boné andSinger (1965) from Langebaanweg, based on a number ofspecirnens.

Boné and Singer (1965) have proposed two subgenerain which aH Afriean fonns may be accommodared:

Hipparion (Hipparion) de ChristolHipparion (Stylohipparion) van HoepenThey place the Langebaanweg material in tbe former

subgenus, but all the other soutbem African fossils in thelalter. Singer and Boné (1966) were inclined lo place allthe known Hippariaon (Stylohipparion) material in thesingle specíes libycum Pomel, but this has not been sup­ported by Churcher (1970), who classitied the Swartkransand Kromdraai material as H. (Stylohipporion) steytlerivan Hoepen.

The Fossil Animals 179

The taxonomie arrangement of Boné and Singer has notgained wide support. Hooíjer (1975) has pointed out thatSlylohipparion is a synonym of Eurygnathohippus andthat H. albertense is a nomen vanum. He regards the formH. albertense baurdi as a full species H. baardí, A fulldescription of the late Phocene Equidae fromLangebaanweg is provided by Hooijer (1976), and he alsoreeords H. cf. namaquense from this site.

In a reeent review, Chureher and Richardson (1978)classify all the advaneed hípparions from Africa in asingle taxon Hipparíon Iybicum. As synonyrns fromsouthern Africa they list Hipparion stevtleri, Euryg­nathohippus comeííanus, and Sty/ohipparion hipkini.This arrangement is followed here.

Hípparion lybicum Pomet. Three-toed horseProvenance: Swartkrans Member 1 and Kromdraai A

Little is known ofthis animal apart from its teeth, but itappears to have been a lightly built equid, appreciably

smaller than the familiar Burchell's zebra. It has threetoes 00 each foot, but whether it was striped or plain issomething we may never know.

There is no certainty about when Hipparion died out inSouth África, although the mcst recent evidence of theanimal is found at Comelia. The fossils are found par­ticularly in Bed b ofMember 1 (Butzer 1974a), a horizonoverlying that ín which artifacts are eoncentrated. Ac­cording lo J. D. Clark (1974), the Comelia artifacts showsimilarities to those of Upper Acheulean affinities at Bro­ken HiIl in Zambia.

Equus (Dolichohippus) capensis Broom. Giant CapehorseProvenance: Sterkfontein Member 4, Swartkrans Mem­ber 1, and Kromdraai A

This was the tirsl species offossil horse lo be describedin South Africa. For sorne time Broom had suggested thata large extinct equid may have existed in the Cape, but in1909 (Broom 1909b) he was able for the first time lo de-

180 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

seribe its dentition. The fossil was found in a slab of sur­face limestone east up on the beaeh at Yzerplaatz in TableBay. It consisted ofmuch ofthe left mandibular ramus ofa horse considerably Iarger than Equus cabal/as Lin­naeus. Further rernains from various parts of South Af­rica allowed Broom (1913b, 1928 and Broom and LeRiehe 1937) to describe the animal in greater detail. On'he evidence ofthe limb bones, Broom (l913b) concludedthat the proportions of E. capensís differed considerablyfrom those of E. cabal/uso E. capensis appears to havebeen more powerfully built, but did not stand so high,probably no more than fourteen hands. It appears to havehad a relatively more massive head.

In his consideration of the nomenc1ature of the SouthAfrican fossil equids, Wells (1959) regards Broom's typeofE. capensis as indeterminable and the name as a nomenvanum. He eonsidered that the correet name for the ex­tinet Cape horse should be E. helmeí Dreyer. In his de­scription of the remains from Sterkfontein. Swartkrans,and Kromdraai, Churcher (1970) does not agree that thetype of E. capensis is indeterminable or that the nameshould be discarded. He applies it to the Sterkfonteinvalley material and suggests that the name E. capensisshould include E. helmei Dreyer, E. ca woodí Broom, E.kuhni Broom, and E. zietsmani Broom as well as sorne ofthe teeth referred to E. horrisi Broom and E. plicatus (vanHoepen). Churcher and Richardson (1978) retain thename E. capensis but place the species in the subgenusDolichohippus,

Klein (19740) has shown that the dísappearance of E.capensis from the southern Cape occurred synchronouslywith that of other fonns such as the giant buffalo,Pelorovís antiquus, the giant hartebeest, Mega/otraguspriscos, and at least five other forms. These extinctionsseem to coincide with terminal Pleistocene environmentalchanges dated at 12,000-10,000 years R.P., and Kleinsuggests that the extinctions may have been caused bothby these changes and by more effeetive human huntingtechniques.

It is not known whether the Cape horse survived after10,000 years B.P. in other parts of southem África, thoughits remains are coming to light at many sítes scatteredover the subcontínent. It seems quite likely that artistsresponsible for rock paintings or engravings could haveseen the horse alive, and it may be possible, from suchartistic records, to establish whether E. capensis hadzebra stripes. At present this is a matter for speculation.

Equus (Hippotigris) quogga Gmelin. QuaggaProvenanee: Swankrans Member 2 and a tentativeidentification at Kromdraai A

The quagga was familiar to early travelers in the south-

em African interior, occurring abundantly in the southemand eastem parts of the great Karroo, with a possibteextension ínto Great Namaqualand (Ansell 1971). It doesnot appear to have occurred in any numbers north of rheVaal River, so that the recorded presence ofthe species inthe Sterkfontein valley caves represents an extension ofthe historically observed range.

It appears that quaggas rnay have survived in theOrange Free State as late as 1878, and we know that thelast individuals in European zoos died in 1872 (London),1875 (Berlín), and 1883 (Amsterdam). A surprisingnumber oí mounted specirnens are still in existence inmuseums as listed and described by Rau (1974).

Although the cranial dimensions oíE. quagga are verysimilar to those of E. burcheJli, there are apparemly con­sistent differenees in the enamel patterns on the occlusalsurfaces of the cheek teeth (H. B. S. Cooke 1943, 1950;Churcher 1970). Remains of E. quagga have been re­corded at tbe Wonderwerk Cave near Kuruman (H. B. S.Cooke 1941) and from near Kroonstad in the Orange FreeState (Shapiro 1943).

Churcher and Richardson (1978) have cast doubt onChurcher's eariier identification of Swartkrans specimensas E. quagga. They state that .. it seems best to regard thepresence oí E. quagga at so early a date and so great adistance from its known range with skepticisrn."

Equus (Hippotigris) burchelli (Gray). Burchell's zebraProvenance: Sterkfontein members 5 and 6, Swartkranschannel fill, and Kromdraai A

This familiar animal has a wide distribution in southemand eastem Africa. Ansell (1971) recognizes six sub-

species and notes that E. b. burchelli (Gray, 1824), whichfonnerly occurred in the Orange Free State, northeastemCape, southem Botswana, and southeastern Transvaal, isnow extinct.

H. B. S. Cooke (1950) expresses the opinion that threeforms described by van Hoepen as E. platyconus, E.slmpíicissimus, and Kraterohlppus elongatus can be re­garded as belonging to E. burchelli. The saine is true ofE.Iylei Dreyer from Florisbad, Remains of Burchell's zebraare a1soknown from the Vaa1 River gravels and a varietyof later P1eistocene sítes.

Order Proboscldea

Family E1ephamidae: E1ephantsOnly one specíes of living African elephallt ís currently

recognized, Loxodonta africana (Blumenbach), a1thougha large number of fossil elephantid genera and specieshave been described from Mrica, creating great confu­sion. Recently, however, Cooke and Maglio (1972), Ma·glio (1970, 1973) and Coppens, Maglio, Madden, and

Beden (1978) have fumished a phylogenetic scheme pro­viding comparative sirnplicity and clarity.

The family Elephantidae is thought to containtwenty-five valid species grouped in two subfamilies, theStegotetrabelodontinae and the Eiephantinae, both oíwhich had their origin in Miocene gomphotheres. It isthought that the Elephantinae arose from lhe Stegote­lrabelodontinae during tbe Pliocene, though neither of thetwo described species of Stegotetrabelodon appear tohave been the direct ancestors.

Wilhin the Elephantinae, three lineages are thought tohave arisco about five million years ago; these are repre­sented by the genera Loxodonta, Elephas, and Mam­muthus. They carne fromPrimelephas, in which the lowerincisors had beco greatly reduced, in contrast to the situ­atíon in Stegotetrabelodon, where lower tusks were typi­cally presento

Fossil representatíves of each of the three generaLosodonta, Elephas, and Mammuthus arefound in SouthAfriea, but only two specimens are known from the cavebreccias. These have been identified as Loxodonta atían­tica (Pomel) from Pit 7, Bolr's Farm by H. B. S. Cooke(1963 and n.d.); as E. ekorensts by Maglio (1973); and asElephas cf. recki (Dietrích) by Maglio (pers, comm.).

Elephas cf. recki (Dietrich)Provenance: Sterkfontein Member 4, known by a singledeciduous lower molar

According to Cooke and Maglio (1972), the Elephaslinease, which arose from a Primeiephas stock 4 to 5

million years ago, showed progressive evolution until itsAfrican extinctioo point in the late Pleistocene. The ear­Hest recognizable form in the lineage is E. ekorensts,known from Kanapoi, Ekora, and Kubi Algi. Here theeranial specializations separating Elephas fromLoxodonla were already apparent: a reduction of thetusks and premaxillae, compression of the skuU in thefacial plane, and an expansión of!he parietals and occip­ítals,

In East Afríca, and presumably in South Africa as well,E. ekorensís evolved into E. recki, which can then betraced ·through four progressive stages (Maglio 1970) inwhich the dentiticn increased its shearing efficiency andthe skull continued the trends that began in E. ekorensis.It appears that E. recki stage 4 gave rise to E. iolensis, aspecies synonymous with E. transvaalensis (fonnerlyPalaeoloxodon) from the Younger Gravels of lbe VaalRiver (H. B. S. Cooke 1960).

The sinaIe deciduous looth from Sterkfontein has beenexamined by Maglio, who assigned it lentatively lo E.recki, sta¡e 1. In lheir 1972 paper, Cooke and Maglio listlhe specimen under E. ekorensis, but whichever name isused lhere can be no doubl thal !he toolh carne from anElephas at an early slage of lhe lineage's prollfessive de­velopment. An ago of 3 to 4 mi11ion yem is indicatel! byCooke and Maglio's (1972) calibracjon diagnun.

The Fossil Animals 181

The Sterkfontein elephant was a juvenile of a formwhich superficially resembled the living Mriean elephant,except perhaps for smaller tusks.

arder Hyracoídea

FamíIy Procaviidae: DassiesThree genera of living hyraxes or dassies are rec­

ognized by Bothma (1971): Procavta Storr, HeterohyraxGray, and Dendrohyrax Gray. They are all restricted toAfrica with the exception of P. syriacus (Schreber), theSyrian dassie, which extends into Israel, Syria, andsoutheastern Arabia. This was apparently the "coney" oflhe Bible.

All living dassies are small, compaet anírnals weighingno more than 4.5 kg, with short necks, ears, and legs. The

toes are flattened except for the second, which carries along curved claw. The feet are well adapted for climbing,with the soles kept darnp through glandular secretions. Alarge gland is also presenl 00 lhe back, surrounded by hairdifferent in color from the rest. Fossils from the Transvaalcases represent two species of Procavía, discussedbelow, and an exceptionally large form, Giganlohyraxmagulreí Kitehing, described from Makapansgat Lime­works (Kilching 1965) and also known from deposits inthe Omo valley. Meyer (1978) suggested that both Pro­cavia and Gigantohyrax probably carne frorn a Pro­hyraxlike ancestor that in tum had been derived from aforro similar to Saghalherium.

Procavta antiqua BroomProvenance; Sterkfoatein Member 4, Swartkrans mem­bers 1 and 2, and Kromdraai A

The species was originally described by Broom (1934)on the basis of a number of broken skulls apparently as­sociated with the Auslralopithecus cranium from Taung.Although lhey resembled P. capensis, the living form, inmany respects, Broom found the fossil dentitions to bemore brachyodont, and he proposed a new subgenus, P.(Prohyrax) antíqua, suggesting lhat this dassie could havebeco ancestral to both P. capensís and P. arboreo (nowDendrohyrax arhoreus A. Smith). In a subsequent de­scription of further specimens from Taung, Broom(19480) dropped the subgeneric designalion and de­scribed. in the same paper, a new species oí dassie fromSterkfontein that he named P. robertsi. This again was abraehyodont foem simílar in size to P. capensis I but nodirect comparison was marle between ¡t and P. antiqua.

Churcher (1956) synonymized P. rohertsi with P. anti­qua and provided a revised definition of lhe laxon. He

182 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

found the skull of P. antiqua to be slightly smaller andIighter than that of P. capensis, with less hypsodont teethand a generally more primitive, unspecialized aspecto

P. anttquo was clearly a common and widespread das­sie in the northem Cape-Transvaal area, and probablyelsewhere. Churcher suggests that it was a member of theIineage leading to the modero Procavio species.

,.Precavía transvaalensís ShawProvenance: Sterkfontein Member 4, Swartkrans mem­bers 1 and 2, and Kromdraai A

This extinct dassie was about one and a half times aslarge as the living P. capensis, as indicated by cranialdimensíons, and apparently was common in the Transvaalduríng the Pleistocene, coexisting with P. antiqua.

The species was described first in February 1937 byShaw on the basis of a partial skutl and mandible fromCooper's site, situated between Sterkfontein and Krom­draai. In October 1936 Broorn (l937c) had announced anew, large Procavía from Gladysvale that he named P.obermeyerae, However, the descripticn was not pub­Iished until March 1937. so that when P. transvaaíensisand P. obermeyerae were found to be synonymous byChurcher (1956), the former name had priority.

At the time of his description, Broom (l937c) pointedout that the new Procavta was the largest true dassieknown. This remained true untíl the description of a verymuch more massive hyracoid, Gigantohyrax maguireifrom Makapansgat, by Kitching in 1965. Remains of P.transvaalensís are now known from a11 the major Trans­vaal cave breccias.

Procavia capensis (Pallas). Cape dassieProvenance: Sterkfontein Member 6 and Swartkranschannel fill

This ís the familiar .dassie of scuthem Africa. widelydistributed wherever suitable rocky habitat oecurs. Theylive in colonies of up to sixty individuals and are verylarge1y diurnal and entire1y vegetariano They tend to con­centrate their droppings in caves and rock shelters, whereJarge accumuIations may build up. The crystalline uríne,known as "hyracium;" has been much prized in the pastas a remedy for various ailments including convulsions,epilepsy, and women's diseases.

Order Lagomorpha

Family Leporidae: Rabbits and haresRabbits and hares were at one time classified as a sub­

order ofthe Rodentia, but they may be distinguished fromrodents in various ways, including the presence of twoupper incisors on either side, while rodents have onlyone. F. Petler (l97Ia) recognizes five genera of extantAfrican lagomorphs, three of which, Lepus Linnaeus,Pronolagus Lyon, and Bunolagus Thomas, occur insouthem Afríca.

No Iagomorph taxon has as yet been described frorn theaustralopithecine-bearing breccias, though H. B. S.Cooke (1963) provisionally lists the two extant forrnsProno!agus randensis from Makapansgat limeworks andLepus capensts from Taung, Sterkfontem, and Bolt'sFarm.

In the present study unidentilied lagomorphs are re­corded from Sterkfontein Member 5. Swartkrans Member2 and channel ñll, and Kromdraai A and B. In all prob­ability the two forms mentioned by Cooke are involved:Pronolagus randensís Jameson, the red rock hare, ashort-eared form occurring in suítable rocky habitats inthe Transvaal, Rhodesia, Botswana, and South-West Af­rica. They frequently live in caves. and for many years apair has lived and bred in the twilight zone of the LowerCave at Swankrans. Lepus capensis Linnaeus, the Capehare, described from the Cape Province but, according toF. Petter (l971a) a species that ranges throughout Africa,Europe, and Asia. It has appreciably longer ears thanPronolagus and apparentIy does not make use of caves.

Order Rodentía

Family Hystricidae: PorcupinesTwo genera of living African porcupines are currently

recognized (Missone 1971): Atherurus Cuvier, thebrush-tailed porcupines, and Hystrix Linnaeus, the com­mon and crested porcupines. No representatives ofAtherurus occur in southern Africa. Two species of Hys­trix are known: H. africaeaustraíis Peters, the comrnonporcupine of southem Africa, and H. cristata Linnaeus,the erested porcupine found in the northern half of thecontinent.

Apart from the forms found in the australopíthecinecaves breccias ofthe Transvaal, only one specíes offossilporcupine has been described from elsewhere in Africa:H. astosobae Bate from Abu Hugar on the Blue Nile inthe Sudan. It is of Middle or Upper Pleistocene age and,according to Maguire (1976), has affinities with H. crís­lafa.

The Transvaal fossils have been equated with the ex­tant H. africaeaustralis and with two extinct forros, H.makapanensis Greenwood and Xenoñyssrix crassidensGreenwood.

Hystrix africaeaustralis Peters. Common porcupineProvenance: Sterkfontein members 4, 5. and 6.Swartkrans members 1 and 2, and Kromdraai A

The common porcupine is extremely important as acollector of bones in caves, as was discussed in chapter 5.It appears lo have existed unchanged for al least threemillion vears, and its remeins are commonly preserved inStone Age sites, incJuding Elandsfontein and various

other localities in the southwestem Cape. Fossils fromthe australopithecine sítes have been studied in detail byMaguire (1976).

Sorne of the material from Makapansgat Limeworkswas described by Bakr (1959) as H. (Hystrix) greenwoodisp. nov. This taxon, however, is not considered to bevalid by Maguire (1976), who equates it with H. af­rícaeaustralís.

Hystrix makapanensis GreenwoodProvenance: Tentative identifications from SwartkransMember 1 and Kromdraai A

The species was originaUy described as H. majar fromthe Makapansgat Limeworks by Greenwood (1955) on thebasis of a left mandibular fragmento 1t had elear aflinitieswith H. africaeaustralis but was about one-thírd larger.Subsequently it was found that the specific name majorwas preoccupíed, having been used by Gervais in 1859fora specimen from Ratoneau Island near MarseiIles inFrance. Greenwood (1958) therefore renamed herMakapansgat taxon H. makapanensis.

Maguire (1976) points out that H. makapanensis is theleast common of the three Makapansgat fossil porcupinefonns. lt is intermediate in size between H. ofricaeaus­tralís and Xenohystrix crassidens, but in other respectswas c1early very like the living porcupines still found inthe area.

Both H. makapanensis and X. crassidens have beenrecorded from the Omo group sediments by Coppens andHowell (1976), suggesting a wide African distribution inearly Pleistocene times.

Class Aves

Order StrulhioformesFamily StruthionidaeStruthio sp. OstrichProvenance: Swartkrans Member 2 and Kromdraai B

The Fossil Animals 183

Ostriches are represented by sorne pieces of bone andeggshell, but these remains are too fragmentary to yieldinformation on whether the fossil anirnals were differentfrom contemporary ostriches,

Ctass Reptilia

Order CrocodiliaCf. Crocodylus niloticus. Nile crocodileProvenance: Krorndraai B

This tentative identification is based on a single tooth,and it would be unwise to deduce much from It.

Order SquamataFamily CordyliidaeCf. Cordylus giganteus. Girdled lizardProvenance: Kromdraai B

The dístríbutíon oC this lizard is restricted today to asmall area of the highveld, and the animal is a elear in­dicator of open grassland conditions. The tentativeidentification from Kromd.raai is based on caudal verte­brae and dermal plates; very little can be deduced fromthese remains about the nature of the fossil Cordylus.

Family VaranidaeVaranus cf, niíotícus, Monitor lizardProvenance: Sterkfontein Member 5

The identification is based on a single vertebra. similarto that of an extant monitor lizard such as still occurs inthe area today.

184 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

cyon and Petrodrornus may reach a head and body lengthof almost 30 cm. AH have long, mobile snouts andstrongly developed hind limbs. Most species are diurnalbut fall prey to owls hunting in the late aftemoon.

Genus Elephantulus Thomas and Schwann, 1906On the basis of remains from Skurveberg, Broom

(1937c) created a new genus and species, Eiephantomys

Order CheloniaFamily Testudinidae

Remains of unidentified tortoises have been recoveredfrom Sterkfontein Member 5, Swartkrans Member 2 andehannel fill, and Kromdraai A and B

Generally, notbing more tban a few píeces of carapaceor plastrón are avaílable, 00 the basis of which Iittle oftaxonomic value can be said.

Phylum Mollusco

Class GastropodaOrder Pulmonatau. Achatina sp. Land snailProvenance: Swankrans members I and 2 and KromdraaiA

Sorne well-preserved shells of these large land snails,still eommon in parts of the Transvaal loday, have beenfound. They await detailed study.

The Microvertebrate Componenl

Order Insectívora

Family Macroscelididae: Elephant shrewsElephant shrews are restricted to Africa ando according

lo Corbet (1971), are represented by fifteen species in thefollowing four genera: Rhynchocyon Peters, PetrodromusPetera, Macroscelídes A. Smith, and gíepñantuíusThomas and Sehwann. They are Iypieally small in­sectivores, although the forest elephant shrews Rhyncho-

Iangi, for elephant shrews having "a molariform secondupper premolar, no third lower molar and no large devel­opment of air cells in the ear region of the skull.·· In theserespects E. langi differed from alJ other known forros ex­cept Elephantulus íruufi Srnith, which Broom proposedshould a1so be placed in his new genus Eíephantomys,Subsequently Broom (1948a) transferred the species long;to Elephantulus, suggesting that two subgenera,Elephnntulus and Elephantomys, were probably war­ranted. In the same papero Broom described a newspecies, Elephantulus antiquus, on the basis of a maxil­lary specimen from Bclt' s Farol.

E. iangi has been recorded from Sterkfontein (Member40r Member 5) by Broom and Sehepers (1946), de Graaff(19600), and T. N. Pocock (1969); from Swartkrans (a1­mosl eertainly Member 1) by Davis (unpublished) and byH. B. S. Cooke (1963), who lists it for Kromdraai A.It is likely lo be presenl in the Kromdraai B mierofaunalsample currently being studied by Davis and Poeock.

In their revision of the elephant shrews, Corbet andHanks (1968) point out that the specific name langi Ispreoccupied and therefore propose the new Dame E.broomi for this taxon. Butler and Greenwood (1976) con­elude that E. hroomi occurred at both MakapansgatLimeworks and Bed 1 of Olduvai, and that it was relatedto the living species E. intufi and E. rupestris.

Remains attributed to E. cf. braehyrhynchus, the livingshort-snouted elephant shrew, have been recorded fromSwartkrans Member 1 and Kromdraai B (Davis, un­published).

Mylomygale spiersi BroomBroom (19480) described the left mandible of a remark­

able macroscelid from a cave about 1.6 km north of theaustralopithecine locality at Taung. Its hypsodont teethappeared unusually rodentlike ando conceming thespeeimen, Broom wrole (19480, p. 8): "Tbere is not, lthink, any reasonable doubt that this fossil mammal is agreatly speeialised Menolyphlan or elephant shrew ofwhich the nearest living ally is Petrodromus, We mayassume as probable that sorne early Pliocene elephantshrew allied to Petrodromus changed its habits and be­carne vegetarían and developed rodentlike moJars. ThisMenotyphlan is one of the most interesting fossil marn­mals discovered in recent years." An isolated premolarfrom Sterkfontein (presumably Member 4 or Memher 5)

has been tentatively assigned to M. spiersi by de Graaff(19600).

Family Chrysochloridae: Golden molesAccording to Mccsrer (1971. p. 1). this family, which is

restricted to Afriea, contains seven valid genera of goldenmoles that are "small to medium-sized, bJind burrowinganimals, with colour varying from pale fawn-grey to blackin different species, always with bright metallic sheen.Adaptations for subterraneanexístence inc1ude a fusiformbody, with no external eyes or ears, well-developedforelegs and claws, no visible tail, close-set fur and amuzzle terminating in a thick, leathery pad."

Although golden moles spend most of their time under­ground, they are occasionaUy taken by owls, suggestingthat they must surface at night. Fossits are not abundant,although remains oí Chrysochloris sp. occur in fair num­bers in the Quartzose Sand Member and Pelletal Phos­phorite Member of the Varswater Formation atLangebaanweg (Hendey 19740). Apart from the two taxamentioned below, the onIy other fossil form from south­em Africa is Chrysotricha hamiltoní, described by deGraaff (1957) from the Makapansgat Limeworks.

Proamblysomus antiquus Broorn, 1941The type specimen described by Brocm (I94lc) carne

from Bolt's Farro and resembled Amb/ysomus, althoughit differed in having a wider temporal region and largetyrnpanic bullae, unlike those of living golden moles. Aspecimen from Kromdraai A has been tentatively as­signed lo this taxon (Broom 19480).

Chlorotalpa spelea BroornAccording lo Meester (1971), the genus Chlorotoípa is

closely related to Amblysomus and contains ñve extant

species. C. spelea was described (Broom 1941c) on thebasis of a single well-preserved skull from the Sterkfon­tein Type Slte (probably Member 4); the skull appearedlonger and narrower than that ofknown living speeies. C.spelea has also been identified from Swartkrans Member1 (Davis, unpublished; H. B. S. Cooke 1963).

Farnily Soricidae: ShrewsAccording lo Meesler (1963), six genera of living Afri­

can shrews are regarded as valid; these are CrociduraWagler, Suncus Hemprich and Ehrenberg, MyosorexGray, Syívisorex Thornas, Paracrocidura De Balzac, andScutísorex Thomas. Only the first three of these are rep­resented as fossils in the australopithecine-bearing caves,together with a species of Diplomesodon mentionedbelow.

The genera Crocidura and Suncus are thought to haveevolved directly from an archaic group, represented bySylvisorex, with prirnitive features in the skulJ and teeth.

The Fossil Animals 185

In addition, sorne membersofthe genus Suncus may haveevolved into Crocídura by the 105s of P", meaning thatCrocídura, as now defined, may have had a polyphyleticorigino

Myosorex, 00 the other hand, appears to represent anentirely different evolutionary Iine perhaps related to theSoricinae, or red-toothed shrews (Meester 1963, pp.11-12).

Genus Crocídura WaglerBroom (19480) described C. taungensís from

Hrdlicka's Cave at Taung, pointing to the resemblance

between this species and the extant C. bicolor. In thesame paper Broom remarked that another species ofCrocidura occurs at Sterkfonteín, and possibly stillanother at Kromdraai, though these had not been satis­factorily identiñed. In bis sludy of South African fossilshrews, Meester (1955) was not able lo substantiate thesuggestion that Crocídura was represented at eitherSterkfonteín or Kromdraai, though H. B. S. Cooke (1963)lists C. ef. bicolor from both sites. It is not clear on whatmaterial his listing is based.

Genus Suncus EhrenbergThree extant species are currently recognized in south­

em Africa: S. lixus (Thornas, 1898), S. varilla (Thornas,1895), and S. infinitesimus (Heller, 1912). The last two ofthese are represented as fossils in the Sterkfontein valleycaves.

Suncus varilla (Thornas, 1895)Meester and Meyer (1972) record this species from

Sterkfontein members 4 and 5, Swartkrans Member 1,Bol!'s Farro, and Witkrans near Buxton.

Suncus i'lfinitesimus (Heller, 1912)The same authors have identified this species from

Sterkfontein members 4 and 5, Gladysvale, andMakapansgat.

Genus Myosorex GrayUntil very recently only two species oflivingMyosorex

were known from southem Afríca: M. varius (Smuts) andM. cafer (Sundevall). A third species has now been dis­covered at Knysna, and íts description is awaited (Mees­ter and Dippenaar 1978).

Two attempts have been made to explain lhe distribu­tion of the original Myosorex species in tenns of pastc1imatic changes (Meester 1958; Brain and Meester 1964).Both species are endemic and are separated by a consid­erable distance from their east and central Mrican allies.They both prefer moist, dense1y vegetated and oftenmountainous habitats. It has been suggested that southemAfrica was colonized twice by Myosorex during periods

186 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

of high rainfall, with the migrations proceeding frorn thelnyanga area oí Rhodesia, along the eastem escarpmentto the Transvaal, and thence farther. During times of drierclimate, it is suggested that the Límpopo valley formed aneffective barrier. The first wave Ied to the establishmentoí M. varius, by way af its ancestral form M. robinsoni.The second wave resulted in colonization by M. cafer.

Myosorex rohinsoni Meester, 1955The type specirnen carne from Swartkrans Member 1

(Meester 1955), but material is also known fromSterkfontein (probably Member 4), Gladysvale, andMakapansgat.

Genus Dipíomesodon Brandt, 1853The genus is not represented in the contemporary Afri­

can shrew fauna, but a new species, D. fossorius, wasdescribed by Repenning (1%5) on the basis offossils fromthe Makapansgat Limeworks. The fossils were found in ablock of breccia prepared at the University of ColoradoMuseum and are thought to be most closely related to theliving D. pulchellum of south-central Russia.

Representatives oí this genus have not yet been foundarnong the Sterkfontein valley microvertebrate as­semblages, but their discovery might well be expected.

Farnily Erinaceidae: HedgehogsNo remains of hedgehogs are known from the cave sites

oí Sterkfontein, Swartkrans, or Kromdraai, but a singleskull from Bolt's Farro was described by Broom (1937b)as Atelerix major and subsequently figured by him(Broom 1948a, p. 9). The specimen was subsequentlytransferred to the taxon Erinaceus broomí by ButJer andGreenwood (1973).

Order Chiroptera

Oceasional remains oí bats oceur among the mieroverte­brate remains, but they may well have resulted from natu­ral deaths of cave residents rather than representing theprey of owls. No detailed deseriptions have been pub­lisbed, but the follcwing taxa have been recorded:Rhinolophus cf darlingi and R. cf augur, together withpostcraníal remains of Vespertilionidae from dump 8 atSterkfontein (T. N. Pocock 1%9); Myotis sp. from Bolt'sFarro (Broom 19480); Rhinolophis cf, capensis from Ro­dent Cave, Makapansgat (de Graaff 19600); R. cf. geof­froyi and Miniopterus cf schrelbersii from the Cave ofHearths, Makapansgat (de Graaff 1960a); Rhinolophuscf. capensis frorn Taung, Makapansgat, and Bolt's Farro(H. B. S. Cooke 1%3); and cf. Mvotis sp. from Bolt'sFarro (H. B. S. Cooke 1963).

Order Rodentia

Suborder HystricomorphaFarnily Bathyergidae. Mole rats

The mole rats are characterized by very Iarge incisors,minute eyes and ears, an insignificant tail, and short Iegs.They are fossorial animals, feeding on roots and bulbsthat they find by burrowing tunnels beneath the surface,throwing up heaps of soil at intervals.

Accordíng to de Graaff (1971), tI1e following five generaof extant mole rats are considered valid: HeterocephalusRüppell, Heliophobius Peters, Bathyergus Illiger, Cryp­tomys Gray, and Georychus Illiger. The number of cheek

teeth is variable, from 4/4 in the southern African generaCryptomys, Bathyergus, and Georychus to 3/3 or 2/2 inHeterocephalus and 6/6 in Helíopñobíus,

Representatives of two extinct bathyergid genera havebeen reported from southem Africa: Bathyergoides neo­tertíarius Strorner, from a Miocene borizon in South­West Africa, and three species of Gypsorhychus­G. dani and G. minor from Taung (Broom 1934, 1948b)and G. makapani from Makapansgat (Broom 1948b).

AlI the bathyergíd remains from the Sterkfontein valleycaves have been referred to the genus Cryptomys.

Geñus Cryptomys GrayOn the basis of sorne cranial material from Skurveberg,

Broom (l937c) described a new specíes, C. robensi, that

differed from living specíes in the shape of the posteriorpar! ofthe mandible. He observed a distinct ridge runningforward on the outer side of the mandíble from the con­dvle.

C. robertsi has been recorded from Sterkfontein, prob­ably Member 4 (de Graaff 1960a) and from Sterkfontein,Swartkrans, Kromdraai A, and Bolt's Farro by H. B. S.Cooke (1%3). Cryptomys sp. is recorded from KromdraaiB (Davis 1959). T. N. Pocock (1%9) lists C. cf.holosericeus frcm dump 8 at Sterkfontein, but de Graaff(1971) regards the species holosericeus as a synonym ofhottentotus, the common living form in the area today.

Suborder MyomorphaFamily MuridaeSubfamily Murinae

The subfamily contains a large number of genera, andthe taxonomy of many is confused. The following taxahave been identiñed by various authors from theSterkfontein valley caves under consideration:

From Sterkfontein, probably Member 4, by de Graaff(l96Oa): Dasymys cr. incomptus, ?Arvicanthus sp.,Pelomys fallar, Rhabdomys cf. pumilio, Aethomys cf.namaquensis, Mastomys natalensis, and Leggadaminutoídes,

From Sterkfontein, dump 8, by T. N. Pocock (1%9):Dasymys sp., Rhabdomys mínor, Aethomys cf. chryso­phi/us, Praomys (Mastomys) cf. natalensis, Praomys sp.,Mus. ef. musculus.

From Swartkrans Member 1, by Davis (1955): Mus cf.minutoídes, Mus cf. triton, Dasymys ?bolti, and Lernnis­comys sp.

Frorn Kromdraai B. by Davis «(1959): Rattus sp., Mussp., Rhabdomys sp., and Dasymys sp.

Davis (1962) has expressed the apioion that Dasyrnysholti, a species named by Broom bu! not formally de­scribed, was directly ancestral to D. incomtus, the extantswamp rat.

Subfamily OtornyinaeRats of this fami1y are characterized by their laminate

molars, the first upper molar with three laminae, the sec­ond with two, and the third with a varying number fromthree to nine. In the lower jaw the first molar has fromfOUT lo seven lamellae, while the second and third lowermolars lypically have only two lamellae each (Roberts1951, p. 418).

Only two extant genera are recognized by Misonne(1971): Parofomys Thomas and Otomys Cuvier. Both arerestricted to Afriea south of the Sabara; the former con­tains the Karroo rats, characterized by inflated buHae andother adaptations to arid environments; the latter con­tains the vlei rats, which are frequently found in marshyhabitals. Individuals of both genera are diurnal but reg­ularly fall prey to owls, so that their remains are com­mon in modero and fossil microfaunal accumulations.Davís (1955) is inclined to regard Myotomys as a subgenusofParotomys, whereas T. N. Pocock (1976)argues that itshould be retained as a valíd genus.

In the past, many mammalogists have been inclined toc1assify the Otomyinae as a subgenus of the Muridae, butthe subfarnily was transferred lo the Cricetidae by Mis­onne (1969), and ihis procedure has been followed in theSmithsonian Institution' s identificatian manual of Africanmammals (Misonne 1971).

Recentiy T. N. Pocock (1976) has descrihed a remark­able rodent from the Pliocene deposíts of E Quarry,Langebaanweg, which appears lo form a link in toothstructure between the Otomyinae and the Murinae. Hehas named it Euryotomys pelomyoldes and described it asa rat or mouse with a typically murine cusp pattern butdiffering from all other known murines in possessing anenlarged and complex thírd upper molar. On the basis ofihis link fossil, Pocock speculated that the otornyineshave evolved from the Murinae wíthin the lasl 4 lo 5milJion years, the ancestral murine being of the Pelomystype,

Two fossil species of the Otomyinae have been de­scribed: Pa/aeotomys gracilis by Broom (1937c) fromSkurveherg, and Prototomys compbelli Broom 1946fromTaung. The latter is known by the type specimen only,but the fonner represents one of the commonest speciesin the microfaunal concentrations from the Transvaalcaves. In his description of tbe type, Broom expressed

The Fossil Animals 187

the oprmcn that P. graciíis was probably near to thecommon ancestor of Otomvs and Mvotomys.

Davis (1962) has speculated that Otomvs (Palaeotomys}graciJislO. (Myotomys) sloggetti form a "chronospe­des" spread over the Pleistocene. o. sloggetti does notoccur in the Sterkfontein area today-two sibling species,O. írroratus and O. angoniensis, take its place. Davissuggests that gracííís may be cIoser to the latter speciesthan to the forrner.

P. gracilis has been recorded from Sterkfontein (prob­ably Member 4) by de Graaff (I960a); from Sterkfontelndump 8 by T. N. Pocock (1969); from SwartkransMember I by Davis (1955); and from Kromdraa¡ B byDavis (1959).

Family CricetidaeMernbers of thís large family of rodents can he sepa­

rated from members of the farnily Muridae only by thestructure of the teeth. The Muridae are distinguished bythe presence of an anterior inner cusp 00 the ñrst Iamellaof M'. In Cricetidae Ibis cusp is absent (Coetzee 1972).

Representatives of the following three subfamilies areto be found among the microfauna from the Sterkfon­tein valley caves: Cricetinae, Gerbillinae, and Den­dromurinae.

Subfamily CricetinaeA single extant genus and species, Mystromys

albícaudatus (Smith), oceurs in África, though the sub-

family is well represented in the northem hemisphere.The genus Mystromys Wagner is oí panicular interest inthe microfauna from the Sterkfontein valley caves, sinceindividuals belonging lo It lend lo he extrernely abundanlas fossils, outnumbering those of murine affinities that arestrongly dominant loday.

M. oíbicaudatus, the white-taíled rat or Souih Africanhamster, is a savanna species of highveld and montanegrasslands, not known north of latitude 25°S bUl widelydístríbuted in the Transvaai, Orange Free State, Lesotho,and the eastern Cape. with a relíct population in thesouthwestern Cape (Davis 1974).

Three extinct species have been named. M. antiquuscame from Taung; íts teeth were figured by Broom(I948a), though a descriplion was never fumished. M.hausleitneri was descrihed by Broom (I937c) as M. haus­lichtnerí, though the spelling was subsequently corrected(Broom 1945a). The type specimens carne from Skurve­berg and differed in skull structure from M. albicauda­tus, particularly in ihe shorter snout and broader pal­ale. Broom (I948a) also descrihed a new subspecies M.hausletmeri barlowi from a fissure filling at Sterkfontein,apparently younger than the main deposít. Davis (1955)has been inclined lo synonymize M. antiquus and M.hausíeímerí barlowi with M. hausleitnerí. From the

Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation188

Makapansgat Limeworks, Lavocat (1956) described anew small form as M. darti.

M. hausleitneri has been recorded from Sterkfontein(probably Member 4) by de Graaff (1%Oa) and fromSterkfontein dump 8 by T. N. Pocock (1969); fromSwartkrans Member 1 by Davis (1955), and from Krom­draai B (Davis 1959).

In his recent review of Afriean fossil rodents, Lavocat(1978) remarks on the abundance of Mystromys in theolder southern African cave deposits. He believes that itwas in the early Pleistocene that the prominence of Mys­tromys declined in favor of the murines.

Subfamily GerbillinaeThe gerbiJs forro a group of nocturnal rodents typically

adapted to arid or semiarid open environments. Theyhave well-developed hind limbs, large ears and eyes, andextrernely well-developed bullae; their molars arecharacterized by transverse oval sections of enamel, eachsurrounding a cement center not in rows of cusps as in theMurinae, nor in layers of transverse enamel as in theOtomyinae. Three genera of living gerbils are currentlyrecognized in southern África (Petter 197Ih): TateraLataste, Desmodtilus Thomas and Schwann, and Gerhil­lurus Shortridge. The first two of these have fossíl repre­sentations in the fauna of the Sterkfontein valley caves.

Genus Tatera LatasteAccording lo Davis (1971). all species ofTalera fall into

one of two groups: the afra group or the robusta group. T.

I .r...~lrI I Y

afra (Gray) is confined to sandy areas of the southwestemCape, while T. robusta (Cretzschmar), Ihe fringe-tailedgerbil, extends from west Africa across tbe sub-Sabaran

.zone to Ethiopia, Somalía, and Tanzania. IncIuded in theafra group are the two southern Afrícan species T.brantsii (A. Smith), the higbveld gerbil, and T. inclusaTbomas and Wroughlon, Ihe Gorongosa gerbil. The onlysouthem African representative of the robusta group is T.Ieucogaster (Peters), Ihe bushveld gerbil.

T.ef. brantsii has been recorded from Sterkfontein(presumably Member 4) by de Greaff (196Oa) and Taterasp. from Sterkfontein dump 8 (T. N. Pocock 1%9), andKromdraai B (Davis (1959). In an unpublished manu­script, Davis (1955) has proposed a new species fromSwartkrans Member 1: T. robínsoní, which, when com­pared with the two living species in the Sterkfontein val­ley today, comes closer to T. brantsii than to T.leucogaster, although it is appreciably smaller.

Genus Desmodíllus Thomas and SchwannA single species is currently recognized: D. auricularís

(A. Srnith), the Namaqua gerbil, characlerized by error­mous development of the tympanic bullae and distributedwidely through the drier western regions of southern Af­rica.

Davis (1955) records D. auricularis from Taung andKromdraai B.

Subfamily DendromurinaeThe dendromurines form a group of small mice

characterized by grooved incisors and well-developed"masseter knobs" OD their maxillae (see Coetzee 1972).Living members in southern Africa are placed in threegenera, Dendromys, Steatomys, and Malacothrix, eachof which has representatives among the fossils from tbeSterkfonteín vatley caves.

Genus Dendromys A. SmithThe cIimbing mice are slender animals with prehensile

tails and fairly large, rounded ears. They spend much of

,,r"'.11 '

their time above the ground, either among the tops oflonggrasses, where they feed 00 seeds, or in shrubs or trees.Tbeir nests are hollow balls of shredded grass wilb holesal each end (Roberts 1951).

Misonne (1971) recognizes five species of living Den­dromys, bul only one fossil form has been named fromsouthem Africa: A. antiquus, figured and mentioned byBroom and Schepers (1946) bul never formally deseribed.The specimen came from Taung, bUI Davis (1955) hassuggested that the taxon may be present in SwartkransMember l. He also records Dendromys sp. from Krom­draai B (Davis 1959), while de Graaff lists C. cf meso­melas from Sterkfontein (probably Member 4) and T. N.Pocock (1%9) has identified D. cf. melanotis amongmicrovertebrates from Sterkfontein dump 8.

Genus Steatomys Peters. Fat miceThe following Ihree species are regarded as valid in

southern Africa: S. minutus Thomas and Wroughton, S.pratensis Peters, and S. krebsii Peters (Meester, Davis,and Coetzee 1964). The common name "fat mouse"comes from the fact that these rodenls tend lo 1ay downappreciable layers of fat and then to estivate for part ofthe year in their grass-lined underground chambers. The

The Fossil Animals 189

preferred habitat appears to be alluvial ftoodplains orsandy areas clase to water.

Steatotnys sp. has beco recorded from Krorndraai B(Davis 1959) and from Sterkfontein dump 8 (T. N. Pocock1969), while S. cf. pratensis is known from Swartkransmember I (Davis 1955).

Oenus Malacothrix Wagner. Large-eared miceOnly one specíes, M. typiea (Smith), is thought lo be

valid by Misonne (1971), with a wide distribution in thecentral and westem arcas of southem Africa. In Bo­tswana, Smithers (1971a) found the preferred habitat to behard ground with short grass around the fringes of pans.Deep inclined burrows are constructed, and the micehave beco found to wander far al night in search of food.

De GraaIf (l96Oa) has proposed a new fossil forro fromMakapansgat, which he designated ?Malacothrix maka­pani on the basis of an isolated left MI that appearedappreciably smaller than that of M. typica.

M. cf typica has beco recorded from SwartkransMember 1 (Davis 1955), and from Kromdraai, presumablyA (H. B. S. Cooke 1963);Malaeothrix sp. was listed fromSterkfontein dump 8 (T. N. Pocock 1969).

Family Muscardinidae. DormiceDormice, widely distributed through Africa, Asia, and

Europe were characterized by bushy tails and soft furoThey are nocturnal omnivores, taking large numbers ofínsects and other small prey. The occlusal surfaces oftheir incisors are V-shaped when viewed from the front;their cheek teeth are very small compared with those ofmost other rodents of equivalent size.

Misonne (1971) recognizes only two genera in the fam­ily: E/iomys Wagner and Graphiurus Smuts, the formerrestricted to North Africa.

The only published record of a fossil dormouse forSterkfontein valley caves is that ofT. N. Pocock (1969),who Iisted Graphiurus sp. from dump 8 al Sterkfontein.

Class Amphibia

Remains of frogs occasionally occur among the mi­crovertebrate remains, but these have not been studied indetail. Bones from two species of frogs have been re­corded from durnp 8 at Sterkfontein, but no further par­ticulars are available (T. N. Pocock 1969).

C/ass Reptilia

Occasional rernains oí lizards are present in the collec­tions, but these have yet to be studied in detail.

Class Aves

The microvertebrate colJection from Kromdraai B con­tains fairLy abundant rernains of birds, which have beenstudied by T. N. Poeoek (1970). Largely on the basis ofpostcranial elements, he identified the foUowing forms:Tvto sp. (barn owl), Francolinus sp. (francolin}, Coturnix(quail), Turnix {button quail), Crex sp. (crake), Apus sp.(swift), Scolopacidae (two species of sandpiper and ruff),Agapornis spp. (larger and smaller species of lovebírd),Stumidae (starlings), cf. Anthus sp. (pipit}, cf. Estrildasp. (waxbilI), cf. Cisticota sp. (grass warbler), ef. ?Passersp. (sparrow), cf. Ploceus sp. (weaver), and Hirundinidae(swaUows of three different sizes).

Pocock interpreted the bird remains as follows (1970, p.4):

Thus thanks to the Pleistocene barn owl sallying forthto prey on the microfauna and retuming to deposit itspellets below the cave-ledge roost, we can form a fairpicture ofthe smaller avifauna at the time ofthe Krom­draai ape-rnan, "Paranthropus" robustus. Quail, but­ton quail, pipits and perhaps also cisticolas were quitecommon, confirming the grassveld character of theenvironment. Flocks of starlings were preyed upon asthey roosted in the streamside or marsh vegetation.Swallows may have fallen victim in the same way, es­pecially if they were non-breeding mígrants, or theymay bave built their mud nests, perhaps colonially, inthe very cave frequented by the owl. Swifts doubt­less also used the cave, nesting in crevices, or perhapscornmandeering sorne of the swallow nests. Non­breeding migrants probably did occur, crakes,sandpipers and ruffs from the steppes and tundra of aglaciated Europe.

An interesting conclusión to emerge from the Krom­draai analysis is that, though small members of the parrotfamily were apparently cornmon, seed-eating passerinessuch as sparrows, weavers, and queleas were not, Pocockspeculated that, during Kromdraai B accumulation times,the niche today filled by abundant seed-eating passerineswas occupied by lovebirds and similar psittacine birds.Such a concept supports the contention that weavers ofthe family Ploceidae are a modero group that may haveentered Africa, perhaps from Asia, in relatively recenttimes and have undergone extensive radiation here. Indoing so they carne into competition with the residentparrots, displaced thern, and reduced their populationnumbers.

Bird remains from the other cave sites await study.They are likely to yield valuable inforroation not onlyebout avian evolution, but also about avifaunal changesduring the Pleistocene,

10 Sterkfontein

A Brief History of Aclivily

During the 18905, prospectors and mmmg companiesshowed interest in the area north of Krugersdorp, ex­tending to the Bloubank River valley. This interest cen­tered particularly around the outcrops oí Black Reefquartzite which, in the vicinity of Kromdraai, pro ved tobe gold-bearing, though ultimately not profitable. Pros­pectors consequently tumed their attention to depositsof pure white travertine that occurred in ancient cavesdissolved in the dolomitic country rock. Such caves fre­quently contained bone-bearing sedimenta, and it wasalmost certainly through Jíme mining that the first fossilscame lo light. On I February 1895a well-known geologist,David Drapee, who was to become the ñrst secretary ofthe Geological Society of South Africa later that year, senta sample of fossilíferous breccia to the British Museum,accompanied by the following information:

Mass of rock containing a number of fragments ofbones, This is from a cave on the farro Kromdraaisituated about 16 miles west of Johannesburg. There isa bed of stalagmite with rnasses of rack in which thebone ís abundant. It is rea11y a cave deposit containingbone breccia or fragments of bone. 1 send you thespecimen for examination as 1 should like to knowwhether the rock is similar to that found in Europeancaves. I am informed that a portion of a human skulland the claw of a líon were found in the cave but Ihave not seco the specimens nor have 1 yet inspectedthe cave though shortly hope lo do so.

Three months later on 8 April 1895, Draper contributedlo the discussion after the inaugural address by the firslpresídent of the newly formed Geological Society ofScuth Africa, H. Exton, He stated that "he had a shorttime ago visited the Krorndraai caves, or, rather, the for­matioos that once had been caves, and he found themmost interesting from a geologícal point of view. Therewas much to discover there. He, therefore, suggested thatthey make a beginning by making up a party lo go andexplore this interesting geological ground" (Draper 1896,p. 11).

In my opinión there is a strong possibility that the fos­siliferous cave deposit from which Draper's sample carne,and which he and other geologists visited in 1895, was lhesuñace outcrop on the Sterkfontein hin. This was thespol being worked for lime al lbe lime by G. Martinaglia,a pioneer prospector of the Witwatersrand goldfield. C.

190

van Riel Lowe (1947) has documented lbis early phase ofaetivity at Sterkfontein and quoted extracts from a letterwritten to him by G. Martinaglía, son of the originalminero It runs as foUows:

My father was prospecting for lime and one day whenI took him sorne refreshments to the quarry where hewas working, he pointed eothusiasticaHy to a darkhole on the side of the quarry and lold me in Afrikaansthat he had just blasted open a "wondergat." Thisrnade a great ímpression 00 my boyish mind and I stillremember well the words spoken on that occasion byrny father. There are only a few people alive to-daywho saw these caves in the days of their originalsplendour before they were destroyed later by com­mercial exploitation.

The date of lbis first discovery of the undergroundcaves has been established as shortly after the JamesonRaid, which ended on 2 January 1896 (van Riel Lowe1947; Malan 1959), but it is obvious that the fossiliferousbreccia outcropping on the surface was known at least ayear earlier than this. The question whether the surfaceoutcrop was, in faet, linked to the Sterkfontein caves hasbeen considered by B. D. Malan as follows (1960, p. 110):

In considering the date of discovery of the Sterkfon­tein Caves, the question arase whether Draper'smention of the Kromdraai Caves on 8lb April, 1895,referred lo the nearby Sterkfonlein Caves, it beíngquite possible that lbe name Kromdraal might havedenoted a large area. On the strength of recollectíons ofpioneers that the Sterkfontein Caves were discoveredafier the Jameson Raid [afier 2 January 1896]andDraper's reference lo "the foundering that had oncebeen caves," it was concluded that Draper's early ref­erence could not have been to the Sterkfontein Caves.This conclusion is now specifically confirmed inDraper's letter lo the British Museum where he clearlystates that the material carne from thefarm Krom­draai, whereas the Sterkfontein caves lie 00 the adja­cenl farm Swartkrans. There can thus now be hardlyany doubt that the Sterkfontein Caves were oot knownwhen Draper reponed on breccias from Kromdraaiearly in 1895.

Malan is certainly right that the underground Slerkfonteincaves were not known when Draper reported 00 the fos­siliferous suñace outcrops, bUl 1 believe the latter areoevertheless more IikeJy to have beeo those 00 the

Ir

Sterkfontein hill, being worked by Martinaglia, than anyother, such as those that exist al Kromdraai.

This view has, 1 find, already beco expressed byBroom, who stated: "what are today cal1ed theSterkfontein Caves had beco known for many years.Tbey are really 00 pan of the original farro Zwartkrans ,and adjoining Krorndraai they have beco referred to asthe Krorndraai Caves; but it seems better to keep themore modero term Sterkfontein Caves for those wherethe Sterkfontein skull was found, and to reserve the DameKromdraaí for the fann where the Krorndraai skuli wasdiscovered" (Broom and Schepers 1946, p. 46).

The Kromdraai B and A deposits both forrned ex­tremely insignificant surface outcrops befare Broom'spaleontological exploitation of them in 1938 and 1947 re­spectively, Draper'a description of "a hect of stalagmitewith masses oí rack in which bone is abundant" is to mymind likely to refer to one of the Sterkfontein hill expo­sures, perhaps that of the "West Pit" where microfaunalremains are specially abundant. Tbe microfaunai listcompiled by de Graaff(l96I) from Draper's sample in theBritísh Museum shows a suite of species that, regrettably,could well have been found in any of the Sterkfonteinvalley site units known at present. Nor are the macro­faunal remains any more diagnostico Oakley wrote oíthem (1960, p. 110), "A number of fossils have beenextracted from the breccia by the acetic acid technique inour laboratory, including remains oí baboon, a smalJ car­nívore, a large horse, porcupine, rodents, a small lizardand bírds."

An interesting conunent 00 the fauna of the "Krcm­draai caves" by M. E. Frames (1898) was elicited by apaper presented to the Geological Society ofSouth Africaby August Prister (1898). Prister presented evidence thathe regarded as indicating that the Pretoria and Witwaters­rand areas had been subjected to glaciation in Quater­nary times. As was to be expected, such an aliegationevoked a good deal of comment, most of it unfavorable toPrister's theory. Frames's cornment (1898, pp. 94-95) ranas follows:

A review oCthe past and present ñora and fauna, as anargument bearing on the theory of a glacial period inquatemary times in South Africa, may not be withoutinterest. ...

To dea1 lboroughly with Ibis aspect oflbe question,it will be necessary to consider the animal remainsfound in the Kromdraai Caves in the Dolomite nearKrugersdorp.

Amongst these are those of the horse species, an­telopes, monkeys, porcupines, rats, bats, etc., and thepresence of the two first in the cave would lead one toinfer that they had been dragged lbere by beasts ofprey. Tbe caVe is now filled with bones, cave earthand staiagmites to a depth of about 15foot, and lbewhole mass is consolidated into hard rock. Tbe latterfeature oaturally points to the antiquity of the caveand its contents. The roof has, in sorne cases, becneroded, so that portions of lbe bone bed are to-dayexposed 00 the suñace. Therefore, assuming that theglacial period waS prior to lbe filling of lbe caves,surely the erosion which removed the roof would haveproved sufficient to obliterate aH traces of glaciationfrom the perishable sandstones and quartzites of lbeMagaliesberg?

Agaio, if the ice age occurred after the caves were

Sterkfontein 191

filled, then the same denudation argurnent wouldapply, with no reasonable explanation accounting forthe remains of living types in the caves. and the ab­senee of arctic types to fill the gap marking the in­tensely cold periodo We cannot believe in the extinc­tion of these animals and their suddeo appearanceagain after the glaciers had melted.

This was the first of a long series of subsequent at­tempts to interpret the fossil fauna of the Sterkfonteinvalley caves in climatic terroso

Whatever the precise locatíon of the original fossili­ferous breccia outcrop descríbed by the early geologists,there is no doubt about that of the Stekfontein under­ground caves, discovered first by Martinaglia. The quarryin which the "wondergat" first appeared was a short dis­tance southwest of what subsequently becarne known asthe Type Site excavatíon. The discovery of the specatularsubterranean caverns al Sterkfontein aroused interestboth locally and overseas. Malan (1959) quoted a reportthat appeared in the little-knownjournalEnglish Mechanicand World 01 Scíence of 27 August 1897:

Sorne wonderful caves were díscovered recently at aplace called Sterkfontein.... Limestone has beenquarried for sorne months in a small kopje, and afteran explosión after sorne blasting operations, a cavityof great depth was left ... a party descended and it wasfound that they had driven through one of the largestof a series of magnificent caves. The spectacle wasone of great beauty, the light carried by the explorersbeing refíected from thousands of stalactites.... Thecaverns, which have not yet been one-half explored,seem to run in tiers down to a depth of lSOfeet.

Republication of this report in the French joumal Cos-mos prompted a Marist brother in Johannesburg (whoseidentíty has not been established) to organize an outing tothe cave with seven of his colleagues and thereafter toprepare an artiele entitIed .. Les grottes de Sterkfontein"for publication in the same journal during 1898. The au­thor of' the article was greatly impressed by the beauty ofthe formations in the caverns and corridors and remarkedon the fossil jaw of a large antelope that was "probablythe prey of a camivore."

The remarkable beauty of the forrnations in theSterkfontein caves had already been perceived bygeologists in the TransvaaJ. In his presidential address tothe Geologicaí Society of South Africa on 22 February1898, H. Exton (1899) referred to the unusual branchingcrystals of aragonite found on the stalactites. He ex­pressed the opinion that, for such aragonite crystals lohave íonned, the caves must at one time have been filledwith hot water.

At a previous meeting of the Society, held on July 12th1897, Draper (1898, p. 63) stated lbat steps had been takento protect the caves for the benefit of the public: "withthis object in view he had approached the owners. andlbey had most willingly acceded to his request to place thecaVes under the protection of their manager, who wouJdhave instructions to prevent any wanton destruction oftheir interior.

Regrettably. such protection was not as long-Jived asrnight have been hoped. and irreparable damage wascaused lo the subterranean fonnations between 1918 and1920. At that time lbe caves belonged to E. P. Binet, whowas leasing 1hem to a Mr. Nolan. Upon expiration of the

192 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Inlerpretation

lease, the owner was not in favor of renewal, a decisionthat so annoyed Nolan that he deliberately dynamitedwhal he eould of the formalions.

Little further interest appears to have been shown inthe Sterkfontein caves tíll after the discovery ofthe Taungskull in 1924. At that time lime mining was being under­taken in the Sterkfontein area by the Glencaim LimeCornpany, the operation being managed by G. W. Bar­low. When the Taung discovery was announced in 1925,acoUection of Sterkfontein fossils was sent to R. A. Dart,who, however, did not investigate the matter further be­cause he thought the deposit was probably relatively re­eent (Dart and Craig 1959).

During 1931 and 1932 J. H. S. Gear, later to beeomedirector of the South African Institute for Medical Re­search, was describing fossil baboons fram Taung andspent a day in the underground caves at Sterkfontein insearch of further fossils. He gave particulars in a letter toP. V. Tobias, written in 1972 and quoled by Tobías(l973h). A ccllection of fossils was, in fact, made andtaken baek lo the Institute for Medical Research. Un­fortunately it has never been located. However Dr.Gear's discussions at the Anatomy Department Junchtable in those days may have led lo further Sterkfonteininvestigations. R. Trevor Jones collected and studieda series of fossil baboon skulls from Sterkfontein in1935, and the following year two other students of Dart's,G. W. H. Schepers and H. Le Riehe, obtained furtherbaboon material at Sterkfontein and showed it lo RobertBroom.

Broom hadjoined the staffofthe Transvaal Museum inAugust 1934 wilb the intention of furthering his paleon­tological studies, which he had previously undertaken inhis spare time. In the hope of finding remains of an adultAustralopithecus. Braom started work at old limestonequarries at Skurveberg, Hennops River, and Gladysvale.When Le Riche and Sehepers showed him fossil baboonskulls from Sterkfontein, Broom arranged to visit the sitethe following Sunday. He was met there by the quarrymanager, G. W. Barlow, who showed him various speci­mens on the table in the rondavel that served as atearoom. The specimens were sold as souvenirs to ínter­ested visitors. In the past, Barlow had worked al Taung,and when Broom asked him whether he had ever seenanything like the Taung child skull al Sterkfontein, hereplied thal he rather thought he hado He promised tokeep a sharp lookout.

On the occasíon of Broorn's third visit to Sterkfonteinon Monday, 17 August 1936, Barlow handed him an en­docraniaI cast and asked, "Is this what you're afier?"Broom replied "Yes, that's whal I'm after." Coneeming

. the specimen Broom wrote as follows (1950, pp. 44-45):

It was clearly the anterior two-thirds of the brain-castof an anthropoid ape or ape-rnan, and in perfect con­dition. 11 had been blasled out that moming. I huntedfor sorne hours for further remains, but withoul suc­cess. But I found the natural cast of the top of theskull in the side of the quarry, and had this cut out. Sowe returned lo Pretoria wilb the fine brain-cast andthe cast of the top of the skull. Next day I was backagain and I had with me Dr. Herbert Lang, Mr. Fitz­Simons of the Museum, Mr. White, my assistant, andthree Museum native boys. Mter a long hunl I dis­covered the base of the skuU00 which the brain-casehad rested, wilb all lbe blocks that had been altaehed

to ít. 1also found parts of the parietals and portions ofthe frontal. When, after sorne weeks ofwork, we hadthe remains of the skull pretty well eleared of matrix,it was found that we had practically all the base, thefront of the brow, and most of the bones surroundingthe eyes, and both upper jaws, which had been badlydisplaced; but the whole face had been considerablycrushed.... To have started lo look for an adult skullofAustralopirhecus, and to have found an adult of atleast an allied form in about three rnonths was a recordof which we felt there was no reason to be ashamed.And to have gone to Sterkfontein and found what wewanted within nine days was even better.

In fael a strangely prophetie remark, made by R. M.Cooper, owner of the caves al that time, had beenfulfilled. In a guidebook to places of interesl aroundJchannesburg, Cooper wrote in 1935, "Come toSterkfontein and find the missing link." Broom went andfound it.

Regular víslts to the Sterkfontein quarry providedBroom with furtheraustraJopithecine and faunal remains.Then in 1937 Barlow stopped work in the Type Site 10­cality and turned his attention to a spot lower down, closelo the subterranean cave's exit. Here further australo­pithecine remains carneto líght before a fall in the price oflime brought mining at Sterkfontein to a close shortlybefore the war in 1939.The site then lay dormanl lilll947.

After the war Gen. J. C. Smuts, prime minister of theUnion of South Africa al the time, asked Broom to con­tinue fieldwork in search of more "míssing links" andpromised financial support from the government. So on JApril1947 Broorn's field team, which had beenjoined byJ. T. Robinson, commenced work in the Type Site, closeto the spot where the first adult Australopithecus hadbeen found nine years before. Important fossils were dís­covered wilbin the ñrst week, and then on 18 April themost complete adult australopithecine skull, that of"Mrs. PIes," was blasted out. Broom described the in­cident thus (1950, p. 64):

just about two weeks frorn the time we started weblasted a large píece of what looked very unpromisingbreccia. It was only a yard below where the type skullhad been found. When the smoke of the blast blewaway, we found that a beautiful skull had been brokenin two. The outer part of the roek had the lop of theskull and all the lower half was exposed in the wall. Asthe top of the skull had been split off we could see intothe brain-cavity, which was lined with smalllimecrystals, I have seen many interesting sights in my longlife, bUIthis was the most thrilling in my experienee .

The Type Site excavations of Broom and Robinsoncarne lo an end in January 1949 when the last of theworkingparty was transferred to Swartkranson the otherside of the valley,

During 1956I was completing a study, "The TransvaalApe-Man-Bearing Cave Deposita," in whieh theSterkfontein site featured prominentiy (Braín 1958).Under the subheading "Cultural Malerial from Sterkfon­teín" I found I had nothing to write and thereupon re­solved to spend the next day, 7 May 1956, checking onceagain that no stone artifacts were to be found in the bree­cia. To rny considerable surprise, breccia exposed in ashallow prospectors' pit, the "West Pit" a few meterswest of the Type Site, contained unquestionable artifacts

fossils. Then in August 1956 the building burned down ina spcctacular manner. On that day 1 was working in theWest Pit arca wh"en. toward noon. a lhunderstorm ap­proached. 1 dl"Ove the Land Rover in under the eaves of¡he rondavcl and waited for the storm to pass, while theAfrican staff sat down to their midday meal in a l"Oomadjoiníng the rondavel. A bolt of lightning struck thelhatched roof of the rondavel, showering the Land Roverwith pieces of buming grass. In great haste 1 moved thevebicle and then entered the building to find that the men,including Daniel Mobelhe and Absalom Lobelo, who hadworked at Sterkfontein sínce 1936 and 1948, had beenknocked from their stools by the bolt and were Iying mo­tionless on the ftoor. Fortunate1y they slowly recoveredand stumbled from the smoke-filled room.

During 1958 the Stegmann family, owners of theSterkfontein fann, donated a 20-morgen area surroundíngthe caves to the Universíty of the Witwatersrand. It be­carne the Isaac Edwin Stegmann Nature Reserve. Theblackened remnants of the rondavel and adjoining roomswere demolished to make way for the new restaurant

Sterkfol1lein 195

building erected by lhe universily. The building now in­dudes the Roben Broom Museu m. opened on I Decem­ber 1966 by Broom's son, Norman Broom. Tourístfacilities to the underground caves hel\ e been taken careof by Basil Cooper, and electric Iighting and recordedtape commentaries have now been instal1ed underground.

The 30th of November 1966 saw [he one hundredthanniversary of the birth of Robert Broom. and as part ofthe celebrations to mark this event, plans were laid in July[966 by P. V. Tobias and A. R. Hughes for a new, long­term program of work al the site. There were fOllr originalobjeclíves (Tobías and Hughes 1969)~

1. A vegetatíon survey of the 20-morgen Sterkfonteinsite.

2. A conlour survey of the same area.3. An archaeological excavation of the overburden,

which would expose the underiying breccia anddolomitic country rock.

4. Systematic excavation of selected parts of the in situbreccia.

Since 1966, work has progressed steadily at Sterkfon-

Fig. 157. Phillip v. Tobias altempts to reposilion the cranium of "Mrs. Pies" in (he Slerkfonteín Type Sile. Augusl1967. Looking on are H. B. S. Cooke. R. A. Dart. and A. R. Hughes.

Fossíl Assemblages frorn the Sterkfontein Valley Caves: Analysis and Interpretation

Sorne Notes 00 fhe Site

As early as 1938 H. B. S. Cooke (938) provided a planand section through the breccia-fiUed cavem at Sterkfon­tein. He visualized the cave filling as consisting predomi­nantly oí stony, bone-poor breccia, but with a (ayer oífiner-grained, richly fossiliferous breccía aboye it. Bed­ding in the entire mass was ínclined, sloping down fromsouthwest to northeast, which suggested to Cooke thatthe cavem had probably been not an occupation site, butrather a receptacle into which the sediment had slumpedfrom a rock shelter at a higher level.

Cooke realized, too, that the underground cave systemat Sterkfontein took the form of "two lines of galleriesdeve~oped along two main fissure directions running ap­pmximately north from their entrances and followingmore or less the general dip of the dolomite" (H. B. S.Cooke, 1938, p. 204).

In the interim since 1938 various workers have Con­tributed to an understanding of the form of the fossil andunderground cavems, as well as to the nature of theirdeposits,

As a result oí his "Extension Site" excavations of1957-58, Robinson (1962) concluded that the breccia inthe fossíl cavem could not be regarded as a single con­formable deposit as previous writers had suggested (e.g.,H. B. S. Cooke 1938; Brain 1958), but that three superim­posed sediments were involved, separated by significantunconformities. He tenned the three deposits the LowerMiddle, and Upper breccias (see figs. 158 and 159) amisuggested that the last two deposits filled spaces beneaththe original cave roof created by periodic slumpmg of theLower Breccia into progressively enlarging subterraneanchambers.

Subdivision of the Sterkfontein deposit on stratigraphíccríteria has been taken furtber by Partridge (1975), whoformalIy designated six members wíthin the SterkfonteinFormation. He points out that the form of both the fossiland underground caves has been strongly controlled bythree sets of fractures, the mosl prominent being paralleland at right angles to the stríke of the dolomitic countryrock, wilb a subordinate set aligned approximatelyno~heast-soulbweslat an angle of about 40 degrees lo thestnke. In the viciníty ofthe Type Site, fractures belongingto the strike and oblique sets íntersect, enclosing aparallelogram-shaped area bounding most of the fossilcavem. According 10 Partridge, the east-west fractureplanes run almost vertically through the dolomite anddetermine the walls of the fossíl cavern, while the irregn­lar ñoor of the cavem may be seen in the "DaylightCave" (now renamed the Silberberg Grotto by P. V.Tobías), almost 20 m below the present surface of thebreccia.

It is in faet !he Daylight Cave, situated below the~uthe~ walI of the fossil cave, that has provided Par­tndge wllb the basís for bis stratigrapbic column. He is oflb~ opinion lbat the breccia previously designated Lower,Mlddl~, and Upper breccias above it by Robinson (1962),resulUng in a far longer profile !han had ever been antici­pated.

Partridge has defined and described his six members asfollows:

Member J

Visible above the dolomite flnor in lbe Daylight Cave, ilconslsts of 0.5-2.0 ID of white. crystalline to cryptocrys­talline calcite. frequently contaminated with fine sedi-

196

tein in accordanee with the planned objectives. The veg­etation survey was undertaken by A. O. D. Mogg and hasresulted in a booklet on the indigenous plants of the area(Mogg 1975); an extrernely detailed survey ofthe area hasbeen made by 1. B. Watt, and a grid network has beenerected; extensive excavatíon of overburden has revealedthat the breccía exposure is a good deat more extensivethan had been anticipated, particularly lo the west of thewest Pit, while limited excavation of in situ breccia hasbeen undertaken, Renewed geologícal ínvestigatíons byT. C. Partridge are discussed in the next section. The areaof paleontological interest 00 the crown of the Sterkfon­tein hill has been surrounded by a securíty fence, and alarge work building has been erected.

The first hominid specimen found during the newfieldwork carne to light on lO June 1968. It consisted ofthe crown of an isolated left upper molar and was fol­lowed by other molars on 25 March and 3 July 1969. Morerecently a spectacular gracile australopithecine cranium,StW 12113/17, was fcund in situ in the Type Site quarry,as well as an adult mandible, SIW 14, and adolescenl rightrnaxilla, SIW 18a-e.

But by far the most important discovery to be made atSter~ontein in recent years occurred on 9 August 1976,precisely forty years after Broom paid his first visir to thesite. It took the form of a remarkably complete hominid

Ro.. o Stll JJ

craníum found by A. R. Hughes in a decalcified pocket ofderoofed breccia about 2 m from lbe spot where stonetools were originalIy found in the Sterkfontein breccia.Fortunately, parts of the skuU were still in place, wea­thering from the wall of the solution pockel and indicatingthat the fossil was preserved in Member 5 breccia. Thes!?"cimen, SIW 53, consists of most oflbe faee and palate,rune upper teeth, a large part of the vaull, and sorne oflbebraincase and mandible. Its discovery was reported in theSouth African louma! of Scíence of Augusl 1976, and adescription of it appeared in Nature of 27 January 1977(Hughes and Tobias 1977). According to Tobias's as­sessment, the new skuU shows (eatures that distinguish itclearly from Australopithecus africanus, whose remainsare found in such abundance in lbe underlying Member 4breccia. These features have aflinities wilb lbe earlyspecies ofHomo, perhaps H. habilis, described originalIyfrom Olduvai Gorge. One is lempled lo think lbe skuUmay have belonged to one of the makers of the stoneartifacts preserved alongside it in Member 5 of theSterkContein Formation.

ment. It has been divided into a lower Bed A, generallyuncontaminaled and sterile of bone, and Bed B, darker incolor and including angular dolomile and chert pieces aswell as some bones.

Member 2

0.5-2.0 m of reddish brown sandy sill resting conformablyon Member 1 and variably calcified. Bone is abundant,particularly in the lower levels, and it is from here that the

Sterkfontein 197

type specimen of the long-Iegged hunting hyena,Euryboas silberbergi (Broom), orjginaIly came (see chapo9). Further carnivore and primate fossils are visible whereMember 2 breccia is exposed in the wall of the DaylightCave.

Member3

Material assigned lo this member forrns most of the verti­cal profile visible in the walls of the Daylight Cave. Par-

WEST PIT ORIGINAL OVERBURDEN SURFACEI_-~I---------

T'tPE SITE

UPPER BRECCIA

MIDOLE BRECCIA

LOWER BRECCIA

@D"........... ....

DOLOMITE

OVERBURDEN

EXCAVATED DEPTH

o -4 8

FEET

12

Fig. 158. Slratigraphy of the upper pan of the Sterkfontein deposit, as visualized by Robinson (1962). Currentterminology for Lower, Middle, and Upper Breccía is Member 4, Member 5, and Member 6.

Fig. 159. T. C. Panridge demonslrates the unCOlÚormity belween members 4 and 5 in the west wall of lhe SteJicfon­tein type site, May 1977.

198 Fossíl Assemblages from the Sterkfontein Valley Caves: Analysis and Imerpretation

tridge considers it (O consist of 10 m of pale reddishbrown. well-calcíñed sediment, ending perhaps with alayer of steeply dipping travertine that has been observedin the "Middle Pir" of the Type Site and that appears tosepárate members 3 and 4.

was once again created which the sediment representedby Member 5 penetrated.

Partridge considere that this sediment is now repre­sentcd by 1-5 m of brownish red, well-calcífied silty sandcontaining chert and dolcmite inclusíons. Bone and stoneartifacts are wel1 represented.

Member4

Breccia assigned to this member occurs particularlyin theType Site excavation and the Lower Site, close to the exitof the underground cave. It is particularly important fromthe poínt of view of the bone accumulaticn discussed inthis chapter. Partridge has subdivided Member 4 brecciainto four beds:

Bed A. 2-3 m of brownish red, ealeified sandy silt thatoverlíes the inclined travertine at the top of Member 3.Sorne bone is present.

Red B. This is essentia1ly a debris cone of fallen roofblocks that had its apex near the eastem end of the south­em walls of the Type Site excavation. The blocks of rocktypically form a mosaie cemented with calcite and in­eluding pockets of calcified reddish sediment that prob­ably filtered down between the rocks after their collapsefrom the cave roof. The current ñoor of the Type Siteexcavation is largely formed of this material; it is not richin bone, and its sterility loo to the tennination of theSterkfontein excavation in 1949.

Bed C. A deposit of dark reddish brown, well-ealeifiedsílty sand resting unconfonnably on the eroded surface ofBed B. Its sediment appears to have been derived fromthe surface via a cave entrance considerably enlarged bythe roof falls that contributed lo Bed B. Tcgether withBed D, this unit constitutes Robinson's Lower Brecciaand was the source of many of the australopithecine andother fossils recovered during the Broom era.

Member6

A rather insignificant deposit, representing the UpperBreeeia of Robinson (1962). It overlies Member 5 bree­cía unconformably in the West Pit area and consists ofdarker, poorly calcified breccia füling spaces beneath thedolomite roof probably created by minor subsidence ofthe breeeia rnass. It is probably eonsiderably youngerthan Member 5 and contains sorne bone, as reflected inthe analysis presented in this chapter.

In addition to sedimentological analyses of samplesfrom the breecia profile, Partridge and Talma (n.d.)undertook stable isotope measurements of carbon andoxygen in carbonates from the deposits. With certain pre­cautions, these measurements may be used as a guide tothe ternperature of the groundwater at the time of carbo­nate depositíon. They concluded that deposition ternper­atures remained constant within 2°C during most of thesediment accumulation period, ahhough an increase of3°C near the base of Member 4 and 4°C in the middle ofMember5 is suggested; the evidence for these increases issomewhat suspect owing to the high Ile concentration inthe sarnples.

Despite many and varied attempts at absolute dating ofthe Sterkfontein deposit and íts fossils, no conclusive re­sults have been obtained. On the basis of cyclic nickpointmigration in the Bloubank valley, and valley flank reees­sion, Partridge (1973) anived al a date of 3.26 millionyears for the ñrst opening of the Sterkfontein fossil cave.

reconstructed secUon

(o)

plan

9 , 2pmeu..

11ICave tllling

• Dolomlte

secUon

rtg. 160. Outline plan(a) and section (b) of the Sterkfontein fossil cave.together withthe suggested fono of a reconslroctedsection al the timeofaccumulation of Member 5 sedimento

Member5

This is the horizon in whieh the Sterkfontein stone ar­tifacts are concentrated; it represenls the Middle Breccia,in Robinson's 1962 terminology. According lo Partrldge'sdefinition it is a mass of calcified sediment that rests un­conformably upon a weathered surface of the underlyingMember 4. Faunal evidenee (Vrba 1975, 19700) suggeststhat the lime interval between the end oC the accumulatíonphase of Member 4 and the beginning of that of Member 5may be considerable-c-perhaps as mueh as half a millionar a minion years. Partridge agrees with Rabinson in aS4

suming that the available space in the fossil cavem wasfilled wilh Member 4 sedimenl; it was only lhrough sub­sidence of this filling into a lower-Iying vOld that space

Bed E. An insignilieanl deposit of laminated ealcilelenses, up lo 75 cm in thickness that eoneluded theMember 4 deposítional phase, frequentIy jusI below thedolomite roof.

Bed D. This unit is separated from Bed e by a markederosional unconformity and consists of 0.5-2 m of palebrownish red, heavily ealeilied silty sand. It is repre­sented in both the Type Site and the Lower Sites and wasthe source oC STS 5, the well-known eranium of "Mrs.PIes," as well as the more recentIy discovered skulldesignated StW 12113/17.

Although this is an interesting attempt, nct everyonewould agree that this kind of geomorphological methodcan be used in precise age estimation.

Tuming to the underground cave at Sterkfontein: a re­markably detailed sludy has been undertaken by Wilkin­son (1973). He has shown that the cave system consists ofthree large chambers: Tourist Cave, Lincoln's Cave. andFault Cave (see fig. 161), linked by a series of fracture­determined passages. The system underlies the mucholder fossil cave system that has already been extensivelyderoofed by prolonged erosion ofthe Sterkfontein hilltop,

The Sterkfontein Bone Accumulations

The analyses presented here were done on collections inthe Transvaal Museum that resulted from two periods ofexcavation:

a. work in the Type Site between 1936 and 1948; allbones are regarded as coming from Member 4;

b. excavations carried out by J. T. Robinson in the"extension site" during 1957 and 1958. The sampleis readily divisible into two components, on thebasis of the enclosing matrix, with origin in eitherMember 5 or Member 6.

The bone accumulations are further divided into macro­and microvertebrate components, the former having pre­sumably entered the cave in a variety of ways, while thelatter was almost certainly derived from pellets re­gurgitated by owls that roosted in the cave.

Remains from Member4: Macrovertebrate Component

The collection consists of 1.895 fossils from a minimum of348 individuals, as detailed in table 76. Particulars of 384bones of bovid origino assigned to size classes but notspecifically identified, are given in table 77. The variousidentiñed animals are depicted in figure 162.

1 shall now presenl data on the fossil material by whichtbe varíous animal taxa are represented.

Class MarnmaliaOrder PrimatesFarnily HominidaeAustralopithecus ofricanus

Many scientists have made their own assessments ofthe Sterkfontein hominid collection, from the poínt of

[3FoUU ca"

o 30~

metr••

Fig. 161. Plan of the Stertáontein underground cave based on a surveyconducted by JustinWilkinson (n.d.).

Sterkfontein 199

view of number of individuals involved and the ages atdeath of these individuals. There have, however, beentwo specific studies on the subject-that oí Tobias(I968h. 1974) and that of Mann (1968, 1975). In his con­sideration oí the age of death among the australopithe­cines, Tobias (I968h) made use ofthe stages of develop­ment, eruption, and attrítion of teeth, although he pointedout tbat such features must be applied "wíth the utmostcircumspection." In this study he defined five agecategories as fol1ows:

Early chi/dhood-from birth up lo the lime of erup­tion of the firsl permanenl leeth.

Lata chi/dhood-up lo the eruption of all teelh ex­cepl for the third molars or wisdom teeth, roughlyending with puberty,

Adolescence-from puberty lo young adulthood,marked by attrition on the premolars and second mo­lars, the approach lo complele developmenl of thethird molars, their emergence and their approach tothe occlusal plane as well as ongoing fusion of thethree components of the hip-bone,

Young adulthood-marked by the complele eruptionof the third molars and the early beginníngs of sigus ofattrition on them.

Adulthood-attrilion on the fully-erupted third mo­lars is moderate or marked, suggesting that these teethhave been in use for sorne time.

Tobias was able to assign 47 specimens from Sterkfon­tein to the five age categories listed aboye. as follows:early childhood, 4; later childhcod, 5; adolescence, 5;young adulthood, 2; adulthood, 30.

In his paleodemographic study of the South Africanaustralopithecínes, Mann made use of similar age in­dicators but assumed that tooth eruption times were com­parable lo those of modem man, narnely, that the ñrstpermanent molar erupts at six years, the second atlwelve, and the thírd al eighteen. Following this basicassumption, it was possible for Mann to estimate ages atdeath, expressed in years, for many of the fossils.

In the following listing of the Australoplthecus af­rícanus material avaílable, Mano's age estimate for eachspecimen is appended. In cases where no estimate inyears is attempted, other categories are used: "Infant,"before MI erupts; "juvenile," before M2 erupts; "ado­lescent," before M3 erupts; "adull, 1," M3 erupted, butlittle or no wear; "aduh, 2." stage 2 wear; "matureadult" or "adult, 3," dentine pits exposed on M3. "Ma­ture" and "imrnature" are more general divisions em­ployed when the condition of the fossil precluded a moredefinite statement (Mann 1975, p. ISO).

Australopithecus africanusMaterial: Craniums in varying degrees of compleleness

TM 1511 + STS 60. The holotype of Plesianthropustransvaalensís Broom, 1936 (TM 1511), ccnsisting of'partof a cranium including left and right maxillae and facialand occipital pieces (fig. 163). The following teeth arepresent: left p3, p4, MI. M2. and isolated M3; right poi, MI,and M'. TM 60 is an associated natural endocranial castoAge estímate, 22 ± 2 years,

STS 5, probably associated with STS 6. STS 5 ís avirtually complete and undistorted cranium, known as"Mrs. PIes," without teeth (fig. 164). lsolated M" and M,(STS 6) were found close by and may be associaled. Ageestímate, ± 35 years.

STS 17, par! of an adult \? cranium with par! of the

200 Fossil Assemblages frorn the Sterkfantein ValLey Caves: Analysis and Interpretation

5T-4

Australopitheeu8 aftlC8nus45 Par.peplo white!

10

Parepaplo broomi91

!1f(Parapaplo jOl\eSI

27

!BitPatapapi<:l lIP. Inda••

63

~Cflrcopithecoidea williame!

17

D6nofelia barlowi 4Megantel'tlOft gr8Ci'a

1Penthera pardus

1

Euryboas ltilbef'bergi4

? Hippotraginl8

I

M.kapanla cf brooml8

~nagelaphus sp. 11ft Bngas!

1

Hya.rb: afrtcHauatralla

•

$yncttfU8 sp. 1

Eleph•• ef rackl1

~Procavla ProclIYia

•r.na......... enUqua• 8

Fig. 162.Animals repreeented by lile fossils in themacrovertebratc componeot ofthe StcrtrOllteÍnMembrer 4 sample.Minimum numbers of individual, are indkated.

3 1 ? (blJ:eua WI. Medlum alcel.aphlnea Reduncll1 7 el arundlnum 1

cf PrOI'\otochoerua .p. 2

Sterkfontein 201

Fig. 163. The firsl auslralopilhecine specimen lo come from Sterkfomein-lhe Iype specímen of PlesjanchrapusIransvaalensis. TM 1511.

Fig. 164. The well-known cranium of "Mts. Pies," STS 5. Although lhe endocraníai space was empty, the skullremained undistorted because ofits protected fossilizalion site, close lo lhe back wall ofthe cave and benealh a srablearea of dolomite roof.

202 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and 1nterpretation

face, palate, and parietals. Teeth: left P3, P'" and MI;right P\ MI, M2, and M:¡. Age estímate, 21:': 2 years.

STS 19, adult cranial base with foramen magnum; leñM' {perhaps same individual as STS 58). Age estimare,26 ± 3 years.

STS 20, fragmentary cal varia and natural endocast.Age not detennined.

STS 25. par! of an adult calvaría, including frontals,parietals, occipital and temporals. Age estímate, "ma­ture.'

STS 26. parts of an adult cranium (two pieces), ínclud­ing left M3. Age estímate, "matute."

STS 58. par! of an adult calvaría, perhaps the sameindividual as STS 19. Age estímate, "mature."

STS 67. calvaria fragmenl witb external auditorymeatus, associated with left M3. Age estímate, "adult,3."

STS 71. right side of a cranium. with P"'. M'-'. Ageestímate, 33 ± 2 years.Maxillae

STS 2. facial fragment of ajuvenile with 1eft dm"? andunerupted C; right dm':". Age estímate, 4 ± I years.

STS 13. facial bones of an old adult with most of theright dentition and part of the left (specimen missing). Agenot determined.

STS 52. face and associated mandible of an adolescent,probably female: 520. maxilla with full dentition, 52b.mandible with full dentition but lacking left ascendingramus. Age estimate, 17 ± 2 years.

STS 63. par! of the face of an old adult with sornefragmentary teeth. Age estímate, "mature."

TM 1512. right maxilla frorn an adult, 1'. C. P', and M'.Age estímate, 21 ± 2 years.

TM 1514, left maxilla from an old adult, teeth worn andbroken, only M3 in reasonable condítion, Age estímate.40 ± 5 years.

TM 1535. C from 1514. Age estimate, as above.STS 8, left maxilla wilh Ml-3 and M' unerupted, Age

estímate, 16 ± 2 years.STS 10. 1eft maxilla witb part of M' and M'. Age

estímate, 39 ± 1 years.STS 12. maxilla witb left p'-'. M'-'. and part of M';

righl P" M1-1• Age estímate, 21 ± 2 years.STS 27. maxillary piece with M2-3 (specimen missing),

Asenot determined.STS 32. maxillary fragment witb right M' and part of

M'. Aseestímate, 24 ± 2 years.STS 35. maxilla witb 1eft p'-'. M'-'. Age estímate,

26 ± 3 years.STS 57. maxil1ary fragment witb 1eft M' and unerupted

crowns of P' and P'. Age estímate, 6~ :!: 1 years.STS 61. right maxillary piece witb p ...... M'. and partof

M'. Aseestimate, 26 :!: 3 years.STS 66. part ofa palate witb fragmentary teeth. Age not

detennined.STS 69, maxillary fragment from ajuvenile individual.

Age estímate, "matare."STS 70. part of a juvenile maxilla witb sorne deciduous

tooth fragments and unerupted anterior permanent teeth.Age estímate, "Immature,"

STS 3009. ri¡ht maxilla of an old adult witb roots ofP",MI-3. Age estímate, "matute,"

STS l. crushed palate of an adult witb left P", MI";right P' and par! of MI. Ase estimate, 13 ± 1 years.

STS 29. par! of an adult palate with right M' and frag­ments of MI and M3. Age estimate, 32 ± 2 years.

STS 42. crushed palate with right P'":", and parts of theleft toothrow. Age estimare, 33 ± 2 years.

STS 53, palate with most of the postcanine dentitionbilaterally. Age estímate, 26 ± 3 years.

STS 64, palate fragment with pieces of teeth. Age notdetermined.Mandibles

TM 1515. mandible ofanold adult with left p,_,. M.-,.Age estimate, "mature."

TM 1516. + STS 50. symphyseal región of mandiblewith left and right C. dm., separate, associated with leftC. (STS 50). Age estímate, 8 ± I years.

TM 1522. right mandibular ramus with roots of M,. Ageestimate, . 'mature."

STS 7. crushed mandible of an old adult male withcomplete dentition except for right e, lost in life; leftascending ramus; par! of right scapula and right proximalhumeros associated (STS 7a-4». Age estímate, 35-40years.

STS 18. par! of a juvenile mandible, right dm.-,. M..and M, (unerupted); left dm, and M,. Age estímate, 8 ± Iyears.

STS 24. incomplete juvenile mandible with left di,_,.de, dm.; right dc, dm¡_2. and MI; also unerupted crowns of11_ 2 and P3and parts of right di2 and left 11, Age estímate,6 ± 1 years.

STS 36. adult mandible in three pieces, left 1.. e-M,;right M,_,. Age estímate, ± 35 years.

STS 38. very fragmentary left mandible piece withpieces of M,.,. Age estimate, 30 ± 2 years.

STS 41.left mandible fragment with M,. Age estirnate ,"adultv Z."

STS 62, mandible fragment with alveoli of incisors,very wom left e and roots of P3'Age estimate, •'mature.'Isolated Maxillary Teeth

TM 1524, roots of a very wom MI, Age estímate,"matute."

TM 1527. right C. Age estímate, "matute."TM 1532. very worn left M'. Age estímate. 33 ± 2

years.TM 1561. crown of right M'. Age estímate, 23 ± 1

years,STS 11, left M' and M' (specimens missing). Age not

determined.STS 16. very worn upper molar. Age estimate, "ma­

ture.'STS 21. crown of upper right molar (M' or M'). Age

estímate, 24 ± 2 years.STS 22. left M' and M'. Ase estímate, 17 ± 2 years.STS 23. par! of righl M'. Aseestimate, 13 :!: 1 years.STS 28 + STS 37. right M'-'and left MI (STS 28); Ieft

M'-' (STS 37). Age estímate, 19 ± 1 years.STS 30, right P' and M'. Ase estímate, 23 ± 2 years.STS 31, par! of M'. Ase estímate, 25 ± 2 years.STS 33. lingual half of a worn upper premolar. Age

estimate, "matare.'STS 39, considerably wom left P" and M'. Age

estímate, "mature."STS 44. left M' and M'. Age estimare, "aduh, 2."STS 45. par! ofan upper molar. probably right M'. Age

not determined.STS 46. most of tbe crown of right M'. Ase estímate,

32 ± 2 years.STS 47, P'. Ase estímate, "mature."STS 48. 1eft C. Age estímate, "mature."STS 54, right M'. Age eSlimate. 38 ± 2 years.

STS 55, P" (a); crown of left M3 (h). Age estimate,16 ± 1 years.

STS 56, left MI, M2, and part of dm2 . Age estimate,9 ± 2 years.

STS 72, part of crown of right Ma. Age estimate, 26 ± 3years.Isolated Mandibular Teeth

TM 1518, right M3• Age estimate, 24 ± ~ years.TM 1519, right M3 , very wom. Age estimate,

± 35 years.TM 1520, left Ms, very wom. Age estimate, 30 ± 3

years.TM 1523, left p•. Age estimate, 16 ± 2 years.TM 1525, roots of a lower premolar. Age not de­

termined.TM 1528, right C. Age estimate, "mature."TM 1533, part of M3 (specimen missing). Age not de­

termined.STS 3, right C, well worn. Age estímate, "mature."STS 4, left M2 (matches STW-Hl). Age estimate,

25 ::!: 3 years.STS 9, right MI with unformed roots. Age estimate,

6 ::!: 1 years.STS 40, lingual part of a lower canine. Age estimate,

"immature..,STS 49, buccal part of crown of P3' Age estímate,

"immature."STS 51, right P3 and unerupted crown of right C. Age

estimate, JO ± 1 years.STS 59, erown of left M3 • Age estímate, 16 ::!: 1 years.STS 1534, 1535, 1374, wom incisors that may or may

Sterkfontein 203

not be australopíthecine. Age not determined.Postcranial Pieces

TM 1513, distal end of left femur. Age estimate, "ma­ture. "

TM 1526, carpal bone: right os magnum. Age not de­termined.

STS 14, parts of an articulated vertebral column ribspelvis, and hind limb: a-f, 6th-1st lumbar verteb~;g-i:12th-9th thoracic vertebrae; k-n. four thoracic vertebrae;o-p. vertebral bodies; q, sacrum; r. left innominate; s.right innominate; t, right proximal femur; u, shaft piece;w-y, rib fragments; z, vertebral and pelvic fragments (fig.165). Age estimate, "mature."

STS 34, distal end of right femur. Age estimate, "ma­ture. "

STS 65, right pelvic bone with most of ilium and partsof pubis; proximal end of left femur (a). Age estimate,"mature" (some doubt whether this specimen is from ahomínid).

STS 68, proximal end of left radius (perhaps cer­copithecoid). Age estimate, "mature."

STS 73, vertebral centra. Age not determíned.With fragmentary remains of this nature it is difficult to

be certain of the minimum number of individuals thatcontributed to the sample. After careful consideralion ofaH possible associations of the pieces avaílable, it isestimated that al least 45 indivíduals are involved.

Table 78 indicates Mann's al1ocation of 47 specimensfrom Sterkfontein Member 4 to age categories. If thosespecimens below twenty years of estimated age at deathare regarded as suhadult, then seventeen out of 47 speci-

Fig. 165. An articulated segment ofvertebral column, pelvis, and other bones ofAust,alopilheclJs africanus (STS 14){rom SteMontein Member 4. The specimen is shown in course of preparation. Photograph by J. T. Robinson.

204 Fossil Assemblages from the Sterkfontein Vallt>y Caves: Analysis and Interpretation

mens, or 36% of the ageable sample, was subadult atdeath, as is shown in figure 166.

Family CercopithecidaeParapapio jonesi (fig. 167a)Material:

37 pieces from a mínimum of 27 individuals.Type specimen: an almost complete 9 cranium with leftC-M3 and right P3_M3.

Craniums in varying degrees of completeness: crushedcranium, ~, probably complete before being shattered,P3_M3 bilaterally, STS 547; base of braincase and pos­terior part of palate, left P4_M3, right M2-3; left side ofjuvenile 9 cranium, C erupting, MI, M2-3 erupting, STS333; maxilIary pieces, STS 250, 284, 287, 367, 368a,372o--e, 390, 441, 456, 458; mandible pieces, STS 276,302, 306,307,313,317,329,334,340,348, 355,381a--e,384, 421, 443, 446, 448, 457, 1925a; isolated lower teeth,STS 418a--e, 485; articulated atlas and axis vertebrae,STS 368b; part of a thoracic vertebra, STS 381e.

The 27 individuals may be divided into age classes asdefined by Freedman (1957). Particulars are given in table79: 2 individuals are rated as "juvenile," 2 as "immatureadult," 7 as "young adult," 12 as "adult," and 4 as "oldadult. "

Parapapio broomi (fig. 167b)Material:100 pieces from a minimum of 91 individuals.Type specimens: partial cranium 0', complete endocast,snout, roots of right M3, STS 564; 9 mandible with C-M3 ,

STS 562.Craniums in varying degrees of completeness: com­

plete cranium, 9, left p 4, M2-3, right M2-3, STS 262;complete 9 cranium except for left side ofbraincase, STS254a; complete ~ cranium, flatteoed laterally, STS 393;almost complete 9 cranium, with Ml-3 bílaterally, STS397; cIUshed and weathered cranium, adult ~, left Ml-3,STS 535; much of a flattened O' cranium, STS 531; rightside of 9 cranium with M3, STS 396; parts of a shattered9 cranium, left p3-4, M1-2, STS 539; part of an immature9 cranium, STS 251. Palates, STS 253, 260,272,297,354,378a-d, 388a-b; left maxíllary pieces, STS 264,266,267,274, 301, 322, 380a--e, 385a--e, 398a--e, 415a, 544; rightmaxillary pieces, STS 277, 325, 379a,b, 383b, 420a, 530;mandible pieces with both rami, STS 258, 261 + 423, 283,296 + 351, 299, 323, 331, 335, 337, 360, 362, 363, 371,374b, 390a,b, 416a,b, 426, 533 + 534, 3035; left mandib1epieces, STS 255, 256, 268, 271, 278, 285, 286, 312, 314,326,328,346,353, 386a, 409, 414a,b, 463,466,557; rightrnandible pieces, STS 270, 280, 289, 298, 309, 311, 338,

, 339,356, 369a,b, 3820, 383a, 411a,b + 425, 434, 469, 542,558; isolated teeth, STS 406, 410a-b, 419. 437, 438,445,484,491,511.

A separation of the 91 individuals into sex and ageclasses is given in table 80; 3 individuals are classed asjuvenile, 10 as irnrnature adult, 20 as young adult, 39 asadult, and 19 as old adu1t.

Parapapio whitei (fig. 167c)Material:14 pieces from a mínimum of 10 individuals.Type specimen: mandible of a young adult 9, left I t-P4;right I,-M3 , STS 563.

Articulated maxíllae and mandible pieces: left and rightmaxilla and mandible pieces associated with axis Ver­tebra, STS 263 + 370; right maxil1a, P4_Ml, mandible, left

'-5 yearsf

"_20""'f21-40 years

o 20 40 60

Percentage 01 sample (n:47)

Fig, 166. The a1Jocation of 47 specimens of Auslfa/opithecus a!ricanusfrom Sterkfontein Member 4 to age-al-<lealh calegones, according lOMann (1975).

1::: ._..L. __~._~

Fig, 167. Examples of cercopilhecoid craniums fmm SterkfonteinMember 4: (a) Parapapio jonesi, STS 565; (b) P. broo;"i, STS 564;(e) P. whilei, STS 568; and (d) Cercopitheeoides williamsi, STS 394<1,

Sterkfontein 205¡¡I

I,-P" right I,-M" STS 389a,b; left maxilla, P'-M', manodible with right P, and left M" STS 548a-<:; palate, s . leftP" M', right P'-M', STS 336; left maxillae, STS 259, 303,424; right rnaxilla, STS 462; mandible with both toothrows, STS 359; left mandible piece, STS 352; right man­dible piece, STS 342, 467.

A separatíon of the 10 individuals into sex and ageclasses is given in table 81. There are no juveniles in thesarnple; 1 individual was classed as írnmature adult, 2 asyoung adult, 5 as adult, and 2 as old adult.

Parapapio sp. indet.A total of 201 cranial and 5 postcranial pieces has been

assigned to Parapapio, but the specimens are not com­plete enough to be specilically identilied. However, 101pieces could be placed, on the basis of tooth eruption andwear, in age classes.Material:Age-detennined individuals10J pieces from a minimum of 38 índividuals; 3 individu­als were rated as "youngjuvenile," 22 as "juveníle," 4 as"immature adult;" 7 as "young adult, " 1 as "adult,' and1 as ..old adult."

Craniums in varying degrees of completeness, STS 364,376,54O,559a-<:, 3048,3057,3076; craniums or maxillaewith articulated mandibles, STS 373a, 387a,b, 395, 546,1046, 3067, 3073; palates, STS 305, 319, 324, 345, 1189,2119; left maxillary pieces, STS 308, 3[5, 375a-<:, 431,488,3047,3052; right maxillary pieces, STS 273, 294, 318,320b, 343, 429, 432, 433, 439, 454, 455, 459, 464, 545,1592,1593,1594,2117,2215; mandibles with left and righttoothrows, STS 281, 351, 468, 1238, 1534,3037; left man­dible pieces, STS 269, 304, 310, 316, 320, 349, 413a, 436,440,442,453,461,1753,3053; right mandible pieces, STS257,293,327,341,358, 399a, 403a, 413b, 427,430,444,447, 449, 450, 451, 560, 3070, 3078; isolated teeth, STS291, 402b, 407a, 412,460, 465, 486, 504, 509, 510, 520,522, 524, 527, 528, 529, 3045.Specimens on which age detenninations are not possible105 pieces from a minimum of about 15 individual s

Craniums in varying degrees ofcompleteness, STS 536,537, 1026, 1052, 3063; maxil1ary pieces, STS 428, 3055,3066, 3072; mandib1e pieces, STS 275, 470, 543, 1544,1817, 2067, 2510, 3043, 3054, 3056, 3061, 3062; isolatedteeth, STS 254b, 292,386, 391a-e, 400a-<:, 401a,b, 402a,403c, 404a,b, 405a-<:, 407b, 408a,b, 417a,b, 422, 452,471,472,473,474,475,476,477,478,479,480,482,487,489,490,492,493,494,495,496,497,498,499,500,501,502,503b,505,506,507,508,512,513,514,515,517,519,521, 525, 526, 556, 1184, 1704, 1760, 1895, 2116, 2123,2348,2477,3036,3038,3039,3040,3041,3042,3044,3046,3049, 3050, 3051, 3064, 3065, 3068, 3071; thoracic ver­tebra píeces, STS 3059, 3060; lumbar vertebra píece, STS415b; proximal radius piece, STS 375e; proximal femurpíece, STS 378d.

Cercopithecoides williamsi (fig. 167d)Material:21 pieces from a minimum of 17 individuals.

Craniums: most of a cranium with full dentitíon, 2.articulated mandible, atlas, and axis, STS 394a-<:; most ofcalvaría without snout, but mandible with P.M, bilater­alIy, STS 541; partial braincase and snout, right MI-',STS 392; left side of braincase and snout, left M,_" STS361; base of calvaría, right M'-', STS 252; palates:juvenile palate with articu1ated mandible, STS290 + 357 + 435; anterior part of o palate, STS 350; right

maxilLae, STS 295, 347; mandible pieces with both rami,STS 300, 366: left mandible pieces, STS 279, 288, 344,532; right mandible piece, STS 282; isolaled teeth, STS516,518,523.

A separation of the 17 individuals into sex and ageclasses is given in table 82.

Two individuals were classed as juveniles, 1 as im­mature adult, 3 as young adult, 10 as adult, and 1 as oldadult.

Cercopithecoid indet.1t was found that 413 cranial and 46 postcranial pieces

were clearly of cerccpithecoid origio but could not beidentified to generic or specífic level, It is highly likelythat most of these specimens carne from individuals thathave already beeo listed in the Parapapio aod Cer­copithecoides categories. At a rough estímate, the speci­meos come from a minimum of 100 individuals.Material:

Cranial pieces with articulated mandibles, STS 423,477, 1013, 1038, 1047, 1178, 1485, 1784, 1816, 2333;craniums in varying degrees of completeness, STS 421,422,430,476,494,943,945, 1002, 1003, 1024, 1028, 1030,1031,1039,1048,1053,1058,1193,1238,1260,1275,1449,1456,1529,1635,1777,2008,2013,2020,2034,2118,2145,2190,2244,2352,2572,3091,3099,3100,3103,3105,3109,3111; calvaría and endocast pieces, STS 217, 400, 405,406,409,412,424,428,434,442,457,458,461,467,484,485, 486, 491, 493, 515, 530, 538, 541, 991, 995, 1000,1002, 1009, 1011, 1012, 1014, 1017, 1018, 1019 + 1074,1021, 1023, 1027, 1032, 1034, 1041, 1042, 1043, 1000a,1050,1054,1055,1059,1063,1065,1066,1067,1070,1071,1072,1075,1077,1078,1079,1080,1081,1082,1083,1088,1109, 1229, 1230, 1235, 1238, 1253, 1256, 1258, 1267, 1306,1310,1334,1342,1361,1442,1459,1471,1474,1486,1499,1506,1532,1536,1547,1588,1615,1639,1675,1688,1740,1778,1810,1841,1955,1957,1960,1963,1965,1969,1970,1983,2009,2010,2012,2015,2023,2032,2035,2038,2043,2045,2051,2111,2143,2212,2215,2220,2222,2223,2224,2262,2264,2377,2388,2406,2450,3092,3093,3094,3095,3098, 3102, 3104, 3106, 3108; palates and maxillarypieces, STS 378, 411, 416, 432, 470, 474, 478, 503, 525,529,544,547,560,992,1001,1005,1008,1025,1037,1051,1089,1188,1217,1271,1305,1311,1343,1455,1494,1516,1541, 1566, 1591, 1640, 1702, 1708, 1770, 1797, 1813, 1837,1890,1917,1929,1939,1961,1966,1995,2002,2015,2018,2044,2071,2079,2106,2113,2142,2161,2205,2208,2265,2303,2328,2344,2350,2375,2376,3097,3101,3113,5394;mandib1e pieces, STS 398, 407, 413, 427, 436, 444, 446,481, 508, 558, 947, 1007, 1010, 1036, 1049, 1056, 1085,1102,1127,1195,1208,1246,1314,1339,1346,1372,1401,1461, 1519, 1533, 1552, 1568, 1575, 1618, 1713, 1728, 1772,1788,1792,1794,1801,1868,2017,2021,2051,2077,2087,2093,2095,2103,2118,2124,2139,2158,2174,2224,2274,2280,2295,2306,2331,2365,2449,2475,2513,2525,2545,2558,2570, 3096,3107; isolated teeth, STS 126, 377a-d,441,452,487,505,528,532,540,549,564,790,803,1022,1093, 1116, 1\60, 1169, 1344, 1387, 1411, 1428, 1472b,1479,1530,1601,1661,1669,1683,1684,1716,1722,1762,1768,1780,1785,1829,1836,1839,1899,1927,1934,1937,1954,1992,2013,2016,2017,2041,2105,2130,2133,2144,2146,2147,2163,2168,2193,2196,2206,2207,2209,2218,2225,2231,2277,2279,2292,2297,2416,2433,2438,2444,2488, 2489, 2508, 2512, 2531, 2543, 2548, 3086; cervicalvertebra, STS 551; pelvis píeces, STS 1000b, 2146; pro­ximal humeros pieces, STS 1458, 2185, 2219; humerosshaft pieces, STS 2308, 1444; distal humeros pieces, STS

1

I

206 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and lnterpretation

1264, 1504, 2074, 2198, 2201, 2380; articulated distalhumeros and proximal radius, STS 2198; proximal radiuspiece, STS 534; distal radius piece, STS 1764; proximalfemur pieces, STS 443, 1069, 1089, 1092, 1094, 1469,1733, 1978, 1992, 2050, 2069, 2259, 2357, 2521, 3109;femur shaft pieces, STS 1846, 2183; distal femur pieces,STS 549, 1614, 1905,2150, 2188; proximal tibia pieces,STS 1860,2230,2562; distal tibia pieces, STS 1204, 1663,2474, 2563.

Order CarnivoraFamily FelidaePanthera pardusMaterial:A single specimen from 1 individual.

Isolated left P\ STS 134.

Dinofelis barlowi (fig. 168)Material:5 pieces from a mínimum of 4 individuals.

The type specimen of Megantereon (Therailurus) bar­/owi, consisting of much of the cranium without the leftmaxilla and mandible, TM 1541; also isolated left C;, with­out root, TM 1542; crushed cranium and mandible, STS131; isolated left 13, TM 1566; left C;, STS 132; left P\ TM1579.

Megantereon gracileMaterial:A single specimen from 1 individual.

The type specimen consisting of a right mandibularfragment with P.¡ and M¡, STS 1558.

Family HyaenidaeEuryboas silberbergi (fig. 169)Material:5 pieces from a minimum of 4 individuals.

The type specimen of Lycyaena silberbergi, consistingof parts of a crushed cranium and mandible, STS \30; leftmaxilla with p'H, STS 127; left maxílla with p::, STS 135;associated with crowns of Pz-~, STS 133; left and rightmandíble fragments, STS 126. The type may be fromMember 2 or 3.

Crocuta crocutaMaterial:A single specimen from 1 individual.

Left maxilla with pH, STS 128, listed by Ewer as 218.

Hyaenid indet.Material:5 pieces from an estimated mínimum of 3 individuals.

Palate fragment, STS 129; mandible pieces, STS 863,867; isolated tooth fragments, STS 880, únnumbered.

Canis brevirostrisMaterial:The type specimen conslstmg of braincase and leftmaxilIa with C;, pl-', MI-2; left mandible piece with P.¡,M'-2' STS 137.

Canis mesomelasMaterial:9 pieces from an esÜmated minimum of 5 individuals.

The type specimen of Canis antiqulls, consisting ofleftand right mandibular rami with deciduous and permanentteeth, TM 1582; cranium and partial snout, STS 142;maxillary fragment, STS 141; mandible pieces, STS 136,138, 139, 2022, 2089; isolated right 13, STS 140.

Camivore indet.Material:32 cranial and 16 postcranial pieces from an estimatedminimum of 8 individuals.

Fig. 168. The lype specimen of Dinofelís barlowi, TM 1541. from Slerkfontein Member 4.

Slerkfontein 207

Maxillary pieces, STS 1591,2113; mandible pieces, TM869, STS 1402, 1538, 1616, 1678, 1876, 1959, 2059d, 2094,2315; toothfragments, TM791, STS858, TM866, STS 870,945, TM 1092, STS 1505, 1518, 1556, 1557, 1604, 1634,1644, 1709, 1909, 1964,2101,2113,2204,3124; axis ver­tebra, STS 1774; 2 articulated lumbar vertebrae, STS1244; scapula pieces, STS 1400b, 1872; humerus pieces,STS 1090, 1832,2176; femur pieces, STS 1700, 2059a,c,2111,2273; tibia pieces, STS 550, 1608; metapodials, STS798b, 2190b.

Order ArtiodactyIaFamily BovidaeDamaliscus cf. sp. 2Material:A single specimen fram 1 individual.

Right mandible piece with MZ- 3' STS 2582.

Damaliscus sp. 1 or Parmularius sp.Material:26 pieces fram a minimum of 7 individuals.

Right maxilla with M'-Z, associated with parts of acranium and base of left hom-core, STS 2368a-{; maxil­lary pieces, STS 1592,2577; mandible pieces, STS 1319a,1563, 1742, 1800a,b, 1980,2027,2135,2208,2581, 2585a,2586; isolated teeth, STS lI00a,b, 1687, 1696, 1743, 1910,1984, 1999, 2005, 2046, 2562, 2580, 2593a.

Hippotragus cf. equinusMaterial:2 pieces from 2 individuals.

Left M2, STS 1630; left M2, STS 2599.

?HippotraginiMaterial:23 pieces from a minimum of 8 individuals.

Palate with left P4_M3, and right M2-a, associatedmandible pieces with left M¡-2' and right PcMa, STS1539a-c; maxillary pieces, STS 1137,1847, 2190a, 2336a;mandible pieces, STS 1438, 1531,1589,1682,2031,2055,22282584; isolated leeth, STS 789, 792, TM 1342, STS1632, 1866b, 1833,2064, 2145a, 2524,2560.

Redunca cf. arundinumMaterial:A single specimen fram 1 individual.

Part of a juvenile palate with right dpmH .

Antidorcas cf. reckiMaterial:5 pieces from a minimum of 3 individuals.

Right maxilla with P3_M\ STS 1435; right mandiblepiece with M l - 2and Ma unerupted, STS 1560; right man­dible piece with M3 , STS 2369; left mandible piece withM2- a, STS 1944; isolated right M2, STS 1325a.

Cf. Megalotragus sp.Material:A single specimen fram 1 individual.

Left M2- 3 , STS 1339.

Connochaetes sp.Material:3 pieces from a mínimum of 1 individual.

Left maxilla, P4_M3, STS 2597; mandible piece with leftMa and right MI> STS 2512b,d; isolated right M2 or M"STS 2200.

Medium-sized AlcelaphinesMaterial:14 pieces fram a minimum of 7 índividuals.

Frontals with bases ofboth hom-cores, cf. Rabaticerasporrocornu1US, STS 2595a; maxillary pieces, STS 1551a,1844, 1852; mandible pieces, STS 714, 1114, 1427, 1445;isolated teeth, STS 1317, 1324, 1333, 1334,2519,2563.

Makapania cf. broomi (fig. 170)Material:22 pieces fmm a mínimum of 8 individuals.

Maxillary pieces, STS 817, 1573, 1721, 1734, 1756,2059b; mandible pieces, STS 952, 1564a, 1879, 1901a,1925,1938, 2362a, 2565,2588, 2593a; isolated teeth, STS1754a, 1824, 1894, 1994,2121,2592.

Tragelaphus sp. aff. angasíMaterial:3 I'ieces from a single individual.

Right maxilla, MI-2, STS 2092; right mandible piece,dpm., MI unerupted, STS 2493; left mandible piece,dpm., M, unerupted, STS 1865.

Syncerus sp.Material:A single specimen from 1 individual.

Part of a juvenile mandible, left dpm 4 , P;rM" STS1936a.

?Gazella sp.Material:2 pieces fram 1 individual.

Right mandible piece with Mz, STS 1996; left mandiblepiece, M,-a, TM 2076.

Antidorcas cf. bondiMaterial:A single speeimen fram 1 individual.

Left mandible piece, P2-M 2, STS 1125.

Antelope size class 1Material:4 cranial and 8 postcranial pieces fram a mínimum of 2individuals, as listed in table 77.

. . 'b-Fig. J69. The Iype specimen of a hunting hyena, Euryboas silberbergi, Antelope size c1ass 11STS 130, from Ihe lower levels of Ihe Slerkfomein deposito Material:

208 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Order RodentiaFamily HystricidaeHystrix africaeaustralisMaterial:6 pieces from a minimum of 5 individuals.

Left maxilla with P\ MI-3, STS 2078; left premaxillaryfragment with part of P, STS 2061; parts of left and rightmandibular rami with dpm., M I - 2 bilatera1ly, STS957a-e; left mandible piece with parts of 11 and p., STS2095; isolated right M3, STS 1622; isolated incisor, STS2540.

The Major Features of tbe Bone Accumulation

The Composition ofthe Fauna Represented by theFossils

The information provided in table 76 and depicted infigure 162 may be further summarized. This is done inTable 83 and figure 171. Ofthe 348 individual animals thatcontributed bones to the accumulation, no fewer than243, or 69.8%, were primates. Of these primates, 45 wereaustralopithecines, while the remainder were representedby three species of Parapapio and one of Cer­copithecoides. Ofthe individual animals, 8.1% were car­nivores, represented by a leopard, four Dinofelis individ·uals, one Megantereon. two species of hyena, and twospecies of jackal.

The artiodactyl component of the fauna was found tomake up 14.7% of the total and carne from 49 bovid indi­viduals, together with two pigs. Dividing the bovid re­mains into size classes, we obtain thc following results:

Order HyracoideaFamily ProcaviidaeProcavia antiquaMaterial:17 pieces from a mínimum of 8 individuals.

Type specimen of P. rohertsi, partial cranium and ar­ticulated mandible, TM 1197, STS 105; almost completecranium, STS unnumbered; palates, STS 109,2140; man­dible pieces, STS 102, 103, 104, 107, 108, 140, 1201,1350h, 1829, 2320; isolated incisor, STS 1878.

Procavia transvaalensisMaterial:9 pieces from a minimum of 5 individuals.

Calvaria pieces, STS 2319, 2361; maxillary piecc, STS1535; mandible pieces, STS 101, TM 1212, STS 2117,2461; isolated incisor, STS 2085.

dible pieces, STS 1830a. 2102 + 1717, 3006, 3127; adultmandible pieces, STS 1853, 3007; isolated teeth, STS1888, 1972,2313, 2316a-c, 3003,3004,3005, 31Z6.

Equus sp.Material:2 pieces from 1 individual.

Isolated incisors, STS llOOh, 2512c.

Order ProboscideaFamily ElephantidaeElephas reckiMaterial:A single specimen from 1 individual.

Isolated dm 1 or dm 2 , probably right, STS 1863.

!(~ .

Pig. 170. Charaeterislie teelh of the remarkable bovid Makapania er.broomi, well represented in the SlerkIonlein Member 4 breceia.

Family SuidaeMetridiochoerus sp.Material:2 pieces from 2 individuals.

MI (?) in maxillary fragment, STS 2395; left mandiblepiece with dpm. and M2 , STS 3074a-b.

Antelope size c1ass IVMaterial:13 cranial and 2 postcranial pieces from a minimum of 2individuals, as listed in table 77.

Order PerissodactylaFamily EquidaeEquus capensisMaterial:18 pieces from a mínimum of 7 individuals.

Left maxilla, P<_M2, STS 3000; left maxilla, p2-<, STS3001; right juvenile maxilla, dpm2-< and part of MI, STS3002; parts of3 maxillary teeth, STS 1571; juvenile man-

Antelope size class 111Material:157 cranial and 89 postcranial pieces from an estimatedminimum of 15 individuals, as listed in table 77.

8J cranial and 30 postcranial pieces from an estimatedminimum of 8 individuals, as listed in table 77.

,class 1,2 unidentified individuals; class H, 14 individuals,including 6 juveniles; class 111, 11 individuals including 4juveniles; and class IV. 18 individuals íncluding 7juveniles.

The extinct Cape horse, Equus capensis, made up 2%oí the preserved fauna, and 13 individual dassies of twospecíes also contributed. Remains of 1 juvenile elephantand 5 porcupines are presento

The Representation ofSkeletal Parts

There is Iittle doubt, 1 think, that the fossil sample fromMember 4 has been strougly and artificially biased infavor of eranial and readily identífíable speeimens. This isa result of the rather haphazard collecting procedure al

the cave between 1936 and 1947. During mueh of thisperíod the site was being mined for lime; it was aIso fre­quented by innumerable casual visitors, many of whomearried away samples of bone-bearing breccia assouvenirs. In addition, Broom understandably con­centrated on those fossils that were potentially eapable ofyielding specific identifications. For these reasons 1 sus­peel that, although lhe fossil sample may provide a rea-

PERISSODACTYLA~

HYRACOIDEAr

AYESI

Sterkfontein 209

sonably accurate reñection of faunal compasilion in theMember 4 bone accurnulation, it cannot be expected lo doso for body-part representation. Many postcranial fossilspresumabLy were simply not collected al all by Broorn'steam and may subsequently have beco picked up on thedumps by souvenir-hunters.

Despite this probable bias, the paucity oí primate post­cranial fossils is striking. To go with the cranial remains ofal least 45 austraJopithecines, there are onIy 7 catalogentries of postcranial bones. One of these, STS 14 (fig.165), represenled articulated bones from a reasonablycomplete torso. but the others represent nothing morethan scattered skeletal fragments. The same is true for thebaboon and monkey remains. Here 769 cranial specimens(sorne of them, admittedly, ísolated teeth) are associatedwith only 55 posteranial pieees. Despite the artificial sam­pling bias, it is difficult to avoid the conclusion that mostof lhe primate posleranial skeletons had disappeared be­fore fossílization. The position in respect to antelope re­mains is only slightly beller: here 49 individual bovids,identified largely on 360 cranial rernains, are representedby only 129 posteranial bones. The 7 equids, 1 elephant,13 dassies, and 5 porcupines have all been identified by

RE~-l_-'-_-r-'-t-r-t-t-r-t-'-r-'--Y

o 20 <OMEII8ER -4

"" o .0 <O

"E"BEII a:ea o 20 '"

IIIE"SER S

ea

PERCENTAOE CONTRIBUTlON ro THE FOSSIL FAUNAS

Fi¡. 171. Percentase contnDutions madeby vanous groups olanimals to the samples of macrovertebrate fossil faunafrom SteJ1d'onlein Members 4, S, and 6.

210 Fossil Assemblages from the Sterkfontein VaIley Caves: Analysís and InteI1'retation

cranial elements only, though some of their other bonesmay well be among the 330 indeterminate fragments, tooincomplete to allow taxonomic placing.

Ohserved Damage lo ¡he Bones

During the analysis of this and other fossil assemblagesfrom the Sterkfontein valley site units, each specimenwas examined for the presence of diagnostic damage.However, the recognition of such damage has been seri­ously complicated for Member 4 fossils by postdeposi­tional events. Almost aH the bones have been distorted orcrushed by compression in the breccia, and most of themalso suffered extensive damage during the mining pro­cess. Broken edges are seldom those that existed whenthe bone was buried, and determining what damage oc­curred before fossilization is extremely difficult.

Details of observed damage are given tn table 84 andmay be summarized as fol1ows:

Porcupine Gnawing. Gnaw marks were observed on onespecimen only-a juveniJe Parapapio mandible.

Carnivore-Inflicted Damage. Two Ausrralopithecusspecimens, a juveniJe mandible and a palate, showragged-edge damage that could have resulted from carni­vore chewing. In both cases the bone is in poor condition,and such a diagnosis is tentative. Two Parapapio mandi­bIes show evidence of camivore damage-one, STS 351,shown in figure 172, bears two punctures on the inner sideof the COI1'us caused-unquestionably, 1 think-by carni­vore teeth.

As detailed in table 84, an antelope mandible and twodistal humeri show traces of camivore chewing, and twodassie cranial pieces have been darnaged in a waycharacteristic of food remains left by large cats.

Traces 01Artificial Bone Alteration. In the course of thisstu dy, 1 observed nothing on any of the bones that, in myopinion, could be attributed conclusively to hominidfeeding or butchering. In the past, R. A. Dart hassuggested that a number of the Sterkfontein craniumsshow evidence of pUI1'0seful hominid violence. 1 shal1now give Sorne details and comments.

In his paper "The Predatory Imp1emental Technique ofAustralopithecus," Dart (1949a) proposed that 3 austra­lopithecine specimens and 18 cercopithecoid skuUs from

1=-

Fig. 172. Parapapio mandible, STS 351, from Sterkfontein Member 4.showing carnivore damage in lhe forro of two punetate marks on lhe innerside of the rnmus.

Sterkfontein showed traces of hominid battering or feed­ing actívíty. In a separate paper (Brain 1972) 1 have com­mented on the evidence for interpersonal violence amongaustralopithecines, but l will reiterate my comments rele­vant to the three Sterkfontein hominid fossils. Endocra­nial casts of three skulls are involved, previously desig­nated as Plesianthropus transvaalensis types 1.2, and 3.They are now c1assified as A. africanus; type 1 is num­bered STS 60, while the other two are unnumbered andwere housed in the Anatomy Department of Witwaters­rand University for many years, although they are nowat the Transvaal Museum. Dart (1949a, p. 38) describedthe specimens and their damage as fol1ows:

Plesianthropus transvaalensis type 1. A fragmentedskul1 (without mandible or much of face) and afronto-parietotemporal endocranial cast lacking theright parietal and occipital regions. The right maxilla,premaxilla and molar are faírly complete and hulecrushed; the left maxilla is better preserved but lcsscomplete. The base and the vault are sufficientIy pre­served to yield accurate reconstruction of the generalcranial dimensions.OBSERVED DAMAGE

The general volume and shape of the endocranialcast have been altered by compression. The squashingof the cast is maxímal anteriorly and the temporalheight is estimated to be reduced by at least 1 cm bythe left temporal bone overriding the parietal at theirsuture. The left temporallobe tip was absent from thecast. Crushing and distortion is evident also in thechiasmatic and inferior frontal regions. The left tem­poral bone was so damaged that it was removedpiecemeal.SUGGESTED CAUSE

A lateral blow on the left temporo-parietal regio n ofthe skulI.

The skul1 from which the endocast carne is extremelyfragmentary, while the compression shown by the castitself reflects the distortion of the calvaria in which itformed. Such distortion is widespread among Sterkfon­tein fossil skulls, both of hominids and of other animals; Ido not believe there is good reason to conclude that thedistortion of this particular specimen has resulted fromdeliberate violence. Dart continues (1949a, p. 38):

Plesíanrhropus transvaalensís type 2. Fronto­parietal temporal endocranial cast. Left fronto-parietalregion deficient and occipital region and base entírelyabsent. The frontal portion of the sagittal suturallinehas been defiected more than 15 degrees to the right ofthe parietal portion of the sagittal suture.OBSERVEO DAMAGE

The bregmatic angles of both parietal bones are theseat of two depressed fractures in the right parietalbone that hinge together on each side oí" an irregularline; the left margin of the fractured area followsroughly the sagittal suture. The lateral margins of thearea are depressed; the right 12 mm and the left 4 mmbelow the remainder of the parietal bones. Runningacross these broken fragments and the frontal bonesrun radiating fracture lines to the left and right lateralorbital margins both of which were fractured trans­versely. Owing to the great distortion of the ríght Syl­vian notch region, the temporal bone overrode theparietal for a centimeter.

SUGGESTED CAUSE

A vertical blow just bebind and to the right of thebregma with a double-headed object.

Apart from the depressed area in the rightparietal bone(fig. 173a), Schepers (in Broom and Schepers 1946)showed that the skull in which the endocast formed wasdistorted in such a way that the anterior part of the rnid­sagittal plane is deftected 15 degrees relative to its pos­terior part (fig. 174).

This kind of distortion is frequenUy seen in skulls pre­served in cave breccia (chap. 7) and undoubtedly oecursafter the skull has been enclosed in the surrounding andsupporting matrix. Examination of the endocast showstbat the skull had been resting upside-down during thefossilízation process. Fine-grained sediment entered, pre­sumably through the foramen magnurn, but did nct fill theendocranial space completely, perhaps because the cal­varía coUapsed after a certain time.

The roof of the calvaria undoubtedly has suffered adepressed fracture as Dart has c1aimed. It has occurred inthat part of the skull that would have been in contact withthe cave floor when it carne to rest before becomingburled and fossilized. Unfortunately we have no record ofthe nature of the breccia that surrounded the skull or ofthe floor on which it lay. Was there, for instance, a stonein the matrix corresponding to the depressed area? Wewill never know, and Dart's claim of antemortern injuryresulting from violence can be neither substantiated norrefuted.

Dart's description of the third specirnen fromSterkfontein (l949a, p. 38) is as follows (see fig. 173b):

Plesíanthropus transvaalensis type 3. Parieto-occipitalendocranial cast fractured (in blasting) horizontallythrough the occipital poles and coronally through theparietaJ region in the vicinity of the vertex.OBSERVED DAMAGE

Fragments of bone in the substance of tbe cast showthat the skull must have been open in sorne región orother (as in the Taungs specimen). The sagíttal andlambdoid sutures have both been sprung. In additionto minor fracture lines, a major line of fracture runscompletely across the cast through the parietal regionsand skirting the anterior margin of the right parietalarea fragment. The postero-medial portion of the leftparietal area is depressed below the ríght parietal areaat the sagittal suture and even more deeply below theamero-lateral portion of the bone both anteriorly andlalerally.SUGGESTED CAUSE

A vertical blow slight!y lo the left of mid-parietalregion with a bludgeon.

Like the preceding endocast, the specimen has lost itscontext relative to the matrix. No associated skuIl isknown, nor is there any infonnation on the nature of thebreccia thal originally enclosed it. The skull was certainlyfractured in the midparietal región, but it is very doubtfulthat it could ever be proved whether the damage occurredbefore or after the burial andfossilization ofthe specimen.

In addition to these australopithecine specimens, Dartselected 18 Parapapio skulls from Member 4 at Sterk­fontein, described the damage lhey had suffered andsuggested how sllch dwnage could have come about. Onlyone case will be quoled here (Dart 1949a, p. 34):

Sterkfontein 211

585 Tvl. Posterior half of an infantile skull (1 mmthick) filled with breccia: species indeterminable.08SERVED DAMAGE

Depressed (4 mm) radiating fracture in left lateralparietal region which caused depression and bucklingof skuU along line of fracture running transverselythrough the middle of the parietal bones. The mainfracture gives the impression of being due to, orexaggerated by, crushing by left thumb, while craniumwas being evacuated oí its contents through anteriordeficiency. The rock filling at the anteríor margin of lbeskull had been broken away before fossilisation.SUGGESTED CAUSE

Destruction of anterior par! of skull wilb a blowfrom front and crushing oí posterior portion on leftsíde between fingers and tbumb during evisceration ofcranial contents.

This and other specimens are figured and described inDart's paper, which the interesled reader can consult, Inmy opinion all the described damage could be attríbutedto causes other than hominid bludgeoning or feeding ac­tion, but 00 the basis of these particular specimens thequestion wilI never be resolved.

Association wíth Stone Artifacts

No stone artifacts have been described from Member 4 atSterkfontein, although a single chert ftake carne lo lightduring the preparation of an Australopithecus maxilla,STS 70. The ñake had c1early been detaehed before beingenclosed in the breccía, and its edges were fresh andsharp. Although it is tempting lo conclude that this ftakeresulted from purposeful hominid activity, it could havebeen detached as a result of a roof-fall.

An interpretation of the Member 4 bone assemblage interms of accumulating agencies will be attempted inchaprer 13.

Clues lo lbe Interpretation of lbe Member 4 FossilAssemblage

The skeletons involved in this assemblage are very in­complete, and the disappearance of many parts beforefossilization needs to be accounted foro We must considerthe possibility that the animals came voluntarily to thecave lo die, or that they fell aecidentally to their deathsinto a natural trap. If Ibis had oecurred, and if the carcas­ses had been protected from scavengers inside the cave,reasonably complete skeletons could have been expected.This, however, is not the case, and other possibleagencies wilI therefore be considered.

Possible Porcupine Collecting lnvoívement

If the presence of gnaw marks on bones is accepted as acriterion for collecting by porcupines, then the fact thatonly 1 bone out of a total of 1,891 shows results of gnaw­ing seems conclusive. As mentioned earlier, the fossilbones are generally not in good condition, and the gnawmarks could have been missed in some cases. However,as was discussed in chapters S and 8, active porcupineinvolvement in abone accumuJation typically results in50% or more of the pieces being gnawed. I have no hesi­tation in saying that the original evidence of gnawing onMember 4 bones appears to have been low and lhal por­cupines played an insignificanl part in lhe collecling.

212 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

,I ,, "...., ...

, II , •, ft ' ,r , •

f '", "',1,\,,

"' ... , ....b....--."a

Fig. 173. Sterlcfontein specimens on the basis of which cJaims of interpersonal violence have been made: (a) Aus­IroJopilhecus endocast 2 showing ilS depressed fracture; (b) endocasr 3 with arrows indicating lines of fracture ordísplacemenl.

, .., ...'... .

"'1·•••• •, ......, .'" ,"'

''': .""'-.',

Fig. 174. Dorsal aspecl of (he Auslroiopilhecus endocast (PIes. 2) rrom Sterkfontein Member 4. The black arearepresenrs the depressed fracture. (o) Distorted forro as when found; (b) with distortion corrected.

r Sterkfontein 213

Possible Hominid Invotvement

As 1 mentioned earlier, no stone or bone tools have yetbeen described from Memher 4. Damage 00 3 australo­pithecine craniums and 18 baboon skulls has beco ínter­preted by Dart as having resulted from purposefulhominid bludgeoning or feeding activity, but, for the fea­sons discussed earlier, 1 find it impossible to be SUTe thatthis damage did not result from compression during fos­silization.

00 currently available evidence 1 am inclined to dis­eount hominid involvement in the bone accumulationprocess.

Possible Camivore Involvement

The extremeIy fragmentary nature of the fossil bones, andthe faet that tooth marks may be observed on some ofthern, means that we must consider camivore involve­ment in the accumulation. As 1 mentioned earlier, posi­tive evidence oí tooth marks on the bones is not abun­dant, but this paucity of evidence can probably be attrib­uted, at least in part, to two causes: my conservativeapproach to the recognition oí such evidence, and thepoor condition of the fossils involved. Had the fossilsbeen in better condition, 1 have no doubt that the in­cidence of recognizably damaged bones would have beenhigher. Evidence summarized ín chapter 8 suggests that,in bone accumulatíons resulting from leopard feeding only30% or fewer oí the pieces may show undoubted toothmarks. Results oí camivore chewing are often more am­biguous than those oí, for ínstance. porcupine gnawing.

The possibility of carnivore involvement in the Member4 accumulation is strengthened by the faet that 8% of allthe individual animals represented are camivores and thecarnivore/ungulate ratio is 28158 or 48.30/o-a remarkablyhigh figure for abone accumulation in an African cave.

Camivores identified .among the fossils consist of aleopard, 4 Dinofeíís cats, a Megantereon, 4 huntinghyenas(Euryboas), I spotted hyena(Crocuta), and 6 Jack­als. Wilb the probable exception of the jackals, aII theother carnivores could well have contributed their foodremains to the bone accumulation.

The Member 4 fossil assernblage is remarkable chieflyfor the great preponderance of primates represented.They aecount for 69.8% of all the animals identified in thesample, A special explanation has lo be sought for theirpresence, and some possibilities are explored in chapter13.

In her study of the fossil Bovidae, Vtba (1975, 1976a)pointed out that the Member 4 antelopes were mostly ofmédium to large síze, while the strong representation ofjuveniles suggests the results oí primary predation ratherthan scavenging. To aceount for the very large bovidsrepresented, she suggested that "the major predatorswere the false and true saber-tooths, and possibly other,as yet undiscovered cats oí suffieient size andJor large­prey adaptation" (Vrba 1976a, p. 67).

Although I agree with this proposition, I would notexclude smaller social carnivores, such as the Euryhoashunting hyenas may well have been, as likely hunters oflarge prey as well as likely accumulators of bones in thecave.

It is unfortunate that the artificial collecting bias re­ferred to earlier has prevented us from obtaining a true

sample of posteranial bones from Member 4. Had such asample been available, it might have been possible toidentify the camivores involved with greater certainty.Food remains of the true saber-toothed cats, for instanee,would presumably have consisted of relatively un­damaged skeletons, whereas those oí the hyenas wouldhave suffered a great deal more.

Sterkfontein Member 4: Microvertebrate Component

Since very little microfaunal material definitely attributa­ble to Member 4 is available in the Transvaal Museumcollections, [ will present no new analysis here. When deGraaff (J960a) undertook his preliminary investigation ofmammalian microfauna from the Transvaal caves, heidentified a number of species from breccia blocks foundin the vícínity of the Type Site. Sorne of these blocks arelikely lo have been derived from Member 4; othersperhaps come frorn Member 5. where the microvertebrateconcentration is greater. The speeies listed by de Graaffare as follows (for details of these fossil animals, seechapo 9): Chlorotalpa speíea Broom, Elephantulus langiBroom, ?Mylomygale spiersi Broom, Myosorex robinsoníMeester, Suncus Hemprich and Ehreoberg sp., My­sotromys bausíeitnerí Broom, Tatera cf. brantsi (Smith),Dasymys cf. incomtus (Sundevall), ?Arvicanthis Lessonsp., ?Pelomys ci.fallax Thomas and Wroughton, Rhah­domys cf. purniíio (Sparrman), Aethomys cf. namaquen­sis (Smith), ?Mastomys ef. natalensís (Smith), Leggadacf minutoídes (Smith), Dendromys cf. mesomelas(Brants), Palaeotomys gracilis Broom, and Cryptomysrobertsí Broom.

Remains from Member 5: Macrovertebrate Component

The collection consists of 1,202 fossils from a mínimum of41 individual animal. as detai1ed in table 85. Particulars of317 bones of bovid origin, assigned to size elasses but notspecifically identified, are given in table 86. The variousanimals identified are depicted pictorially in figure 175.

Data on the fossil material, by which the various animaltaxa are represented, will oow be presented.

Class MarnmaliaOrder PrimatesFamily Hominidaecr. Horno sp. (or Austraiopithecus africanus)Material:

SE 255, fragmentary maxilla of ajuvenile with par! ofdm', dm", and crown oí MI. Age estímate, 6 ± 1 years,

SE 1508, isolated right M'. Age estímate, adu1t.SE 1579, part of the crown of left M' or M'. Age

estímate, indeterminate.SE 1937, isolated left C. Age estimate, adultoSE 2396, lingual half of left P'. Age estimate, adoles­

eent.SE 2398, tooth fragment (specirnen rnissing). Age

estimate, indetenninate.

Family CercopithecidaeGen. et sp. indet.Material:2 pieces from a single individual.

Maxillary fragment with unerupted molar, SE un­numbered; par! of an isolated canine, SE 350.

214 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Order CarnívoraFamily FelidaeCf Panthera leoMaterial:3 pieces from a single individual.

Isolated left P, SE 15000; left M" SE 1500b; right distalhumerus, SE 524.

Cf. Megantereon sp.Material:3 pieces, probably Crom a single individual.

Complete right radius with pathology of the distal end,SE 670; right calcaneus, SE 1692; left distal tibia, SE1832.

8T-5

Family HyaenidaeProtetes sp.Material:A single specimen.

Piece of right mandible, SE 562.

Family CanidaeCanis ef. terblancñeiMaterial:A single specimen.

Left maxillary piece with canine and alveoli, SE 125.

Carnivore indet.Material:

HolnO ~.

3Cen:oplthecold Indet.,

~ II/C. Iff .,.CMls cf terblanchel Prot.l.. IIP.cf PIonthefa '" el Meg.nl ••_ 'P. t

1, •

lkllfl lnctet.1

ef ConnoChnt.. .p.,

~~Makllpenl. cf b.OOtni

T..,rotraou- cf or'JI.1

H,.triJl. cf If,lc:_stnli.,

A_ índet.1

F"1I. 17S. Animalsrepretented by theCossils inthe macrovertebnlte component oftlle Steztiontein Member 5 samplc.Minimum numbers of individuals are indicatcd.

12 pieces from a minimum of 3 individuaIs.Maxillary piece, SE 1154; mandible pieces, SE 56.

1629; tooth fragments, SE 1172. 1477. 1883; humerospieces , SE 1075, 1627; metapodial pieces, SE 1520a.h.

Sterldontein

Makapania cf. broomiMaterial:A single specimen from an adult individual.

Par! of a left M", SE 1425.

215

Order ArtiodactylaFamily BovidaeDama/iscus ef. dorcasMaterial:4 pieees from a minimum of 2 individuals (t juvenile,adult).

Left maxilla with dpm":", SE 1185; Ieft mandible piece,M,_" SE 1728; isolated rigbt M', SE 1218, 1770.

Domalisrus cf. sp. 2Material:7 pieces from a minimum of 4 adult individuals,

Left maxilla with P', M'-', SE 794; rigbt maxilla withMH, SE 588; left mandible piece, P"M, and rigbt M"SE 1233; left mandible piece, M.-" SE 1614; left M', SE1334; rigbt M', SE 1381; rigbt M" SE 1754,

Damaliscus sp. 1. or Parmularius sp.Material:A single specimen from an adult individual.

Left mandible piece, Ps-M" SE 192.

Medium-sized AleelaphinesMaterial:7 pieces from a mínimum of 5 individuals (t juvenile and 4adults),

Rigbt mandible piece, M,_" SE 464; rigbt mandib1epiece MH , SE 1763; isolated left dpm¿ SE 2133; rigbtM" SE 1828; right M, and M" SE 627; right M" SE 1424;rigbt M" SE 535.

Cf. Connocñaetes sp.Material:A single specimen from an adult individual.

Isolated rigbt M" SE 2601.

?HippotraginiMaterial:A single specirnen from an adult individual.

Part of an isolated left M" SE 1125.

Antídorcas ef. reckiMaterial:5 specimens from a mínimum of 3 individuaIs (1 juvenileand 2 adults),

Left mandib1e piece, dpm¿ M,_., M, unerupted, SE1258; left .mandible piece, p._" SE 125; left mandiblepíece, P..--M" SE 535; left mandible piece P..--M" SE1855; right mandible piece, dpmc.¿ SE 1313.

Oreotragus majarMaterial:3 pieees from a single adult individual.

Crushed palate with left P'_M'; rigbt P'-M'; parts ofboth mandibular rami, left P,-M,; right M._" M 8361a,b(ex Bemard Price Institute).

Taurotragus cf. oryxMaterial:A single specímen from an aduIt individual.

Isolated left M" SE 196.

Antelope size class 1Material:1 cranial and 19 postcranial pieees from an estimatedmínimum of 3 individuals, as listed in tabIe 86.

Antelope size class 11Material:29 eranial and 122 postcranial pieces from an estimatedminimum of 6 individuals, as listed in table 86.

Antelope size class 111Material:32 cranial and 108 posteranial pieces from an estimatedminimum of 5 indivlduals, as listed in tabIe 86.

Antelope size class IVMaterial:1 eranial and 5 postcranial pieces from a minimum of 2individuals, as Iisted in table 86.

Family SuidaeSuid, gen. et sp. indet.Material:2 pieees from a single individual.

Mandible piece with M" SE 1069; tooth fragment, SE1624.

Order PerissodactylaFarnily EquidaeEquus ef. burclteíliMaterial:18 eranial and 5 posteranial pieces from a minimum of 3individuals (1 juvenile and 2 adults).

Isolated left P', SE 2752 (also numbered 1602); isolatedtooth fragments, SE 3, 63, 74,116,189,214,245,274,401,687, 1246, 1382, 1391, 1628 1629, 1855. 2397; proximalfemur piece, SE 1440; articulated distal femur and prox­imal tibia, SE 1605; distal tibia piece, SE 1795; carpalbone, SE 100.

Order HyracoideaFarnily ProcaviidaeProcavia sp.Material:3 pieces probably from a single individual.

Craníal fragments, SE 744; part of an isolated incísor,SE 1655; distal humerus fragment, SE 723.

Order LagomorphaFamily LeporidaeLagomorph, gen. et sp. indet.Material:A single specimen.

Parts of a mandible, SE 222.

Order RodentiaFamily HystricidaeHystrix ef. ofrícaeaustralísMaterial:6 cranial and 1 possibJe posteranial pieee from a minimumof 2 individuals (1 juvenile, 1 adult).

2J6 Fossil Assemblages from the Sterkfonteín Valley Caves: Analysis and Interpretation

Isolated teeth and tooth fragments, SE 411, 1249, 1255,1262, 1263. 1269; left calcaneus , may be porcupine , SE2024.

Clasa AvesBird indet.Material:] craniaJ and J postcranial piece, possibly from 1 individ­ual.

Shattered cranium, SE 1342; proximal ulna piece, SE2225.

Class ReptiliaOrder SquamataFamily VaranidaeVaranus ef. niíotícusMaterial:A single isolated vertebra, SE 817.

Order CheloniaFamily TesludinidaeGen. et sp. indet,Material:A single carapace piece, SE unnumbered.

Sterkfontein Member S: The Major Features of the BoneAccumulation

The Composition of the Fauna Represented by theFossils

The information provided in table 85 and depicted infigure 175 rnay be further summarized. This is done in thesecond columns of table 83 and of figure 171. The overal!composition of the preserved fauna in Member 5 differsmarkedly from that of lbe underlying Member 4. Whereasprimates dominated in the Member 4 assemblage, ar­tiodactyls are by far the most important group in Member5. In fact, while 56% of the individual Member 4 animalswere cercopithecoids, the same percentage of Member 5indivhíuals are bovids. The dífference between the twoassemblages clearIy results from a fundamental change inaccumulation pattern between the two members, a matterto be discussed in chapter 13.

The few primate remains present in the Member 5 col­lection came from 3 hominid individuals and 1 baboon.Tbe hominids, 2 adults and a child, have been variouslyclassified as Australopithecus and as Horno. and the ba­boon remains are loo fragmentary lo allow firmidentification.

Camivores are represented by a single lion, a Megan-. lereonlike sabertoolb, an aardwolf, and ajackal. The 22

bovid individuals may be placed into siz.e classes: class 1,represenled by 3 adull klipspringers, OreolTagus mqjor;class II by 10 individuals, including 2 juveniles, ofDamaliscus, Parmularius, andAntidQrcas,· dass III by 6individuals, including 1 juvenile. of medium·sizedalcelaphjnes, includingConnochaeles; class IV by 3 adullindividuals that ¡nelude an eland and a Makapania.Among the artiodactyl fossils are remains of a single un­idenlified pig.

The rest of the animals made a small contribution to theoverall faunal pieture. Of interest is the presence ofmonitor lizard (Varanus) and tortoise remains-animalsthat feature consistently in the diet oC Stone Age hunteTpgatherers of more recent times.

The Representation 01 Skeletal Parts

Must oí the animals are very poorly represented and havebeen identified on HUle more than scraps. Yet this is a truereñection of the position in Member 5, since the sampleresulted from systematic excavation by Robinson andfrom careful acetic acid preparation of many of thespecimens. The only animals whose remains are rea­sonably abundant are the bovids (table 86). In size elassIl, 10 individuals contributed 151 bones, and 6 class IIIanimal s are represented by 140 pieces. The skeletal partstable shows that there is a scatter of fossils from aU partsofthe skeleton, ranging from the head lo the foot borres.This does not necessariJy imply that the animaIs werebrought back whole to the cave, but rather that a widerange of parts, at any rate, found their way there.

Observed Damage to the Bones

From the point ofview ofestirnating damage to individualbones, the Member 5 sample is mueh more satisfactorythan the Member 4 equivalenl. The fossils had beencarefully excavated and prepared so that damaged edgesand surfaces couId be studied with more definitive results.Details are provided in table 87.

Porcupine and SmalJ-Rodent Gnowing. Fifteen bonepieces showed clear evídence of porcupine gnawing, and17bore small-rodenl loolh marks. 1am not able to suggestwhat species of rat or mouse was responsible for the latterrnarks, though it appears to have been an animal about thesize of a gerbil, An excellent exarnple of a rodent-gnawedbone from Member S has been figured by Robinson(1959).

Carnívorc-Lnfiicted Damage_Clear evidence of carnivorechewing was noted on 23 bovid bones (see table 87 fordetails) as well as on 14 bone f1akes and fragments, Prob­able cemívore damage was seen on 7 more specimens.

Traces 01 Artificial Bone Alteration, Abone tool,fashioned on a flake and bearing two worn facets (SE 612,ñg. 176a) came from the Member 5 excavation and wasdescribed by Robinson (1959). To my mind there is Iittledoubt that this is a genuine artifact as Robinson hasclaimed and that the facets were worn through use onskins or other similarly soft surfaces. A second boneñake, SE 1000, has wom edges that may have been artiñ­cially caused. A third specímen, SE 1524, is shown infigure 176b; it appears lo be a piece of a bovid horn-core,threc surfaces of which have been wom lo a taperedpoint. lt may well have becn employed for lbe same pur­pose as SE 612.

A specimen showing convincing evidence of hominidaction js SE 1729 (Iig. 176c), lhe Sharl of a small bovidhumeros bearing three cIear chop marks that were pre­sumably caused by the edge of a stone too1.

Several specimens show oval openings up to 2 em indiameler in bone shafts. These are ralber larger than lhepunctate marks usually eaused by carnivore teeth andmay have resulted from hominid use of pointed objecls.

Association with Artifacts

The extension site excavation conducted by Robinsonduring 1957-58 yielded 284 stone artifacls (Robinson1957, 1962). These were in silU in Member 5 and directiy

Sterkfonteín 217

earnívores-a lion, a saber-toothed eat, an aardwolf, anda jackal, while the camivore/ungulate ratio proved to be4126, or 15.4. Despite sueh pointers to mínima! earnivoreinvolvement, at least 37 bones show c1ear carnivore toothmarks, indicating that carnívores were undoubtedly in­volved in the history of sorne of the bones.

From her study of the bovíd fossils, Vrba (l976a) con­c1uded that, on account of the low juvenile percentageand the distribution of antelope indívíduals through aHweight classes, the assemblage was predominantly a non­primary, scavenged one. She suggested that, on the evi­dence of the bovid food remains, the Member 5 homínídswere scavengers rather than effective hunters. In the lightof evidence 1 have assembled sinee Elisabeth Vrba wrote,1 have no reason to dispute her conclusion, implicationsof whích are discussed further in chapter 13.

Remains Crom Member 5: Microvertebrate Component

T. N. Pocock (1969) provided a list of microfauna pro­visionaUy identífied from sieved material recovered fromSterkfontein dump 8. Aeeording to the síte plan given byTobías and Hughes (1969), dump 8 was a long, narrowone extending along the northem margin of the West Pitand Extension Locality; it is thus almost certainly derivedfrom Member 5 breccia.

The fol1owing taxa were identified by Pocock;minimum numbers of individuals are given in par­entheses: Elephantulus langi (31); Myosorex sp. (9); Sun­cus er. orangiae (6); Suncus cf. gracUis (4);Chrysochloridae (postcranial material); Rhinolophus er.darlingi (2);R. er. augur (1); VespertiJionidae (postcranialmaterial); Leporidae (1); Mystromys hausleitneri (286);Tatera sp. (31); Dasymys sp. (2); "Rhabdomys minor"

ba

Possíhle Hominid Involvement

The 1,202 bones in this sample were directly assocíatedwith 284 stone artifacts and at least 2 bone tools. Thisimp!ies a ratio of 4.2 bone pieees to 1 stone artifact. Atleast 1 bone piece shows indisputable chop marks causedby a sharp-edged tool.

The remarkable density of artifacts in the excavatedpart of Member 5 strongly suggests that the cave wasintensive!y occupied during this accumulation phase. Itwould therefore be remarkable if the bone pieces as­sociated with the artifacts did not represent hominid foodremains.

Clues to the Interpretation oC the Member 5 FossilAssemblage

Possible Porcupine ColJecting Involvement

The porcupine-gnawing íncídenee is 15 bones out a totalof 1,202, or 1.3%. By this criterion poreupines may beexcluded as sígnificant collectors ofthe Member 5 bones.

assoeiated with the bones under diseussion. They havebeen deseribed by Mason (l962h) and by M. D. Leakey(1970) and consist mainly of eore or ehopping tools with­out the flakes, which must have been detaehed elsewhere.

Only 4 of the 41 animals represented in the fossíls were

Possihle Carnivore Involvement

t '"

Fig. 177. A selection oflhe very numemus stone artifacts from Slerlcfon­teín Member 5. eore and chopping tools predorrunate, the ftaJc:es fromwhích do not appear lO have been detached in lhe cave.

{ .. ­

1'.

Fig. J76. Examples of bones fmm Slerkfontein Member 5 showing eví­dence of artificial a1teration by homínids: (a) abone 100\ described byRobínson (]959) in whích facels have been wom on abone ftaJc:e, SE 612;(b) a piece ofbovid hom-core, SE 1524, wom lo a poínl; (e) Ihe shaft of asmaJl bovid humeros, SE 1729, showíng chop marks presumably causedby the edge of a slone too1.

218 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

~.nl.lope el_ 1/1,

H~~'''' '"~,..

Pr_loI el.-,

~~, ~ ~~ .,. __ Antelope ea- I

t ,

Fig. 119. Animal! represented by (he fossüs in tbe macrofaunal compo­nent 01 the Stcrkfontein Member 6 semple. Minimum numbers of individ­uals are indicated.

PEtllCEltTAGE AEPtIIESENTATtON

Fig. 178. Thepcrccntaae representation oCanimals ol varíous kinds in themicrovertebratc component al the Sterkfontein Member 5 fossil sample;644 individual animals are invclved, and lile semple is domlnated byshrews. dendromtUines. and tbe whíte-taííed rat, Mystromys.

Antelope size class IIIMaterial:4 eranial and 15 posteranial pieces from a minimum of 4indivíduals as listed in table 90.

Order ArtiodaclylaFamily BovidaeAntidorcas bondiMaterial:3 craníal píeces from a mínimum of 2 adult individuals.

Right mandible piece, M,_" SE 690; isolated 1eft M',SE 829, 875.

Antelope size class IIMaterial:3 cranial and 34 postcranial pieces from an estimatedrninimum of 4 individuals, as listed in table 90.

Damaliscus cf, dorcasMaterial:A single specimen, right maxi11a, P'---M', SE 1318.

Remains from Member 6: Macrovertebrate Component

The sample discussed here is derived from Robinson's1957-58 Extension Site excavatíon. On the basis of theappearance of the enclosing matrix it is fairly easy toseparate fossils derived from members 5 and 6. This sep­aration has been made, and the sample consists of 454fossils from a mínimum of 19 individual animals, as listedin table 89. The various identified animals are depicted infigure 179.

Particulars of 66 bones of bovid origin, placed in sizeclasses bUI nol specifically identified, are given in lable90.

Details of the fossil material by which the various ani­mal taxa are represented will now be presented.

Antelope size elass 1Material:10 postcranial pieces from a minimum of 2 individuals, aslisted in table 90.

Phylurn ChordalaClass MammaliaOrder CarnivoraCamivore indet.Material:3 cranial and 6 poslcranial pieces from an estimatedminimum of 3 índíviduals.

Mandible piece, SE 2104; tooth fragments, SE 2290,2329; distal humerus, SE 864; ?libia shaft, SE 2059;metapodial pieces, SE 1140,2062,2095; phalanx, SE 73\.

(6); Aethomys cf, ehrysophi/us (2); Praomys (Mastomys)cf. natalensls (8); Praomys sp. (6); Mus. cf. musculus(2); Dendromys cC. melanotis (36); Malacothrix sp.(19); Steatomys sp. (6); Otomys cf. gracilis (43);Graphiurus sp. (1); Cryptomvs cf. holosericeus (30); un­identiñed remains of frogs, lizards, and birds.

In the course of his Extension Site excavations of1957-58, Robinson recovered samples of breccia rieh inmícrofaunal remains. These carne from the upper levels ofMernber 5 breecia on the western side of the West Pit.Sorne of this breccia in the Transvaal Museum collectionwas prepared by the acetic acid method in the course oftbis study. The sample was found to contaín remains froma minimum of 644 individual animals as listed in table 88.The faunal composilion of Ibis sample is presenled pícto­rially in figure 178. As is apparent, the fauna is dominatedby Mystromys, dendromurines, and shrews.

arder PerissodactylaFamily EquidaeEouus cf. burcheííiMaterial:6 cranial pieces from a minimum of 2 individuals.

Isolated tooth fragmenls, SE 686, 693, 704, 794. 818.unnumbered.

Order HyracoideaFamily ProcaviidaeProcaviu d. capensisMaterial:3 cranial and 2 postcranial pieces from 1 individual.

Left mandible piece, SE 806; isclated teeth, SE 928,1050; radius, SE 995; pelvis piece , SE 703.

arder RodentiaFamily HystricidaeHystrix africaeaustralisMaterial:2 fragments of incisor, probably from a single individual,SE 92[, 994.

Class Avesarder FalconiformesFamily AccipitridaeBird, Iarge raptorMaterial."2 postcranial pieces, probably from a single individual.

Piece of an ulna, SE 1888; isolated cIaw, SE 1.

The Major Features of the Bone Accnmulation

The Composition of the Fauna Represesued by theFossils

The information provided in table 89 and depicted infigure J79 may be further summarized. This is done in thethird columns of table 83 and figure 171. Member 6 is arather insignificant stratigraphic unit representing the in­fillingof a low space between the top of Member 5 and theoverhanging dolomite roof. Of the 454 bone pieces from itin the present sample, 290 were bone ftakes. The restwere found to come from a minimum of 16 individualanirnals, 8 of which were bovids. Among these, onlyAntldorcas bondi and Damaliscus cf. dorcas wereidentifiable.

Representation 01 Skeletal Parts

A sample of this size cannot be expected 10 provide muchinfortnation about skeletal part representation. As table90 shows, the 8 bovid individuals were represenled by 59postcranial pieces, including parts from the lower Iimbsand feel. A far larger sample would be required befareconclusions could be drawn about butchery practices orcarnlvore feeding pattems.

Sterkfontein 2[9

Observed Dumage to the Rones

As with bones from other site unit s, cach piece fromMember 6 was examined for diagnostic damage traces.Results were as follows (table 91):

Porcupine and Small-RodentGnawing. Eight bone piecesshowed clear evidence of porcupine gnawing, and 4 boresmall-rodent tooth marks.

Camtvore-lnfíícted Damage. Clear evidence of carnivorechewing was noted on 3 bovid fragments and 5 otherfragments. Probable carnivore damage was apparent on 2other pieces.

Traces of Artificial Bone Al/era/ion. Cut marks werenoted on a mandible of Antidorcas bondi (SE 690), and 3bone flakes showed wear and rounding that may havebeen artificially induced.

Assocíatíon with Artífacts

No artifacts are known to have been definitely associatedwith the Member 6 bones.

Cjees to the Interpretation of the Member 6 FossilAssemblage

Possihle Porcupine Collecting Involvement

Eight bone pieces out of 454 showed evidence ofporcupine-gnawing, a percentage of 1.8. By this criterionporcupines may be excluded as significant contributors tothe collection.

Possíble Homínid Invoívement

No artifacts have yet been found in Member 6, but at leastone bovid mandible shows cut marks that were certainlycaused by sharp-edged lools. Thal 290 of the 454 bonepieces were bone flakes is strongly reminiscent of thesituation in Stone Age human food remaíns.

Possible Camivore tnvolvement

Parts of about 3 unidentifiable carnivores were found,which means that the carnivore/ungulate ratio is 3/10, or30.0%. Clear evidence of camivore chewing was ob­served on 8 bones.

The sample is too sma11 to allow firm conclusions, but itis obvious that both hominids and camivores have beeninvolved to sorne extent in the taphonomic process.Perhaps further exeavation of Member 6 breeeia wilIelucidate the situation.

Remains from Member 6: Microvertebrate Component

No microvertebrate remains from Member 6 were avail­able al the time of the study.

11 Swartkrans

A Brief Histo.-y of Activity

During 1948 the University of CaJifornia's African expe­dition was searching for australopithecine fossils in theTransvaal. In September of that year the expedition'sleader, Wendell Phil1ips, approached Broom and offeredfinancial help for exploiting a new fossil locality. At thetime Broom and J. T. Robinson were excavating in theSterkfontein Type Site, but they agreed to move part oftheír work force across the valley to Swartkrans, wherethe presence of bone-bearíng breccia had been known forsorne time. The quarrying of both sites was placed underthe supervision of Mr. van der Nest.

The Swartkrans 10cality was ímmediately productíve,and within the first week the mandible of what appearedto be a new species of robust australopithecine, Paran­thropus crassidens Broom, had come to light (fig. 193).During the subsequent months many valuable finds weremade, including the jaw of an early fonn of true man,found by J. T. Robinson on 29 April 1949. This wasnamed Telanthropus capensis Broom and Robinson butsubsequently was transferred to Horno (fig. 187).

By March 1949 most of the California expeditíon mem­bers were back in the United States, and no further fundswere forthcoming from this source. The Swartkrans ex­cavation was neverthe1ess continued throughout 1949, at

Fig. 180. Searching for fossíls among blocks oC breccia in 1948. Broom may be seen in [he background. Photographby J. T. Robinson.

220

Transvaal Museum expense, until financial probJemsbrought the work to an end in November. By this lime asubstantial seam of pure travertine (fig. 181) had beenexposed along the north wall of the excavation. This at·tracted the attention of a local lime miner named Fourie,who started blasting at Swartkrans in December 1949, tothe dismay of the Transvaal Museum tearn, which waspowerless to stop it. The mining continued throughout1950 and 1951 and resuJted in the finding of sorne spec­tacular specimens, such as the well-known cranium SK 48(fig. 144) and the beautiful mandible SK 23 (fig. 194).

Robert Broorn died in Pretoria on 6 ApriJ 1951, butRobinson was abJe to recomrnence paleontological workat Swartkrans with financial assistance from the NuffieldFoundation. This continued into 1953 and yielded manyvaluable specimens. Thereafter the site was abandonedfor twelve years.

Active research was conducted at the TransvaalMuseum between 1953 and 1957 on various aspects ofthefauna and geology of Swartkrans and other SterkfonteinvaIley caves. This resulted in many papers by variousspecialists, as well as two museum memoirS-<:ln the den­tition of the Australopitheeinae (Robinson 1956) and onthe geology of the cave deposits (Brain 1958).

In April 1965 paleontological work was resumed atSwartkrans, with generous support from the Wenner­Gren Foundation for Anthropological Research in NewYork and, subsequently, from the South African Councilfor Scientific and Industrial Research (Brain 1967a). Theimmediate aim of this work was to restore order after thechaotic mining episode of 1950-51, but it soon becameapparent that 1951 was not the first time the site had beenmined. An extensive dump on the hillside below the cavecontained thousands of tons of rock blasted from the sitebefore paleontological work started there in 1948. Thiswas unexpected, since the site had shown littie obviousevidence of such aetivity. It was only when attention wasgiven to clearing the cave itself ofrubb1e that the extent ofthe first mining operations becarne apparent. Not onlyhad considerable quantities of rock been blasted out, butthe quarry had been back·filIed with rubble before theminers left it.

Local residents were asked about the identity of thefirst miner, and firsthand information was provided by M.Bolt, owner of Bolt's Farm adjoining Swartkrans. Shewas able to teH us of a prospector named Knowlan who

Fig. 181. The very large stalagmitic boss thal originaUy separated theOuter Cave from the Inner Cave al Swartkrans. A hominid cranium isvisible in the breccia adjacent lo the stalagmile.

Swartkrans 221

had been active in the Swartkrans area during the early1930s and who had also worked at Taung. An aged Afri­can was then found who had actually been employed byKnowlan during the Swartkrans quarrying, probablyabout 1932. He remembered the back-filling after theblasting, which enabled the miners to ¡aya cocopan trackacross the main outer cave area when they turned theirattention to the deep shaft at the eastem end of the site.

The project of removing and sorting the miner's rubblehas been extremely tedious, involving seven years oflabor (Brain 1973). It has resulted in the finding of manyfossils, none of them, regrettably, in situ; moresignificantly, however, it has clarified the structure of thecave and the stratigrapruc relationships of its filling.

A comprehensive study of the fossil bovid remainsfrom Swartkrans was undertaken between 1969 and 1973by E. S. Vrba (1975, 1976a). This contributed greatly to aproper understanding of the age relationships between thebreccias in the Outer Cave. Previously it had beenthought (Brain 1958) that most of the Outer Cave fillingwas formed by breccia of a single time unit. Gradually,however, it became apparent that at least two depositswere involved, differing greatly in age and faunal compo­sition (Brain 1976a). These were formaIly named mem­bers 1 and 2 of the Swartkrans Formation by K. W.Butzer (1976), who had been undertaking detailed strati­graphic studies at the site for sorne years.

The University of Witwatersrand acquired theSwartkrans farm and cave site in 1968. Through thegenerosity of the Bemard Price Institute of the Univer­sity, it has been possible for Transvaal Museum opera­tions to continue at Swartkrans with gratifying results.

Although the first evidence for co-existence of an earlyform of Horno with robust australopithecines atSwartkrans came to light in 1950 (Broom and Robinson1950), further significant information was forthcomingtwenty years later (Clarke, Howell, and Brain 1970).After R. 1. Clarke's assembly ofa composite cranium (fig.192) it was possible to confirm that the two hominids hadcoexisted during Member 1 times.

Current work at Swartkrans aims to define the extent ofthe breccia mass in the Outer Cave and 10 elucidatestratigraphic relationships within this. Carefully con­trolled excavation is for the first time providing speci­mens whose provenance is accurately known. Swartkranshas suffered from the disadvantage that, although numer­ous hominid speeimens have come from the site, very fewof them can be accurately positioned within the cave andits stratigraphy.

Since the manuscript of this book was completed, newexcavations have shown that the eastem end of the OuterCave is extensive, containing a major sediment accumu­lation tnat entered through a shaft aboye the southeastemcave wall. l'he main Member 1 sediment mass enteredthrough an opening aboye the north wall, and the re­lationship of the two bodies is now being studied.

Sorne Notes on the Site

The Swartkrans cave eomplex is on the southeastem as­pect of the Swartkrans hill al an altitude of 1,480 m. Atrigonometric beacon on the crown of the hill stands at1,497 m, and the leveI of the Bloubank streambed belowthe cave is 1,454 m. The hill líes in a zone where intensefaulting has occuned; many of the outcrops are of ashatterrock consisting of angular fragments of chert;

222 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

produced by the faulting, recemented in a dolomític ma­tríx. Making use of a large-scale aerial photograph, thenature of the rock outcrops are plotted, particularlywhether they consisted mainly of undisturbed dolomite,undisturbed chert, or shatter rock. It was found that widebands of fault-shattered rock traverse the hill, but thecave system has developed in a thick dolomitic band,comparatively free of chert bands. The aerial photographshowed two lineaments in the country rock intersecting atapproximately right angles, and the cave system has de­veloped close to the intersection of these two lines. Twonatural shafts-the "eastem shaft" within the Swartkranscave system itself and the "southwestem shaft" about 80m distant-are both aligned with the lineaments.

The Swartkrans cave system consists of what have be­come known as the Outer, Inner, and Lower caves. TheOuter Cave has no roof and is elongated east-west; itstretches for about 45 m as shown in figure 184. lts north­em and western walls are clearly delimited by the out­cropping dolomitic country rack, but for most of its length

the southern wall is undefined. Here the dolomite wall isnol apparent on the surface, and further excavalion isbeing undertaken to establish its posítion.

The Outer Cave is separated from the Jnner and Lowercaves by the "f1oor bJock"-a vast piece of dolomite thatseparated from the roof at the western end of the originalsolution cavem and divided the space there into threeseparate volumes. The f100r block hinged downward fromthe northem wall of the cavem so that the gap betweenits upper surface and that of the remaining roof is greateston the southem side-a vertical measurement ofabout 3 m.The block was extensively cracked by its fall, and thecracks have been filled with white travertine. In the areaseparating the Inner Cave from the Outer Cave, a verysubstantial boss of travertine was built up (fig. ]81) be­fore the first direct opening of the cave to the surface.This aH but sealed the connection between Ihe two caves.

The Inner Cave retains its dolomitic roof, which is asmuch as 10 m thick. Downward, the Inner Cave leadsbeneath the floor block to the Lower Cave, typically low

Fig. 182. Looking down on lhe conlents of lhe Ouler Cave al Swartkrans in 1969. Miners' mbble is being clearedfroro lhe outcrops of fossitiferous breccía.

roofed at the westem end but opening toward the east intoa Jess confined chamber, hung wÍth stalactites. Stillfarther to the east the Lower Cave is connected with thesuJiace by an irregular vertical shaft about 9 m deep.

In a cave system Iike the one al Swartkrans, where thevertical depth of a chamber frequently exceeds its width,

\~\.2~ ../~ .~\~': \. ¡

"';¡aL /'''~$~~ .a-~ .

,1.......

Swartkrans 223

drawing an adequate plan is difficult. In the Outer andInner caves, a series of five horizontal lines was paintedaround the cave walls, using a surveyor's leve!. On each¡ine a ¡arge number of points were surveyed, and plots ofeach ¡¡ne were used to build up composite plans and sec­tions.

Fig. 183. Excavations camed oul during 1978 tila! aim to define lhe southeastern lilTÚts ofthe Swartkrans deposit. Inlhe pícture are George Moenda, foreman al [he síte, and Bella, "lhe spiril of Swartkrans."

" -':':.::­=:::::=,~:.--::=-

'=--_..-_..-----

~INNER CAVE

Fig. 184. A plan of [he Swartkrans cave syslem.

o 5metres

10

224 Fossil Assemblages from the Sterkfontein VaUey Caves: Analysis and Interpretatíon

An irregular!y shaped cavern has been dissolved in thedolomite below the level of standing water.Stage 2

The level of standing water has dropped through ínci­sion of the Bloubank River valley in the vicinity of thecave. A large piece of dolomite, destined to become the"fioor block," has become detached, hingeing down fromthe north wall.Stage 3

Indirect conneetion with the surface has led to ventila­tion of the cave; travertines, in the fonn of stalactites anda massive boss of stalagmite over the f100r block, are de­veloping. Planes of weakness in the doIomite aboye the

In reading the discussion that follows, refer to the verticalsections in figure 185. These are reconstructed andsomewhat ídealized sections through the western part ofthe Outer, Inner, and Lower caves, running southeast tonorthwest. They do not take into account the newly dis­covered talus cone fil1ing the eastem end of the OuterCave and penetrating beneath the mass of Member 1 inthe westem end. This deposit is now being excavated,and results ofthe current study wil1 certainly infiuence thescheme presented here.Stage 1 (lig. 185)

Suggested Stages in the Formation of the Cave System

8

iMEMBER 2

GffiillMEMBER 1

mi TRAVERTlNE

.OOLOMITE

o 5! , ! ! , ,

METRES

185. Reconstrucled sections lhrough lhe weSlern end of lhe Swartkrans cave syslem showing suggesled slages ofdevelopmenl.

Outer Cave are being enlarged by solution , leading to thedevelopment of avens.Stage 4

A vertical shaft has developed, Iinking the Ourer Cavewith the surface. It appears to have had suitable ledgesaround it that served as owl roosts. The Outer Cave hasstarted to fill up with bone-bearing sedirnent destined tobeco me Member 1 oí the Swartkrans Formation.Slage 5

Accumulation oí Member 1 sediment has completelyehoked the shaft.Stage 6

Over a long period, perhaps as much as a miIlion years,the calciñed Member I sediment was severely eroded bywater passing through the cave. Progressive denudationoí the hillside above the cave continued.Stage 7

Another shaft has once again linked the Outer Cavewith the surface, this time somewhere close to the cave'ssouthern margino Through the shaft carne sedlment des­tined to beeome Member 2 breccía; it infiltrated theeroded surfaee of Member I and gradually filled the re­maining space in the Outer and Inner caves. The LowerCave remained empty.Stage 8

As time passed, the Outer Cave's roof was erodedaway, exposing the cave filling 00 the hillside. A systemof irregular channels developed through both breccíamembers, serving as stormwater drains from the surfaceto the Lower Cave and thenee to the subterranean res­ervoirs in the dolcmite. These ehannels were partly filledwith later sediments.

Stratigraphy

In the course of an early study of Swartkrans stratigraphy(Brain 1958), researchers appreciated that breceias oftwotypes and ages existed in the cave. These were termed theOuter Cave pínk breccia and the Inner Cave stratifiedbrown breecia. The former, being the source oC the abun­dant australopithecine fossils, was obviously the older,and the two were separated by a clear unconformity.

More recently, on the basis of paleontologieal andgeological evidence (Vrba 1975, 19700, Brain 1976a), itbecame clear that a considerable volume of the OuterCave fiUing consisted of breecia younger than theaustralopithecine-bearing matrix. The more recentbreecia was browner than the older pink equivalenl butgenerally lacked the layering of the "slratified brown bree­cía" previously described from the Inner Cave. Wherecontact between the two Outer Cave breccias could beobserved, it was invariably unconfonnable and often intri­cate, Fortunately it has preved possib1e lo sepárate thetwo breccias faírly easily in the hand specimen.

On the basis of his detailed sedimentological study,Butzer (1976) has fonnal1y designated lbe older andyounger deposils as Member 1 and Member 2 of lbeSwartkrans Formation. Member 1 ineludes the basaltravertine and a breccia mass that was the source oí the"Member 1" fossil assemblage described in lhis book.

At the time of Butzer's analysis, the extensive filling inthe eastem end of the Quter Cave had not been uncoveredor recognized. The unexpected deposit will be describedelsewhere.

Conceming Member 2, BUlzer (1976) noled lhal Ihe

Swartkrans 225

properties ofthe sediment differed somewhat between theOutcr Cave and the Inner Cave. but he did not propaseany stratigraphic subdivisions, He coneluded that theMember 2 material in the Outcr Cave suggested a cavernwith direct access te a doline in process of active en­largement, while the Inner Cave facies of Member 2 rep­resented a "ñltered," sorted variant of the Outer Cavematerial, lacking both the coarsest fraction (trapped in theOuter Cave) and much of the clay (fíushed out tbrougb theLower Cave).

As 1 mentioned earlier, the Outer Cave filling has beentraversed by a series of irregular solution channels. Thesemake their way down from the original surface of theexposed breccia to the Lower Cave. In places the chan­neis are completely or partially filled with sediment invarious degrees of calcífication, and these fillings arelikely to vary in age. A volume of about 2 m] of un­consolidated sedimento undisturbed by lime-mining oper­ations, has been excavated. This yielded a large numberof bone fragments, details of which are given in the sec­tion on "Channel FiU," but no artifacts. The positionsand inclinations of the channel system through the west­em end of the Outer Cave filling are shown in figure 186.

Stratigraphic Relationships of tbe Homo Remains

During the first year of paleontological work atSwartkrans, 1948-49, abundant remaios of robust aus­tralopithecines were found. These were interspersed withsporadic finds of a more advanced hominine, originallydesignated Teíaruhropus capensís Broom and Robinsonbut subsequently transferred to Horno sp. (for particulars,refer to the seetion on Horno sp. in chapo 9).

The original "Telanthropus" find, made on 29 April1949 (fig. 187), was preserved in a browner matrix thatdiffered in appearance from the main mass ofaustralopithecine-bearing breccia. Conceming the man­dible SK 15, together with lwo associaled teeth and prox­imal radius, Broom and Robinson wrole (l949a, p. 322):

Though this [the mandible, SK 15] was discovered inthe same cave as the large ape-rnan, it is clearly ofconsiderably later date. In the main bone breccia ofthe cave deposit there has been a pocket exeavatedand refilled by a darker lype of matrix. The pocketwas of very limíted extent, being only about 4 fl by 3ft, and about 2 ft in thickness. The deposit was re­markably barren, there being no other bones in it ex­cept the human jaw and a few remains of very smallmammals. We are thus at present unable to give anage to the deposit except to say that it must be consid­erably younger than the rnain deposito If the main de­posit is Upper Pliocene, not improbab1y the poekelmay be Lower Pleistocene.

When Robinson (l953b) wrote his paper "Telanthropusand Its Phylogenetie Significance," he described the sec­ond specimen (maxillary fragmenl SK 80) from lhe mainQuter Cave breccia and inclined to the view that aH the"Telanthropus" specimens, including the controversialSK 15 mandible, were contemporaneous with the austra­lopilhecine fossils. Meanwhile, K. P. Oakley undertook aflourine test 00 bone from the SK 15 pocket, and com­pared the results with those on bones from the undoubtedaustraJopithecine breccia. No significant difference inftourine content was found between the two samples,

Fig. 187. The Telanihropus mandible, SK 15, from Member 2 breccia inlhe Outer Cave.

suggesting that any large age difference was unlikely pro­vided the flourine method could reliably be applied incalcified sediments, which has unfortunately not yet beenestablished. Oakley (1954a,b) was of the opinion thatthe "Telanthropus" and robust australopithecine remainsfrom Swartkrans were probably contemporaneous.

Another piece of evidence that contributed to Robin­son's (I953h) suggestion that the SK 15 pocket was of thesame age as the main deposit carne from sediment studiesundertaken at the time (Brain 1958). Grading ofthe SK 15pocket sediment was found to be broadly similar to that ofthe adjoining breccia mass, although the carbonate con­tent was found to be abnormally low.

When the significance of the calcified channel fills in theSwartkrans Outer Cave was first appreciated in 1974(Brain 1977), 1 speculated that the SK 15 pocket couldwell have represented part of such a filled channel sys­temo Shortly thereafter, however, 1 reached the conclu­sion (Brain 1976a) that the Outer Cave did in fact containa large volume of brown Member 2 breccia that, in thecontact zone, had iniiltrated the eroded surface of theMember 1 mass.

It is unfortunate that the "Telanthropus" pocket nolonger exists and cannot be repositioned in the cavestratigraphy as we now understand il. My current opin­ion, however, is that the SK 15 mandible and associatedfragments should be allocated to Member 2, while theother remains of Horno sp.-that is, the mandible pieceSK 45, the composite cranium SK 846h/847/80, thejuvenile cranium SK 27, and upper premolar SK 2635­are undoubtedly derived from the later component ofMember l.

Age Relationships of the Swartkrans Members

Repeated attempts to obtain abso1ute dates for theSwartkrans deposits and fossi1s have so far been fruitless.Fairly recently a series of samples was collected forpaleomagnetic evaluation, but the results proved ambigu­ous (Brock, McFadden, and Partridge 1977). An attempt

Swartkrans 227

by Partridge () 973) to date the first opening of each of theaustralopithecine caves by measurements of cyclicnickpoim migration and valley flank regression gave afigure of 2.57 million years for the first opening of theSwaltkrans cavem. Although this figure ís interesting,there is a good deal of disagreement among specialists onwhether a geomorphological method of this kind is reli­able.

On the basis of her study of fossil bovids, Vrba (1975,1976a) suggested that remains from Member ) (the latercomponent of Member ) as we now know) could be re­ferred to the Swartkrans faunal span with an age indica­tion of between 1 and 2 miHion years. On bovid faunalgrounds Vrba also divided the Member 2 assemblage intoolder and younger groups termed "earlier and later SKB"and spanning, in aH probability, part of tiJe past half mil­líon years. In the present study 1 make no attempt toseparate the Member 2 fossil assemblage into older andyounger components. 1 would have no confidence indoing so, since very few of the fossils have come fromcontrolled excavations, and stratigraphic information istherefore lacking. The stratigraphic relation of the OuterCave breccias is exceedingly complex owing to periodieerosion of earlier breccias within the cave, followed byinfilling of the irregularly eroded surface by later sedi­menl. Examples of the result of this process may be seenin figure 189. In a, a f1attened australopithecine skuU isshown that had much of its braincase removed by erosionafter being embedded in Member 1 sediment. The spacewas then filled with browner Member 2 material. In b theintricate relationship between Member 1 and Member 2breccias can be seen.

Evidence of Stone Culture at Swartkrans

Site clearing during the late 1960s resuIted in the dis­covery of a series of stone artifacts embedded in brecciablocks that had been blasted from the site by lime miners.Although the stratigraphic origin of none of these wasknown in detail, it was certain that they all carne from theOuter Cave, which at that time was thought to contain asingle breccia mass. The available sample of 30 artifactswas described by M. D. Leakey (1970) on the assumptionthat they aH carne from breccia of robust australopithe­cine age. As discussed aboye, the initial assumption thatall Outer Cave breccia was of one depositional phase andage proved incorrect (Brain 1976a). Both Member 1 andMember 2 were, in faet, well represented in the OuterCave, and it was therefore necessary to try to assignthe artifacts to the members in which they originally lay.This was a difficult and unsatisfactory task, since in sornecases very little or no adherent breccia had been left onthe tools.

From the characteristics of the matrix, and consideringthe artifacts described in detail by Leakey, we may con­elude that only one tool in this collection is definiteiy fromMember 1 breccia (fig. 190a). This was classified as a"heavy-duty scraper" and described as follows: "aquartzite flake, 84 mm long and 88 mm wide, struck froma cobblestone. Approximately 50 percent of the circum­ference has been steeply but irregularly trimmed. Theedge is chipped and blunted by use on the dorsal aspectand there is also a crushed notch, 12 mm wide and ap­proximately 4mm deep." The specimen shows a certainamount of general abrasion as if it had been transpOlted inthe streambed sorne distance before it entered the cave.

228 Fossil Assemblages from the Sterkfonteín Valley Caves: Analysis and Interpretatíon

Six other specimens in the original collectíon couldhave originated in Member 1 breccia, though the quantítyof adhering matrix is too small to allow eertain allocation.These inelude a second "heavy-duty scraper," SK 3960,a "bífacial side chopper," SK 3967, and three unnamedtools, SK 3946, 3961, and 7878. AlI the remainíng tools,íneluding the well-made diabase eleaver SK 3962 (fig.190h), appear to have come from Member 2 breccia.

The position regardíng the occurrence of stone cultureat Swartkrans has c1early been most unsatisfactory,though it is improvíng now as a result of carefully con­trolled excavation.

The Swartkrans Bone Accumulations

The analysis presented here has been done on the fossílcollections in the Transvaal Museum that resulted from

the early work at the sile, 1948-53, as well as from thecurrent operations, 1965-75. The fossils have been di­vided inlo lwo groups, on the basis of their enclosíngmatrix, as to their origin in Member I or Member 2. Athird group from an excavated channel filI (see aboye) hasalso been considered. The bone accumulatíons are furtherdívíded ínto macro- and microvertebrate components.The former component must have entered the cave in avariety of ways, and the latter was almost certainly de­rived from pelIets regurgitated by owls tha! roosted in thecave.

Remains from Member 1: Macrovertebrate Component

The coIlection consists of 2,381 individual fossils from aminirnum of 339 individual animals belonging to 41identified taxa. Details are provided ín table 92 and

Fig. 188. A faee in (he Outer Cave ñlling showing a solution channel lraversing ¡he older Member 1 breecia. Theehannel is partly lilled with younger Member 2 sedimenl.

Swarrkrans 229

Class MammaliaOrder PrimatesFamily HominidaeHorno sp.Material:

SK 80 + 846b + 847 + 45 (lig. 192), maxillary piecefram an old individual with heavily wom left F and part ofp3; 2 isolated teeth (SK 80); part of left temporal withpartially preserved auditory meatus, most of mastoid pro­cess, and the petrous portion, SK 846b; cranial pieceswith left MJ, S K 847; right mandible pieces with M,- 2 andalveolus M3.

SK 27, flattenedjuveníle cranium with left 12, dm', P\unerupted, M'; right MI, unerupted C;, p3, and M2 re­moved; SK 2635, isolated left pJ-~, showing strong lingualwear.

Australopi,heclIs robustusMaterial:Craníums in varying degrees 01 completeness

SK 46, cranium from an adult, prabably female, se­verely crushed; left half of braincase, part offace, palate,and maxilla with left p4, MI-3; right p3-', MI-3. Ageestimate, 34 :': 3 years.

SK 47, partial cranium from an adolescent probablyfemale; most of base and palate well preserved, with leftM'-2, right P~ (unerupted), MI-2 and M J erupting. Ageestimate, 13 :': 2 years.

SK 48 + SKW 7 (fig. 144), reasonably completecranium from an adult, probably female; palate with leftP~ (roots), MI-3; right C, p3, P~ (roots), M'-3 (roots).SKW 7 consísts of crowns ofright P~, MI-z. Age estimate,20 :': 1 years.

SK 49, badly crushed cranium of an adult, withparieto-occipital píece separate; left and right pJ-4, MI-J.Age estímate, 19 ± 2 years.

SK 52, cranium of an adolescent, probably male, withlower part of the face, part of the right side and craniaJbase; left 12, p3-~, MI; right P, P3-4, MI, and uneruptedM3. Age estimate, 16 ± 1 years.

SK 54, part of a juvenile calotte with two depressedfractures near lambda (fig. 197a). Age estimate, im­mature.

SK 79, anterior part of an adult cranium with much ofthe face and palate; left P3~\ MI- J; right p3-', M[-J (fig.189a). Age estimate, 32 ± 2 years.

SK 83, most of a badly damaged adu!t cranium; rightp-z (roots), M3; left e, p3-4, MI-3. Age estimate, 32 :': 2years.

SK 821, part of a face, including left pJ somewhatwom. Age estimate, 26 .:t 2 years.

SK848, fragment of the audítory region of a cranium,showing the extermal auditory aperture and glenoid fossa.Age estímate, mature.SK 859, most of the occiput and parts of the parietals ofa young juvenile. Age estimate, irnmature.

SK 878, small piece of caloue with tooth fragments.Age estímate, mature.

SK 1585, approxímately right half of an endocramalcast. Age estimate, indetermínate.

SK 1590, fragments ofpalate wíth right F, e, P:l-\ MI;crushed and distorted mandible in very poor condition;femoral head and parts of an innominate. Age estímate,27 ± 2 years.

SK 14003, part ofa crushed skulJ with left MI, right M'(part), M2-~. Age estimate, 32 ± 2 years.

a

Fig. 189. Specimens showing Lhe interface between Member 1 andMember 2 breccias. (a) A ftatlened auslralopilhecine cranium. SK 79.which had been embedded in Member I sediment. Pan of the specimenwas lhen eroded away and the space fitled with browner Member 2 maleorial. (o) A piece of breeeia showing [he intricale form of Ihe MemberJ1Member 2 inlerface. The surface of [he Member I breccia was erodedbefore infilJing wilh Member 2 sedimenl occurred. A hominid 100Lh, SKW10, is preserved in the older breccia and is indicated by lhe arrow.

particulars of644 bovid bones, assigned to size classes butnot specifkally identified, are given in table 93. The vari­ous identified animals are depicted in figure 191.

Details of the fossil material by which the various ani­mal taxa are represented will now be presented.

Fig. 190. (a) The only anifael positiveJy known lO have come fromMember 1 breceia: a quanzite ftake, SK 3963. (o) A diabase cleaver and(e) biface from Member 2.

230 Fossil Assemblages from thc Sterkfontein Valley Caves: Analysis and Interpretation

SK826a + 877 +SKW31 + SK843 + 846a, maxillaryfragment with very worn and broken left p:¡-~, MI, ''.(826a); part ofa maxilla with right P" M'-' (SK 877); partof left M~ in maxillary fragment (SKW 31); left corpus ofmandible with slightly wom M I - 2 and M:¡ unerupted (SK843); isolated right M, (SK 846a). Age estímate, 14 ± 2years.Maxillae

SK 11 + lO, part of an adult faee with left P":', M'-',right P'-" MH (SK 11); left mandible piece with parts ofM,_, (SK 10). Age estimates, 27 ± 2-31 ± 2 years.

SK l2a,h, associated maxilla and mandible of an el­derly male; maxilla with left P (root), P' (roots), P'. andM' (roots); right P'-" MH; mandible with left and right

SK'l

P, (roots), P¿ M I _ :I • Age estimate, 33 ± 3 years.SK 13--14, part ofan adolescent face with len p:¡--\ and

MI, and M" (unerupted}; right p:¡-~, M 1-2, M" (unerupted}(SK 13); isolated left M2 with pronounced dental caries(SK 14). Age estimate, adoJescent.

SK 21-21a, maxilla with left e, P3-\ MI-2; M3 sepárate(21a). Age estímate, 37 ± 2 years.

SK 55a,b, juvenile maxilla with Ieft P-2, e, P", dm",M1-2; right 1'-2, C. P", dm- (55a); part of associated man­dible with left P3 , dm¿ M I _ 2 ; right dm¿ Ml - 2 , and M:¡{unerupted). Age estimare 13 :t 2 years.

SK 57, fragmentary palate with left P'-" M'-'; rightP3-\ and MI. Age estímate, 25 :t 2 years.

SK 65 + 67 + 74c, adult maxilla wíth left 1'-', e, P'-'

,,

canla me&Ornelas4

H.rion .teJl.lerl1

Galltropoda indet.

"

Prole'" , ••nu.alen.l.I

Megantereon ~.

1Dlnofeli. 11).,

HptrlllafrlcauUslralill

3

.~Aye. indet. 1

-.u.'ralopllhecu......1....e.

PIIl'lt .....e pIIrd...a

"

Horno 119.3

h "" !!!t ,f!!3 Oocuta eroeuta 1 86

Fig. 191. Animals represeuted by the fossils in the macrovertebrate component ofthe Swartkrans Member I sample.Minimum numbers of indi"iduals are iedicated.

(SK 65); isolated right c: (SK 65a); isolated right I ' (SK67); isolated righL pi (SK 74c). Age estimate, 26 :!: 2years,

SK 66, juvenile maxilla with roots of right d!:;, dm ' - z,

incompletely developed crowns of P-z. Age estimate,41,1 ± I years.

SK 83 la, adult maxilla with left M2-3 (not related ro SK831). Age estimate 28 ± 2 years.

SK 838a + 102, par! of a juvenile maxílla with rightdm2 , M' (SK 838a); crown of left M' probably associated(SK 102). Age estimate 8 ± 2 years.

SK 839 + 852, part of a juvenile palate with damagedleft and right di, dm2 , and MI; probably associated withjuvenile mandible in poor condition, left dm 1-2 and rightdm, (SK 852). Age estimate, 5 ± I years.

SK 845, part of an adult face, crushed, with left 12, e,P"-4, MI-2; right p3-', MI. Age estimate, adult.

SK 881 + 882, part of a maxilla with left P"-\ M';probably associated with part of a left upper molar (SK882). Age estimate, 26 ± 2 years.

SK 1512, part of a paIate with tooth roots and premolarfragment. Age estimate, mature.

SK 1592, part of a palate with broken right P" MI-3.Age estimate, 32 ± 3 years.

SK 1595, part of a juvenile maxilla with left dm2 andpieces of erupting teeth. Age estimate, 4 ± 1 years.

SKW 11, maxilIa, in situ against west wall of cave;unprepared.

SKW 12, part of a maxilla with left pJ-4, MI-J. Age notestimated.

SKW 33, part of a maxilla with right MI-2 (previouslycataloged as SK 14129a). Age estimate, 14 ± 2 years.Mandibles

SK 6 + 100, the holotype of Paranthropus crassidens(fig. 193), part of a mandible with left P3- 4 , M ,- z, andunerupted M 3 ; right P4 (abnormal), MI - 2 and M3 (un­erupted) (SK 6); isolated right PJ (SK 100), Age estimate,15 ± 1 years.

SK 23, mandible of adult, probably fema1e, with fulldentition in excellent eondition (fig. 194). Age estimate,19 ± 2 years.

SK 25 + 832, part of a juvenile mandible with left p.(erupting), M,-z; right p., M I - z (SK 25); probably as­soeiated with crown of left MI (SK 832). Age estimate,9 ± 3 years.

Swartkrans 23 [

SK 34, almost complete mandible of a large adult,probably male, in two halves, left P,. M'-3: right 1,-2, e,P'I-" M 1-'" Age estimate, 23 ± 3 years.

SK 37, part of a mandible with left M2 and small pieceof M,. Age estimate, 12 ± I years.

SK 6], part of a juvenile mandible with left and rightdi'_2' de, dm,_z and right MI erupting. Age estimate,6 ± 2 years.

SK 62, part of a juvenile mandible with left di z, de,dm'_2' and M, unerupted; right 1

"de, dm'_2' Age

estimate, 6 ± 1 years.SK 63 + 89a + 90 + 91, almost complete juvenile

mandible with left and right de, dm l _ 2 , MI> and M 2 (un­erupted, right unerupted crowns of 11 and e (SK 63); un­wom erowns of left and right MI (SK 89a); left dm2 (SK90); right dm I (SK 91). Age estimate, 7 ± 1 years.

SK 64, part of ajuvenile or infant mandible with rightdml-2' Age estimate, 2Y2 ± ~ years.

SK 74a, part of an adult mandible, showing possiblemental eminence, with left P3 - •• M I - 2 ; right 12 , P3 - .. M.-2 .

Age estimate, 28 ± 3 years.

Fig. 193. The holotype of Paranlhropus crassidens, SK 6. fromSwanlaans Member l.

p

Fig. 192. Pan of a cranium from Member I aLtribuled lo Hamo sp.: Fig. 194. A remarkabJy complete and well-preserved mandible of A IIS-

SK 80 + 846b + 847 + 45. Ira!opilheclIs mbllsllis, SK 23, from Member 1.

232 Fossil Assemblages from the Sterkfontein Valley Caves: Analys¡s and Interpretation

SK 81, part of a mandible in poor condítion with leftP, 4, M l _ :l ; right L. C. P:I _ 1o MI a- Age estímate, 26 == 1ycars.

SK 96, part of a juvenile rnandible with roots of left dm.,unerupted left e and P'I. Ase estimate , 6 ± 1 years.

SK 438, part of a juvenile mandible with left unerupteddm.. Age estimate, 2 ± lh years.

SK 841a, mandible fragmeot with left dm, and part ofM, (not related to 84lb). Age estimate, 2 ± v.. years,

SK 842 + 869, part of a juvenile mandíble with roots ofdm , (SK 869); left dm, (SK 842). Age estimate, 2v.. ± v..years.

SK 844, part of an adult mandible in poor conditicn, leñM t - :¡. Age estímate, 25 ± 2 years.

SK 862, part of a mandíble witb roots of right P, andMI, M 2- 3 • Age estímate, adult.

SK 876, part of a erushed mandible in pool' eondition,left e root, P3-~' M'_3; right 12, C-Ma. Age estímate.28 :t 4 years.

S K 858 + 861 + 883, part of an adult mandible with left11- 2> e, P:¡-4. Ml_2~ right L-2. e, Pa- h MI_a (SK 858);mandible fragment with parts of M,~, (SK 861); part ofright M, and angle of rnandible. Age estímate, 19 ± 2years.

SK 1514, mandibJe fragmeüt in very poor conditionwith roots of MI and part of the erown of M2 • Ageestimate, mature.

SK 1586, most of a pooríy preserved mandible witb left11, MI_a; right 11- 2 , M1- :¡ . Age estirnate, 27 :t: 2 years.

SK 1587a,b, part of a maodible with roots ofleft P3 , P"M.-, (a); isolated rigbt M, (b). Age estirnate, 17:t Iyears.

SK 1588, maodible fragment witb roots of right P" P"MI' Age estímate, 13 ± 1 years.

SK 1648, anterior part of a mandible in poor conditíon,left MI; right M I - 2. Age estimare, 29 ± 2 years.

SK 3978, infant mandible with left and right dm.-, aodleft M, in crypt (fig. 197b). Age estímate, 2v.. ± v.. years.

SKW 5a,b, adult mandible in two parts, unprepared,Appears to be adult,Isolated maxillary teeth (for age estimates oC isolatedteeth, see Maoo 1975)

Incisors: SK 68, left JI; SK 69 and 73, left P and right JI;SK 2, right 1'; SK 40, SK 3, aod SK 4, leftl', rigbt 1', andright C; SK 70, left 1'; SK 71, rigbt 1'.

eanioes: SK 38, right e SK 85a aod SK 93, right Candleft C; SK 86, left C; SK 92, rigbt C SK 95, left C; SK1596, C.

3d premolars: SK 24, left 1"; SK 28, left P'; SK 33, rigbtP"; SK 44, right P'; SK 101, left P'; SK 822, left P'crowo;SK 823, right p:l; SK 867, p3, incomplete crown; SK14001, left P'; SKW 32 (SK 14128), right P'.

4tb premolars: SK 32, right P'; SK 39, right P'; SK 99,left P' crowo; SK 824, left P' crowo; SK 825, left P'crowo; SK 856, right P' aod M'; SK 1589, right P' uo­erupted crown.

1st molars; SK 17, right M'; SK 35, left M' fragment;SK 829, left MI; SK 833, left M'; SK 849, rigllt M'; SK872, left M'; SK 1591 and SK 16, left M' and M'.

2d molars: SK 42, right M'; SK 98, left M' crowo; SK834, rigbt M'; SK 837, ligbt M'; SK 868, rigbt MH frag­meots; SKW 33 (SK 14129a), rigbt M' aod fragmeot ofM'.

3d molars: SK 31, rigbt M'; SK 36, right M'; SK 41, leftM"; SK 105, left M' crown; SK 835, left M'; SK 836, leftM"; SK 870, left M'; SK 1524, left M', uoerupted crowo;SK 3975, left M"; SK 3977, rigbt M'.

Isolated mandibular teethIncisors: SK 741>. rigbt 11 •

Canines: SK 29, right C; SK 87, right e crown; SK 94.right C; SK 884, left C and fragment of mastoid.

3d premolars: SK 30, left P,; SK 72, Ieft P,; SK 831,right P" not related to 831a; SK 850, right P,; SK 857,right P,; SK 1593, right P,.

4th premolars: SK 7, right P,; SK 9, left P,; SK 88, leftP,; SK 826h, left P" not related to 8260; SK 827, leñ P,;SK 830, left P,.

1st molars: SK 20, left M,; SK 104, right M" uneruptedcrown; SK 828, leñ M,; SK 838b, left M" not related to8380; SK 864, probably M" severely darnaged; SK 1594,right M" buccal half; SK 3974, right M, crown.

2d molars: SK 1, left M" complete; SK 5, left M,; SK3976, left M" complete.

3d molars: SK 19, right M" probably: SK 22, right M,;SK 75, right M" unerupted crowo; SK 840, Ieft M,; SK84lb, left M,; SK 851, right M,; SK 855, right M,; SK871, left M" unerupted crown; SK 880, left M" as­sociated with two toorh fragments; SK 885, left M:J.lsolated tooth fragments

SK 863, 865, 866, 873, 874, 875, 879, 14000, 14080;SKW 6, 8a,b 10, 14, 15, 30.Postcraníal bones

SK 50. pelvis bone from adult, probably male; most ofright innominate, pubis and iJiac erest rnissing. Ageestímate, adult.

SK 3155b, right iJium with complete acetabulum. Ageestimate , matute.

SK 82. proximal end of right femur from a mature adult.Age estímate, mature.

SK 97, proximal end of right femur from a mature adult.Age estimate, rnature.

SK 860, distal eod of right humerus (considered bysorne to be cercopithecoid). Age estímate. mature.

SK 853. isolated lumbar vertebra. Age estímate, im­matare.

SK 854, axis vertebra. Age estímate. mature.SK 3981a,b, last thoracic vertebra rcjr tast lumbar Ver­

tebra, (b). Age estímate, adult.SK 84, 1st left metaearpal. Age estímate, mature.SK 85, part of 4tb left metacarpal. Age estímate, ma­

ture.SKW 27 (SK 14147), adult 5tb left metacarpa!. Age not

estimated.Estimating the minimum number of robust australo­

pithecine individuals whose remains are represented intbe Swartkraos sample is not easy, and figures will vary.Mano (1975) was able to deduce ages at deatb for 113specimens; the distribution of these specimens in agee1asses is giveo io table 94 and graphically depicted iofigure 195. Tbe mean age at deatb for the sample is 17.2years; if 20 years is taken as representing the advent ofmaturity, Ibeo 68% of Ibe sample was immatuTe at dealb.There is no question thal sorne of the specimens in thissample oí 112 carne froro the same indíviduals, and myestimate of the actual number of individuals ¡nvolved inthe Member 1 sample is 87. However, the proportions ofindividuals in each of the age e1asses was probably verysimilar to those io Mann's sample of 113.

Family CereopithecidaeParapapio jonesiMaterial:29 eranjal pieees from a minimum of 8 individuaJs.

d paJate with JI-C bilaterally, SK 588a; crushed eal-

,

:;~

Swartkrans 233

Percentage of aample

(n:1I3)

Fig. 195. Histograms showing estimated eges-at-deatb of 113 au­stralopithecine specimens fmm Swartkrans Member l. Data from Mann(1975).

varia and faee with right M3, SK 2127; brainease with panof left M3, SK 556; maxilla with left II_M2 and right 1I • SK573h: left maxillae, SK 543 and unnumbered: rightmaxillae , SK 442, 462, 14151; ~ mandible with full denti­tion , SK 573a; ~ mandible with P"-M,, bi1aterally, SK414; mandible with left dm l _ 2 , M l - 2 , and right de, dm¿and part of M" SK 418; left mandible pieces , SK 433, 437;isolated associated teeth, left C. r., MI, and right r., SK588b; isolated teeth, SK 468, 493, 501, 502, 505, 511, 512,533, 601b, 612,635, 1835, 14164.

The 8 individuals have been plaeed in age classes(Freedman and Brain 1977) as indicated in table 95. Thesample eontains 2 juveniles only.

Papio robínsoníMaterial:121 eranial pieees from a mínimum of 38 individuals,

The type specimen, consisting of the snout and parts ofthe orbits ofa young adult o, with left e, P'-M", and rightP'-M", SK 555; almost complete ~ skull, SK 558; left halfof a skutl with P:LM:l, SK 557; complete muzzle and par­tial braincase, SK 560; complete muzzle with P:1_M3

bilaterally, SK 562; anterior part of muzzle and palate,SK 565; palate with MH bilaterally, SK 566; flattenedcranium, SK 590; shattered skull, SK 602; left side ofcranium, SK 2202; parts of a fragmentary eranium, SK14163; maxillary pieces, SK 436, 439, 447, 449, 456. 458,466,469, 476, 497,536,537, 538,540,544,549,552, 571b,608, 614, 629, 631, 215\, 2164, 2171, 2180; mandiblepieces, SK 406, 407, 408, 409, 410, 416, 417, 419, 420, 421,423,425,427,429,430,431,435,445,446,453,457,459,460,463,464,550, 568,570, 571a, 572,596,609,610,615,617, 622, 2161, 2175, 2319, 2694. 2924, 3211b. 14083,14152, 14156; isolated teeth, SK 465, 469, 475, 477, 478,482,483,484,485,486,499,500,508,509,516,518,520,

Dínopithecus ingensMaterial:57 cranial pieces from a minimum of 26 individuals.

Complete skull, v. with deciduous and permanentdentition, SK 554: 2 skull with P' and M" bilaterally, SK600; crushed ~ skull with P'-M' bilaterally, SK 603; righthalf of a skull with P'-M", SK 553; crushed skull withoutteeth, SK 599; maxillary pieces, SK 440, 441,443, 542a.545,546,548, 571b, 574,576(1, 578a.b. 604a. 630, 1518,14004; mandib1e pieces, SK 401, 404, 413, 415, 422, 424,428,455,470,589,2407; isolated reeth, SK 473, 474, 487,489,492,498,503,510,513.514,526,527,528,532,577,585,586,618, 628a,b. 636, 1583,2163,2169,2625,14157,14159.

The various individuals have been placed in age classes(Freedrnan and Brain 1977) as indicated in table 98. Thereare 4 juveniles in the sample.

Theropith ecus danieliMaterial:31 eranial pieces from a minimum of 17 adult individuals.

The type specimen, consisting of a )' muzzle withPt-M" bilaterally, left mandible piece with e-M:\, andright mandible piece with P,rM", SK 563 + 402 + 405;part ofa shattered cranium with left P-M:l, right JI, e-M:l •

SK 561; maxillary píeces. SK 448, 461, 464, 567, 575h.c.593, 597, 1607, 2148, 2\81, 2193; mandible pieces, SK403. 411, 426, 432, 491, 530, 569, 575a. 2177; isolatedteeth, SK 479, 495, 507, 521,58\,2158,2172,3529, 14162.

The various individuals have been placed in age cíasses(Freedrnan and Brain 1977) as indicated in table 97. Nojuveniles are present in the sarnple.

Cercopithecoid indet.Material:134 cranial and 31 postcranial pieces from ao estimatedminimum of 28 individuals.

Calvaria pieces, SK 559, 634, 1584a,b, 1599, 2105,2120,2123,2124,2130,2131,2132,2134,2136,2137,2138,2139,2140,2141,2174,2176,2183,2189,2190,2195,2199,2201,2203,2204,2205,2206,2208,2210,2211,2213,2216,2775,2976,3241,3256,3260,3266,3280,3282,3290,3450,3863, 3885, 5582, 7144, 7198, 7284, \4025, 14160, 14161;maxillary pieces, SK 601, 2170, 2173, 2196, 2352b, 2758,2842,3522,7032,7462,9345, 14154; mandib1e pieces, SK541,594,2125,2128,2143,2152,2159, 2160, 2165, 2179,2182,2188,2221,2222,2427,2710,2858,2891,3187,3258,3272, 4019, 7333; isolated teeth and tooth fragments, SK467,471,472,480,481,490,494,504, 506,523,525,584,627,1861,2014,2\44.2145,2149,2154,2162,2168,2192,2194,2276,2467,2524,2692,2789,2890,2912,3\89,3309,3430, 3502, 3516, 3602, 3678, 3895, 6996, 7094, 7274,14158; pelvis piece, SK 3434; proximal radius, SK 1815,1816. 1867, 1870; radius shaft piece, SK 2813; proximalulna, SK 591; proximal humeros, SK 2601, 14029: distalhumen, SK 1506, 2 unnumbered; proximal femur, SK607, 1513, 1823, 2595, 2810, 2836, 8065, 14015, 14024;distal femur, SK 605, 1500, 1817, 3254, 5236, 2 un­numbered; proximal tibia. SK unnumbered; phalanx, SK6759.

522,529.530,531,534.582.587.595.598,611,613.619,620.621, 623, 625, 62o, 2146. 2t~7. 14153.

The various individual s have been placed in age ctasscs(Freedman nnd Brain 1977) as indicated in table 96. Thesample contains 4 juveniles.

4020o

t- s,•.,,*

Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation234

Order CarnivoraFamily FelidaePanthrra pardusMuterial:29 cranial and 3 postcranial pieces from a minimum of 12individuals. of which I was juvenile and 1 irnmature.

The type specimen ofP. pardus incurva, consisting of acranium and mandible in three parts, SK 349: parts ofcrushed craniums, SK 353. 354. 1932; almost completebraincase, without muzzle, SK 14010; left zygoma andcranial base, SK 14185; palates, SK 351, 352; left maxil­lae, SK 355, unnumbered; right maxilla with P" SK 5%0;three parts of a mandible with left L.l_:¡' e, P3-4, M" andright P,_" SK 1866 + 343a.I>: left mandibular pieces, SK339,341,342,344,2797,6892; right mandible pieces, SK340,345,350,357,1806,7132; isolated teeth, SK 347, 356,358,7088; distal humeri, SK 1639, 1899; distal radius, SK3060; metapodial fragment, SK 2605.

Dinofelis sp.Material:2 cranial and 2 posteranial pieces, probably from 1 indi­vidual.

Right mandible fragment with P~ and MI, SK 335; lso­lated right e, SK 372; metatarsal pieces , SK 1848. 1860.

Meguntereon sp.Material:A single cranial piece.

Right mandible fragment with PH , SK 337.

Family HyaenidaeHvaena brunnea disparMaterial:7 cranial pieces from a minimum of 3 individuals-2adults and J immature.

The type specimen consisting of a right maxilla withpH and M', SK 326; left maxilla with P' erupting, SK331; ríght maxilla with P'-', SK 332; left mandible frag­ment with M" SK 329; isolated right 1", SK 330; right P',SK 333; left P" SK 327.

Crocuta crocuta venustu/aMaterial:3 cranial pieces from 2 adult individuals.

The type specímen consisting of 3 mandible pieees withright P2 - . and M" SK 317; left mandible fragment withP,-M" SK 318.

Crocuta eracula ultraMaterial:9 eranial pieces from a mínimum of 4 adult individuals.

The former type specirnen of C. crocuta angel/a, con­sisting of much oí a craniurn without rnandible, exten­sively damaged on left side, SK 319; parts of a eraniumand mandible, 5 pieces, SK 320; right maxílla with P'-"SK 323; leñ mandible fragment with P,-M" SK322 + 325; right mandible fragment with P2- , amd M" SK321; isolated P', SK 1807; left P" SK 324; right P" SK2452.

Hyaenictis forfexMaterial:Parts of a single specimen.

The former type oí Leecvaena forfex. consisting of acomplete cranium with full dentition, muzzle crushed

dorsally. SK 314; associated with left and right mandibu­lar rami. SK 315, 316.

Euryboos nitídu!a, type B, advancedMatcsiaí:9 cranial pieces from a mínimum of 5 individuals.

The former type of Lycyaenü nitidula, consisting of aright mandibular ramus with P:I_'~' SK 301; left maxilIarypieces , SK 307, 312; right mandible piece with e, PZ- 4>

SK 304; isolated left P'. SK 310,379; right P', SK 311; leftP", SK 309. 313.

Euryboas nitidula, type A, primitiveMaterial:5 cranial pieces from a minimum of 2 individuals.

Right maxillary piece with p:.!~~, SK 305; right maxillarypiece with p:¡-\ SK 306; left mandible piece with e, P Z- h

SK 302; left rnandible fragment with M" SK 336; rightmandible piece with PrM I , SK 14005.

Euryboas niíidula, type uneertainAla/erial:4 cranial pieces, possíbly from a single individual.

Isolated right~, SK 1853; left C. SK 1831; left P" SK308; left P" SK 303.

Hyaenid indet.Material:16 cranial pieces from an estimated mínimum of 4 individ­uals.

Parts of 2 erushed skulls, SK 2732, 14082; mandiblepieces, SK 1840, 2300. unnumbered ; isolated teeth andtooth fragments, SK 1818, 1827. 1830, 1844, 1856, 1857,1858, 1872, 1873,2695,2922.

Proteles transvaalensisMaterial:2 cranial pieces, probably from a single individual.

Leñ maxillary fragment, SK 1851; mandible fragmentwith roots of 2 postcanine teeth, SK 3173.

Farnily CanidaeCanis mesomelas papposMaterial:9 cranial pieces from a mínimum of 4 adult Individuals.

Complete cranium, SK 375; right maxillary fragmentwith M', SK 370; left mandible pieees, SK 361, 362,14107; right mandible fragment, SK 363; isolated right C.SK 1868; left P" SK unnumbered; M', SK 3284.

Vulpes pulcherMaterial:2 cranial pieces from 2 individuals.

Left mandible fragment with C-M" SK 376; left man­dible piece with alveoli of premolars and pan of M.; SK12707.

Camivore indet.Material:28 eranial and 28 posteranial pieces from an estimatedmínimum of 15 individuals.

Maxillary fragment, SK 14042; mandible pieees, SK1608, 1805,2769,3366,5927; isolated teeth, SK 535, 1874.2119,2166,2236,2353,2395, 2415h. 2550,2675,2708,2721. 2743, 2911, 2913, 3661, 3695, 3798, 6944, 7082,14130; axis vertebra, SK unnumbered ; scapula piece , SK

h643; proximal humerus, SK 1894; proximal radius, SK.1.1J7, 10535; proximal fcrnur , SK 606, 1802. J269. 8014;distal tibia, SK 1879, two unnumbered; metapodialpieces , SK 1812, 1822, 1832, 1854, 1876 + 1859, 2270,2815,3265, J69O, 5063, 8025, 12017; phalanges, SK 1569,1834. unnumbered.

Order ArtiodactylaFamily BovidaeDarnaliscus sp, 1 or Pamularius sp.Material:8 cranial pieces from a minimum of 1juvenile and 3 adultindividuals.

Part of a palate, left MI-

2 and other tooth fragments,SK 3832; left mandible piece , MH , SK 3127; left mandi­ble piece, MH , SK 2957; left mandible fragment, M" SK2064; mandible piece, dpm , M.-" SK 10500; right rnan­dible piece, P4- M 1• SK 1999; right mandible piece, P4 ,

MH , SK 2000; right mandible piece , Mh " SK 3135.

Rabaticeras porrocornutusMaterial:3 cranial pieces from 2 adult individuals.

The type specimen consisting of left and right frontalpieces with incomplete hom-cores, SK 321 ta,b; almostcomplete braincase with both orbits and hom-core bases,SK 14104 + 2620 + 2865.

Medium-sized alcelaphínes, including Rabattceras den­titionsMaterial:97 cranial pieces from a minimum of 4 juveniles and 18adults,

Left and right mandible pieces associated with horn­eore pieces, SK 3213a-d; left juvenile maxilla, SK 2274;right juvenile rnaxillae, SK 1633,3115; Left adult maxillae,SK 2032, 2076, 2314, 2662, 2950, 2989, 3056, 3111, 3153,12193; right adult maxillae , SK 1616, 1624,2092,2107,2114,2116,2239,2257,2286,2318,2326,2336,2510,2987,3013,3053,3087,3108,3118,3126,3142,3207,14117; leftjuvenile mandible píeces, SK 2082, 14214; right juvenilemandible pieees, SK 2964; left adult mandible pieces, SK1613e. 1656a, 2478, 2492, 2523, 2974, 2983, 2992, 3040,3067, 3146; right adult mandible pieces, SK 1623, 1961,2083,2287,2316,2529,2971,2991,2985,2996,3002,3004,3043, 3046, 3089, 3125, 3141, 3151; isolated deeiduousteeth, SK 3498, 12633; isolated permanent teeth, SK1991,2006,2048,2049,2068,2232,2269,2278,2296,2302,2364,2406,2426,2438,2448,2457,2526,2527,3050,3081,5941,14056, 14124.

Cf. Connochaetes sp.Material:70 cranial and 4 postcraniaJ pieces frorn a minimum of 6juvenile and 13 adult individuals.

Cranial piece associated with a vertebral fragment, leftastragalus, calcaneus, and left naviculocuboid attached topiece of proximal metatarsal, SK 3812a-f; fragmentarypalate with left and right M'~', SK 2224; left and rightmandible pieees with left MH , right M,_" SK 3100; leftjuvenile maxillae, SK 3229, 14207; rightjuvenile maxillae,SK 2066, 5991; left adult maxillae, SK 1634,2966,3066,3097, 3128, 5946; right adult rnaxillae, SK 2061, 2097,2225,2261, 259lc. 2686,2982,3080,14120; leftjuvenilemandible pieees, SK 1618, 3090, 5203, 5906 + 4479; rightjuvenile mandible píeces, SK 2973, 7315; left adult rnan-

Swartkrans 235

dibJe pieees, SK 1630,2065,2354,3091,3104,3131,3134,3156, 72t6; right adult mandible pieces , SK 2069, 2352a.2358,2697,2986,3010,3045,3052,3061.3068,3105,3137,6073; isolated deciduous teeth, SK 4089, 5185; isolaledpermanent teeth, SK 1652,2025,2054,2109,2110, 2284,2379,2482,2483,2498,2667,2749,3008,3018,3041,3047,3102,4244.

Cf. Megalotragus sp.Material:8 cranial pieces from 3 adult individuals.

Left maxillary fragment with P'-M', SK 3031; leftmandible pieees, SK 1944, 2063, 2081, 3099; right mandí­ble pieees, SK 2118, 3132,14113.

?Hippotragin¡Material:2 cranial pieces from a single juvenile individual.

Right maxilla with dpm", MH, SK 3107 + 3139.

Redunca cf arundinumMaterial:A single cranial piece from a juvenile.

Right maxilla with dpm":', SK 3533.

Pelea cf. capreoíusMaterial:2 cranial pieces from a juvenile and an adult individual.

Left maxilla with dprrr', MI-2, and erupting premolars,SK 2682; right mandible pieee with M,,, SK 2378.

Antidorcas cf. reckiMaterial:J2 cranial pieces from a mínimum of 4 juveniles and 2adult individuals.

Fragmentary palate, left M' and right M'-', SK 3033;right juvenile mandible, dpm, MH , SK 3094; left maxil­lary fragment, SK 3169; right maxilla, M'~', SK 2567;mandible with M.-, biLaterally, SK 2310; left mandiblepieee, SK 2545; right mandib1e pieees, SK 1931, 2235,3095, 3501; isolated right dprn, SK 2079; left M" SK3759.

?Gazella sp,Material:8 craníal pieces from a minimum of 1juvenile and 5 adultindividuals.

Incomplete snout and palate, left P2_M2, found in con­taet with hominid innominate, (SK 3155bJ, SK 3155a;palate with left and right P'-M', SK 14063; complete pal­ate, right P'-M'; left P'-M', SK 3012; left maxillae,SK 2495, 2972; right maxilla, SK 3261; right mandiblepiece, SK 10440; isolated left M', SK 14068.

Antilopine or Neotragine indet,Material:3 cranial pieces from 3 adult individuals.

Right mandible píece, P,.-M" SK 3019 + 2509; rightmandible piece, M'_3, SK 2665; right mandible pieee withM" SK 3025.

Oreotragus cf. majarMaterial:A single cranial specimen.

Left mandible pieee, PH , M" SK 14059.

236 Fossíl Assernblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Syncerus sp.Material:13 cranial pieces from 1 juvenile and 2 adult individuals.

MaxilJa fragment, right M", SK 2077; leftjuveníle man­dible piece, dpml_.J, Mr_2' SK 3064; left adult mandiblepieces, SK 2491, 3130; right adult mandible pieces, SK1972, 2968; isolated teeth, SK 1983, 2028, 2480, 2517,3034, 3059, 3074.

Tragelaphus cf. strepsicerosMaterial:12 cranial pieces from 1juvenile and 3 adult individuals.

Part of a right horn-core, SK 3171; right juvenilemaxilla, dpm>", SK 2271; left adult maxillae, SK 2281,2681; right adult maxillae, SK 2095, 3000, 14112?; rightmandible piece, P.,. MI~3' SK 3110; isclated teeth, SK2304, 2541, 2576, 3023.

Cf. Makapania sp.Material:10 cranial pieces from a mínimum of 3 adult individuals.

Left maxillary piece with P" M"" SK 3150; rightmaxillary pieces, SK 2759, 3005, 3065; left mandiblefragment, SK 2965; right mandible pieces. SK 1627,3113;isolated teeth, SK 2373, 2693, 2849a.

Antelope size clase IMaterial:25 cranial and 11 postcranial pieces from an estimatedminimum of S individuals, as listed in table 93.

Antelope size class 11Material:117 eranial and 113 postcranial pieces from an estimatedminimum of 20 individuals, as listed in tabJe 93.

Antelope size ctass IIIMaterial:199 cranial and 162 posteranial pieces from an estimatedminimum of 25 individuals, as listed in tabJe 93.

Antelope size ctass IVMaterial:15 cranial and 2 postcranial píeces from an estimatedmínimum of 2 individuals as listed in table 93.

Family SuidaeMetridiochoerus andrewsiMaterial:8 eranial pieces from a minimum of 2 juvenile and S adultindividuals.

Anterior pan of juvenile mandible with Ieñ dilo de,dpm,... SK 381; juvenile mandible piece with 1eft dpm,right dpm,.•, and part of M" SK 380; maxillary fragmentand part of ?right M', SK 393 + 391; cranial fragmentwith parts ofleft and right M" SK 394; isolated 1eftdpm",SK 14240; isolated left M" SK 387; right M" SK 388;?Ieft M" SK 392; fragment of M', SK 2088, 2380; part ofM" SK 390; right M" SK 389.

Suid indet.Material:9 cranial pieees from an estimated minimum of 3 individu­als.

Isolated tooth fragments, SK 384, 2456, 2677, 2811,3445,4743,6018.9423, 14148.

Order PerissodactylaFamily EquidaeEquus capensisMaterial:9 cranial pieces from a minimum oí 1juvenile and 5 adultindividuals,

Palate with parts of P'-M' bilaterally, SK 3983; leftmandible fragment with M" SK 2584; left mandible piecewith fragmentary mclars, SK 1619; right mandible piece,P.-" SK 1626; isolated teeth, SK 2104, 3159, 3311, 3993,3996.

Hipparion libycumMaterial:3 isolated teeth from a single old individual.

Left P', P', and right P" SK 3278 + 3982 + 2307.

Equid indet.Material:14 cranial and 3 postcranial pieces from an estimatedminimum of 4 individuals.

Mandible fragment, SK 1945; isolaled teerh and toothfragrnents. 2018, 2070, 2481, unnumbered, 2539, 3250,3530, 3534, 3830, 7146, 7357, 9095, 12029; distal femur,SK 18%; lateral rnetepodial, SK unnumbered; navicular,SK 6691.

arder HyracoideaFamily ProcaviidaeProcavia antiquaMaterial:35 cranial pieces from a mínimum of 16 individuals.

Craniums in varying degrees of completeness, SK 143,145,172,2794. 6Oq, 6072,14142,14239; maxiIJary pieces,SK 135, 163, 164,J78, 204,1993,3698, 1414Ob: mandiblepieces, SK 128, 136, 138, 140, 161, 162, 182, 205, 1924,3220, 3285, 3449, 4076, 5969, 14090, 14134, 14140a; íso­lated teeth, SK 2493, 3367.

Procavio transvaalensisMaterial:20 cranial pieees and 1 posteranial piece from a minimumof 8 individuals.

Craniums with articulated mandibles, SK 184, 196;craniums in varying degrees of completeness, SK 188.207, 209, 211, 2768; maxiIJary piece, SK 208; mandiblepieces, SK 1120, 113, 117, 126, 201a, 2935,4216, 14135;isolated teeth, SK 3223, 12709, 14141, 14144; 1eft distalhumeros, SK 5255.

Hyracoid indet.Material:6 cranial píeces from a mínimum of 2 individuaJs.

Partial cranium, SK 3244; maxillary pieces, SK 2126,3510; mandible pieces, SK 2889, 2934; part of an isolatedincisor, SK 3153.

arder RodentiaFamily HystricidaeHystrix afrícaeaustraíisMaterial:6 cranial pieces from a minimum of 3 adult individuals.

Incornplete palate with left M'" and right M', SK 3082;right M' and M', SK 2466; isolated left 1', SK 4315; ?right1" SK 2875; left M', SK 14236.

Hystrix ?makapanensisMoteriaJ:

An isolated left P-t, SK 14237, has been tentatively re­ferred.

Class AvesBird indet,Material:

A complete cranium frorn an as yet unidentified bird,SK 1001.

Phylum MnlluscaClass GastropodaOrder PulmonataLand snail, cf, Acñatína sp.Material:15 shells in varying degrees oí completeness.

SK 2405, 2627, 2723, 2817, 3186, 3333, 3343, 3353,3608, 4814, 4958, 7476, 7488, 8090, 9192.

Swartkrans 237

The Major Features of the Bone Accumulation

The Composition ofthe Fauna Rcprescnted bv theFossils

The information provided in table 92 and depictcd infigure 191 may be further summarized. Th¡s is done in thefirst columns of table 99 and figure 1%. Of the 339 anirnalsthat contributed bones to the assemblage , 179, or 52.9'1&,were primates. These were almost equally divided be­tween hominids and the four species of baboons. Next innumerical Importance were 82 bovid individuals coveringa wide range of body sizes. In size class 1,5 adulta arerepresented: in class 11, 20 individuals including 8juveniles are found; in class IU, there are 50 individualswith 11 juveniles, and class IV is represented by 7 indi­viduals, among which are 2 juveniles.

The camivore component in the Member 1 fauna isirnportant. with 36 individuals making up 10.6% of thetotal. Cats are represented by J2 individual leopards, a

LAaOMORPHA~

ROOENTlA~.~...

HYl:trlcld••~

AVO·tI

REPTILlA.

• 20MEMBER I

o 20 40MEMBER 2

•• o 20 40 6Q

CHANNEL FILL

PERCENTAGE CONTRIBUTlON ro rne FOSSIL FAUNAS

Fig. 196. Percentage comríbutíons mece by various groups of anima1s lo samples of macrovertebrate tossil faunafrom Swartkrans members I and 2 and from channel fill.

238 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interprctation

single Dinolelis and a MegalltC'rC'on. Rcmains of 3 brownand 2 spotted hyena individual s have been found, to­gether with those of 8 long-legged hunting hyenas. Partsof 1 aardwolf, 4 jackals, and 2 foxes are known.

Other fauna! elements arc 7 equids, 24 dassies, and 3porcupines.

The Representarion 01 Skeletal Parts

The Member 1 sample has undoubtedJy bcen artificiallybiased in favor of identifiable cranial fossils, but this biascould not account, 1 think, for the almost total absence ofprimate postcranial parts. Study of the assemblagestrongly suggests that the bodies of the hominids and ba­boons had largely disappeared before fossilization of theskuUs, or parts thereof, took place. In the case of Aus­lralopithecus robustus, for instance, only 11 postcranialbones are associated with 218 cranial pieces, although theJatter do include isolated teeth. The situation among thebaboon fossils is even more extreme, where only 31 post­cranial pieces accompany 372 cranial pieces.

Most of the carnivores, too, are known by cranialspecimens only, and this is also tme of the dassie compo­nent of the assemblage.

For the bovids, skeJetal representation is certainlymore comprehensive. In the various size classes, relativenumbers of craniaVpostcranial pieces are as follows: classI, 25/11; class n, 117/113; class III, 1991162; class IV,15/12. Details of the parts actually preserved may befound in table 93, but the numbers are too small to Jet usdiscern a pattem. In size classes Ir and IU parts frommost of the body are represented, albeit in small numbers.

Observed Damage to the Bones

As with the assemblages from other site units in theSterkfontein valley caves, each piece was examined fordiagnostic damage marks. Details of such damage aregiven in tabIe 100 and may be summarized as fotlows:

Poreupine and Small-Rodent Gnawing Marks. The in­cidence is extremely low in this assemblage, there beingone specimen showíng porcupine gnaw marks and onewith small-rodent gnaw marks.

Carnivore-Inflicted Damage. Clear camivore damage wasobserved on 3 australopithecine specimens: two punctatemarks on the juvenile calvaria SK 54 (fig. 197a) to bediscussed further; ragged-edge damage on the juvenilemandible SK 3978 (fig. 197h); and tooth marks on theinnominate SK 3155h. Positive camivore damage wasalso seen on 10 bovid fossíls.

Damage that could very probably be attributed to car­nivore action was observed on the Hamo mandible SK45, on 14 australopithecine specimens, and on 1 cer­copithecoid calvaria (table 100).

Traces 01Artificía/ Bone Alteration. None was observed.

Association with Slone Artifacts

As 1 mentioned earlier, only one stone artifact may besaid with positive assurance to have come from Member Ibreccia. It is somewhat abraded, as if water-wom, andthus may not be in true association with the bone as­semblage.

C!ues lo lhe Interpretation of the Member J FossilAssemlllage

Possih/C' Porcllpine Col!ecring 111l'oll'e/llellt

Only 1 bone, out of 2,366, shows porcupine gnaw marks.a percentage of less than 0.1. By this criterion porcupinesmay be excluded as significant collectors of the Member 1bones.

Possihle Hominid Involvement

As I mentioned earlier, onJy one artifact has definitelycome from Member 1 breccía thus far, and this showssigns of stream abrasion before incorporation in the de­posit. No bone damage clearly attributable to hominidshas been observed. On the current evidence, therefore.hominid involvement in the accumulation process is notindicated, though further excavation of Member I brecciamay change this conclusion.

Possihle Carnivore Involvement

Fossils from Swartkrans Member 1 are generaJly in bettercondition than those from Sterkfontein Member 4, thoughmany have aIso suffered badly from rough handling dur­ing mining or excavation. 1 have no doubt that my list ofspecimens showing carnivore tooth marks and otherdamage would have been a good de a] longer if all thefossils had been removed from the breccia with the carethat is essential for proper taphonomic assessment. 1t issignificant that 18 of 29 specimens from Member 1 that I

6...'

Fig. 197. (a) Part of a juvenile austra10pithecine cranium, SK 54, wirh apunctate maJ'k in each of the parietal bones. (b! A juvenile aus­tralopilhecine mandible, SK 3978. showing raggcd-edge damage lo. ilS

ramus.

have listed as bearing traces of carnivore damage (tabletOOl are from hominids , It has not been possible to pro­ccss the rnajority of Swartkrans fossils with the care thathas beco devoted to the hominid finds , ami so details ofdamage most of tbe bones may bave suffered are stillobscured by enclosing breccia.

In some cases the absence of skeletal parts is as in­dicative of carnivore action as is the presenee of toothmarks. Member 1 hyracoid or dassie remains come from amínimum of 26 animals. This estimare is based on 61 era­nial specimeos that are associated with only one posteta­nial bone, lo chapter 4 a description is given of thecharacteristic way leopards consume dassies. The entirebody is typically caten and only skull pieces remain, aswith the Swartkrans hyracoid fossils.

A remarkable range of carnivores is represented in theMember I fossil assemblage , as was discussed earlier.The carnivore/ungulate ratio is 39/%, or 37.5%, a highfigure for abone accumulation in an African cave.

An explanation has to be sought for the presence ofvery numerous hominid and cercopithecoid remains: thisis further explored in chapter )3.

After her study of the fossü bovids, Elisabeth Vrba(1975, 1976a) pointed out that, though many of the an­telopes represented in Member I could bave been killedby leopards, about half of them were too large to be re­garded as typicalleopard prey. Moreover, the proportionof juveniles did not rise as the body size of the speciesincreased. For this reason Vrba suggested that false andtrue saber-toothed cats could also have contributedsignificantly to the aeeumulation.

We should not lose sight of the fact that, in addition tothe felids known from Member l. hyenas are also wellrepresented. Remains of 3 brown and 2 spotted hyenashave been found together with those of at least 9 huntinghyenas of the genera Hyaenictis and Euryhoas. It wouldbe remarkable indeed if these had not contributedsignificantly to the bone accumulation. I suggested earlierthat the hunting hyenas may well have been social carni­vores that used the caves as breeding lalrs. Withcooperative effort they would have been able to hunt thelargest bovids and could easily have brought parts of theirprey back to the cave for their cubs.

In conclusion I have no hesitation in confinning myearlier suggestion (e.g., Brain 1970) that the Member Ibone accumulation resulted very largely from carnivorefeeding. In my opinion, both cats and hyenas are Hkely tohave been involved, as is further discussed in chapter 13.

Remains from Member 1: Microvertebrate Component

Acetic acid preparation of australopithecine and otherfossils during the early phases of Swartkrans paleen­tologieal work produced a by-product of microfaunalbones. In about 1960 the available collecnon, which hadbeen sorted and provisionally identified by J. T. Robin­son, was submitted to D. H. S. Davis, who drew up adetailed report (1955) that, regrettably, was never pub­Iished. The manuscript, referred lo here, -was found inthe Transvaal Museum fijes. The following taxa wererecorded:

Order InsectivoraFamily MacroscelididaeElephantulus (Nasilio) er. brachyrhynchus (2 individuals)E. (Elephanlomys) langi (4 individuals)

Swartkrans 239

Family ChrysochloridaeAmblysomns tChlorotulpa} speleu (number not recorded)

Order LagomorphaLepns sp. nov. (?) (1 individual)

Order RodentiaFamily BathyergidaeCryptomys robertsi (2 individuals)

Family MuridaeSubfamily MurinaeMus cf. minutoídes 06 individuals)M. cf. tríton (1 individual)Dasymys ?bo/ti (1 individual)Lemniscomys sp. nov. (about 5 individuals)

Subfamily OtomyinaeOtomys (Palaeotomys) gracilis (9 individuals)

Family CricetidaeSubfamily DendromurinaeSteatomys cf. pratensis (2 individuals)Dendromys ?antiquus (5 individuals)Malacothrix cf. typica (2 individuals)

Subfamily CrieetinaeMystromys huusleitneri (24 individuals)

Subfamily GerbillinaeTatera roblnsoni sp. nov. (3 individuals}

Sorne elements of the microvertebrate assemblage ,such as the soricids, reptiles, and birds, were not includedin the analysis , and it is no longer possible to be sure ofthe fauna! eomposition of the sample as a whole. For thisreascn two new samples were prepared from SwartkransMember 1 breccia, one from the "rodent breccia" thatoutcrops around the entrance to the "eastern shaft" inthe Outer Cave, the other from the "orange breccia,"designated Member le by Butzer (1976). It is likely thatthe "rodent breccia" represents a very early phase in theaccumulation of the Member 1 material and that Memberle represents the terminal phase. The two samplestherefore appear to provide infonnation about the faunain the vicinity of Swartkrans at each end of the Member 1time span, On the olber hand, the original sample studiedby Davis would have been intennediate in age betweenthe other two.

The samptes from the "rodent breccia" and fromMember le were found 10 contain remains from amínimum of 271 and 256 individual animals respectively.The faunal composition of the two samples is given intable 101. The composítíon is remarkably similar in eachcase, the only important difference being that dendro­murines are more abundant in the "rodent breccia'sample than in the Member 1e sample; the reverse is truefor remains of lizards. Further taxonomic study of thesamples is required, although a broad picture ofthe faunalcomposition has emerged. This is shown in figure 198.

Remains from Member 2: Macrovertebrate Component

Evidence is accumulating that Member 2 is a rather het­erogeneous stratigraphie entity, embracing breccias andcalcified channel fills of several ages. CoJlectively thesedeposits are aH younger lhan Member I and, for lhe time

240 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and lnterpretation

Phylum ChordataClass MammaliaOrder PrimatesFamily HominidaeHorno sp. (fig, 187)

Order CamivoraFamily FelidaePanthera aff. leoMaterial:Left mandible piece, juvenile with de and dpm.: left andright mandible pieces associated with 2 thoracic verte­brae, SK 359; isolated teeth, SK 360, 1865, 1869, 11818;metapodial pieces, SKW 463, SK 3133, 6684, 6747.

Cercopithecoid indet.Material:Calvaria fragment, SK 3145; maxillary pieces, SK 3446,4974, and 14150; mandible pieces, SK 2121. 2150, 4012,5165; tooth fragments, SK 2185,14202,24567; left femur,SK 5910; humerus , SK 59lc; distal tibia, SK 32110; andphalanx, SK 4t45.

Proteles cristatusMaterial:Right maxillary piece with C, SKW 122; left mandiblepiece with 2 postcanine teeth, SK 11400.

Family HyaenidaeHyaena brunneaMaterial:Almost complete cranium, SK unnumbered; right maxil­lary piece with P' and part of P'; left mandible piece withP.-., SK 328.

Family CercopithecidaePapío sp.Material:Articulated parts of a <3 maxilla and rnandible , SK 547; <3left maxilla, SK 444; <3 left mandible piece, SK 434; ~

palatal piece with broken teeth, SK 592; right mandiblepiece, SK 14100; right mandible piece from an irnrnatureindividual, SK 616.

Felid indet.Material:Metatarsal, SK 6669; phalanx, SK 6736.

Panthera pardusMaterial:Left mandible piece with part of e and P" SK 346; rightmandible piece with part of e, PH , and M" SK 348; leftdistal humerus, SK 18\0.

Material:An adult mandible, SK 15, with left M 1- :¡ and right Mz_ :¡ :

associated with it, part of right P h SK 43; left P;¡, SK 18a;and the proximal end of a radies, SK J8h.

Cercoptthecoídes wiíííamsiMaterial:<3 palate with right C-M" and left P'-M" (former typespecirnen of Cercopithecoides molletti, SK 551); maxíl­lary fragment, SK 2135; 3 mandible pieces, SK 412, 579,624.

ac"..

""S''".

Dend"""'IIII....

'....."_.". ~~.

1o>Oe1. iiiL

being, the fossils they have yielded are considered as aunit. Fieldwork is under way that should elarify the re­lationships of the deposits in the Outer Cave and provideinformation on the valid subdivisions of Member 2.

At present the Member 2 fossil sample consiste of 5,894specimens from a mínimum of 258 individual animals,details of which are provided in table 102. The variousidentified animals are depicted in figure 199. The samplewas found to inelude 3,811 specimens of bovid origin,mainly Irom the postcranial skeleton, tbat could not bespecifically identified. They were placed in bovid sizeelasses (see chapo 1 for definition of the classes); particu­lars are provided in tables 103 and 104. The remains de­tailed in table 104 consist of 2,860 specimens from an­telopes classed as Ha, or animals with body weights lowin the class II range, AH appear to have come from thesame species, and there is little doubt that the taxoninvolved is Antidorcas bondi, whose identified cranialremains are similarly abundant. The signiticance of thisstrong representation of A. bondi, together with im­plications of the observed damage to bones frcm Member2, will be considered further.

Details of the fossil material representíng the variousanimal laxa wilJ now be presented.

PERCEN1',\GE REPRESENTATION

Fig. 198. The percenteae representation ofanimals of various kinds in themicrovertebrate componen; of the Swartkrans Member I sample. A totalof 527 animals is involved.

Hyaenid inder.Material:Isolated premolar, SK 14189; proximal end of a metapo­dial, SK \0365.

Swartkrans 241

Family Mustelidae .Mel/ivora aff. sivalensís

Material:. . SK 6918; isolated right M" SKRight mandible piece,~ d I f e SK 1226.1829; isolale e t •

Family Canidae''í' mesomelas

Cani . . s SKMaterial: . SK 371' maxillary piece ,Almost complete eraOl:';ed, 3512: 10388, 10529, 1~~;~;367, 368, 2771, u~~~637 2266,2756, 4035, 6~~, 3719:mandible preces, 11426' isolated teeth, ,

10658, 10997, 1¡~8'1I815, i1823, 11927; dislal,~um;7<~~4:9t:31~, \1

4017; dista:!~~~~" :~i~~~ :~Z9~, :~'597.10760; metapodial preces, S

Otocyon reckiMateriaí:MaxiJIary piece ,nurnbered.

. ht M SK un-SK 5972; isolated rtg l'

Cf. Lycaon sp.

Material: e SK 1843.1solaled right _'

Family Viverridae .Cf. Herpestes sangumeus

~Otoeyon red,,; 1

~Melll...or. cf. I; ....!ensis 1

el. lyeeon ep. 1

Protel... e.1et«tlls,

..... 7;~~

CynlCli$ penlclllala 1

Horno ep. ,

~~~~Penth«e an. leo 3

EqUlR eepen.l. 4 Equue qlla9S1e SI P~rue entiql,lU8

~ c ....·----,-;"~ _.,,. .-.". ber z••pl• C"'~,," ....., , f the Swartkrans rnem

b te component oStnrtNoep. 1 b fossils in tite macroverte ra

. Animals represented y . d' ated In eech case.Fig. 199. be f ndividuals IS IR ICTbe mínimum num r ot r

-~~ rt~. __o' "~("'~?l • " ....,..... •

i, .n.~-~=~ ~; h,, ......." " ,"

ef. KobU. elll,..prymnu8

-;".,~-- " ~ ~• el ......'-" ._-, el .,,,,'

n-uelaJ1h118 5 . nUl'Olrag... .

242 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Material:Right mandible piece with E, P" P" and M" SK 11184.

Cynictís penicillataMaterial:Right mandibular ramus, type specimen. SK 377.

Carnivore indet.Material:Calvaria fragment, 1; mandibular pieces, 4; isolated toothfragmenta, 8; postcranial pieces, 43.

Order ArtiodactylaFamily BovidaeDamaliscus cf. dorcasMaterial:Maxillary pieces, SK 2116, 3123, 5996, 10941, 12003;mandible pieces, SK 6037, 10421, 10867, 11238, 11777,11889, 11939; isolated teeth, SK 4015, 4056, 4574, 5397,9341,9897,10653, 11271, 14111.

Damaliscus sp. 2 (?niro)Material:Hom-core piece, SK 2862; maxillary pieces, SK 1520,2516,3129,3148,4036,4044,5954, 10521, 10797, 11244,11404, 11504, 12485h; mandible pieees, SK 4016, 4219,5123, 5180, 5920, 5979, 7050, 7335, 7716, 11003, 11390,11827,11851, 14054; isolated teeth, SK 1971,2003,2017,2242,2540,3306,4065,4075,4572,5023,5172,5208,5942,5999, 6000, 6014, 6029, 6032, 7791, 8007, 10841, 10906,11117,11178,11391,11477,12145,14122,14205.

Medium-sized aJcelaphines (including cf Beatragus sp.)Material:Maxillary pieces, SK 1523, 2010, 6064, 10917, 11124,12000; mandible pieces, SK 2978, 2993, 3003, 3083, 3143,4013, 4020, 6057, 12201, 14048, 14211, 14212; iso1atedteeth, SK 2341, 3251, 5900, 5949, 5951, 5967, 5977, 5978,6008, 6090, 11199:

Beatragus sp.Material:Frontal pieee with base of a leñ horn-core, SK 14183.

ce. Connochaetes sp.Material:Isolated left M, or M" SK 6004; isolated left M" SK6059.

Cf. Mega/o/ragus sp.Material:Left maxillary pieee with M' and M', SK 14218; rightmandible pieee with M" SK 1953; isolated right M" SK3249,

Hippotragus cf, nigerMaterial:Maxillary píeces, SK 2663, 5926, 5993, 14046; mandiblepieces, SK 1428, 1977, 1992,2072,2285,2548,2954,3032,6001, 6002, 6005, 6124, 14214, 14242, 14243; isolatedteeth, SK 1947, 1980, 2355, 5909, 8010, 11641.

Cf Kobus ellipsiprymnusMaterial:Left mandible piece with P" M" and M" SK 2960; iso­lated left M', SK 11297.

Pe/ca cf. coprcnlnsMotcrial:Parts of a complete but disintegrated skull, SK 2735a,h;an almost complete palate , SK 2990; maxillary pieces, SK2923,4030,4040,6087,14049,14116; mandible pieces , SK1429.1995,2246,2273,2311,2455,2468,2981, 3015, 3035.3042, 3085,6047, 10694, 11221, 14055, 14241; isolatedteetb, SK 2090, 2308, 3124. 4029, 4087, 9911, 10741,12531.

Antidorcas australis and/or marsupialisMaterial:Horn-core pieces, SK 1428, 3011, 3071, 7281,7436,9524,10597, 14216: maxillary pieees, SK 1%0, 2115, 2730,30550, 4022, 5418, 5427, 5995, 6051, 11287; mandiblepieces, SK 2027, 2253, 2293, 2362, 2381,2479,2535,2664,2685,2702,2953,2956,2%1,2979,3037,3057,3075,3116,3138,3838,4006,4043,4081,4305,5154,5175,5958,5982,9201, 11724, 12051, 12056, 12125, 14169; isolated teeth,SK 2547, 3941, 4021, 4039, 4054, 4068, 11073, 11683,14064, 14070, 14123.

Antidorcas bondiMaterial:A partial t¡! braincase with right orbit and both horn­eares. mandible pieees, 4 thoracíc vertebrae, 2 rib pieces,parts ofboth scapulae, proximal metapodial and navícuto­cuboid, SK 14126a-p; right ¿ frontal with completehom-core. right maxilla with M:! and M:\ SK 3152; cal­varia pieces, SK 2948, 3084, 3152, 5488; frontals andhorn-cores, SK 1223,2640,2641,2647,2722,2781,2880,2946,2949,3001,3192,3253,3571,3773,4010,4534,5115,5277,6924,6959,6973,6987,7010,7115,7254,8544,9221,9321, 9393, 9658, 10402, 10634, 10702, 10943, 11070,11105, 11433, 11748, 11862 + 11596, 12689, 141250,14167, 14176, 14215, 14221; palates, SK 3014, 3103; leñmaxillae, SK 1640, 1930,2046,2115,2328,2531,3048,3112,4059,5938,7079,7426,10670,10724,11031,11036,12671, 14047; right maxillae , SK 1958,2366,2384,2439,2984,3117,3122,3147,4240,5910,5914,5975,5976,5992,7694,7703, 10350, 11557, 12578; left mandible pieces, SK1921,1957,1%5,1987,2020,2096,2113,2306,2375,2574,2578,2952,2958,2962,2%3,2977,2998,3030,3049,3054,3073,3092,4032,4041,4042,4046,4051,4445,4497,4570,5130,5143,5354,5704,5901,5922,5929,5944,5959,5962,5974,5984,5988,6021,6043,6052,6075,6080,6081.6088,6093,6116,7435,7521,7698,9385,10038, 10417, 10489,10577, 10622, 10663, 11084, 11099, 11272, 11389. 11933,12067, 12135, 12623, 12677, 14050, 14051, 14064, 14225,14226, 14227; right mandible pieees, SK 1628, 1920,2030,2085,2250,2291,2315,2321,2338,2351,2409,2490,2518,2532,2568,2705,2970,2999,3006,3009,3029,3062,3076,3079,3140,3144,3841,4011,4024,4049,4074,4086,5059,5155,5204,5880,5899,5905,5907,5908,5934,5945,5956,5986,5987,6023,6045,6076, 11412, 11986, 12273, 12472,12526; isolated teeth, SK 2051, 2067, 2264, 2277, 2289,2292,2367,2372,2387,2393,2399,2414,2417,2465,2474,2486,2506,2530,2553,2720,3248,3931,4023,4025,4061,4062,4063,4064,4071,4080,4285,4626,4633,5057,5404,5731,5882,5890,5902,5956,5990,6038,6044,6084,6095,6101, 6106, 6108, 6109, 6117, 6118, 7559, 7920, 10278,10520,10555,10601,10611,10804,11068,11122,11167,11168,11345,11506,11514,11561,11609,11637.11801,11899, 11946, 12324, 12501, 12596, 12628, 12630, 12669,12753, 14066, 14069, 14072.

". '-,

Oreotragus cf. majorMaterial:An isolated right horn-core, SK 14243.

Oreotragus cf. oreotragusMaterial:Right mandible pieces, SK 1631, 4052; left mandiblepiece, SK 11405.

Raphicerus cf. campestrísMaterial:Almost complete skull, SK 1515; maxillary piece, SK12363; mandible pieees, SK 2024, 2108, 2719, 4287, 5930,14060.

Cf. Raphicerus sp.Material:Left mandible pieee, SK 2040; hom-eore pleces, SK 7880,14170.

Ourebia cf ourebíMaterial:Maxíllary pieees, SK 5892, 5893; mandible pieees, SK6995, 14168; isolated tooth, SK 4060.

Tragelaphus cf. scriptusMaterial:Mandible píeces, SK 2329,3114,4261, 14052, 14205.

Tragelaphus cf. strepsicerosMaterial:Maxillary pieee, SK 3098; mandib1e pieees, SK 1941,2500, 3086, 6860; isolated teeth, SK 1989, 2230, 5888,5923, 10848, 14012.

Tragelaphus sp, afI. angasiMaterial:Mandible píece, SK 2980; isolated tooth, SK 4028.

Taurotragus cf. oryxMaterial:Isclated left M" SK 1417\.

Antelope size class IMaterial:239 pieees, as Usted in table 103.

Antelope size eiass HaMaterial:2,860 pieces, as listed in table 104.

Antelope size class \lbMaterial:458 píeces, as Usted in table 103.

Antelope size class IIIMaterial:245 pieces, as Usted in table 103.

Antelope size class IVMaterial:9 pieces, as listed in table 103.

Family SuidaePhacochoerus modestus

Swartkrans 243

Material:Palate with left and right M' and M', SK 4005; mandiblewith right e, P4. M._2 , left M2 and M3 associated with leftand right M', SK 382, 385, 386; left and right maxillarypieces, SK 6030, 14131a-e; right maxilla and right mandi­ble piece, SK 2359, 5989.

Suid indet.Material:Mandible piece, SK 8604; eanine fragment, SK 3825b.

Family GiraflidaeSivatheríum maurusiumMaterial:lsolated and unerupted molar, probab1y left M', SK14045.

Order PerissodactylaFamily EquidaeEquus quaggaMaterial:Parts of a palate and mandible with full dentition, SK3997a-z; palate with assoeiated mandible, SK 3998a,b;mandible pieces, SK 2243, 3989, 3994, 3999; isolatedteeth, SK 2339, 2424, 2736, 3166, 3986, 3987, 3988,3991,3992, 3995; two earpal bones associated with left mela­carpals, 1st and 2d phalanges, SK 4000a-:f.

Equus capensísMaterial:Mandible piece, SK 1942; isolated teeth, SK 1632, 2626,3160, 3164, 3984, 3990, 14133; two distal melapodialpieces, 2d phalanx and sesamoid, SK 4002a-d; left as­tragalus, SK 4001.

Equid indet,Material:Tooth pieees, SK 1837,2047, 2052, 2317, 2537, 3190, 4403,5000, 5119, 5120, 5134, 5167, 5275, 5287, 5375, 8014,10380, 10735, 12088, 12445, 14222; carpal, SK 4721;metapodial pieees, SK 5070, 9469; astragalí, SK 4924,6257; 2d phalanx, SK 7423.

Order HyracoídeaFamily ProeaviidaeProcavia cf. antiquaMaterial:Calvariae, SK 139, 144, 147, 148, 149, 174, 1615, 1959,6111, 14081; maxillary pieees, SK 123, 129, 130, 134, 150,154, 159, 166, 169, 173, 176,215, 1620,4094,4628,4663,5940,6017,6899,9295,10458,11110,11183,11507,11565,12525, 12675; mandib1e pieces, SK 114, 118, 125, 127,131,132,133,137,142,150,158,167,170,171,175,180,181, 185, 202, 210, 1614, 1621, 2212, 2536, 2746, 2951,3026,3027,4017,4204,4647,5948,6015,6023,6085,6102,6110,7205,7338,7342,7501,10017, 10534, 10539, 11169,11353, 11518, 11602, 11809, 11930, 12680, 12702, 14143,14146;isolated teeth, SK 168,190,3324,3361,3425,3454,3476,6853, 10295, 11204, 14145, 14165; distal humeri, SK4150,9802, 11306, 11308, 11411, 11687, 11955, 14136.

Procavta transvaalenstsMaterial:Complete eranium with mandible, SK 216 + 111a,b;maxilla with assoeiated mandíble, SK 183a,b; calvaria,SK 112b, 115, 157, 199,200,212,217,2197,7061; maxil-

244 Fossil Assernblages from the Sterkfontein Valley Caves: Analysis and Interpretation

lary pieces , SK 193. 194a.h. 206,3479,6031,7420; man­dible pieces. SK In, 151. 152. 153, 155, 172, 187, 191.213.214.2106,3474: isolated teeth. SK 2899, 3243; as­sociatcd postcranial boncs: right scapuJa, SK 192; rightdistal humeros, SK 198; right radius and ulna and carpalbones , SK 197; left distal radius and ulna and carpals , SK201.

Procavia sp.Material:Calvaría pieees, SK 3301, 3327, 3730, 3732, 3736; maxil­lary pieees, SK 3196. 3716, 7093, 7262, 7529, 9411; man­dib!e pieees, SK 2508, 3304, 4260, 4568, 7135, 7651, 9555,10376, 14126, 14200; isolated teeth, SK 3775, 3796. 5886,7627; humerus shaft, SK 10117; distal humerus, SK 4409.

Order LagcmorphaFamily LeporidaeLagomorph gen. et sp. indet.Materia!:Craniums with associated mandibles, SK 3345, 14097;calvaría pieces, SK 1617, 3299 + 3734, 3326; rnaxillarypiece, SK 3485; mandible pieees, SK 2895, 3226,3354 + 3394 + 3604, 3391, 3407, 3818,11071.

Order RodentiaFamily HystricidaeHvstrix afrícaeaustmtisMateríat:Maxillary piece. SK 2835; mandible pieces, SK 3063,6945, 14235; isolated teeth, SK 2040, 2045, 7155, 14238.

Class AvesOrder StruthioformesFamily StruthionidaeStruthio sp.Material:lndet. bone pieces, SK 7728, 8633; eggshell pieees, SK7171,14126.

Class ReptiliaOrder CheloniaFamily TestudinidaeChelonian indet.Material:Humeros pieces. SK 9866, 11947; carapace pieces, SK4678,4775.7739,9835, 12342, 12408.

Phylum MolluseaClass GastropodaOrder PulmonataLand snail, cf. Achatínu sp.Material:Shells in varying degrees of completeness, SK 2609, 2660,2887,3185,6946,7628,8126,8213,9264,11537.

Indeterminate fragmentsA total of 723 pieces constitutes this category,

Bone flakesMaterial:330 pieces with lengths as listed in table 102.

CoprolitesMaterial:Parts of two carnivore coprolites, SK 7542, 7667.

Swartkrans Member 2: The Major Features of the BoneAccumulation

The Compositíon of thc Fauna Re prescnted hy thcFossíls

The information provided in table 102 and depicted infigure 198 may be further surnmarized. This is done in thesecond columns of table 99 and figure 196. The faunalcornposition of this assemblage forms a striking contrastto that from the older Mernber 1 at Swartkrans. Australo­pithecines have disappeared ; a single Horno is repte­sented together with only 8 individual cercopithecoids.The collection is dominated by no fewer than l60 individ­ual antelopes, 118 ofwhich fall iota síze ctass Il. The classhas been further subdivided into Ha and lIb. the forrnercontaining the smaller antelopes, particularly Antidorcasbondi, of which at least 70 individuals are represented,

In size class J, 12 individuals of klipspringer, steenbok,and oribi are found, while 27 individuals of class In an­telopes are represented. OC these, 17 are juveniles. Re­mains of 3 adult individuals of larger, class IV bovidshave also been found.

Possils of at least 37 individuals of two species of das­sies have been recovered, together with parts of at least21 camivores. These inelude 3 lions, 2 leopards, 2 brownhyenas, 1 aardwolf, 8jackals. and various smaller animalsas depicted in figure 198.

The Representatíon 01 Skeletal Pares

Particulars of the various skeletal parts making up theassemblage are given in the various tables. A most strik­ing and valuable feature of the accumulation is the collec­tion of 2,860 bones of class na antelope. which withoutdoubt may be associated with the 70 Antldorcas bondiindividuals listed on the basis of cranial fossils. Tbesebones come, perhaps entirely, from the younger compo­nent of Member 2. which occurs as an irregular sheet oflightly calciñed breccia in tbe central area of the OuterCave (see aboye). The sample is large enough to provideuseful insights into those parts of the springbok skeletonsthat survived and those which disappeared before fossili­zation.

On tbe basis of the Iisting of skeletal parts provided intablé 104, it has been possible to work out percentagesurvival figures for many of these parts, assuming that amínimum of 70 individual springbok contributed to thesample. Results are given in table 105 and are shown pie­torially in figure 200. Taking certain cranial parts as repre­senting l()()tk survival, a very considerable variation inrepresentation occurs from one area of the skeleton to thenext. The vertebral colurnn and rib cage have sutTeredparticularly severly, together with vulnerable segments ofthe limb bones such as the proximal humeros and prox­imal tibia. Many of the lower limb segments are particu­larly well represented. The significance of the survivalpattern shown by the A. bond! skeletons will be consid­ered shortly.

Observed Dümage to the Rones

Details of specific damage observed on bone pieces fromMember 2 are given in tabIe 106 and may be summarizedas follows:

Porcupíne and Small-Rodent Gnaw Martes, The typicaJgnaw rnarks of porcupines were observed on 12 pieces.

1¡

t¡,

and 163 specimens bore rnarks caused by the incisors ofsmall rodents.

Carnil'ore·/njlicled Dümage . As listed in table 106, nofewer than 291 bones, out of a total of5,884, showed clearevidence of carnivore damage , and another 123 bore lesspositive traces. The great majority of the damaged pieceswere bovid skeletai parts , and, from the interpretativepoint of view, bones from class IIa are of particular inter­est here. As detailed in the table, carnivore damage wasobserved 00 17 vertebrae, 6 scapulae, 9 pelvis pieces, 55humeri, 15 radii, 46 femurs, 19 tibiae, 3 calcanei, and 39rnetapodials from bovid class lIa.

Traces 01 Artificial Bone Alteration, Cut marks, certainlyproduced by a sharp-edged artifact, have been observedon the mandibJe of an oribi and on 2 other bones. Un­doubted chop marks were also found on a limb boneshaft.

Association with Artifacts, The rather unsatisfactorysituation regarding stone arufacts from Member 2 hasaJready been discussed. A single biface has been found insítu in the younger component ofMember 2 (fig. 188), and24 other tools were embedded in Mcmber 2 matrix,though it is not known from what part of the deposit theycarne. Until carefully controlled excavation is undertakenit wilJ not be possible to say whether tools occur as ascatter throughout the depth of the Member or whethercertain layers reftect more intensive human occupation.

Cines to the lnterpretation of the Member 2 FossilAssemblage

Possibíe Porcupine Colíectíng Invoivement

Porcupíne gnaw marks were observed on 12 bones out of5,884 specimens, giving a gnawed-bone percentage of0.2%. By this crítcrion, porcupine involvement in the

Fig. 200. A diagram of the skeleton of the extinct springbok. Antidorcas/Jundi, of which remains from 70 individuals have been recovered fromSwartkrans Member 2 breccia. The percentage survival of various skeletalparts in rhe sample is indicated. Further details aTe given in rabie 105.

Swartkrans 245

collecting process may be discounted. A larger number ofbones. 163, showed traces of small-rodent gnawing. Sinccmany of the bone pieces weighed more than the estimatedlivcweight of the rodents that gnawed them. it is likclythat the gnawing was done where the bone s were found.

Posstbte Hominid Involvement

The Acheulean tools occasionally found in Member 2clearly indicate that the cave was intermittently occupiedby hurnans during the accumulation of this deposito Thisis borne out by the fact that a few of the bone pieces showunmistakable cut and chop marks. As I rnentioned earlier,Member 2 is not a homogeneous stratigraphic unit, andthe fossils under consideration here come from older andyounger breccia components. It is not irnpossible that thetocls were concentrated in only one part ofthe profile, butthis cannot be establíshed without careful excavation.

Possible Camívore Involvement

Of the 258 individual animals identified from Member 2,21 are carnivores, giving a carnívore/ungulate ratio of21/179, or 11.7%. The presence of remains from 3 lions issomewhat unexpected, since these cats normally havelittle to do with caves. Camivores represemed in themember that might well have contributed to the accumu­lauon process are leopards and brown hyenas; the othersmaller species are. to my mind, more Iikely to haveformed part of the prey.

From her study of the fossil bovids, Elisabeth Vrba(1975, 1976a) pointed out that the Member 5 bovid as­semblage was characterized by animals of low liveweight,there being 130individuals in classes I and II out ofa totalof 160. Moreover, of the rernaining 30 larger bovids, nofewer than 17 were juveniles. These facts led Vrba tosuggest that the collecting agents were either carnivoresof fairly small prey-adaptation, hominid hunters, or both.

Let us first consider who, or what, could most likelyhave been responsible for the Aruídorcas bond; compo­nent of the fossil fauna. At least 70 individuals of thissmall springbok are involved, of which 17 are juveniles.The evidence I have presented indicates that, althoughparts of the skulls and lower legs are very well repre­sented, the vertebral columns and rib cages have in manycases disappeared, Evidence of camivore damage to thevarious skeletal parts is widespread and convincing, Jhave little hesitation in suggesting that the springbok werekilled by carnivores and consumed within the catchmentof the Swartkrans cave entrance as it was in Member 2times. The pattern of damage and part-survival is verysimilar to what 1 described in chapter 4 where leopardswere observed to feed on impala. To my mind, the obvi­DUS choice for the killer of most of the Member 2springbok was the leopardo

As with Member I fossils, the hyracoid oc dassie re­mains from Member 2 are very suggestive of leopard in­volvement. A minimum of47 animals are involved, repre­sented by 160 cranial pieces, with which are associutedonly J4 postcranial bones. The presence of skull pieces,with an absence of postcranial bones, could well haveresulted from the kind of felid feeding described in chap­ter 4.

Despite this conclusion. the evidence is c1ear thathuman hunters were also involved in building up theMember 2 bone assemblage, and it would be very sur-

246 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

prising if the cave had not been used as a brown hyenabreeding lair as well. Numerous remains of small carni­vores could very well have come from prey that brownhyenas fed to their young in the cave.

Remains from Member 2: Microvertebrate Component

Microfaunal remains are surprisingly rare in Member 2and, although a large volume of breccia has been dis­solved in acetic acid, the yield of small bones has not beensufficient to pennit a faunal analysis.

Remains from the Channel Fill: MacrovertebrateComponent

As 1 mentioned earlier, the sample of bones to be dis­cussed here carne from approximately 2 m' of un­consolidated sediment excavated from a solution channelpassing through both Member 1 and Mernber 2 in theOuter Cave. The positíon of the excavated area is in­dieated on the plan (fig. 186).

In aH, 837 bone pieees were reeovered (table 107), in­cluding an australopithecine maxillary fragment that hadobviously weathered out from the wall of the channel. Itis likely that sorne of the other bones in the sample werelikewise derived from older breccia in the walls. The ani­mals represented by the fossils are depicted in figure 20t.

Among the bone pieces were 178 specimens of bovidorigin, largely postcranial, that were not specificallyidentified but were simply assigned to antelope sizeclasses (as defined in chapo 1). Particulars are given intable 108.

The material by which each identified taxon is repre­sented is now presented.

Phylurn ChordataClass MammaliaOrder PrimatesFamily HominidaeAustralopithecus robustusMaterial:A single cranial piece, weathered from Member I brecciaof the ehannel waH.

Left maxilla with P"" MI-' and part of M', SKW 12.

(ll\~~- ........ ,

""- ~<1 , e '•••_ .

.~," ~-~c""""""u...... I .....-

~ ::!!.EOUUl ;" '_EL" .,. CA"""" •

Fig. 201. Animals represented by the fossils in the macrovertebrate corn­ponent of the Swartkrans channel fill sample. Minimum numbers of indi­viduals are indicated in each case. The single Australopithecus individualis represented by a maxillary fossil derived from Member 1breccia fonningthe wall of the channel.

Order CamivoraFamily FelidaeCL Pantlicra purdusMaterial:t cranial and 1 postcranial piece, probably from a singleindividual.

Isolated carnassial, SKW 2669; distal radius, SKW2859.

Family CanidaeCanis rnesomelasMaterial:3 cranial pieces from a minimum of 2 individuals.

Piece of maxilla with articulated mandibular fragment,SKW 2616; maxillary fragment, SKW 2463; mandible frag­ment, SKW 3054.

Camivore indet.Material:12 postcranial pieces from an estimated minimum of 3individuals.

Distal humeros, SKW 2809; ulna shaft, SKW 2866;proximal fémur, SKW 2838; distal tibia, SKW 2825,2873;metapodial pieces, SKW 2841, 2670; pelvis fragments,SKW 2436, 2456; ealcaneus, SKW 2647; phalanx, SK W2399.

Order ArtiodaetylaFamily BovidaeMedium alcelaphine, inciuding Damattscus sp. and Con­nochaetes sp.Material:17 craníal pieces from a mínimum of 5 individuals.

Isolated left dpm¿ SKW 2638; M', SKW 2466; left M",SKW 2462, 2484; right M" SKW 2479; tooth fragments,SKW 2464, 2471, 2472, 2478, 2483, 2490, 2645, 2652,2695,3059, 3095, 3181.

Antidorcas bondíMaterial:6 cranial pieces from a minimum of 1 juvenile and 2adults.

Left mandible pieee with dpm.. and M" SKW 2465;isolated right dpm; SKW 2668; left mandible piece, P,(roots). P,." M,." SKW 2455; right mandible pieee withM" SKW 2494; isolated right M" SKW 2486, 2605.

Antidorcas cf. marsupialisMaterial:5 cranial pieces from a mínimum of 2 adults.

Hom-eore pieces, SKW 2495, 2628; isolated left M"SKW 2467, 2631; M', SKW 2603.

Antelope size class IMaterial:4 cranial and 37 postcranial pieces from an estimated 3individuals, as listed in table 108.

Antelope size class 11Material:33 cranial and 73 postcranial pieces from an estimatedminimum of lO individuals, as listed in table 108.

Antelope size class IIIMaterial:6 cranial and 24 postcranial pieces from an estimatedminimum of 3 individuals, as listed in table 108.

Antelope síze cfass IVMaterial:A single' postcranial piece , detailed in table 108.

Order PerissodactylaFamily EquidaeEqulIs cf. burchellíMaterial:4 cranial and 3 posteranial pieces, probably from 1 indi­vidual.

Tooth fragments, SKW 2459, 2468, 2498, 2602;metapodial pieces, SKW 2360,2867; phalanx, SKW 2819.

Order HyracoideaFamily ProcaviidaeProcavia cf. capensisMaterial:11 cranial and 16 postcranial pieces from a minimum of 5individuals.

Maxillary fragmente, SKW 2449, 2476, 2492, 2609,3128; mandible pieces, SKW 2606, 2623, 2640, 2652; iso­lated teeth, SKW 2644, 2655; distal humen, SKW 2678,2820, 2822, 2823, 2826, 2827, 2834, 2835, 2842, 3005;humeros shaft, SKW 2831; proximal radius, SKW 2879;femur shaft, SKW 2836; distal tibia, SKW 2843; cal­caneus, SKW 2549; phalanx, SKW 2833.

Order LagomorphaFamily LeporidaeGen. et sp. indet. HareMaterial:3 cranial pieees from 2 individuals.

Maxillary pieces, SKW 2618, 2649; molar fragment,SKW 3198.

Class AvesOrder GalliformesFamily NumididaeCf', Numida sp. Guinea fowlMaterial:5 postcranial pieces, probably from 1 individual.

Proximal ulna, SKW 2846; tibiotarsus píece, SKW2806; femur shaft piece, SKW 3004: tarsometatarsuspieces, SKW 2839, 2899.

Class ReptiliaOrder CheloniaFarnily TestudinidaeGen. el. sp. tndet. TortoiseMaterial:3 carapace pieces, probably from 1 individual.

SKW 2493, 2619, 3166.

Indeterminate fragrnents193 pieces.

Bone f1akes375 pieces as listed in table 107.

Swartkrans Channel Fin: The Major Features ef tbe BoneAccumulation

The Composition 01 the Fauna Represented by theFossils

The information provided in table 107 and depicted infigure 20] may be further sumrnarized. This is done in the

Swartkrans 247

third columns of table 99 and figure 196. This smal1 sam­ple , derived from about 30 anirnals , is dominated byborres from 17 individual antelopes , 3 of which were ofsize class 1. 10 from class II, 3 from c1ass 1I1, and 1 fromc1ass IV. Second in number to the bovids were 5 dassies,followed by 3 carnivores and 2 hares.

Representation 01 Skeíetaí París

Particulars of the bovid skeletal parts are glven in tablelO8. The sample is too smalI to reveal much of a feedingor butchery pattern, although in the first three size classesa fair scatter of parts throughout the skeletons does ap­pear to be preserved, from skull to feet.

Observed Damage to the Bones

Details of specific darnage observed 00 bone pieces fromthe channel fill are given in table 109 and may be surn­marized as follows:

Porcupine and Small-Rodent Gnaw Marks. Porcupinegnawing was observed on l specimen ; srnall-rodent gnawmarks appeared on 2 bones.

Camivore-Infíicted Damage. Clear traces of this kindwere fouod 00 4 bones. Iess positive traces appeared on 2others.

Traces 01 Artificial Bone Altera/ion. A probable chopmark was found on a bovid horn-core, and one bone flakehas wom edges, suggestive of human agency.

Associutíon with Artifacts. No tools were found directlyassociated with these bones, though flakes have been ob­served in other similarly loase channel fills of theSwartkrans cave system.

Clues to the Interpretation ef the Channel Fin BoneAssemblage

The assemb1age must be regarded as one of mixed prov­enance, since sorne specimens, such as the australopith­ecine maxiHa SKW 12. have weathered out from olderbreccias forming the channel wall.

Absence of any porcupine-gnawed pieces but oneexcludes porcupine collecting as an important agency inthis case. The bones show indicetions of both human andcamivore involvement, and the presence of leopard re­mains suggests that these cats may have used the ramify­ing channels as a lair. On the olher hand, that 375 of the837 bone pieces proved to be bone ftakes points to pos­sible human food remeíns. As with bones from most ofthe other Sterkfontein valley site units, the assemblagesuggests involvement of more than one accumulatíngagency.

Remains from the Channel FiII: Microvertebrate Remains

No microfaunal remains were present in the excavatedsediment.

12 Krorndraai

A Brief History of Activity

It was al Krorndraai that the first robust australopithecineknown to science was discovered. Circumstances sur­rounding that discovery are best told in the words ofRobert Broom (Broom and Schepers 1946, p. 3):

One Wednesday in June , 1938,1 visited Sterkfon­tein, when the caretaker, Mr. G. W. Barlow, showedme a palate with a molar tooth that had been pickedup. [ immediately purchased it from him; but he didnot seem inclined to tell me how he had obtained it, orwhere it had been picked up. It was manifestly thepalate of an anthropoid allied to the Sterkfontein skull,bu! apparently different. Sorne teeth had been freshlybroken off, and there were other evidences of freshfractures. The rnatrix was different from that in theSterkfontein caves, and 1 felt sure it had come fromelsewhere. 1 returned to Sterkfontein on Saturday,when 1 showed the palate to the Kafirs, who worked inthe quarry; but none had seen the specimen before.Barlow was away. Determined to get to the bottom ofthe mystery 1 was again at Sterkfontein 00 the Tues­day following, when 1 insisted on Barlow telling mewhere the specimen had come from. He told me that itwas a school boy, Gert Terblanche, who had picked itup at Kromdraai, two miles east of Sterkfontein. Gert,who was about 15 years of age, acted as a guide tovisitors to the caves on Sundays. Of course 1 im­mediately set off in pursuit of Gert. 1 found he líved ona small farm about two miles away, and 1 went to hishorneo

Gert was at school, another two miles away; but 1saw his mother and sister; and his sister took me tothe top of an adjacent hill where the specimen hadbeen broken out of a weathered outcrop of bone bree­cia. 1 picked up two or three Dice fragments, and acouple of teeth: bUI the sister told me that Gert hadfour beautiful teeth with him at school; and she wassure he had sorne other nice pieces hidden awaysomewhere.

1 went off to the schoot. About a mile of the waywas so rocky that it was impossible to go by car. Itwas playtime, about 12,30 prn., when 1 arrived. 1 sawtbe principal, and told him what 1 had come about.Gert was found, and drew from his trouser pocket fOUTof the most wonderful teeth ever seen in the world'shistory. Two of these fitted on to the palate which 1had in my pocket. The other two were teeth of theother side.

248

Gert told me that there were sorne more bits on thehillside hidden away. As tbe school only broke up al 2o'ctock 1 suggested to the principal that 1 would give alecture to the teachers and children, on how caveswere formed, and how animal bones got into them. Hewas delighted. So] lectured to four teachers and about120 children for over an hour. As it was now almost 2o'clock the principal dismissed tbe classes: and Gertcarne away with me, and took me to the place on thehill where he had carefully hidden away a nice lowerjaw. 1gathered every fragment ,

When 1 retumed to Pretoria, and c1eaned up andjoined the fragmente. 1 found 1 had most of the palate,much of the left side of the face, almost the whole ofthe left zygomat¡c arch, and the left side of tbe base ofthe skull, a considerable portian of tbe parietal, andothe greater part ofthe right mandible, with most oftheteeth. Alrnost the whole of the right side of the skullhad been weathered away with the back of the occiputand the top of the brain case; but otherwise the skullwas well preserved, and quite uncrushed.

By later sieving the grcund 1 was fortunate in findinga number more teeth; and we now have nearly everyfragment that was there when Gert broke the skull outof the matrix with a hamrner. Believing this Kromdraaiskull to be very different from the Sterkfontein form 1have called it Paranthropus robustus,

After the original Kromdraai discovery in 1938, thetempo of paleontological work in the Sterkfontein valleyslowed for a while. A fall in the price of lime broughtquarrying at Sterkfontein to a halt in 1939, and soonthereafter World War JI broke out. Nevertheless, in Feb­ruary 1941 Broom sent a small work party to Kromdraaifor a short period to investigare breccia adjacent to thespot where the type skull had been found. This work re­sulted in the discovery of a juvenile mandible with excel­lent deciduous dentition. The work was stopped when theunderlying stony breccia appeared to be sterile.

Paleontological work at Kromdraai A, or the Fauna!Site, appears to have been restricted to the period Janu­ary to April 1947, before Broom resumed his work atSterkfontein. A good deal of breccia was blasted from thesite, and promising blocks were taken back to the Trans­vaal Museum, where they were processed over a numberofyears. The resulting fossil collection is described in thischapter.

In an attempt to obtain further information about theextent and nature of the Kromdraai B deposit, 1 carried

Kromdraai 249

out fieldwork there fmm March 1955 until May 1956.Very [illle so[id breccia was removed, but decalcifiedbreccia along the north wall of the cave was excavated toa depth of 4.9 m, or 16 ft (fig. 203). The excavation re­vealed that the deposit was considerably more extensivethan we had thought, and that the stratigraphy was in­clined fmm east to west. Bones recovered from the de­calcified breccia, including robust australopithecine teethand pan of a pelvis, are included in the analysis to bepresented here.

A new excavation of the Kromdraai B deposit (fig. 204)was initiated by E. S. Vrba of the Transvaal Museumtoward the end of 1976. Cores have been drilled atselected spots along the length of the deposit and haverevealed a considerable depth and an interesting succes­sion of breccias. Samples have been taken forpaleomagnetic orientabon, and the results are awaitedwith interest. At this time Vrba and her team are prepar­ing the first australopithecine specimen (fig. 205) to come

Fig. 203. Excavalions conducred al Kromdraai B by lhe aurhor in 1955.rhe ai m being lo define lhe northem margin of lhe cave deposil.

Fig. 204. Renewed excavations al Kromdraai B in 1977. under rhe direc­rion of Elisabelh v roa.

Fig. 205. Elisabelh Vrba and David Panagos wilh lhe firsl in Silu auslraJo­pithecine specimen lO come from lheir new excavalions al Kromdr.aai B.

,\

1

\

,, ,,

Fig. 202. Reproduction of two drawings by Robert Broom of the lypespecimen of ParanlhroplIs robllslIIs. fossil and reconsrruclion. FromBroom 1950.

250 Fossil Assemblages from the Sterkfontein ValIey Caves: Anaiysis and Interpretation

from the new excavation. It is an especially significantfossil in that ji may be tied precisely to the rather complexstratigraphy of the site, which is now starting to unfold.

Sorne Notes on the Krorndraai Sites

The Kromdraai cave sites He clustered on the south sirleoí the Bloubank River, 1,750 m east-northeast ofSterkfontein. The Kromdraai A and B deposits both con­sist oí calcíñed fillings oC narrow solution galleries, paral­lel to one another, trending east-west and no more than 17m apart (see fig. 206). The dolomitic country rock appearsreasonably undisturbed, dipping in a generally northerlydirection al angles of 11°_25° in the immediate vicinity oíthe fossil sites. lt appears that at Krorndraai, as atSterkfontein, the country rock has developed a rec­tangular joint pattem trending north-south and east-west.The two rnain fossiliferous deposíts at Kromdraai fillcavities that resulted from enlargement of east-west Jine­aments, while others such as Kromdraai C (KC in fig. 206)are aligned north-south. Kromdraai A and B caves haveboth lost their dolomite roofs through erosion, and thevertical extent of the KA deposit is not known. Inclina­tion of the stratigraphy, downward toward the west in theKromdraai B deposit strongly suggests that the originalopening to the cave was al a higher level and at the east­ern end. It is not impossible that the Kromdraai e depositis an extension oC the KB accumulation to the east.Vrba's current excavation will clarify this.

The alignment and proximity of the KA and KB cavessuggest that they may have formed two paraUel galleriesof a single cavem system. In the past it has been custom­ary to associate the fauna from KA with the hominidsfrom KB and to assume contemporaneity. Fauna! studiesnow suggest that the deposits were almost certainly notcontemporaneous (e.g., Freedman and Brain 1972; Hen­dey 1973; Brain 1975b). It therefore appears that thefillings of the two parallel and adjacent galleries wereintroduced at different times. It is not clear whích isdefinitely the older.

The Kromdraai ABone Accumulation: MacrovertebrateComponeDt

The collection consists of 1,847 fossils from a minimum of194 individuals, as detailed in table llO. Particulars of702

~~'-"r"'" ••

bones of bovid origin, assigned to size classes but notspecifically identified, are given in table 111. The variousidentified animal s are depicted in figure 207. Data on thefossil material by which the various animal taxa are repre­sented will now be presented.

Class MammaliaOrder PrimatesFamily CercopithecidaeParapupío jonesiMaterial:Right maxilla with M hí. KA 175; right mandible piece withM._ s, KA 176.

Papio rohinsoniMaterial:Lefl maxillary piece with M' and M". KA 160.

Papio angusticepsMaterial:Type specimen, an almost complete? cranium, withoutmandible, KA 194; snout with drn', dm", and MI bilater­a1ly, KA 155; snout with parts of both toothrows , KA 156;snout with left P4_M3, and right M3, KA J57; maxillary andmandible pieces, associated with parts ofa distal humeros,KA l68a-d; right maxilla with P3_Ml, assoctated withparts of an articulated hand, KA 1670-[; right mandible.piece with P;rM3. associated with 2 isolated teeth, distalhumeros, proximal tibia, and part of a phalanx, KAJ66a-g; left side ofa cranium with drn', and MI, KA 151:right maxillary piece, KA 161; mandible pieces, KA 163.165,179.

Gorgopitherus major (fig. 208)Matertaí:Type specimen, left M' and M', KA 193; completecranium, without mandible, KA 192; maxillary pieces,KA 153, 154, 169, 524, 605; mandible pieces, KA 150,152,173, 198, 1148; isolated teeth, KA 170, 180, 182, 183,620.

Cercopithecoid indet.Material:Calvaria pieces, KA 976, 1689, 2454; maxillary pieces,KA 159.944, 1l42, 1421, 1434,2333; mandible pieces, KA158, 164, 167, 171, 172, 177,613, 1240, 1254. 1489, 1538,2327,2331,2373; isolated teeth and tooth fragments, KA181, 184, 185, 186, 187, 188, 189, 708, 1l71, 1231, 1327,1439,1562,1627,2332,2337,2340,2341,2345,2348,2349,2361,2362,2367,2375,2384; thoracic vertebrae, KA 745,2379,; lumbar vertebra, KA 507; proximal femur, KA1658, 1798; distal femur, KA 190, 191, 699.

Fig. 206. Plan of the Kromdraai A, B, and e deposite, based 00 a tevelsurvey by the author in 1973. A sectíon through the eastem eod ofKromdraai A is also shown.

i. . ..íñiin..

KIJ-~'. ~.-t....

O""O ......0 ~

Family FelidaePanthera leoMaterial:Left maxillary piece with P', KA 67; isolated left p'. KA68; right P" KA 69; right M" KA 70.

Petís crassidensMaterial:The type specimen, which consists of right maxillarypiece with P' and P" KA 87; also left maxilla and rightmandible piece referred to by Broom (1948). but nowmissing.

Krorndraai 251

Fclis sp."Ha/erial:Isolated 1eft P'. KA 1544.

left scapula and proximal humeros: and various otherfragments , KA 64.

Homothcrium sp.Material:Isolated 1eft and right M" KA 6611,b.

Megantcreon eurynodonMaterial:The type specimen. which consista ofmost ofthe craniumwith articulated mandible , severelv decayed; articulatedatlas, axis, and parts of other cervical vertebrae; head of

Dinofelis píveteaui (fig. 105)Material:The type specimen, which consists of a completecranium, KA 61; right mandible piece with P~, KA 62: leftmandible piece with Mio KA 63.

Family HyaenidaeCrocuta crocutaMaterial:The type specimen of C. crocuta ultra, which consists of

•

Vulp.. puletler 1

7 HippotrsSlIftl

C.nla me.olll.l.s e

HlpputrwgUl et. equiftW;1

cf Cohlmba .... 1

Homotherium 1

'p.

Phl'ODChoer".~1

1rOinoleH. piveteeui

y­il tHetpe.t.. rnnot.. ,

Hyslrlx el m.k.pen.....I.I

Equ". burehem 5

~Croaurehu.

Hystrilt .lric..sue,r'lIs3

t.L'"ornorph

ind.t,

Pelee el....preoiu. Antidl>rCe. bondl

". ~-~r'f ~-_.~_..Equu, ".pen." 23

,f1t'l( ~~ ~P.pl!> rob,"soni 1 11 Gorgoplthieu. m.jl>r e

p~ Jt'f F~Fo' .. «0""0"' , '"

~Croeul. erocut. 3

•Hyaena o.lIu 1

HlpperlDn lKeytlerl,....proesvl. tnulsv'sl.nsis Proc:evl.

• • ...tlqua10

Fig. 207. AnimaIs represented by rhe toseüs in the macrofaunal componen! of the Kromdraai A sample. Minimumnumbers cf individuals in each case are Indicated.

252 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

many separate parts: snout and right zygoma, posteriorpart of braincase, left and right mandible pieces, variousisolated teeth, thoracic vertebra and 7 vertebral disks, 8Jib pieces, proximal libia, 10 metapodial pieces, 3 carpalbones, 22 phalanges, KA 58; the type specimen of C.spelaea capensís, an almost complete cranium, KA 56;left mandible piece with e, Pz, and p.!> associated with 2skull pieces, KA 57a-('; isolated right M" KA 59; isolatedp.l , KA 60.

Hyaena bellax (fig. 209)Morería!:An almost complete cranium with articulated mandible,the type specimen, KA 55.

Hyaenid indet.Material:Right maxillary fragment with pH, KA 538; part of abraincase, KA 1664; isolated teeth, KA 861, 954, 985,1429, 1437, 1525.

Family CanidaeLycaon sp.Materíal:Former type specimens of Canis atrox, isolated right M"KA 1288; left M', KA 1556.

Canis mesomelasMaterial:The type specímen of C. mesomelas pappos, which con-

Fig. 208. A reroarkably complete cranium of Gorgopilhecus majar fromKromdraai A.

Fig. 209. The type specimen of Hyaena bellax. an extinct specles relaledto the brown hyena, froro Kromdraai A.

sists of left and right mandibular rami, isolated right P,and right M,. KA 73a,h: complete cranium and half­mandible, KA 7[; crushed skull, KA 88; maxillary pieces,KA 77, 85, 691, 1118,1297; mandible pieces, KA 74, 75,766, 986, 1456, 2086; isolated teeth, KA 76, 78, 79, 948;left scapula head, KA 72.

Canis terblancfteiMaterial:The type specimen, consisting of most of a crushedcranium with mandible, KA 1290.

Vu!pes pulcherMaterial:The type specimen, conSlstmg of the left mandibularramus with C-M 3 , KA 1289.

Family ViverridaeHerpestes meso/esMaterial:The type specimen, consisting of a cranium and mandible,3 cervical vertebrae, left humeros, right proximalhumerus, distal radíus, and ulna fragments, 4 metapodialsand phalanges, KA 86.

?Crossarc!lCIS transvaalensisMarería!:A snout with both toothrows, KA 1569.

Viverrid indet.Material:A cranium and endocast, KA 1488; right mandíble piecewith P~ and part of MI, KA J508.

Camivore indet.Materia!:Calvaría pieces, KA 68J, 1677; maxillary pieees, KA 676,1580, 1600, 2360; mandible pieees, KA 1444, 1570; atlasvertebra, KA 1164; seapula piece, KA 1723; ribs, KA 741,1226; humeros, KA 642; distal radius, KA 1145; femurpieces, KA 547, 637, 643, 849,1208,1696; astragalus, KA756; metapodial pieces, KA 629,634,795,920,943,1037,1054,1126,1148,1773,1788,1815,1842,1875; phalanges,KA 553, 556,593,693,932,956,996, 1612, 1647.

Camivore coprolitesMaterial:2 ofjackal size, KA 1420; hyena size, KA 732,1490,1796.

Order ArtiodactylaFamily BovidaeDamaliscus sp. 1 or ParmulariusMaterial:Complete but shattered eranium with articuJated mandi­ble, atlas, and axis, KA 1592 + 1716; right side ofcranium with articulated mandible, atlas, and axis, KA731; upper part of cranium, snout, and hom-core bases,right p4 and M3, KA 1601a--c; cranial piece and almostcomplete left mandible, KA 931a,b; maxillary and mandi­ble pieces, KA 709a-g; base of right hom-eore, KA 540;maxillary pieces, KA 514, 525, 526, 564, 587, 635, 646,697,731, 750a, 762,877,897,960,970, 991a-d, 994,1127,1619a, 1750, 251Ib. 2512a; mandible pieces, KA 541,566,569, 576,578,646a, 700,725,728,749,751,758,770,776,833, 855, 858, 867, 913, 929, 935, 951, 969, 1004, 1010,1101, 1102, 1122b, 1134, 1136, 1153, 1183, 1198,'1204,1246,1296,1484,1516,1553,1587,1635,1653,1691,1668,

16870. 1739. 1747, 1774, 1775a. 1784, 1803, 1827a, 2353,2450, 2511, 2608a, 2611, 2613; isolaled teeth , KA 516,565,583,585,587,606,608,611,632,656,659,665.668,673, 692b, 703,718,743,780,822,832,836,838,841,862,868.872,898,916,918,920,922,926,947, 1041, 1095.1096, 1097b. 1104,1108, 1113, 1115, 1117, 1140, 1168,1170a, 1173, 1174, 1186, 1195, 1211, 1244a, 1258". 1272.1306.1344,1460.1631,1636,1642,1646.1660,1681,1686,1688,1693,1701.1711.1737.1745,1751,1752,1776,1800,1824, 2481, 2483, 2612.

Medíum-sized alcelaphinesMaterial:Maxillary pieces, KA 794", 924, 1067, 1273, 1781,2410,2500; mandible pieces, KA 542, 680a, 1I56a,b, 2453,2497, 2514; isolated teeth, KA 532, 650, 771, 825. 906,992.1022.1151.1177.1225.1235.1291.1309,1520.1671,2424, 2457. 2460, 2461. 2466, 2609.

cr. Connochaetes sp.Material:Maxillary pieces. KA 1584a. 2513; mandible pieces, KA740.827,883.981. 1069, 1147, 1278; isolated teeth, KA615, 782a,b. 865. 1066, 1293.2409.2449.2482.

Cf. Megaiotrogus sp.Material:Left mandible piece with M,_" KA 1371; isolated left M,.KA 1292.

Híppotragus cf. equínusMaterial:Left mandible pieee with M,. KA 2491.

?HippotraginiMaterial:Parts oí an unprepared and unnumbered skull.

Redunca cf. arundinumMaterial:Ríght mandible pieee with P" KA 1349.

Pelea cf, capreolusMaterial:Maxillary pieces, KA 527. 1164; mandible pieces, KA903, 1149, 1285, 1766b,c.

Antidorcas reckiMaterial:Almost complete skull, KA 1779; maxillary and mandiblepieces, KA 603a-f; frontal pieces with hom-core bases,KA 1567, 1577; maxillary pieees, KA 765, 881a. 901,925,964c, 1046,1111. 1213a. 1310, 1517b, 2501,2610; mandi­ble pieces, KA 506, 769, 821, 842, 964a.b. 1002, 1093,1114, 1123a..... , \205, 1517a. 1632, 1867. 2474; isolatedteeth, KA 520, 679, 864, 881b, 1119, 1278a.b, 1352a.1453d. 1639, 1817, 2512b.

Antídorcas bond;Material:Maxillary pieces, KA 1157, 2465; mandible pieees, KA537, 648, 1162, 1163, 1676, 2172, 2464, 2508; isolatedteeth, KA 999, 2472.

cr. Raphicerus sp.Material:Right mandible pieee witñ P,--M" KA 1152.

Kromdraai 253

Syncerus sp.Material:Isolated teeth, right M'. KA 1630; right M", KA 752; leftM', KA 1451; right M" KA 1268.

Trugeíuphus cf. scriptusMaterial:Rigbt mandible piece, P,--M" KA 2498.

Tragelaphus ef. strepsicerosMaterial:Left maxilla, dpm'', M'-', KA 644a; right mandible. dpm,KA 856; isolated teeth, KA 1269, 1298, 1616, 2451b,2459, 254Ia-<:.

Taurotragus ef. oryxMaterial:Mandible píece, left M" KA 1303; ísclazed right M'. KA2469.

Antelope size class IMaterial:7 eranial and 56 postcraníal pieces, as listed in table 111.

Antelope size class nMaterial:214 eranial and 299 postcranial pieces, as listed in table111.

Antelope size class 111Material:49 eranial and 73 postcranial pieces, as listed in table 111.

Antelope size class IVMaterial:2 cranial and 2 postcranial píeces, as listed in table 111.

Family SuidaePhacochoerus modestusMaterial:The type specímen, consisting of a juvenile skull andmandible lacking the braincase and occíput; dprn":" andMI in wear, M2 unerupted, KA 89.

Metridiochoerus andrewsiMaterial:The former type specimen of Notochoerus meadowsi,consisting ofcraniumwith articulated mandible, completeexcept for posterior part of calvaria.

Order PerissodaetylaFamily EquidaeHipparion libycumMaterial:Isolated right M,.

Equus burchellíMaterial:Isolated teeth: right M', KA 1213; left M', KA 1909; rightM" KA 1130, 1906; left M" KA 1908.

Equus capensisMaterial:Maxillary pieces, KA 712, 1179, 1898; mandible pieees,KA 685, 1906a-<:, TM 1250; isolated teeth, KA 534a-<:.575, 6IOa-<:, 623, 624a--<l. 729 + nsi, 735. 787a,b,817a-g, 824, 86Oa, 869, 892, 899a...... 9IOa,b, 927a,

254 Fossil Assernblages from the Sterkfontein Valley Caves: Analysis and Interpretation

939a-f, %6,972a-e, 1086a-e, 1091, 1102a-c , 1162,1624,1648, 1684, 1703, 1713, 1772, 1879, 189511-j, 1896, 1897,1899, 1900, 190Ia,b, 1909, 1910, 1902a-d, 1903, 190411,1905a-(', 1911a, b, TM 1123, 1187,2445,2507; femurpieces, KA 508, 557, 1008, 1112; phalanges, KA 796,1505.

Equid indet.Material:Maxillary pieces, KA 1217; isolaled teeth and tooth frag­ments, KA 641, 1044, 1129, 1158, 1192, 1234, 1377, 1652,1596, 1708, 1743, 1789, 1846, 2475, 2478, 2489; proximalfemur pieces, KA 557,1008, 1112; distal femur piece, KA508.

Order HyracoideaFamily ProcaviidaeProcavía antíoua

Material:Most of a skull with 8 articulated vertebrae, KA 807;partíal cranium, KA 3; calvaria pieces, KA 36, 1488; pal­ates and maxillae, KA 11, 12, 15, 16,21,22,28,31,34,39,54; mandible pieces, KA 6, 7, 8, 14, 17, 18, 19,24,27,33,44,49, 1724; isolated incisor, KA 13.

Procavia transvaaíensísMaterial:Complete craníum without mandible, KA 48; cal variapiece, KA 1173; maxillary pieces, KA 23, 25, 35, 53;mandible pieces, KA 1, 2, 4, 5, 9, io, 26, 40, 43; isolatedteeth, KA 29, 37,42,47; distal humerus, KA 50.

Phylum MolluscaClass GastropodaOrder PulmonataCf. Achatina sp.Material:Land snail shells, KA 736, 1265, 1458, 1529. 1810,5991.

The Major Features of the Bone Accumulation

The Composition o/the Fauna Represented by theFossils

The information provided in table 110 and depicted infigure 207 may be further summarized. This is done in thefirst columns of table 112 and figure 210, In terrns ofnum­bers of individual animals, the fauna represented by thefossils is dominated by bovids (48%), folIowed by equids(15%), carnivores (12%), and baboons (11%).

The antelope rernains carne from anirnals of widelyvarying sizes: in size class 1, 5 adult steenbok are repre­sented: in class lI, parts oC 56 individuals are involved, ofwhich 22 were juveniles; in class lII, 23 individuals, in­cluding JOjuveniles, are present; and in the largest sizegroup, IV, 7 individuals including 3 juveniles, are repre­sented.

More individual equids are represented at Krorndraai Athan in any other site unit; at least 23 individuals of thelarge extinct Cape horse, Equus cupensís, are involved,of which 3 were juveniles. The 5 Burchell's zebras in­cluded 2 colts, and the single Hipparion was mature.

The Krorndraai A assemblage comprises a remarkablevariety of carnivores, including a Iion, 2 Feíís species, 2

0'1020310.0500102030<10

PERCENTAGE CONHIlIlUTION TO tHE FOSSIL , ...UMAS

fig. 210. Percentege contributions made by various groups oí animals tothe samples oí macrovertebrate íossil fauna from Kromdraai A and B.

Hyracoid indet.Material:Almost complete cranium, KA 51; mandible piece, KA1822; isolated incisors, KA 1216.

arder LagomorphaFamily LeporidaeLagornorph indet.Material:Complete cranium, KA 588; cranial piece, KA 1560;mandible pieces, KA 1233, 1316.

arder RodentiaFamily HystricidaeHystrix afrícaeaustralisMaterial:Rightjuvenile mandible with 1,dpm¿ M.-., KA 1432; leftmandible piece, P..-M" KA 757; isolated teeth, KA 674,690, 1546, 1607.

Hystrix ef. makapanensisMaterial:lsolated right M', KA 1912.

Class AvesCf. Coíumba sp.Material:Part of an eggshell, KA 1167.

Class Reptiliaarder CheloniaFamily TestudínidaeCf. TestudoMaterial:Carapace piece , KA 1198a.

KRO..DR......1 B

saber-toothed cats and a false sabertooth, 2 species ofhyena, 4 different canids, and 2 mongooses.

Among the baboons, I species of Parapapio is involvedand 2 of Papío. as well as 8 individuals of the remarkableanimal Gorgopi/hecl/s majar (fig. 208). Qther animals ín­elude 6 individuals of the large dassie, Procavia Ira115­vaalensis, and 10 of the smaller form P. alltiqua.

The Represen/alion 01 Skelelal ParlS

As wil! be seen in the Iistings under individual taxa, mostidentifications are based on cranial remains, and theremay have been some artificial bias toward such parts inthe coIlecting process. Like the Sterkfontein Member 4and Swartkrans Member 1 assemblages, the Kromdraai Asample is likely to give a truer reflection of species pres­ent than of skeletal part representation. For this reason Iam not inclined to attempt much in the way of deductionsfrom the presence or absence of skeletal parts.

It is worth noting, however, in the bovid size class n,that the 56 individuals are represented by 299 postcranialpieces (table 110 that cover virtually the whole range ofskeletal parts. Furthermore, these parts are not typicallythe most resistant ones, there being more proximal hu­men, for instance, than distal ones (17: 12). In the tibialfossils, 10 proximal ends are found in association with 4distal ends. We may conclude that destructive action onthe bones was not intensive, resulting in survival of re­sistant elements only, but rather unintensive, permittingpreservation ofparts with low survival ratings. Rapid bur­ial by incoming sediment is also indicated.

Observed Damage to Ihe BOlles

Like the fossils from other site units, Kromdraai Aboneswere individually examined for characteristic damagemarks. Results are presented in table 113 and may besummarized as follows:

Porcupine and Small-Rodent Gna11' Marks. Only 5 fossilsin the assemblage showed evidence of porcupine gnawing,and 2 bore traces of small-rodent tooth marks.

Carnivore-Inflicted Damage. Clear evidence of carnivoredamage was observed on 19 fossils, and damage probablyattributable to camivores was seen on 12 more speci­mens. Undoubted camivore damage was seen on a Gor­gopithecus mandible, on a wide range of antelope bones,and on craniums of bolh species of dassie.

Traces 01 Artificial Bone Alteration. None was observed.

Association with Artifacls. No stone artifacts have yetbeen found in the breccia from which the fossil samplecame.

Clues to the Interpretation of the Kromdraai A FossilAssemblage

Possible Porcupine Collecting lnvolvement

Only 5 out of 1,841 bone pieces were found to beporcupine-gnawed-a gnawed-bone incidence of 0.3%.One may confidently conclude that porcupine involve­ment in the collection of the bones was minimal. The fewpieces gnawed by small rodents were probably worked

Kromdraai 255

upon where they lay in the cave, before being covered bysediment.

Possible HOlllillid lllvolvement

We have no evidence at present, in the forrn of eitherartifacts or artificial alteration ofbone, to suggest hominidinvolvement in the bone accumulation.

A Possihle Death Trap'

The Kromdraai A assemblage is unusual among the otherSterkfontein valley site units in that articulated skeletalparts (see fig. 211) are comparatively common. This couldindicate that the Kromdraai A cave opening served as anatural death trap into which the animals fel!. If this wereso, one would expect the skeletons to be more completethan they are. Furtherrnore, that sorne of the bones showclear signs of carnivore chewing suggests that the site wasmore probably a camivore lair than a death trap.

Kromdraai A: A Carnivore Lair?

It would be difficult, I think, to avoid the conclusion thatthe site had served as a camivore feed ing retreat, abreeding lair, or both. It is always tempting to simplifywhat must obviously have been a complex situation andto make a dogmatic statement on the course of events. Iwill try to resist this temptation. The carnivores knownfrom the site consist of at least 5 species of felid, 2 speciesof hyena, 4 different canids, and 2 mongooses. The mon­gooses and similar canids may be excluded as significantaccumuJation agents; they are more likely to have forrnedpart of the prey of the larger predators. The evidencesuggests that some of the animals were killed and eaten inthe cave by felids, perhaps sabertooths, that did com­paratively liule damage to the skeletons, while other,more fragmentary remains were left by hyenas. The felidsmay have been responsible for the articulated and rela­tively undamaged bovid specimens that charactenze theKromdraai A assemblage, together with the typical1ydamaged dassie craniums, while the sparse and frag­mented remains of equids could well represent partsbrought back and discarded by hyenas.

In the course of her study of lhe fossil bovids, ElisabethVrba (1975, 1976a) pointed out that in size class 11, 38% ofthe Damaliscus or Parmularius individuals (now rec­ognized as Parmularius by Vrba 1978) were juveniles,suggesting that they had been consum.ed by primary pred­ators within the cave. Interestingly enough, Vrba found

Fig. 211. Many of lhe fossils from Kromdraai A consist of aJ1iculatedskelelaJ parts. Two examples are shown here.

256 FossíJ Assemblages from the Sterkfontein valley Caves: Analysis and Interpretation

that, as the bovid liveweight rose in the Kromdraai Aassernblage, the proportion of juvenile individuals repre­sented did not mercase. It appeared that. whatever pred­atar had becn involved. it was powerful enough to copewith adult individuals of very large prey species. Suitablepredators known from Kromdraai would have beenHomotherium, Megantereon, and Dinofelis, and Felíscrassídens may also have been involved in hunting thesmaller prey, though we know nothing of this cal' s habits.That 11% of the animals represented al the site were ba­boons suggests that the predators involved took baboonsas a significant pan of their prey spectrum.

As 1 have mentioned elsewhere in this work, the activehunting of large prey does not necessarily mean thatsingle powerful predators must have been involved, So­cial carnivores, hunting in packs or clans, may certainlyoverpower large prey animals. Among the Kromdraaicamivores, both spotted hyenas, Crocuta crocuta, and aclose relative oí the brown hyena, Hyaena bellax, arerepresented. Both may well have hunted cooperativelyand have brought back parts of their kills or theirscavenged meals to the Kromdraai cave lair.

The camivore/ungulate ratio (as defined in chapo 3)preves to be 25/123, or 20.3%. This is not as high as mightbe expected in an assemblage from a brown hyena feedinglair (see chapo 4), but it is higher than is typical in StoneAge human food remains. The figure is consistent with amixed assemblage collected by both felid and hyenid oc­cupants of the cave.

the northern wall of the deposit, which was explored to adepth of 4.9 m (16 ft). Microvertebrate bones were foundin very large numbers, and these are still being studied byD. H. S. Davis and T. N. Pocock. Largerbones were notas abundant, but almost 5,000 were found, and these havebeen grouped according to their depth in the excavation:Layer 1: surface-fB m (O-{j ft)Layer 2: 1.8-3.7 m (6-12 ft)Layer 3: 3.7-4.9 m (12-16 ft)

The srratiñcation of the breccia is fairLy steeply inclinedfrom east to west, which means that the horizontal de­lineatíon indicated aboye has little meaning. Fortunatelythe new excavations being conducted by Vrba are alreadyproviding in situ fossils whose stratigraphic relationshipsare precisely known.

The sample being considered here consists of 4.985specimens from a minimum of 94 individual animals, aslisted in table 115and depieted in figure 213. Partículars of197 bones of bovid origin, assigned to size c1asses but notspecifically identified, are given in table 116. Included inthe fossil sarnple were 2,887 bone flakes; particulars oftheir lengths are given in table 117.

Data on the fossil material by which the various animaltaxa are represented will now be presented:

Phylum ChordataClass MammaliaOrder PrimatesFamily HominidaeAustralopithecus robustus

Fig. 212. Percenrage representation of animaJs of various kinds in themicrovenebrate component oCtbe Kromdraai A samote. A total of 273individual animals is invoíved, and the sample is dominated by dendro­rnurines and the white-tailed rat, Mystromy.s.

,.ERCEIITIIOE IIE,.RESE'''lIr1Of11

n'I'2

The Kromdraai ABone Accumulation: MicrovertebrateComponent

As 1 mentioned in chapter 10. sorne breccia samples con­taining microvertebrate remains from the •.KromdraaiCaves" were sent by D. Draper to the British Museum in e"....-.~

1895. The sarnples were rediscovered in the museum col- 1.-~lections by L. S. B. Leakey in 1958. as reported by Oak­ley (1960). Preparation of the breecia blocks in aeetie acidaUowed de Graaff (1961) to tist the animal speeies repre­sented, and it now appears likely that the .. KrorndraaiCaves" referred lo by Draper were, in fact, the surfaceexposures on the Sterkfontein hill. The sarnples studiedby de Graaff carne, in all probability, from Member 5 ofthe Sterkfontein Formation.

The breccia from Krorndraai A contains a fair scatter ofmicrovertebrate bones, though these do not seem to be asconcentrated as in parts of Sterkfontein Member 5 andSwartkrans Member 1. A large sarnple of Kromdraai Abreccia has recentIy been prepared in acetic acid, and 1have provisionally identified the microvertebrate bones,though further taxonomic work is needed before a fullspecies list can be compiled. The sample was found tocontain remains of at least 273 individual animals, aslisted in table 114 and depicted in figure 212. 1t is e1earthat the owls that roosted in the Kromdraai A cave werefeeding largely on dendromurines and 00 rats of the genusMystromys.

The Kromdraai B Bone Accumulation: MacrovertebrateComponent

The bones that fono the sample analyzed here are thosethat carne from the site between 1938 and 1956. The as­semblage does not include any of the newer material re­sulting from Vrba's current excavation. The fieldwork of1955-56 concentrated mainly on decaJcified breccia along

" "

Kromdraai 257

TM 1517, holotype of Paranthropus I'olm.'·/u.v. consíst­ing of left side of cranium mcluding rHl"icllll~, temporals ,face, and palate; left maxilla with P,I 1, MI 2; isoletedcrowns of right P"", M""; mandible plccc with right P:J-\MI-a, and cast of right C; isolated len l" ': proximal endof right uina, distal end of right humcrus: left metacarpal11; right talus; proximal phalanx (manus); proximalphalanx and distal phalanx (pediN¡ (/111. 214). Ageestímate, 19-20 years.

TM 1603, left M" crown from sam.lndividual as 1517.TM 1536, ¡nandible of a juvenil. with ri¡¡ht 1" di" de,

drnl _ 2 • MI; left de, dm., Age estímate, 2!h % ~ years.TM 1600, rnandible fragments with left P", M,.". Age

estímate 19 ~ 1 years,

TM 1601, isolated teeth, left dm.. crowns of C. P:l - h

and left MI. Age estimare, 2 :!:: l years.TM 1602, right maxillary piece with roots of Ml-;¡ and

portion of palate. Age estímate, mature.TM 1604. isolated left dm¿ Age esnmate, 6 :!: 1 years.TM 1605, ilium, including part of acetabulum, but with­

out iliac crest, Age estímate, mature,Summarizing the results of his age estimations, Mann

(1975) concluded that 2 individuals fall into the 1-5 yearage class, 2 mto the 6-10 year class, and 2 into the 16-20year class, the latter being regarded as mature.

Family CercopithecidaePapio robtnsoni

KB

Cereopitheeoidea williamale

Australopithecua rObustu•

•

Panther. pudua 2

Dl"0'eI15 ap. 1

Vive"a &p. 1Herpestes sp. 2

Proteles ap. 1

Antelope lJize cia•• IV

~el Lepva ep. 2

el Te.tudo ap. 3

el Crocod,lue nllotleu. 1 ~Slruthio .p. 1

\Indel. r.ptor 1

Fig. 213. Animals reprev:nt.erl: tr¡ (t ..~ihl in the macrovenebrate componcnt of the Kromdraai B sample. Minimumnumbers of individuals are Ir'A~;~ in exh case.

258 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation ,

Fig. 214. (a) Histograros showing the percentage abundance of boneftakes of various lengths from Kromdraai B; the anaIysis is based on2,887 pteces. (h) A similar analysis of a1I 4,985 fragments of bone in thesample from the deposito

Material:Layer 1: Juvenile specimens

Part of a juvenile maxilla with right dm':", KB 687;fragments of maxilla and mandible with deciduous andunerupted permanent teeth, pieces numbered individuallyKB 688-703; isolated right di" KB 243, 270; right di" KBunnumbered; right de, KB 244, 269, 274, 282, 289; left de,KB 252, 259; de fragrnent, KB 271; left dm, KB 272, 285,286; right dm.. KB 234, 249, 260, 273, 279; left drn, KB261; right dm, KB 251, 281; left di', KB 226; right di', KB239,283; left di', KB 267, unnumbered; left and right de,each with two roots, a very odd occurrence for a primate.KB 233, 224; left de, KB 277, 278; right d~, KB 228, 255,258; left drn', KB 241, 734, 755, and unnumbered; rightdm", KB 245, 265, 284; drn', side uncertain, KB 280; leftdm', KB 738, 759; right drn", KB 747, 756, 767.Layer 1: Adult specimens

Mandible fragment with parts of M,_" KB 21; anteriorpar! of a mandible with P, bilaterally, KB 683; left 1', KB728; right 1', KB 268; 1', side uncertain, KB 736; left 1"KB 237; right 1', KB 238, 248, unnumbered; ríght e, KB230; right P', KB 749; right P" KB 246, 735; right M" KB242, 742, 750; left M', KB 764; right M', KB 739, 748,762,763; left M" KB 682, 732; right M" KB 685; left M',KB 231, 765.Layer 2

Isolated left de, KB 3390; right de, KB 3428; right 1"KB 3357; right e, KB 2919, 3346; right P" KB 3416; rightM" KB 3360.Layer 3

Isolated right di', KB 3225; right di" KB 3117; rightdm., KB 3233; left dm, KB 3115; left 1', KB 3112; left e,KB 3223; molar crown, KB 3222.

The individuals may be allocated to age classes as in­dicated in table 118. Ofthe 18 individuals, 11 proved to be

Papio angusticepsMaterial:Layer 1

Part of juveníle mandible, unerupted mesial incisors,KB 684; isolated left dm", KB 236, 254; right drn", KB253; mandible with left and right e-M" KB 680; mandiblepiece, right M,_" KB 686; mandible piece with rightM,_" KB 197; mandible piece, ~, left PrM" KB 196;isolated left 1', KB 229, 2365; left P" KB 746; right P" KB288; left P" KB 740; left M', KB 250,251; right M', KB751; right M" KB 257.Layer 2

Isolated left de, KB 3427; left M', KB 3363.Layer 3

Maxillary piece with left p3-4, and o canine socket, KB3118; maxillary fragment with left M', KB 3120; isolatedright dm., KB 3123; right <3 e, KB 3107; right P" KB3227; right M', KB 3109; left M" KB 3114.

The individuals may be allocated to age classes as in­dicated in table 118. Ofthe 14 individuals, 3 proved to bejuveniles. Fuller particulars may be had from Freedmanand Brain (1972).

Cercopithecoid sp. indet.Material:Layer 1

71 cranial and 194 postcranial pieces, as listed in table119.Layer 2

12 cranial and 37 postcranial pieces, as listed in table119.Layer 3

19 cranial and 34 postcranial píeces, as listed in table119.

?Megantereon sp.

Cercopíthecoídes wílliamsíMaterial:Layer 1

Almost complete calvaría, wíthout snout, KB 195; iso­lated left M', KB 225; right M', KB 256; unerupted uppermolars, KB 745, 754; unerupted lower molars, KB 730,758.Layer 3

Left mandible fragment with dml_2, socket for MI, KB3108 + 2230; isolated right M" KB 3124; isolated uppermolars, KB 3113, 3119, 3121, 3229.

The individuals may be allocated to age classes as in­dicated in table 118. Of the 5 individuals, 2 proved to bejuvenile. Fuller partículars may be had from Freedmanand Brain (1972).

Order CamivoraFamily FelidaePanthera pardusMaterial:6 postcranial pieces from a minimum of 2 adult individu­als.Layer 1

Distal radius, KB 2885; proximal ulna, KB 2901;navicular, KB 2903.Layer 3

Left astragalus, KB 3259; phalanges, 3249, 3252.

juveniles. For fuller particulars see Freedman and Brain(1972).

•• t'l12•

•

"

,....

..

Material:9 pcstcranial pieces, probably from 1 individual.Layer 2

Right astragalus, KB 2942; right calcaneus, KB 2946;metapodial pieces, KB 2932, 2933, 2935, 2937, 2939, 2948;2d phalanges, KB 2934, 2938, 2940.

?Dínofeíís sp.Material:A single postcranial piece.Layer 3

Distal end of a metapodial, KB 3248.

Family HyaenidaeHyaena cf. brunneaMaterial:1 eraniaJ and 2 posteranial pieees, probably from 3 indi­viduals.Layer 1

Isolated M" KB 295.Layer 2

Distal metapodial, KB 2936.Layer 3

1st phalanx, KB 3250.

Proteles sp.Material:A single eranial specimen.Layer 2

Anterior part of right mandibular ramus with alveoli,KB 2945.

Family CanidaeCanis sp.Material:7 craníal and 2 postcranial pieees from a mínimum oí 4individuals,Layer 1

Right maxillary fragrnent with incisors and caninesocket, KB 292.Layer 2

Part of M" KB 2947; isolated canine, KB 3320; frag­ment of proximal metapodial, KB 2930; 2d phalanx, KB2931.Layer 3

Maxillary fragment, KB 3255; isolated incisor, KB3258; isolated canínes, KB 3246, 3253.

Family ViverridaeHerpestes sp.Material:2 cranial pieces from 2 individuals.Layer I

Isolated left P" KB 290.Layer 2

Mandible fragment, KB 2944.

Viverra sp.Material:A single postcranial pieee.Layer 3

Distal humeros, KB 3258.

Camivore indet.Material:16cranial and 26 posteranial pteces, as listed in table 120.

Order Artiodactyla

Kromdraai 259

Family BovidaeConnochaetes sp.Material:6 cranial pieees Irom a minimum of 2juveniles and 1adukindividual.Layer 1

Left mandibular ramus with dm l _ 2 , right dm3 • KB 382,383, 388; fragment of left hom-core, KB 376.Layer 2

Isolated right dpm, KB 3009; M" KB 3006 + 3008.Layer 3

Isolated left dpm", KB 3226; right dpm", KB 3178;hom-core fragment, KB 3187.

Antidorcas cf. reckiMaterial:6 cranial píeces from a minimum of 2 or 3 individuals,including t juvenile.Layer 1

Hom-core fragments, KB 79, 377, 381.Layer 2

Isolated lower molar, KB 2999.Layer 3

Isolated right M" KB 3224; fragment of hom-corebase, KB 3190.

Cf. Antidorcas bondiMaterial:4 eranial pieces from a mínimum of 2 adult individuals.Layer I

Hom-core pieces, KB 372 + 373, 375.Layer 3

Horn-core piece, KB 319\.

Gazel/a sp.Material:1 cranial piece from 1 individual.Layer I

Frontal piece with left hom-core, KB 380.

Bovid incertae sedisMaterial:3 pieces from 1 adult individual.Layer 3

Hom-core fragments, KB 3188 + 3193 + 3195.

Antelope size class 1Material:11 cranial and 26 postcranial pieces from a mínimum of 6individuals, as listed in table 116.

AnteJope size class 11Material:26 cranial and 72 posteranial pieces from an estimatedminimum of 10 individuals, as listed in table 116.

Antelope size class IIIMaterial:15 cranial and 45 postcranial pieces from an estimatedminímum of 8 individuals, as Usted in table 116.

Antelope size class IVMaterial:2 postcranial pieces from t individual, as Usted in table116.

Family SuidaePhacochoerus modestus

260 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

Material:I cranial piece from a juvenile individual.Layer 3

Isolated right M', unerupted, KB 3276.

Order HyracoideaFamily ProcaviidaeProcavía sp.Material:A single postcranial piece.Layer 1

Distal humeros fragment, KB 7.

Order LagomorphaFamily LeporídaeCf. Lepus sp.Material:2 cranial and 4 postcranial pieces from a mínimum of 2individuals.Layer 1

Isolated molar, KB 1476; atlas vertebra, KB 706;proximal humerus, KB 1819; distal humerus, KB 562;rnetapodial, KB 705.Layer 3

Mandible fragment, KB 3254.

Ciass AvesStruthío sp.Material:A single piece of eggshell, KB 1082, from Layer 1.

Bird indet. Large raptorMaterial:2 postcranial pieces from a single individual.Layer 2

2 terrninal phalanges, KB 3031, 3366.

Class ReptiliaOrder SquamataCf. Cordylus giganteusMaterial:8 postcranial pieces, probably from 1 individual.Layer 1

6 pointed seale plates, probably from the base of thetail, KB 2617; caudal vertebrae, KB 704, 707.

Order CrocodiliaCf Crocodylus nilotícusMaterial:A single craniaI pieee.Layer 3

lsolated tooth, KB 3235.

Order CheloniaCf Testudo sp.Material:7 postcranial pieces from a minimum of 3 individuals.Layer 1

Proximal femur, KB 10; limb bone pieces, KB 263, 583,2732; plaslron piece, KB 142; earapace piece, KB 903.Layer 3

Coracoid, KB 3237.

CoprolitesMaterial:I ofmongoose size, KB 39; 3 ofjaekal size, KB 37, 38, 40.

Indeterminate fragments

Material:838 preces from Layer 1; 212 pieces from Layer 2; 224pieces from Layer 3; giving a total of 1,274 fragments.

Bone flakesMaterial:2,887 píeces as detailed in reble 117.

Kromdraai B: The Major Features of the BoneAccumulation

The Composttion of the Fauna Representad by meFossils

The infonnation provided in table 115 and depicted infigure 213 may be further surnmarized. This is done in theseeond eolumns of lable 112 and figure 210. Whereas thefauna from Kromdraai A was dominated by bovids, thatfrom Kromdraai B is similarly dominated by primates­austraLopithecines, two species of baboon, and a colobidmonkey. Bovids are welI represented by 6 individuals inclass 1, \O in class II, 8 in class III, and 1 in class IV.Carnivores are represented by 15 individuals, and acrocodile has been identified on the basis of a single tooth.

The Representatíon 01 Skeletal Parts

The most charaeteristic feature of the sample from thedecalcified breccia is the extreme fragmentation of thebones, making detailed skeletal par! studies very dífficult.What information 1 have been able to extract is presentedin the various tables. 1t is worth noting that the austraío­pithecíne individual now regarded as the type of A.robustus was represented not only by the skull, but alsoby fragments of an arm, hands, and feet. The cer­copithecoids were represented by 210 postcranial pieces(table 119) from most parts of the skeleton, and the sameis true of lhe bovid posteranial remains (fig. 208). Un­fortunately, the samples are too smalI and the fragmenta­tion too great to allow worthwhile conclusions on the rea­sons for the presence or absence of skeletal parts in thesample.

Observed Damage to tbe Rones

The bones are generally so fragmented that it is diffieult tobe sure of the damage lbey may originally have suffered.Whal effects have been observed are detaíled in table 121.A single piece was found to bear traces of porcupine­gnawing, and probable carntvore tooth marks may beseen 00 the australopithecine pelvis piece TM 1605 and 00

3 other bones.In connection with the Australopithecus robustus type

skulJ, TM 1517a, a claim has been made for deliberarehominid violence. After initial description by Broom, thespecimen was entrusted to G. W. H. Schepers for a studyof Its endocranial contours. It was his task to remove thematrix that filJed the left parietal región of the brainease,and in this conneetion he wrote (Broorn and Schepers1946, p. 174):

While excavating this matrix a large ñint-Iíke rack wasfound embedded in it. The parietal bone had beendriven into the endocranial cavity ahead of this rock.It eould not be preserved as it was necessary toundercut the rock by destroying the bone. otherwise itwould have been well-nigh impossíble to remove it.

The presenee of this rock is evidence suggestive ofthe claims that the Homunculi represented by the

r Kromdraai 261

australopithecoid and Plesianthropoid fossils, wereskilled enough to employ missiles or weapons for de­fenslve. offensive and predatory purposes.

In his papee 00 the predatory implernental technique ofAustralopithecus, Dart (19490) provides a photograph ofthe internal aspect of the Kromdraai specirnen with thestone to which Schepers refers still in place. From this itcan be established that the stone was oval in outline,about 3.5 cm long and 3 cm across. Its thickness cannotbe judged but appears to have beco between 1 and 2 cm.The breccia matrix from which the specimen carne is veryrieh in chert blocks and fragments derived from thedolomite country rock by normal weathering. These beara superficial resemblance to flint, and ít is extrernelylikely, though it cannot be proved, that the stone Schep­ces observed was a piece of chert. On the basis of theestimated size cf the piece, its weíght would have beenabout 75 g. If a stone of this size were to have penetratedthe parietal bone of a living australopíthecine, it wouldhave had to be hurled at a very considerable velocíty.

Since only a piece from the left side of the craniurn hasbeen found, it is impossibJe to establish whether the cal­varia was more complete when it carne to rest in the cave.One can argue that the entire right sirle of the cranium wasmissing before fossilization and that the stone, togetherwith the associated matrix , simply filled in the hollowskull fragment, the stone causing damage to the parietalduring the consolidation oí the breccia.

In my opinión, Schepers's c1aim for deliberately in­ñicted injury to the Krorndraai skull is based on an un­critical appraisal of the evidence. A natural explanationfor the observed damage can readily be found.

As is detailed in table 121, 70 bone pieces showedrounding and wear as if they had been water-worn. Withthem were a number of pebbles of similar appearance. Asis apparent from the tableo the great majority of wornbone pieces were between 1 and 3 cm in length. I am at aloss to explain tbe wear on tbese bones unless it wascaused by a stream flowing through the cave. Perhapssimilarly wom bones will be found in situ in the excava­tíon being conducted by Vrba; ifso, it should be possibleto decide what mecbanism caused the wear.

1975b) 1have suggested that the extreme fragmentation ofthe bone pieces (see fig. 214), and the fact that 57.9% of allpieces were bone ftakes, indicated that horninids had bro­ken up the bones with stone hammers. In fact, the condi­tion of the Krorndraai B bones is very reminiscent of thatto be seen in Later Stone Age human food remains asdescribed in chapter 3.

Since originally writirtg on the subject, however, I havehad serious reservations about the reasonableness oí myconclusión. Could the fragmentatlon, for instance, havebeen a result of the breccia decalcification process? Ifthey had been broken by hammerstones, why are notmore artifacts found in tbe breccia? These questions can­not be answered at presento though I suspect they will beas the current Krorndraai excavation progresses,

Possíble Carnivore Invoívement

That remains of two leopards, of probable Megoruerecnand Dinofelis individuals, and of three brown hyenas arefound in the assemblage strongly suggests ínvolvernent bythese animals in the bone accumulation. Moreover theextremely high camivore/ungulate ratio oí 15/26, or57.7%, adds support lo the suggestion.

Conclusion

vrba's new excavation will confirm whether or not theobserved fragmentation of the bones was a result of de­calciñcatíon. If it were, then the indications of hominidinvolvement would diminish ccnsiderably. In the interim1 suggest that predation by cats and hyenas could bestexplain the origin of the fossil assernblage.

The Kromdraai B Bone Accumulation: MicrovertebrateComponenl

The excavatíon of 1955-56 produced a wealth of micro­faunal bones which were submitted to D. H. S. Davis forstudy. Preliminary results on bis analysis of rodent re­mains from the upper levels of the excavation have ap­peared (Davis 1959, 1%2) and may be tabulated as follows(Davis 1959, p. 151).

22

%42

3613l-í3l-í

22

66

1174411

Number135Cricetínae (Mystromys)

Elephant shrews (Elephan-tulus)

Otomyinae (Otomys)Bathyergidae (Cryptomys)Murinae (Rattus, Mus,

Rhabdomys, Dasymys)GerbiJIinae (Torera)Shrews (Crocidura, Suncus,

Myosorex)BalsDendromurinae (Steatomys,

Dendromys} known to occur

Davis remarked on the exceptionally bigh proportionsof Mystromys and Elephantulus in the fossil sample-farhigher than those he had observed in modem owl pellelaccumulations from the Sterkfontein valley.

More recentJy T. N. Pocock (1970) has pubJished lb~

results ofhis analysis ofbird remains from the KromdraaiB microfaunal accumulation. His faunallist and the inter­pretatíon based on it has a1ready been discussed in lbesection of chapter 9 deaIing wilh lhe class Aves.

Association wítñ Artlfacts

Throughout the excavation of decalcified breccia alKromdraai B, 1 examined all available stones for ~igns ofartificial fracture. The result was thal 1 unquestionableftake of chert was found (figured in Brain 1958, p. 98),together with 3 less convincing chert ftakes and a brokenpebble of quartzite. No artífacts have yet been found inthe solid breccia.

Cines to lbe Interpretation of the Kromdraai B FossllAssemblage

Possible Porcupine Collecting Invoívement

Since only one bone out oí a total of4.981 shows evidenceoí porcupine gnawing, the involvement of porcupines ascoJlecting agents can be discounted.

Possible Hominid lnvolvement

A single convincing flake has been found in decalcifiedbreccia, but no artifacts are yet known from in situ bree­cia. On grounds of associaled artifacls, hominid involve­ment is thus very dubious. In previous writing (Brain

13 A Note on Taung and Makapansgat

The project described in this book is concemed with fossilassemblages from the Sterkfontein valley caves; yet 1wish to make a few passing cornments 00 the assemblagesfrom Taung and Makapansgat, the two other southemAfrican sites that have yielded australopithecines, sincethese have a bearing on the questions al issue.

Taung

The original Australopithecus cranium, together withother faunal rernains, carne lo light as a result of traver­tine quarrying at Buxton-Norlim, which by 1924 had de­stroyed the actual site where the hominid skuU had lain.A few years after his announcement ofAustralopithecus,R. A. Dart described the bone accumulation as follcws(1929, p. 648):

Examination of the bone deposít al Taungs shows thatit contains the remains of thousands of bone frag­ments. It is a cavern lair or kitchen-mídden heap of acamivorous beast. It is not a water-borne deposit andthe Taungs remains could not have been washed intothe cavem from the surface, The bones are chieftythose of small animals like baboons, bok, tortoises,rodenls, bats and birds. Egg shells and crab shellshave also been found. The fauna is one which is notcharacteristic of the lair of a leopard, hyena or otherlarge carnivore, but is comparable with the cave de­posits formed by primitive mano The deposit was,therefore, fortned by primitive man or by Australo­pithecus, an advanced ape with human carnivoroushabits. As no human remains have been found there,and as no Australopithecus remains bave been foundelsewhere in known Pleistocene deposits, I am of theopinion that the deposit was formed by the Taungssub-man himself.

Since Dart wrote this account, we have leamed a gooddeal abaut the nature ofbane accumulations in caves, andI have íittle doubt that Ibis Taung assemblage included amicrovertebrate owl-collected component as weU as amacrovertebrate component that is likely to have foundits way to the cave in a variety ofways. The association ofeggshells, crab carapaces, and a diversity of small animalremains is very suggestive of the possibility that the cavehad been used as an eagle owl breeding place. As wasdescribed in chapter 6, the Cape eagle owl, Bubo capen­sis, has the habit of nesting in caves, frequently clase towater, and ofaccumulating crab carapaces and other food

262

remains around its nest site. Eggshells in such an as­semblage are then, as Iikely as not, derived from thehatched eggs ofthe owls tbemselves. Remains ofbaboonsand bucks will obviousIy represent a different accumula­tion pattern, to be discussed in a moment.

During the late 1940s the California Africa expeditionteam made a very thorough study of the Buxton-Norlimtravertines (Peabody 1954) and excavaled fossiliferouscave ñllings adjacent to the spot where the Austraíopíthe­cus skull had been found. Using all methods al his dis­posal, Peabody tried to reconstrucllhe stratigraphy of theoriginal cave, concluding that there had been four de­posits, each resuIting from a different accumulationphase. From oIdest lo youngest, these were:l. A fossiliferous, red sandy limestone, represennng a dryclimatic phase.2. A much purer calcareous deposit, encJosing parches offossiliferous sandy limestone and indicating a wet ac­cumulation periodo3. A sterile deposit of red sand.4. An unconsolidated black earth wíth Middle Stone Ageremains.

Peabody concluded that the auslralopithecine skull hadbeen preserved in the second, wet-phase deposit.

More recently Butzer (l974b) has reinvestigated rheBuxton-Norlim travertines and the sediments enclosingthe fossils. He .has confirmed that the australopithecinespecimen díd, in faet, come from the wet-phase deposits,whereas most of the faunal remains, including the numer­ous baboon skulls, had been preserved in the older, dry­phase sediments. Thus, unfortunately, the taphonomicinterpretatíon of the fossil assemblage as a whole hasbeen greatly complicated,

Most of the numerous baboon skuUs which carne fromTaung in the early days have lost their context relative toone another and lo the strata in which they lay. In hispaper "The Predatory Implemental Technique of Aus­tralopithecus ," Dar! (I949a) lists 21 baboon craniumsfrom Taung, many of which he considered to show an­lemortem damage indicative of deliberare hominid blud­geoning. The depressed fracture in the right frontal boneof specimen 5365 (South African Museum catalog; thespecimen was made available to me by courtesy of Q. B.Hendey) is shown in figure 215. About it Dar! wrole(I949a, p. 27):

5365 cl. Skull and poslerior (circumorbital) región of ajuvenile (calvaria 3 mm thickness) face of in-

A Note on Taung and Makapansgat 263

Fig. 215. A baboon skuU. Soulh African Museum cataJog number 5365, from Taung, showing a depressed fracturein [he right frontal bone.

determinate sex: no teeth presento Probably Parapapioafricanus, Gear. Nature of the damage: A depressed (8mm depth) fracture (30 x 27 mm) of right frontal bonesevering frontal bone along the mid-line into the rightorbit and along the right fronto-parietal suture, shat­tering the medial halves of the orbital roofs and par­tially crushing down the left orbit and muzz1e. Thefrontal bone is distorted and cracked on both sides,but especially on the right side where the fracturedportion is hinged along its parietal and sphenoidalsutural margino This hinged region and the v-shapedisland of bone 1eft standing above the obvious depres­sion of the cranium shows that the implement used tosmash it was double-headed. The parietal bones alsoexhibit splits radiatiog from the site of the principalfracture. Remaioder of braincase intact: no evidenceof opeoing the skull.

Estimated cause of the damage: A direct downwardblow upon the right frontal bone and muzzle deliveredfrOl11 the front with a v-shaped doub1e-headed objecthaving vertical internal borders or sharp margins (e.g.,ungulate humerus) and measuring approximately 30mm between the two heads (e.g., epicondylar ridges).

The damage to this and several of tbe other baboonskutls from Tauog is very interesting, and it is not sur­prising that Dart's imagination was stirred. The skull,5365, described above, had been completely fiUed with astone-free sandy matrix, but, in addition to the depressedfracture, the entire calvaria shows signs of compressionalstress. We will uofortunately never know whether a hardobject, such as anothcr fossil, lay io contact with the de­pressed area during fossilization and, in my opioion, willnever be certain whether the damage resulted from anantemortem injury or from postdepositional compression.

Although aH the baboon skulls recovered early from thesite are now preserved as isolated specimens, it is fortu­nate that one Iarge block of fossiliferous breccia, or atleast a cast of it, is still in existence. The block was exca­vated by the 1948 California expedition in Pit 5b and isnumbered 56831 io the catalog of the Museum ofPalaeontology ofthe University ofCalifomia at Berkeley.The block, which has been mentiOn"ed and figured byFreedman (1965), is shown in figure 216. It contains anumber of relatively undamaged baboon skulls sur­rounded by postcranial skeletal pieces that show in­disputable evidence of carnivore damage. I would nothave the slightest hesitation in concluding that these fos­si/s represent food remains of a camivore, very probably

Fig. 216. A block of breccia from Pil 5b al Taung. showing numerousbaboon skuUs and camivore-damaged postcranial bones.

264 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

ba

involvement could be invoked, although in a denselypacked mass of lOO,OOQ bone pieces, or more, the chanceof certain bones' entering the cavities of others must beconsidered. If hominids had been involved, it would beinteresting to know from which part of the very extensivecavern system such artificially modified pieces came. Hadthey perhaps been left close to the cave entrance andsubsequently transported to the inner recesses of thecavern by resident porcupines or hyenas?

Over the years, a picture has developed in my mind ofhow the Limeworks cave may have looked when thebones were accumulating there. l visualize an extensiveamphitheater that had resulted from a collapse of part ofthe cavern system's roof, while fram this amphitheateropenings to the cavem system we know today led down­ward. l visualize, too, a permanent water hote in the am­phitheat.er, perhaps at the point where the Makapansgatstream descended into the subterranean chambers. Pi­naUy, I visualize large numbers of animals regularly vis­iting the water hole and some of them being killed thereby camivores that perhaps íncluded hominids. Theirbones would lie about in abundance within the catchmentarea of the cavern's mouth. Some would be modified byaustralopithecines, all would be worked oVer by scaven­gers, and large numbers would be transported to the ¡nnerrecesses of the cavem by breeding hyenas and residentporcupines. While lying in the much disturbed sandaround the fringes of the water hole, sorne of the bones

Fig. 217. James Kitching demonstrates bones from (he Makapansgat graybreccia !hat show signs pos sibly attribulable lo hominid activity

Fig. 218. Examples of modified bones from lhe Makapansgat gTay breeeia:(a) bovid metapodials indented on one surfaee only: (b) ShaflS of limbbones ínlo wh.ich other bones have been lhrusl.

a leopard, that was preying on baboons and using the caveas a feeding retreat.

That the baboon skulls carne, in all probability, fram adifferent accumulation phase fmm that of the australo­pithecine skull means that taphonomic deductions validfor the baboon fossils are not relevant to the hominidfossil. Unless a new cave deposit in the Taung travertineis located and excavated with dUe attention totaphonomic detail, the origin of the bone accumulationsthere will remain a matter of conjecture.

Makapansgat

l am aware that taphonomic studies of the Makapansgatfossil assemblage are being conducted by Alun Hughes,Judy Maguire, and James Kitching, and for this reason lhave not done any research of my own there. l have sim­ply made use of the accounts published by Dart (e.g.,1957a) in drawing certain deductions.

As I demonstrated in chapter 2, the remarkable dis­proportions in bovid skeletal parts that Dart observed fortbe first time in his gray breccia fossil sample need not beinterpreted as the resuIt of artificial hominid selection.The assemblage is clearly composed of the more resistantskeletal elements--those with high survival ratings asoutlined in chapter 8-which are typically left whenske1etons are subject.ed to a good deal of destruction thatmight include carnivore action.

At the time of Dart' s original study, it seemed reason­able to assume tl1at hyenas did not accumulate significantnumbers of bones in cave lairs. As was discussed inchapter 4, however, current evidence indicates thathyenas do accumulate bones at their breeding lairs, and, ifthese lairs are in caves that have served as bone­preservation sites for thousands of years, the ultimateaccumulations could be very considerable indeed. Someof the specimens Dart described, such as the "skullbowls," for instance, are in my opinion unquestionablythe product of hyena gnawing.

Most of the gray breccia fossils have been extractedfrom blocks of rack blasted from the site by lime miners.The few pockets ofbone-rich gray breccia l have seen stillin situ at Makapansgat had accumulated in what werethen low recesses beneath the dolomite mof, where thisstepped down considerably around the edges of the majorcaverns. Such places were almost certainly dark or dimtylit and were too low to have been suitable for hominidoccupation.

Observations recently made in Israel on striped hyenas,H. hyaena, indicate that dark, low-raofed caves are reg­ularly used as breeding dens and that bones are broughtback to such places in very considerable numbers (seechapo 4). H. hyaena is known to occur as fossils atMakapansgat and could weIl have been responsible forintroducing many of the bones preserved in the gray brec­cia.

Although I cannot agree that the vast majority ofbonesfrom the gray breccia show signs of hominid modification,Dart and Kitching (fig. 217) have shown me some speci­mens bearing signs that could be interpreted as such.Examples are bovid metapodials (fig. 218a) that have beenindented on one suIface only. Had such damage beencaused by the teeth of a carnivore, one would expect tosee comparable damage on the opposing suIface as well.A second category consists of bones that have been thrustinto the cavities of other bones (fig. 218h). Again, hominid

would acquire the wear and polish so characteristic ofcertain specimens in the gray breccia assemblage (seechapo 2 for a discussion on this "pseudotool" productionprocess).

Like all other fossil assemblages in caves, theMakapansgat bones couLd be taphonomical1y interpretedwith assurance only ir they were excavated with due re­gard to subtle detail. 1 have no doubt that if an in situ

A Note on Taung and Makapansgat 265

deposit of bone-rich gray breccia could be stripped of itsoverburden and ir the individual fossils could then bechipped out as they lay in a carefully controlled grid sys­tem. it would be possible to assess with confidence theaccumulation pattem that originally operated. Such a taskwould be difficult, but it would be highly rewarding in theinterpretation of a situation that has excited the imagina­tion of paleontologists for years.

14 Who Were the Hunters and Who the Hunted?,

Who Killed the Australopilbecines? A Lineup oC PotentialCamivore Culprits

00 the basis oí evidence presented in earlier chapters, 1have beco led lo the conclusion that the australopithecineremains found in the three Sterkfontein valley caves werenot from animals that went voluntarily lo the sites to die.They carne. rather, from hominids that had been taken tothe caves by camívores and eaten there, the fossils repre­senting discarded food remains. 1 believe the same is trueoí the cercopithecoid remains that are abundantly as­socíated with those of the australopithecines. If this inter­pretation is true, we have lo account for the deaths oí thefollowing primate individuals represented in the three siteunits:

Sterkfontein Member 4: australopithecines, 47 individ­uals (17 subadult); cercopithecoids, 198 individuals (± 29subadult).

Swartkrans Member 1: australopithecines, 87 individu­a1s (58 subadu1t); cercopithecoids, 89 individuals (27 sub­adult).

Kromdraai B: australopithecines, 6 individuals (6 sub­adult); cercopithecoids, 37 individuals (17 subadult).

(In calculating these figures, "subadu1t" australopithe­cines are regarded as those with an estimated age-at-deathof twenty years and less, although this age barrier may beregarded by sorne as too high; "subadult" cer­copithecoids are taken to inelude those in the "juvenile"and "immature adult" age classes.)

Whal carnívores could have killed and partly eatenthese primates? The possibility always exits that someunknown predator, whose remains are not represented atthe sites, could have been involved. I admit this possibil­ity but realize that further discussion about it would befruitless. Of the camivores known by remains in the rele­vant site units, felids and larger byaenids must be consíd­ered as potential culprits; I aro inclined lo regard theaardwolf (an aberrant smaller hyaenid), the jackals, andthe mongooses as most likely prey of the larger carni­vores,

As ligure 219 shows, the lineup of potential primate­killers represented in each of the three relevant site unitsis remarkably similar. In each case Panthera pardus, Di­nofeíís sp., and Megantereon sp. are represented, thoughidentification of the latter two in Krorndraai B are sorne­what tentative. In each site unit, hyaenid remains alsohave been found, those of "normal" hyenas----the brownand!or spotted-at all three shes, and those of "hunting"

266

hyenas at Sterkfontein and Swartkrans. None of thesecamivores can be positively exeluded as potential killersoí australopithecines or cercopithecoids, and the possiblerole of each will be considered further.

The Swartkrans Leopard Hypolbesis

Sorne years ago I speculaled (Brain 1970) on how theabundant australopithecine remains may have found theirway into the Swartkeans cave. It was clear to me then thatthe very fragrnentary hominid fossils probably repre­sented carnivore food remains, and since leopards werewell represented in the Swartkrans assemblage it seemedreasonable to assume that they had been involved. I hadmade the observation that when leopards are harried byspotted hyenas (whose remains were also known fromSwartkrans) the leopards are obliged to take their preyinto a tree or other inaccessible place. If they fail to dothis, they are likely to lose their meal to the hyenas, whichare generally dominant in a competitive feeding situation.In woodland areas the leopards simply take their prey upinto the nearest tree, but on the open highveld, whichappears to have been largely a grassland in SwartkransMember 1 times as well, trees were certainly less abun­dant. In the dolomitic areas of the Transvaal highveld,large trees of the genera Celtis and Ficus are typicallyassociated with the shaftlike openings of caves. In a gen­erally treeless hábitat, the cave entrances provide shelterfrom frost and fire to saplings that would not readily SUf­

vive on the exposed hillsides. So the very trees availablefor leopard prey storage were those that overhung theshafts leading downward to the underground fossilizationsítes, I therefore speculated that, if in SwartkransMember 1 times leopards were preying 00 australopithe­cines and baboons, they may well have fed on tbem insucceeding generations of trees that overhung theSwartkrans shaft. Food remains falling from the lrees,Cortuitously passing the waiting hyenas and gravitatinginto the subterranean cavern, ended up as the fossils bywhich the Swartkrans australopithecines are now known.This situatíon is stylistically depicted in ligure 220.

The leopard hypothesis apparently was supported bythe nature of the damage observed on tbe cranium of anaustralopithecine child, SK 54 from Swartkrans, whichhas been described and ligured in chapter II (see fig.197a). lt consisls of two holes, almost certainly caused bythe canines oí a carnivore, in the parietal bones. As isshown in figure 221, the spacing of these boles, 33 mm

STER

KFO

NTEI

NM

EMBE

R4

Il«

-tiA.~~

~~

198

471

14

14

SWAR

TKRA

NSM

EMBE

R1

l.

1I

Il«

A~¿.

""-~

~~

8987

121

15

9

KROM

ORAA

IB

Il«~

""'-'~

~37

611

21

13

CERC

OPIT

HECU

IOS

AUST

RAlO

PITH

ECIN

ES11

lEOP

AROS

SABR

E-TO

OTHS

FAlS

ESA

BRE

-TO

OTH

SNO

RMAL

HYAE

NAS

HUNT

lNG

HYAE

NAS

Pan

ther

ap

ard

us

Meg

ante

reon

sp.

Din

ofel

issp

.H

yaen

a&

Cro

cuts

spp,

Eur

yboa

ssp

.11

Fig

.21

9.0

0th

ele

ft,

prim

ates

from

thre

esi

teun

itsw

bose

deat

hsm

ust

be

acco

unte

dfo

r;on

the

righ

t,a

lineu

poí

pote

ntia

lca

rniv

ore

culp

rits

.A

sw

illbe

appa

rent

,th

ese

are

rem

arka

bly

sim

ilar

inee

chca

se,

F"_

_'_«

'___

">

r.f

t"L_

_"'''':::.':~_~'.~_-=''''''''''''''~'''

268 Fossil Assemblages frorn the Sterkfontein Valley Caves: Analysis and Interpretatian

apart, precisely matches that of the lower canines of aleopard, SK 349, from the same deposit al Swartkrans. Asdescribed earlier (Brain 1970), the damage could havebeen caused in the manner portrayed in figure 222, wherea leopard picked up by tbe head the ehild ít had killedand dragged it to a secluded feeding place.

If the damage was infticted by the canines of an adultcarnivore, then the spacing suggests that the killer was aleopard rather than any of the other known camivores(see Brain 1970 for lables and eanine-spaeing graph). Thepossibility cannot be exc1uded, however, that a juvenileof a larger carnivore was responsible.

Although this leopard hypolhesis seemed plausibleenough when it was formulated, and still does in certainrespects, I have had reservations about its validity 00 twocounts.

Reservation 1

From Elisabeth Vrba's 1975 study of fossil bovids fromthe Sterkfontein valley caves, it became apparent thatrnany of the fossils carne from animals too large to haveformed typicalleopard prey. The relevant evidence, pre­senled in detail in ehaplers JO, 11, and 12, is graphieallydepicted in figure 223. Class III bovids oulnumber their

Fig. 220. A stylized impression of the Swartkrans Ic:opard hypothesis. Anaustralopithecine is being consumed in a Ce/lis tree overhanging theentrence to a subterranean dolomite cave. while hyenas of several specieswait hopefulIy bejow.

class IJ equivalents in both Sterkfontein Member 4 andSwartkrans Member 1, and they form an irnportant part ofthe assemblage at Kromdraai B as well. On the graph, thissituation is contrasted with that in four samples of ob­served leopard prey, details of which are provided inchapter 4. Whereas the leopards in the four study areasconcentrated almost exc1usively on antelopes in the c1assII weight category, the three Sterkfontein vatley as­sernblages under review here contained bones fromsignificant numbers of heavier antelopes. Apart fromthese bovid remains, sorne of the fossils were also derivedfrom other large animals, such as the extinct Cape horse,Equus capensís. adult specirnens of which were certainlyloo bulky lo have been transported lo the cave byleopards.

To explain the presence oC remains of c1ass III and IVbovids al Sterkfontein and Swartkrans, Vrba (1975)suggested that. in addition to leopards, false and truesaber-toothed cats had been involved in the bone ac­cumulation process. By extrapolation, one would be in­c1ined to believe that these Jarger cats may also have par­ticipated in the primate-predation process.

Now that it is possible to view the fossiJ assemblages asa whole, 1am convinced that contributions will have beenmade nol only by the cats, but also by the hyenas. Thefossil assemblage frorn Swartkrans Member 1 contains ahint that two separate predation patterns may have beeninvolved in its accumulation. The first pattern is indicatedby the bovid remains in which, of the 82 individuals rep­resented, 25 carne from the weight classes I and 11, while57 were derived from heavier antelopes. Of these largeranimals, J2 were juveniles, which could conceivably havebeen caught by leopards, but 45 were adults that muslhave fallen prey to carnivores adapted to larger prey.

Turning to the primate component of the fossil fauna, adifferent pattern is discernible. Here five species are in­volved that clearly varied a good deal in Jiveweight, as isindicated by these tentative estimates: Parapapio jonesi,25-40 lb, or 11.3-18.1 kg; Papio robinsoni, 35-1lO lb, or15.9-36.3 kg; Theropithecus danieli and Dinopithecus in­gens, 60-100 lb, or 27.2-45.4 kg; Australopíthecusrobustas. 8~150 lb, or 36.~8.0kg. There appears lo bea correlation between the estimated liveweight of theseprimates and the percentage of subadult individuals rep­resented in each taxon, Figures are as follows: Parapapiojonesi, 8 individuals, 2 subadult: subadult percentage 25;Papío roblnsonl, 38 individuals, 9 subadult: subadult per­centage 23.7; Theropithecus danieli, 17 individuals, 6subadult: subadult percentage 35.3; Dinopithecus ingens,26 individuals, 10 subadult: subadult percentage 38.5;Austraíopíthecus robustas, 87 individuals, 58 subadult:subadult percentage 66.7 (if the subadult/adult transitionis taken as 16 rather than 21 years, then tbe number ofsubadult individuals would be 42 and the subadult per­centage 48.3).

These dala are depícted graphieally in figure 224, fromwhieh it wil1 be seen that, as the estimated liveweights ofthe primates increase, so also do the percentages of sub­adult individuals. A possible interpretation is that thepredator ínvolved selected, by choice, primates withliveweights of aboul 3~100 lb, or 13-45 kg. This pattemis very differenl from that suggested by the anlelope re­mains, where the majority oí prey animals hadliveweights well over 200 lb, or 90 kg.

While the primates eould well have fallen prey loleopards, the heavier bovids could represem the prey of

the larger cats, Dínofelis or Megantereon, or of sociallyhunting hyenas, particularly Euryboas or Hyaenictis. Thebovid remains could equally well have been brought backby brown or spotted hyenas, which appear to have fre­quented the cave.

The conclusion 1draw here is that, if the cave had beenused by severa] species of cat as well as severaI differenthyenas, it would be remarkable indeed if the bone ac­cumulation did not include contributions made by each ofthese camivores. The indications are that this was, infact, the case.

Who Were the Hunters and Who the Hunted? 269

Reservaríon 2

My second reservation about the validity of theSwartkrans leopard hypothesis concerns the pre­ponderance of primates in the prey sample. As figure 196indicates, more than 50% of all the animaJs in the Member1 macrovertebrate assemblages are either hominids orcercopithecoids. This is a remarkable state of affairs, andone is prompted to ask why, if leopards were the mainpredators of these primates, they were concentrating sointensively on this kind of prey. Studies of contemporary

Fig. UI. Sorne evidence for the Swartkrans leopard hypothesis: ¡he spacing of holes in the parietal bones of ajuvenile australopilhecine cranium, SK 54, marches closely the spacíng of rhe lower canínes of a leopard, SK 349,from lhe same deposil.

270 Fossil Assernblages from the Sterkfontein Valley Caves: Analysis and Interpretation

leopards, described in chapter 4, suggest that primatestypically form only a small part of leopard diet, except inparticular circumstanccs such as the baboon sleepingsituation 00 Mount Suswa.

Evidence presented in chapter 11 suggests thatleopards had beco the main predators of the very numer­ous small springbok, Antídorcas bondi, whose remainsdomínate the Swartkrans Member 2 assemblage. Theleopards appear to have used the cave itself, or rreesoverhanging its entrance, as a feeding lair, to which theybrought back the bodies of their prey. The composition ofthis assemblage conforms closely to what might be ex­pected for leopards hunting over a grassland aree in nor­mal circumstances. Antelopes in size class II were thedomínant prey, and baboons constituted only 3% of ani­mals represented.

The question, and crux of my second reservation,therefore is this: Why, if the Swartkrans leopards hunteda "normal' speetrum ofprey in Member2 times, did theyconcentrate so heavily on primates during the accumula­tion of Member 1? 1 believe the answer to this questionwill elucidate the situation not only for the Swartkransaustralopithecines, but also for the primate-dominated as­sernblages in Sterkfontein Member 4 and Kromdraai B.

There appear to be two possibilities. which 1 wi11 dis­cuss as hypotheses A and B.

Hypothesis A: A Specíatized Predator 01 Primates? Itwould be possible to aceount for the abundance of pri­mate remains at the three caves if these places had servedas lairs for specialized predators that huntcd baboons andhominids to the virtual exclusión ofother prey. No livingAfrican camivores are known to do this, except in specialcircumstances such as those described in Hypothesis B,below. Nevertheless, we cannot exclude the possíbilítythat sorne of the extinct camivores may have been soadapted.

Hominids and baboons must be regarded as dangerousprey for any predator-nol because the individual primateis a formidable adversary, but simply because an attackon one individual is likely lo precipitate retaliation by thewhole band unless the predator operares at night or si­lently takes a straggler from the fringes of the group.

In the Sterkfontein valley context the potentíally spe­cialized predators of primates were the false sabertooths,Dínofelis, the true sabertooth, Megantereon, and thehunting byenas of the genera Euryboas and Hyaeníctis.Tú my mind, Dinofelis, would have been well suited forlhe task: it was c1early a cal lhal hunted by stealth, it had

¡

Fig. 222. A reconstrucnon to accouat for tbe damage observed on theskuU of an eustraloplthecíne child, SK 54. Tbe lower canines appear tohave perforaeed the parietals while the uppers gripped rhe chtld's rece.From Brain 1970.

•

Fig. 223. The percentage representation oC bcvíds of different weightdasscs in four Icopa«! prey samples, described in the text, and in thefossíl assembleges from Sterkfontein Member 4, Swartkrans Member 1,and Kromdraai B.

Fig. 224. Histograms. sbowingthe percentagc abundancc ofsubadult indi­viduals among primates of different adult líveweights, whose fossils arepreserved in Swartkrans Member l.

...._­U...o"U-I"',

--",""",.,.-,,,-- ~.

80-150'" .·eel<l

TlIe,,,,,/I/NC:U.

DltlOlI'1flleC".

--10 20 30 50

U ... • • lW UII

BODY WEIGHT

-----_.

,,~~11 11I IV

~I

,.. -..........."

"ii "<

~• ..•

powerful forequarters to hold the prey in position andlong canines to ensure rapid killing. A combination ofrobustjaws and a well-developed crushing component inthe dentition would have allowed Dinofelis to eat aH partsof a primate skeleton except the skull. The hypothesisthat Dinofelis was a specialized killer of primates is per­suasive.

The true sabertooth, Megantereon, was perhaps lesssuited to the role. Its extremely long canines may havebeen more effective 00 prey larger than australopithe­cines or baboons where there was less danger ofdamagingthe sabers 00 bones of the prey skeleton, Its postcaninedentition was so highly adapted for slicing meat that itcould have caused very little damage to even a primateskeleton. The extensive damage observed in the fossilprimate skeletons, particularly the disappearance of post­cranial elernents, is unlikely to have resulted fromMegantereon action alone. Superimposition of hyena ac­tíon upon that of the sabertooths would have to be in­voked,

We kncw regrettably little about the behavior of theextinct huntinghyenas. They appear, from the abundanceof their fossils, to have frequented caves, perhaps whenrearing their young there, They were cursorial hunters,presumably runningdown their prey after the manner ofwild dogs. It is not impossible that they had acquired thehabit of hunting primates and of taking parts of this preyback to theír breeding lairs.

In conclusion, it would be unrealistic 1 think, to ignorethe possibility that both Dinofelis and the hunting hyenascould have been specialized primate-killers. But the time,regrettably, is long past when they might be observed asthey went about their business.

Hypothesis B: A Sleeping-Site Situation? In the firsthypothesis 1 considered the possibility that predators ac­tively went out to hunt primates and brought them back tothe cave lairs. The main postulate of the second hypothe­sis is that the primates actually carne to the caves by theirown volition and were killed and eaten there by carni­vores that were exploiting the situation in an opportunis­tic manner. Such a situation has already been described inchapter 4 where, on Mount Suswa in Kenya, baboonsregularly come to sleep around the edges of sinkholes.Resident leopards kili thern there and drag the bodies intosubterranean lava tunnels 10 eat them undisturbed.

The Transvaal highveld is characterized by cold winternights, and it is known that baboon troops living in thisarca use caves as sleeping places during the wintermonths. Over a number of years 1 have made observa­tions at a dolomite cave 12 km north-northeast ofSwartkrans, 00 the farm Uitkomst, which is used by ababoon troop as a winter sleeping site. The cave opensinto the side of a rocky hillside by way of a high, narrowdoorway (fig. 225) partly obscured by a large Ficus tree.Inside, a steeply inclined talus slope leads down into alarge cavern whose form is shown in figure 226. Thecavem is dominated by a gigantic block of dolomite thathas parted from the roof and come to rest on the ftoor.The ñat top of this block, determined by a dolomitic bed­ding plane, is where the baboon troop normally sleeps,and an extensive deposit of baboon droppings has ac­cumulated around the base of the block, as indicated onthe plan. Lighting in the sleeping area is verysubdued-oo a bright day it is just possible to make out

Who Were the Hunters and Who the Hunted? 271

the internal contours of the cave. 1 have seen a troop ofabout 30 baboons use the cave on nights between Mayand September; duríng warmer months they use an openc1iff face about I km to the northeast. The baboons typi­cally enter the cave shortly before sunset and leave itafter sunrise. On one occasion 1 hid inside the cavem,making my presence known only after the baboons hadtaken up their sleeping places. Although pandemoniumbroke out in the cave, the baboons could not be inducedto leave the place in the dark.

In view of a baboon's characteristic fear of the dark,sorne observations by Chrís Gow (1973) are of specialinterest. He described a sinkhole 30 m deep in Tertiarylimestones near Bredasdorpin the southeastem Cape thatis regularly visited by a baboon troop. At the base of thesinkhole a dark passage extends \30 m to a subterraneanstream, and this running water attracts the baboons,which are prepared to brave the darkness to reach it.Afier they drink, the baboons generally spend the night inthe cave.

The current distribution of baboons on lbe Transvaalhighveld is certainly regulated by the availability of suit­ably sheltered sleepíng sites (L. P. Stoltz, pers. comm.,and personal observation). Areas devoid of such sites donot support resident troops.

So it was, perhaps, during australopithecine times. 1have suggested in chapter 3 that Sterkfontein valleyhominids may have participated in a seasonal round thattook them into warmer bushveld areas during the coldestmonths and brought them back onto the highveld in thespring. Allbough baboons do not do this today, australo­pithecines may have, in which case they probably usedcave sleeping sites only on cold nights in spring and au­tumn.

Erosion has removed the original entrance areas of theSterkfontein, Swartkrans, and Kromdraai caves, whichmust have been intact when the australopithecines in­habited the valley. Yet the abundance of owl-pellet­derived microvertebrate bones in the cave deposits ar­gues in favor of dimly lit entrance areas with suitable owlroosts beneath their overhanging roofs. Such places maywell have attracted baboons and hominids as sleepingplacee, while the deeper recesses of the caves were usedas lairs and feeding retreats by the cats who preyed uponthe primates.

Conclusions

Sorne of the Swartkrans evidence suggests that leopardspreyed on primates there. The same may well have beentrue at the caves of Sterkfontein and Kromdraai. lfleopards were invclved. however, the overwhelming pre­ponderance of primate remains suggests that a sleepingsite was being exploited rather than that the leopardswere hunting normally in open country.

If the abundance of primate remains in the SterkfonteinvalLey caves can be attributed to opportunistic predation atsleeping sites, as 1 think it can, it is probable that morethan one species of cat was involved in the predaticnprocess. Besides leopards, it is higly likely that Dinofeliscats also participated in this process.

The suggestion that Dinofelis and the huntíng hyenaswere active, selective predators of primates cannot beproved or disproved-the rnatter therefore remains open.

A

Who Were the Hunters and Who the Hunted? 273

It is to be expected that each of the three bone ac­cumulations has resulted from the independent activity ofcats and hyenas. From the point of view of taphonomicinterpretation, this complicating fact is as unfortunate asit is true.

The Hunted Becomethe Hunters

Al Sterkfonteín, the interface between the top of Member4 and the bottom of Member 5 represents a time intervalcrucial in the course of human evolution. During thisinterval the gracile australopithecines disappeared fromthe Transvaal scene and the first men appeared. In thisinterval, too, the evolving men mastered a threat to theirsecurity that had been posed by the cave cats overcountless generations. During Member 4 times the catsapparently controlled the Sterkfontein cave, draggíngtheir australopithecine victims into its dark recesses. ByMember 5 days, however, the new men not only hadevicted the predatcrs, but had taken up residence in thevery chamber where their ancestors had been eaten.

How the people managed this is not recorded, but itcould surely have been achieved only through increasingintelJigence reftected in developing technology. It istempting lo suggest that the maslery of fire had alreadybeen acquired and that this, together with the develop­ment of crude weapons, tipped the balance of power intheir favor. The tipping of this balance represented, 1think, a crucial step in the progressive manipulation ofnature that has been so characteristic of the subsequentcourse of human affairs. It was a step the robust australo­pithecines apparently failed to take, and their extinctionwas doubtless hastened by predators they were powerlessto control.

Yet the evidence, which flowed first from the re­searches of Elisahelb Vrba, does nol suggest that lbe firslmen to live at Sterkfontein were more than amateurs athunting. The nature of their antelope food remains, pre­served in the upper levels of the cave, suggests that theydepended heavily on the kills of professional camivoresbefore lbey progressively developed their own prowess as

Fig. 226. The Uitkomst baboon sleeping cave in plan and section. Thebaboon troop spends winter nights on the fallen roof blocks inside themain cavem.

~ fallan roof bloc::ks

lI'a concentration ofbllboon droppings

cave entranceA

o

274 Fossil Assemblages from the Sterkfontein Valley Caves: Analysis and Interpretation

hunters. This interesting and significant conclusion is cor­roborated by the taphonomic studies described in thisbook.

Postscript

In many ways, rhe first part of this book is more satisfac­tory than the second. It explores the potential usefulnessof bone accumulations in caves as indicators of pastevents. For the Sterkfontein valley assemblages used intbe study, only a small part of this potential could berealized because the fossils had not been recovered withthe regard for subtle detail that is a prerequisite for reli­able taphonomic reconstructíons, We may hope theguidelines that have emerged in the course of this studywill help ensure that fossil assemblages as yet unearthedwiU not lack the detailed information so vital for theirproper evaluation.

II,•!

Appendix: Tables

Table 1. Separation of the Extant Southem African Bovid Fauna into Size Classes, Based 00 Liveweights

Weight Range Weight Range

Species Lb Kg Species Lb Kg

Antelope class J Antoíone cíass IIIDikdik, Madoqua kirki 10-11 4.5-5 Lechwe, Kobus leche 170-285 77-130SUDi, Nesotragus moschatus 10-15 4.5-7 Nyala, Tragefaphus angasi 200-250 91-114Blue duiker, Cephalophus monticola 12-16 6-7 Sitatunga, Tragelaphus spekei 200-250 91-114Cape grysbok, Raphicerus melanotus 15-20 7-9 Tsessebe, Damaliscus lunatus 258-350 117-58Sharpe's grysbok, Raphicerus Red bartebeest, Alce/aphus

sharpei 15-20 7-9 bucelaphus 232-380 106'-72Red duiker, Cephalophus natatensis 20-30 9-14 Lichtenstein's hartebeest,Klipspringer, Oreotragus oreotragus 21-36 10-16 Alcelaphus Iichtensteini 322-450 146-205Steenbok, Raphicerus campestris 24-33 11-15 Kudu, Strepsiceros strepsiceros 330-651 150-2%Common duiker, Sylvicapra grimmia 24-45 11-21 Black wildebeest, ConnochaetesOribi,Ourebia ourebí 30-42 14-19 gnou 350-400 158-82

Defassa waterbuck, Kobus defassa 350-450 158-205Antelope cíass 11 Waterbuck, Kobus ellipsiprymnus 350-600 158-272Springbok, Antidorcas marsupialis 40-115 18-52 Gemsbok, Oryx gazella 400-524 182-238Mountain reedbuck. Redunca fuí- Sable, Hippotragus niger 450-580 205-{,4

vorufula 50-60 23-27 Blue wildebeest, ConnochaetesGray rhebuck, Pelea capreoíus 50-60 23-27 taurinos 450-602 205-74Bushbuck, Tragelaphus scriptus 50-182 23--83 Roan, Hippotragus equinos 491~58 223-99Blesbok, Damaliscus dorcas 70-180 32-81Impala, Aepyceros melampus 80-15L 36-69 Antelope class IVReedbuck, Redunca arundinum 98-228 45-104 Buffalo, Syncerus caffer 808-1,841 367-837Puku, Kobus vardoní 124-84 56-84 Eland, Taurotragus oryx 870-2,078 396-945

Tab1e 2. Details of the Inhabitants of Eight Villages onthe Kuiseb River, South-West Africa, in March 1966

Human HumanVillage Males Fernales Do8' Goats

Ossewater 5 6 8 460Natab 2 4 7 188Soutrivier 10 11 3 177Klipneus 7 4 6 200Swartbank 3 4 3 125Itusib 6 6 3 44Ururas 17 25 6 320Rooibank 11 12 4 240

Total 61 72 40 1,754

275

276 Appendix: Tables ~Tabie 3. List of Goat Bones Collected at Kuiseb River Hottentot Villages

~.

Skeletal Part Numbers Skeleral Pan Numbers

Skul/ 5i2 Radins DnJ u/na 207Homs and cores 385 Complete bones 3Cranial fragmenta 70 Proximal ends 62Maxillary fragmenta 57 Distal ends 19 "

Shaft fragrnents 123

1Mandible 188Complete half-mandibles 38 Femur 115Mandible fragments 150 Proximal ends 18

Distal ends 9Loose teem 15 Shaft fragmenta 88

Yertebrae 115 Tibia 237

1tst cervical (atlas) 12 Proximal ends 132d cervical (axis) 14 Distal ends 72Other cervical 12 Shaft fragments 152 1•Thorecic 21 !Lumbar 31 Metocorpol lOO )

Sacra! 1 Complete bones 8 ICaudal O Proximal ends 24Fragments 24 Distal ends 15 \Shaft fragments 53Ribs 174

Met atarsaí JOIScapula 59 Complete bones 9Head portien 28 Proximal ends 30Other fragments 31 Distal ende 11

Shaft pieces 51Pelvis 55Acetabular portian 34 Asrr€lgalus, complete 16Other fragments 21

Catcaneus. complete 14Humeras 196Proximal ends O Pñatanges, complete 21Distal ends 82Shaft fragments 114 Bane fíakes 248

Total 2,373

Table 4. Estimated Ages for Kuiseb River Goats Based00 Tooth Eruption in Left and Righ1 Half-Maodibles

Number of Goats

Age Class Left Side Right Side

Under 6 months 1 O

9--12months 17 2315-30 months 7 6More thaa 30 months 28 35

Total 53 64

Tab

le9.

Ov

eral

lA

nal

ysi

so

fB

on

eF

rag

men

tsfr

om

Po

mo

ng

we

Cav

e,M

ato

po

HU

ls,

Zim

bab

we

Tor

tois

eO

strt

chSc

apul

aan

dL

imb

Bon

eM

isce

llane

ous

Car

apac

ean

dE

ggsh

ell

Lan

dSn

ail

Lo

ng

-Bcn

eL

evel

SkuU

piec

esV

erte

brae

Rib

sP

elvi

sPi

eces

Foot

BO

lles

Frag

men

tsP1

astr

onP

iece

sPi

eces

Mus

selS

hell

FIak

esTo

ta1s

I10

413

417

2117

655

2811

212

613

0O

299

1,20

22

461

1321

140

2026

6474

2O

104

511

314

529

459

7645

082

100

321

37O

O39

31,

957

440

920

415

014

533

120

423

839

066

O2

3,38

75,

526

521

114

611

593

217

149

223

189

SO

12,

289

3,63

86

189

9356

5810

812

815

417

52

OO

1,73

12,

704

797

2614

1984

25SO

498

OO

661

1,03

38

2514

4S

3016

101

3128

OO

257

SIl

9SI

95

IS34

938

2218

3O

365

569

106

3O

111

113

42

1O

63lO

ST

otal

1,28

392

443

34S

41,

581

689

981

1,35

736

613

63

9,54

917

,756

Not

e:T

otal

num

ber

ofbo

nefr

agm

enta

:17

,756

.

....."

',,-

JI'

t5

(ti

ti

...

Tab

lero

,L

eng

ths

of

Bo

ne

Fla

kes

fro

mV

ario

us

Lev

els

inth

eP

om

on

gw

eC

ave

Dep

osi

t,Z

imb

abw

e

Len

gth

(in

inch

es)

Lev

e!(>

-11-

22-

33-

44-

5

N%

N%

N%

N%

N%

1+

212

631

.321

553

.357

14.1

51.

23

4712

,026

066

.174

18.8

92.

33

.84

804

23.7

2,28

167

.326

87.

932

.92

.15

+6

738

18.4

2,54

963

.456

813

.916

34.

11

.37

+8

+9

207

16.1

798

62.2

249

19.4

262.

03

.310

46·

352

82.5

57.

91

1.6

11.

6T

otal

1,92

66,

155

1,21

123

610

N

5-{

j

% .3

N 403

393

3,38

74,

020

1,28

363

9,54

9

Tot

al

% 99.9

100.

099

.910

0.1

100.

099

.9

.""T

able

11.

Min

imu

mN

um

bers

of

Ind

ivid

ual

An

imal

sR

epre

sen

ted

inP

re-

Tab

le12

.M

imin

umN

umbe

rso

íInd

ivid

ual

Ani

mal

sR

epre

sent

edin

Mid

dle

Mid

dle

Sto

ne

Ag

eL

evel

sat

Po

mo

ng

we

Cav

e,w

ith

Tb

eir

Per

cen

tag

eC

on

trib

u-

Ston

eA

geL

evel

sat

Pom

ongw

eC

ave,

wit

hT

heir

Per

cent

age

Con

trib

utio

nsto

tio

ns

loth

eD

iet

of

the

Peo

ple

the

Die

to

fth

eP

eop

le

70%

of70

%of

Tot

alW

eigh

tT

otal

Wei

ght

Tot

alW

eigh

tP

erce

ntag

eT

otal

Wei

ght

Per

cent

age

Ani

mal

NL

bE

ach

(lb)

Lb

Kg

ofT

otal

Díe

tA

nim

alN

Lb

Eac

h(l

b)L

bK

gof

Tot

alD

iet

Ant

eíop

ecí

ass

1A

nte

lope

cla

n[

Dui

ker

235

70.0

49.0

22.2

1.89

Dui

ker

535

175.

012

2.5

55.4

1.29

Stee

nbok

229

58.0

40.6

18.4

1.57

Ste

enbo

k4

2911

6.0

81.2

36.7

.85

Gry

sbok

218

36.0

25.2

11.4

.97

Kli

pspr

inge

r4

2911

6.0

81.2

36.7

.85

Klip

spri

nger

229

58.0

40.6

18.4

1.57

Gry

sbok

218

36.0

25.2

11.4

.26

Ant

elop

ecl

ass

III

An

telo

pe

cias

s1I

Wat

erb

uck

147

547

5.0

332.

515

0.5

12.8

6R

eedb

uck

413

554

0.0

378.

017

1.0

3.97

Tse

sseb

e1

325

32s.

o22

7.5

102.

98.

80B

ush

bu

ck1

116

116.

081

.236

.7.8

5Sa

ble

151

551

5.0

360.

516

3.1

13.9

4A

nte

lope

clas

sJI

lO

ther

ma

mm

a/s

Tse

sseb

e5

325

1,62

5.0

1,13

7.5

514.

711

.96

Das

sie

159

135.

094

.542

.83.

65W

ater

bu

ck3

475

1,42

5.0

997.

545

1.4

10.4

8H

are

35

15.0

10.5

4.8

.41

Sabl

e2

515

1,03

0.0

721.

032

6.2

7.59

wer

thog

114

014

0.0

98.0

44.3

3.79

Wild

ebee

st1

526

526.

036

8.2

166.

63.

87Z

ebra

180

080

0.0

560.

025

3.4

21.6

5K

ud

u1

491

491.

034

3.7

155.

53.

61E

xtin

ctho

rse

11,

000

1,00

0.0

700.

031

6.7

27.0

6R

oan

157

557

5.0

402.

518

2.1

4.23

Por

cupi

ne1

3030

.021

.09.

5.8

1A

nte

lope

cías

sIV

Leo

pard

1E

land

11.

474

1,47

4.0

1,03

1.8

466.

910

.86

Spr

ingh

are

17

7.0

4.9

2.2

.19

Oth

erm

amm

als

Can

efa

!1

22.

01.

4.6

.05

Das

sie

!I9

99.0

69.3

31.4

.73

Rep

tile

sW

art

bo

g4

140

560.

039

2.0

177.

44.

12T

orto

ise

43

12.0

8.4

3.8

.32

Zeb

ra3

800

2,40

0.0

1,68

0.0

760.

217

.66

Mon

itor

lizar

d2

714

.09.

84.

4.3

8E

xtin

ctho

rse

21,

000

2,00

0.0

1.40

0.0

633.

514

.71

Snak

ein

det.

21

2.0

lA.6

.05

Har

e2

510

.07.

03.

2.0

7B

irds

Bus

hpig

118

018

0.0

126.

057

.01.

32O

stri

cheg

g-

--

--

-Po

rcup

ine

130

30.0

21.0

9.5

.22

Roc

kpi

geon

1I

1.0

.7.3

.03

Rat

el1

2020

.014

.06.

3.1

5In

vert

ebra

tes

Spr

ingh

are

17

7.0

4.9

2.2

.05

Lan

dsn

all

--

--

--

Rep

tile

sT

otal

453,

431

3,69

5.0

2,58

6.5

1,17

0.3

99.9

9T

orto

ise

53

15.0

10.5

4.8

.11

Mon

itor

Liz

ard

37

21.0

14.7

6.7

.15

Snak

e¡n

det.

31

3.0

2.1

1.0

.02

Plat

edIiz

ard

1I

1.0

.07

.3.0

1B

irds

Ost

rich

egg

Roc

kpi

geon

11

1.0

.07

.3.0

1ín

vert

ebra

tes

Riv

erm

usse

l1

Tot

al74

6,94

713

,592

.09,

514.

44,

305.

110

0.00

280 Appendix: Tables 1¡

Minimum Numbers of Individual Animals.~.

Table 13. Represented in LaterStone Age Levels al Pomongwe Cave. with Their Percentage Contributions tothe Diet of the People

7CY70 ofTotal Weight

Total Weight PercentageAnimal N Lb Each (lb) Lb Kg of Total Diet

Anteíope class 1Steenbok 6 29 174 122 55.2 1.80Klipspringer 5 29 145 101 45.7 1.49Duiker 5 35 175 122 55.2 1.80Grysbok I 18 18 13 5.9 .19Anteíope class IIBushbuck 116 116 81 36.7 1.18Anteiope class IJISable 4 515 2,060 1,442 652.5 21.26Wildebeest 4 526 2,104 1,472 666.1 21.70Tsessebe 2 325 650 455 205.9 6.70Other mammalsDassie 65 9 585 409 185.1 6.03Hare 13 5 65 46 20.8 .68Warthog 5 140 700 490 221.7 7.22Zebra 3 800 2,400 1,680 760.2 24.77Leopard 2Antbear I 130 130 91 41.2 1.34Bushpig I 180 180 126 57.0 1.86Sprínghare 1 7 7 5 2.2 .07ReptilesTortoise 20 3 60 42.0 19.0 .62Monitor lizard 10 7 70 49.0 22.2 .72Plated lizard 4 I 4 2.8 1.3 .04Snakeindet. 3 I 3 2.1 1.0 .03Python 2 10 20 14.0 6.3 .21BirdsRack pigeon 4 I 4 2.8 1.3 .04Vulture 2 7 14 9.8 4.4 .14Francolin I 2 2 1.4 .6 .02Goose I 5 5 3.5 1.6 .05Ostrich eggtnvertebratesRiver mussel 2Land snail

Tola/ 168 2,901 9,164 6,784 3,069 99.97

Table 14. Percentage Contributions of Various AnimalGroups to the Diet of Stone Age People al PomongweCave

Percentage Representation

Animal Group Pre-MSA MSA LSA

Antelope class I 6.0 3.3 5.3Antelope class 11 4.8 1.2Antelope class In 35.6 41.7 49.7Antelope class IV 10.9Othermammais 57.6 39.0 42.0AH other animais .8 .3 1.9

Total 100.0 100.0 100.1

Appendix: Tables 281

Table 15. Skeletal Parts of Dassies, Procavía capensis and Heterohyrax bruceí, from Al] Levels of thePomongwe Cave

Mínimum number of individuals Ulna 66Based 00 maxillary fragments: 74 Complete bones 5Based 00 mandible fragments: 76 Proximal ends 56Based 00 distal humen: % Distal ends 5Composite minimum number in alllevels: 114

Radius 21Total numher of bone fragmente /,/92 Complete bones 6

Proximal ends 6Skul/ 473 Distal ends 9Maxil1ary pieces 187Mandible pieces 199 Femur 85Isolated teeth 16 Proximal ends 38Miscellaneous skull pieces 71 Distal ends 18

Shaft pieces 29Yertebrae 1iJ6Atlas 18 Tibia 45Axis 9 Complete bones 2Other 79 Proximal ends 30

Distal ends 13Scapu/a, pieces 38

Fibuia, fragments 3Pelvis, pieces 36

Phalanges. complete 3Ribs, píeces 32

Pieces 01tib bone shaft 77Humerus 207Proximal ends 21 Overatt total 1,192Distal ends 186

Table 16. Lengths of Bone Flakes from Various Levels at Bushman Rock Shelter, Eastem Transvaal

Flake Sizes Flake Sizes(inehes) (inehes)

Level 0-1 1-2 2-3 3-4 4-5 5-6 Total Level 0-1 1-2 2-3 3-4 4-5 5-6 Total

1 211 151 20 383 27 9 20 24 4 1 582 51 47 5 3 106 28 8 43 19 3 4 77

Total 262 198 25 3 489 29 8 20 4 3230 1 6 1 8

3 48 73 11 134 31 32 22 6 614 32 4 11 2 175 1 I 33 6 4 3 136 35 35 8 78 34 2 6 7 3 187 38 53 12 1 104 35 7 22 10 2 418 46 51 15 I 113 36 1 2 39 33 94 44 6 178 37 1 6 2 1 10

10 18 32 4 54 38 5 11 13 2 3111 19 34 13 3 70 39 6 3 912 35 66 24 2 127 40 2 2 413 22 31 19 1 73 41 1 5 2 814 39 74 34 6 154 4215 24 59 49 16 148 43 1 2 316 34 138 90 17 279 Tolal 81 189 97 21 5 393t7 26 48 38 6 11918 7 24 20 6 5819 2 2 420 6 8 1 2 1721 1 3 6 3 1322 5 13 3 2123 12 12 17 2 4424 6 4 3 1325 1 3 2 626 6 14 13 33 Overol!T0((11 458 873 427 76 6 I 1,841 Total 801 1,260 549 100 7 6 2,723

282 Appendix: Tables

Table 17. Lengths of Bone Flakes from VariousCultural Horizons at Bushman Rock Shelter

Flake Sizes(inches]

Horizon 0-1 1-2 2-3 3-4 4-5 5-6 Total

BantuN 262 198 25 3 1 O 489% 53.6 40.5 5.1 .6 .2 O 100.0LSAN 458 873 427 76 6 1 1.841% 24.9 47.4 23.2 4.1 .3 .05 100.0MSAN 81 189 97 21 O 5 393% 20.6 48.1 24.7 5.3 O 1.3 100.0

Total 801 1,260 549 lOO 7 6 2,723

Table 18. Occurrence of Various Animal Species al Bushman Rock Shelter and Parts by WhichThey Are Represented

Animal

Tortoise (sp. indet.)

Zebra (Equus burcheiíi¡

Reedbuck (Redunca arundinum)

Warthog (Phacochoerus aethíopicus¡

Hartebeest (cf. Alceiaphus)Wildebeest(Connochaetes taurinus¡Dassie (Precavía capensis)

Monitor lizard (Varanus sp.)

Impala (Aepyceros melampus)Kudu (Tragelaphus strepsíceros¡Sable (Hippotragus niger)Duiker (Sylvícapra grimmia)Steenbok (Raphicerus campestris)

Hace (Lepus sp.)

Python (Python sebae}Extinct horse (Equu.r capensís}Tsessebe (Damaliscus Junatus)Smallcamivore (sp. mdet.IBird (ef. Bubo sp.}Mussel (Unio caffer [7])Ostrich eggshell

Land snail (Achatina zebra [1) )

Total Numbercf Individuals

34

JI

9

8

666

6

44433

3

21111I

67 pieces

38

Layer Number(numbeeoí individuals in brackets)

1 (1); 2 (1); 3 (1); 6 (1); 7 (1); 8 (1); 9 (1);10 (1); JI (1); 12(1); 13(1); 14 (1); 15 (1);16 (1); 17 (1); 18 (1); 19 (1); 20 (1); 21 (1);22 (1); 23 (1); 24 (1); 25 (1); 26 (1); 27 (1);28 (1); 29 (1); 31 (1); 32 (1); 34 (1); 35 (1);38 (1); 39 (1); 43 (1)13 (1); 15 (1); 16 (1); 17 (1); 21 (1); 22 (1);23 (1); 24 (1): 27 (1); 29 (1); 30 (1)8 (1); 9 (1); 10 (1); 16 (1); 17 (1); 18 (2);21 (1); 24 (1)16 (1); 18 (1); 19 (1); 23 (1); 24 (1); 26 (1);28 (1); 41 (1)11 (1); 12 (1); 17 (1); 18 (1); 26 (1); 43 (1)9 (1); JI (1); 16 (1); 22 (1); 26 (1); 27 (1)9 (1); 22 (1); 23 (3); 41 (1)

1 (1); 2 (1); 3(1); 15 (1); 21 (1); 23 (1)

7 (1); 9 (1); 15 (1); 16 (1)15 (1); 17 (1); 23 (1); 28 (1)15 (1); 16 (1); 17 (1); 27 (1)6 (1); 7 (1); 18 (1)6 (1); 8 (1); lO (1)

1 (1); 6 (1); 9 (1)

26 (1); 27 (1)41 (1)9 (1)23 (1)1 (1)JI (1)1 (1); 2 (2); 3 (2); 6 (2); 7 (3); 8 (1); 9 (1);10 (3); lJ (1); 12 (2); 13 (2); 14 (6); 15 (5);16(14); 17(3); 18(2);21 (1);23(1);24(1)25 (1); 26 (2); 27 (1); 28 (4); 29 (2); 31 (4)1 (2); 2 (1); 3 (1); 5 (1); 6 (1); 7 (1); 8 (1);9 (5); 10 (1); 11 (1); 12 (1); 13 (1); 14 (1);15 (1); 22 (1); 23 (1); 25 (1); 26 (1); 27 (1);28 (1); 29 (1); 30 (1); 31 (2); 32 (1); 33 (1);34 (1); 35 (1); 36 (1); 37 (1); 38 (1); 39 (1);40 (1)

Number oC Skcletal Parts Usedin Identification

Carapace: 588Pectoral and pelvic girdles: 26Limb bones: 59Vertebrae: 1

Teeth: 24

Teeth: 15Mandibles: 2Teeth: 8

Teeth: 8Teeth: 6Mandibles and maxi1lae: 6Limb bones: 1Mandibles: 3Vertebrae: 5Teeth: 6Teeth: 4Teeth: 5Teeth: 4Teeth: 2Mandible: IMaxiUae: 1Limb bones: 2Vertebrae: 2Teeth: ITeeth: 1Teeth: ILimb bone: IHingejoints: 1Fragments of sheU: 67

Fragments of shell: 490

,Appendix: Tables 283.~'

Table 19. Minimum Numbers of Individual Animals Represented in the MiddleStone Age Layers al Bushman Rock Shelter, with Their PercentageCon tribu tions 10 the Diet of the People

7rHo ofTotal Weight

Total Weight PercentageAnimal N Lb Each (lb) Lb Kg of Total Diet

Antelope class /lfHartebeest 386 386.0 270.2 122.3 6.84Wíldebeest 526 526.0 368.2 166.6 9.31Kudu 491 491.0 343.7 155.5 8.69Sable 515 515.0 360.5 163.1 9.12Other mammalsZebra 3 800 2,400.0 1,680.0 760.2 42.50Warthog 2 140 280.0 196.0 88.7 4.96Exrinct horse 1 1,000 1,000.0 700.0 316.7 17.71Dassie 1 9 9.0 6.3 2.9 .16ReptilesTortoise 10 3 30.0 21.0 9.5 .53Python I 10 10.0 7.0 3.2 .18BirdsOstrich egglnvertebratesLand snail

Total 22 3,880 5,647 3,952.9 1,788.7 100.00

Tab1e 20. Minimum Numbers of Individual Animals Represented in the LaterStone Age Layers al Bushman Rock Shelter, with Their Percentage Contributionslo the Diel of the People

70%ofTotal Weight

Total Weight PercentageAnimal N Lb E.eh (lb) Lb Kg of Total Diet

Antelope class 1Duiker 3 35 105.0 73.5 33.3 .61Steenbok 3 29 87.0 60.9 27.6 .51Antelope class 11Reedbuek 9 135 1,215.0 850.5 384.8 7.09Impala 4 106 424.0 296.8 134.3 2.47Anteiope class 111Hartebeest 5 386 1,930.0 1,351.0 61I.3 11.26Wildeheest 5 526 2,630.0 1,841.0 833.0 15.35Kudu 3 491 1,473.0 1,031.1 466.6 8.60Sable 3 515 1,545.0 1,081.5 489.4 9.02Tsessebe 1 325 325.0 227.5 102.9 1.90Otñer mammalsZehra 8 800 6,400.0 4,480.0 2,027.1 37.35Warthog 6 140 840.0 588.0 266.1 4.90Dassie 5 9 45.0 31.5 14.3 .26Hare 2 5 10.0 7.0 3.2 .06Small camivore 1 3 3.0 2.1 .95 .02ReptilesTortoise 22 3 66.0 46.2 20.9 .39Varanus 4 7 28.0 19.6 8.9 .16Python 1 10 10.0 7.0 3.2 .06BirdsOstrich egglnvrnebratesRiver musselLand snail

Total 86 3,525 17,136.0 11,995.6 5,427.9 100.01

284 Appendix: Thbles

Table 2). Percentage Representation ofVarious AnimalGroups in the Middle and Later Stone Age Levels alBushman Rock Shelter

Percentage Representation

Animal Group

Antelope class 1Antelope class 11Antelope class 111Antelope class IVOther mammalsAII other animals

Total

Middle Stone Age

34.0

65.3.7

100.0

Later Stone Age

1.19.6

46.1

42.6.6

100.0

Table 23. Occurrence and Percentage Representatíonof Unidentifiable Bone Fragments in Each Layer at theWillan Large Rack Shelter

Bone Fragment Length(inches)

Layer 0-1 1-2 2-3 J-4 4-5 Total

N 181 37 4 1 223% 81.2 16.5 1.8 0.4 99.9

2 N 2.518 126 4 2.648% 95.5 4.7 .2 100.4

3 N 22.993 1.677 90 7 1 24,768% 92.8 6.7 .3 .03 .004 99.8

4 N 6,252 6.252% 100.0 100.0

Total 31,944 1,840 98 8 33,891

·""T

able

24

.O

ver

aJl

An

aly

sis

of

Bo

ne

Fra

gm

ents

fro

mth

eW

ilto

nL

arg

eR

ack

Sh

elte

r

Lev

els

Hea

rth

Hea

rth

Hea

rth

Hea

rth

Hea

rth

Ske

leta

lP

arts

12A

2B2C

3A3B

3C3D

12

33E

45

3F3G

67

3H31

3i4

Tot

al

.B

ovid

Rem

ains

23

Hor

n-co

res

12

46

Cra

niaJ

frag

men

ta1

2M

axil

lary

frag

men

taI

14

Man

dib

lefr

agm

ents

11

17

311

33

993

1%T

ooth

frag

men

ts1

192

726

71

44

12

9A

tlas

vert

ebra

e1

14

11

2II

Axi

sve

rteb

rae

1I

14

15

23

17

30O

ther

cerv

ical

vert

ebra

e1

12

11

17

48

11

443

Tho

raci

cve

rteb

rae

33

11

14

56

17

32

36L

umba

rve

rteb

rae

2¡

15

25

12

57

Cau

dalv

erte

brae

381

113

156

927

203

Inde

term

inat

eve

rteb

rae

116

121

14

1033

114

41

3¡

12Sc

apul

a2

¡3

4P

elvi

s1

201

237

11

2s12

3R

ibs

56

14

94

133

12

Stem

umI

11

Hum

eros

.pro

xim

al1

11

210

Hum

eros

,dis

tal

11

31

11

4R

adiu

s,pr

oxim

al1

21

11

6R

adiu

s,di

stal

14

31

121

Uln

a,pr

oxim

al3

1I

12

22

32

113

Fem

ur,p

roxi

mal

21

I1

11

11

Fem

ur,d

istal

3T

ibia

,pro

xim

al1

21

12

6T

ibia

,dis

tal

11

313

1810

24

21

1714

3C

arpa

ls,t

arsa

ls,s

esam

oids

38

72

121

217

39

24

13

539

Ast

raga

lus

31

21

33

26

11

118

Cal

cane

us2

21

13

1M

etac

arpa

l,di

stal

11

Met

atar

sal,

dist

al1

282

134

31

210

117

Met

apod

ialf

ragm

ents

215

41

44

89

16

601

714

214

811

2626

7Ph

alan

ges

927

217

68

721

36

1911

122

13

911

4T

erm

inal

phal

ange

s3

68

17

72

184

1

Mam

maJ

ian

,no

nbov

idre

mai

ns1

Cra

nial

frag

men

ts1

25

14

18M

axíll

ary

frag

men

ts1

21

21

I3

18M

andi

ble

frag

men

ts2

51

21

15

22T

ooth

frag

men

ts1

81

21

11

Atl

asve

rteb

rae

1L

umba

rve

rteb

rae

12

114

Inde

term

inat

eve

rteb

rae

22

61

26

Con

tin

ued

onfo

llow

ing

pa

ge

Tab

le2

4.

Con

tin

ued

Lev

els

Hea

rth

Hea

rth

Hea

rth

Hea

rth

Hea

rth

Ske

leta

lP

ans

12A

28

2e

3A3

83C

3DI

23

3E4

53F

30

67

3H31

3!4

Tot

al

Sca

puIa

12

I4

Pel

vis

13

116

Rib

s4

62

32

Hum

eros

,pro

xim

al1

13

33

620

Hum

eros

,dist

al2

31

23

Rad

ius,

prox

imal

11

514

Uln

a,pr

oxim

alI

32

I1

37

Fem

ur,p

roxi

mal

22

Tib

ia,p

roxi

mal

1I

IT

ibia

,dis

tal

1I

Car

pals

and

tars

als

I1

I6

Ast

raga

lus

13

32

111

Cal

cane

us3

2I

4M

etap

odia

lfra

gmen

ts3

33

738

PbaJ

ange

s3

13

63

I6

2I

Term

inal

phal

ange

s1

33I

2427

517

9420

168

905

Mic

rofa

unal

frag

men

ta40

3832

1622

4029

3422

Bir

dre

mai

ns1

35

Bir

d,k

elet

alp

an

,1

411

221

Ost

rich

eggs

helI

frag

men

ta1

3

Rep

tile

rem

ains

Tor

tois

e,ca

rapa

cean

d35

23

1224

823

04

2411

160

3851

62.

586

plas

trón

frag

men

te22

8810

010

2511

325

311

87

196

5628

I12

132

104

558

215

Tor

toís

e,o

ther

skel

etal

parts

1415

91

138

79

613

635

52

415

1Sn

ake,

skel

etal

part

s10

41

188

345

6I

Liz

ard,

skel

etal

part

s1

Fish

I2

Fis

h,sk

elet

alpa

rts

tnve

rteb

rate

rem

ains

13

Mar

ine

mol

lusk

frag

men

ts2

1533

623

14

920

4T

erre

stri

alm

olJu

skfr

agm

ents

225

1020

2531

28

26

I2

118

Fres

hwat

erm

ollu

skfr

agm

ente

4020

517

93

330

155

1018

136

101

19

2715

61I

830

Cra

bfr

agm

ents

318

234

613

384

40

lnde

term

inat

efl

akes

an

dfr

agm

ents

223

1.11

21,

317

219

1,97

42,

061

2,43

82,

232

205

1,10

71,

703

4,28

898

479

3,07

32,

730

3137

21,

202

567

208

6,25

233

,891

Mis

ceJl

aneo

us

skel

etal

part

s50

149

8521

142

106

162

251

3113

635

570

827

307

243

588

5348

3948

03,

036

Tot

al39

61,

602

1,66

928

72,

314

2,46

53,

140

2,91

325

61,

600

1,84

55,

770

125

558

3,93

83.

787

4253

81,

539

757

304

7,78

443

,629

~"'\

"'""

"""O

,&tn

W'I

'lr

."WlO

ffl'"_O~

288 Appendix: Tables

Table 26. Inventory of Skeletal Parts on Which Identifications Were Based in the Wilton Remains

Animal

Blue duiker

Gray duikerVaal rhebuck

BushbuckDomestic cow (?)

Large alcelaphine(wildebeest)

Common dassie

Scrub hare

SkunkSmall camivoreChacma baboonBushpígSmall rodents

Insectivores

BirdBarn owl (1)Ostrich eggshell

Part

I hom-core, 2 tooth fragments (3A); 2 teeth,1 distal rnetatarsal (38); 1 tooth (3D); 2 bom­cores, I tooth (30); 1 tooth (4)I tooth (1)

1 maxillary fragment (1); 2 teeth (4)

I tooth (38)1 terminal phalanx (1)

3 teeth (4)1 distal humerus, J occipital (1); 1 maxilla,

7 teeth (2A); 1 maxilla, 1 mandible , 2 distalhumen(3A); 1maxilla, 4mandible fragments(38); I maxilla (3E); 2 distal humeri (3F); 3maxilla, 1 distal humeros (30); 2 mandiblefragmenta, 1 proximal femur OH); 4 max­itlae, 3 mandibles, 1 tooth, 6 distal humen,1 proximal femur, l proximal ulna (4)

1 distal humeros (1); 1 proximal humerus(ZA); 1 maxilla, 2 teeth, 2 distal humen (3A);

] mandible (3B); I maxilla, I tooth (3e); l

mandible (3D, Hearth 2); I maxiña, 3 distal

humen (3E); ] distal humerus (3F); 2 max­illae, 1 distal humerus (30)

1 tooth (1)1 tooth (2A); 2 teeth (3E)1 tooth (28): 1 tooth (3A)l tooth fragment (3G); 4 tooth fragments (4)Sku\J parts: 3(1): 6(2A); 4(28); 2(2C): 14(3A); 5(38); 6(3C); 5(3D); 2(3D, Hearth 2):I(3E); 1(3F); 18(3G): 2(3G, Hearth 7); 11

(3H); 2(31); 7(4)Sku\J parts: 1(1); 1(2A); 1(28); 1(38); 1(3D);1(30, Hearth 2); 2(3G)l maxilla (2A)l c1aw (3G); 3 claws (4)Fragments: 1(2A); 3(28); 4(3D. Hearth 3):1I(3E)

Animal

Tortcise

Puff adder

Snake (indet.)

Lizard (indet.)

Fish (indet.)

Crab (indet.)

Freshwater snails(indet.)

Land snails

Marine mollusks

Part

II limb bones (1); 15limb bcnes (ZA); 7limb

bones (ZB); 1 limb bone (2e); 9 limb bones(3A); 81imb bcnes (3B); 7 limb bones OC);

81imb bones (3D); 7 carapace fragments(3D,

Hearth 1); 4 limb bones (3D, Hearth 2); 56carapace fragments (3D, Hearth 3); 22 limbbones (3E); 3 carapace fragments (3E,

Hearth 4); 1 limb bone (JE, Hearth 5); 71imbbones (3F); lZlimb bones (3G); 21imb bones

(3G, Hearth 7); 8 Jimb bones (3H); 4 limbbones (31); 3 limb bones (3J); 39 limbbones (4)

I maxiJIa (3D)

1 rib (2C); vertebrae: 10(2A); 4(28); 18(3A);8(38): 34(3C): 4(30); 6(3D, Hearth 2); 13(3E); 6(3F): 35(3G): 5(3G. Hearth 7); 2(3H);4(4)J mandible (3D)

1 vertebra (3D); I vertebra (4)

Fragments: 3(1): 4(3A); 6(38); 1(3G,Hearth 6): Mandihles: 3(2A); 1(28); 32(3C):13(3\); 3(30, Hearth 2): 1(3D, Hearth 3);29(3E); 2(3E, Hearth 4); 2(3E, Hearth 5); 25(3F); 22(3G); 6(3G, Hearth 7): 6(3H); 5(3\):3(31): 6(4)

Fragments: 37(2A); 20(2B): 5(3A): 17(3B):9(3C); 3(3D); 3(30, Hearth 2): 2(30, Heartb3); 8(3E): 2(3F); 6(3G); 1(3G, Hearth 6);

1(31); 9(4)Fragments: 2(2A); 25(3A): 10(38): 20(3C):25(30): 31(3D, Hearth 2); 15(3D, Hearth3): 33(3E): 6(3E, Hearth 5): 23(3G); 3(3G,Hearth 7); 4(3H)2(3A)

Note: Stratigraphic units are given in parentheses.

Appendix: Tables 289

Table 27. Minimum Numbers of Individual Animals Represented in the LaterStone Age Layers al Wílton Large Rack Shelter, with Their Contributions to theDiet of the People

7<Yró ofTotal Weight

Total Weight PercentageAnimal N Lb Each (lb) Lb Kg of Total Diet

Antelope clase IBlue duiker 5 14 70.0 49.0 22.2 L14Sp. indet. 36 25 900.0 630.0 285.1 14.70Antelope class IlGray rhebuck I 55 55.0 38.5 17.4 .90Bushbuck 1 116 116.0 81.2 36.7 1.89Sp. indet. 24 118 2,832.0 1,982.4 897.0 46.24Antelope class UfWildebeest 1 526 526.0 368.2 166.6 8.59Sp. indet. 2 418 836.0 585.2 264.8 13.65Other mammalsRodents 35Insectivores 7Dassie 11 9 99.0 69.3 31.4 1.62Hace 9 5 45.0 31.5 14.3 .73Small camivore 2 3 6.0 4.2 1.9 .10Babeen 2 70 140.0 98.0 44.3 2.29Bushpig 2 ISO 360.0 252.0 114.0 5.88ReptilesTortoise 40 3 120.0 84.0 38.0 1.96Snake Indet. 13 1 13.0 9.1 4.1 .21Putf adder I 2 2.0 1.4 .6 .03Bi,dsOstrich eggBarn owl 2Sparrow 1FishSp. indet. 2 2 4.0 2.8 1.3 .07InvertebratesRiver crab 51River mussel 17Land snail 18Marine mollusk 1

Total 284 1,547 6,124.0 4,286.8 1,939.7 100.00

Table 28. Percentage Contríbutions of Various AnimalGroups (o the Diet of the Later Stone Age Inhabitants ofWilton Shelter

Animal Group

Antelope c1assIAntelope class IIAntelope class UIAntelope class IVOther mammalsAH other animals

ToIal

Percentage Representation

\5.849.022.3

10.62.3

100.0

290 Appendix: Tables

Table 29. Lengths of Bone Flakes from the Later StoneAge Levels al Fackeltráger Shelter, South-West Africa

Length (inches) N %

0-1 9()9 56.31-2 600 37.82-3 65 5.03-4 12 .84-5 2 .1

Tolal 1,588 100.0

Table 30. Animal Taxa from lhe Later Stone Age Layers al Fackeltráger Shelter, South-West Africa, and PartsUsed in Their Identification

,\11\

Raphicerus campestris, steenbok: 9 individualsMaxillary pieces. 3; mandible pieces, 2; calvaria frag­ment. 1: isolated teeth and tooth fragments, 5; distal tibiafragment, 1; pelvis pieces, 2

Antelope class 1, cí. Raphicerus: 6 IndlvidualsMandible píeces, 3; rocth fragmenta, 3; horn-core píeces,2; atlas vertebra. 1; proximal femur, 1; astragali, 4; prox­imal metapodials. 2; distaJ metapodials, 12; phalanges.?

Antídorcas marsupiails, springbok: I individualIsolated molar 1

Antelcpe class Il, cf. Antidorcas: 8 individualsMandible fragmento 1; tooth fragments, 9; vertebral frag­ment, 1; distal humerus, 1; proximal ulna, 1; distalradias, 1: astragalus, t; sesamoid, 1; distal metapodials,4; phalanges.Z

Antelope class 111, ef. Oryx: 2 individualsTooth fragmenta, 3; calvaria píece, 1; accesscry carpal. I

Procavta capensis, dassie: 17 adults, 1juveuileMaxillary pfeces, 9; mandible peces. 19; isolated molar.1; iscleted inctsor pieces, (4; proximal humen, 2; distalhumen. 16; proximal ulna, 1; pelvis pieces, 3; distalfemur, 1: femur sbeñ, l ; metapodial, 1

Lepus capensis, scrub hare: lO indivídualsMaxilla, 1; mandible pieces, 2; isolated teerh and toothfragmente, 16

Petromus typicus, rock rat: 6 individualsMaxillary pieces, 2; mandible pieces, 5

Etephantulu.s sp .. elephant sbrew: 3 individualsMandible pieces, 3

SmaJl carnivore, ce. mongoose: 2 individualsMandible fragments, 2

Large carnívore, cf. leopard: 1 individualMetapodial fragmenta, 2

Diceras bicornis, black rhino: 1 individualTooth Iragments, 5

Chelonian indet.. tortoise: 17 individualsCarapace and plastron fragments, 140; vertebrae. 4:fragmenta of pectoral and pelvic girdle, 8; limb bones, 47:miscellaneous fragments, 9

Varanus cf aíbíguloris, monitor Iizard: 10 IndlvidualsMaxillary fragment, I ~ mandible pieces, 6~ vestebrae. 11

Lizard. indet.: 2 individualsMaxillary fragment, 1: mandible píeces, 2

Snake, indet.: 2 individuaJsvenebraec z

er. Columba guinea, rock pigeon: 8 individualsLimb bones, IS; eggshell píece. 1

Birds indet., larger species: 7 individualsLimb bone preces, 11; phalanx, 1: eggshell piece, J

,~i Appendix: Tables 291

Table 31. Minimum Numbers ofIndividual Animals Represented in the Later Stone Age Layersal Fackeltráger Shelter, South-West África. with Their Percentage Contributions to the Diet ofthe People

70700fTotal weíght Percentage

Total Weight Percentage cf DietAnimal N Lb Each (lb) Lb Kg of Total Diet without Rhino

Antelope ctass 1Steenbok 9 29 261.0 182.7 82.7 5.67 10.9Indet. (cf.

steenbok 6 29 174.0 121.8 55.1 3.78 7.2Antelope class 11Springbok 78 78.0 54.6 24.7 1.69 3.2Indet. (cf.

springbok) 8 78 624.0 436.8 197.6 13.54 25.9Anteiope c1ass UIIndet. (ef.

gemsbok) 2 450 900.0 630.0 285.1 19.54 3704Other mammalsDassie 18 9 162.0 113.4 51.3 3.52 6.7Hare 10 5 50.0 35.0 15.8 1.09 2.1Rock rat 6 1 6.0 4.2 1.9 .13 .3Elephant shrew 3 1 3.0 2.1 1.0 .07 .1Mongoose 2 1 2.0 1.4 .6 .04 .1Large carnivore IBlack rhino 1 2,200 2,200.0 1,540.0 696.8 47.75ReptilesTortoise 17 3 51.0 35.7 16.2 1.11 2.1varanus 10 7 70.0 49.0 22.2 1.52 2.9Lizard 2 1 2.0 lA .06 .04Snake 2 1 2.0 lA .6 .04BirdsCe. rock pigeon 8 8.0 5.6 2.5 .17 .3Indet. (larger

spp.) 7 2 14.0 9.8 404 .30 .6Total 113 2,896 4,607.0 3,224.9 1,459.1 100.00 99.8

Table 32. Percentage Representation ofVarious AnimalGroups in the Diet of Inhabitants of the FackeltrágerShelter, Sourh-west Afriea

With Rhino Without Rhino

Animal Group N % N %

Antelope c1ass I 15 9.5 15 18.1Antelope class 11 9 15.2 9 29.1Antelope class III 2 19.5 2 3704Antelope class IVOther mammals 41 52.6 40 9.3Al! other animals 46 3.2 46 5.9

Total 113 100.0 112 99.8

Tab

le3

3.

Rep

rese

nla

tio

no

fY

ario

us

Mar

nm

alia

nO

rders

inth

eH

um

an

Fo

od

Rem

ain

sfr

om

Var

iou

sS

ou

thern

Afr

iean

Cav

eS

ites

Pom

ongw

eB

ushm

anR

ock

Wil

ton

Fac

kel

trág

erN

elso

nB

ayD

ieK

elde

rsJ

Kla

sies

t(t

his

stud

y)(t

his

stud

y}(t

his

stud

y)(t

his

stu

dy

)(K

le!n

1974

b)(K

lein

1975

b)(K

le!n

1975

b)

---

Mam

mal

ian

Ord

erN

%N

%N

%N

%N

%N

%N

%

Pri

mat

es,

oth

erth

anH

orno

--

--

22.

1-

-23

3.1

--

7l.

lC

arni

vora

41.

61

1.5

22.

13

5.2

385.

128

l.l

386.

1A

rtio

dact

yla

8535

.748

70.6

7275

.026

44.8

344

46.2

246

10.0

415

66.0

Per

isso

dac

tyla

104.

211

16.2

--

11.

73

A8

.39

1.4

Hy

raco

idea

114

47.9

68

.811

11.5

1831

.014

319

.286

3.5

518.

1L

ag

om

orp

ha

187.

62

2.9

99A

1017

.23

A20

08.

11

.2R

oden

tia

(lar

gefo

nn

s)6

2.5

--

101.

31,

866

75.8

304.

7P

hcli

dota

,T

ubul

iden

tata

,P

robo

scid

eaI

.5-

-3

A-

-3

.5P

inni

pedi

a-

--

-16

522

.226

1.1

65lO

AC

etae

ea-

--

-12

1.6

4.2

91.

4T

otal

238

100.

068

100.

096

100.

158

99.9

744

99.9

2,46

410

0.1

628

100.

0

--

De

Han

gen

Kla

sies

,O

ther

Sit

esS

cott

'sC

ave

And

ries

kraa

lI

Ela

nds

Bay

(Par

king

ton

and

Pogg

enpo

elR

edcl

iff

(Kle

in19

700)

(Kle

inan

dS

cott

1974

)(H

en

de

yan

dS

inge

r19

65)

(Par

king

ton

1976

)19

71)

(Kle

in,

n.d

.)

Mam

mal

ian

Ord

erN

%N

%N

%N

%N

%N

%

Pri

mat

es,

oth

erth

anH

orno

82.

1-

-I

1.5

3lA

I1.

18

1.0

Car

nivo

ra13

3.3

3(1

2)5

.910

14.5

188A

2(1

4)2.

335

4.5

Art

ioda

ctyl

a20

953

.542

82A

3144

.911

654

.214

15.7

500

63.9

Per

isso

dact

yla

143.

6-

-4

1.9

11.

I14

318

.3H

yra

coid

ea47

12.0

23.

99

13.0

2511

.664

71.9

719.

1L

ago

mo

rph

aI

.33

5.9

1623

.28

3.7

66.

716

2.1

Rod

enti

a(f

arge

fonn

s)6

\.52

2.9

3516

.3I

l.l

7.9

Pho

lido

ta,

Tub

ulid

enta

ta.

Pro

bosc

idea

I.3

31.

42

.2P

inni

pedi

a88

22.5

Cet

acea

41.

02

.9T

otal

391

100.

151

100.

169

100.

021

499

.889

99.9

782

100.

0

Not

e:M

inim

umnu

mbe

rso

fín

divi

dual

sha

vebe

enus

edin

allc

ases

exce

ptR

edcl

iíf

whe

renu

mbe

rsof

iden

tifi

edbo

nes

are

liste

d.

red

Table 34. Percentage Representation of Various Mam­malian Orders in Food Remains from Thirteen SouthemAfrican Cave Sites Known to Have Been Occupied byStone Age Peoples

Appendix: Tables 293

Table 35. Bone Remains Found in the Vicinity of SixSpotted Hyena Breeding Deos in the Timbavati Area,Eastem Transvaal

Table 36. Occurrence of Remains of Various Animal Taxa in Droppings, Regurgitations,and Bone Collections al Breeding Deos of Spotted Hyenas in the Timbavati Area, EastemTransvaal

Scats Regurgitations Bones al Deos

N

Taxon N % N % (pieces) %

MAMMALlA

Primatesrapto ursinas, baboon 4 .5 5 1.9

CarnívoraHerpestinae, mongoose .\

PerissodactylaEquus burchelli, zebra 36 4.5 9 3.4 18 14.6

ArtiodactylaGiraffa camelopardaíis, giraffe 304 38.3 43 16.3 20 16.3

Bovid size class 1Svlvicapra grímmia, duiker 17 2.\ 7 2.7Raphicerus campestrís, steenbok 2 .3 4 1.5

David size class 11Aepvceros meiampus, impaJa 292 36.8 \59 60.2 59 48.0

Bovid size class 111Cannochaetes taurínus, wildebeest lOS 13.6 \9 7.2 16 13.0Kobus ellipsiprymnus, waterbuck 8 1.0 I .4Tragelaphus strepsiceros. kudu \ .4 \O 8.1

RodentiaHvstrix africaeaustralis, porcupine 2 .3

LagomorphaLepus saxtilis, hare 2 .3 4 \.5

REPTIUATestudo pardalis, mountain tortoise .1 4 \.5Serpentee índet., snake 2 .8

AVES

Indet., unidentified bird 4 .5 2 .8UNIDENTlFIED TAXA 9 1.1

Total 794 100.0 264 100.1 123 100.0

Source: Data from Bearder (1977).

294 Appendix: Thbles 1Table 37. Food of Spotted Hyenas in the Serengeti and Ngorongoro Areas, as Determined by s: '?Direct Observation ,

Serengeti Ngorongoro .,;

Carcasses from Carcasses from Carcasses from Carcasses fromKills + Scavenging KiJls Only Kills + Scavenging Kills Only

.1Prey N % N % N % N %~:k -~

Wildebeest, adult 199 38.8 85 38.4 131 44.1 109 44.6 .LiWildebeest, calf 64 12.5 33 14.9 74 24.9 69 28.3 ,":1

, ~

Zebra, adult 55 10.7 19 8.6 36 12.2 29 11.1 .,~-

Zebra, foal 13 2.5 4 1.8 18 6.1 18 7.4Thomson's gazelle, adult 101 19.7 32 14.5 5 1.7 -~

Thomson's gazelle, fawn 49 9.6 30 13.6 10 3.4 7 2.9Granes gazelle, adult 4 .8 I .5 5 1.7 I .4Grant's gazelle, fawn 3 .6 3 1.4 1 .3 1 .4 ,Topi, adult 2 .4 I .5Topi, juvenile 3 .6 3 1.4Kongoni, adult I .2 I .5Waterbuck, adult I .2 I .5 .3 .4 1Eland. adult 2 .4 I .5

',~Buffalo, adu1t 3 .6 .3 .4Impala, adult 1 .2 .5 nWarthog 4 .8 .5 v iHare I .2 .5 .3 .4Springbare I .2 .5Domestic stock .3Porcupine .3 .4Bat-eared fax I .2 .5 .3 .4Golden jackal 2 .4 .5Lion 1 .3Hyena 5 1.7 .4Puff adder 1 .3 .4Ostrlch eggs .2 .5Termites 3 1.0 3 1.2Afterbirth 2 .4 1 .3

Total 513 100.2 221 100.6 297 99.8 244 99.9

Source : Data from Kruuk (1972).

Table 38. Animal. Found lo Have Been Killed bySpotted Hyenas in the Kruger Natíonal Park

Prey

ImpalaWaterbuckWildebeestKuduBushbuckBuffaloScaly anteaterReedbuckSharpe's grysbokWarthogZebra

Total

Source: Data from Pienaar (1969).

N

1102421195221111

187

%

58.812.811.210.22.7l.ll.l.5.5.5.5

99.9

Table 40.

Table 39. Lengths of Muscles Involved in the Closure of Crocuto Jaws, withPercentage Changes of Muscle Components at Two Jaw Openings

Jaws Half Open Jaws Fully OpenMuscle Jaws al RestComponent (length, cm) Length. cm % Cbange Length. cm % Change

Temporalis I 50 64 28.0 82 64.0Temporalis 2 76 97 27.6 116 52.6Temporalis 3 80 98 22.5 117 46.3Masseter I 60 72 20.0 76 26.7Masseter 2 57 68 19.3 75 31.6Masseter 3 70 78 11.4 80 14.3

Composition of Bone Accumulations Associated with Hyaena brunnea

A. Composition of 39 bcne pieces from tbe vicinity of fiveH. brunnea dens on the farm Tweeputkoppies

80S taurus, domestic OX: 2 adults, 2juveniles; 21 píeces2 skuUs, 3 mandibles, 1 scapula, 8 metacarpals, 1 meta­tarsus, 3 femurs, 3 tibiae

Tragelaphus strepsiceros, kudu: 2 adults; 3 pieces1 metacarpal, 2 hooves

Aepyceros melampus, impala: 2 aduus, 1juvenile: 8 pieces1 hom, 2 femurs, 5 metacarpal fragments

Sylvicapra grimmia, duiker: 1 adult; 1 piece1 femur

Phacochoerus aethiopicus, warthog: 3 adults: 4 pieces2 vertebrae, 1 rnetacarpal, 1 hoof

Canis mesomelas, black-backedjackal: 1 adult: 1 piecet skull and mandible

Source: Data from Skinner (1976).

Small carnívore, ?mongoose: I adult; 1 piece1 partial cranium

B. Ccmposition of6 bone pieces from aH. brunnea breedinglair on the farm Leeufcntein

Bos laurus, domestic ox: 1 aduJt; 1 pieceIsolated teeth

Phocochoerus aethiopicus, warthog: I adult: I pieceI mandible

Equus sp., horse or zebra: 1 adult; 1 piecelsolated teeth

Papio ursinus, baboon: I aduh; LpieceUnidentified bovid, 2 fragmenta

Table 41. Anaiysis of Bones CoUected around H. brunnea Dens in the Kaiabari Natíonal Park

Bos laurus, domestic ox: 1juvenile; 7 pieces7 pieces from a single craníum

Ovis/Capra sp., sbeep or goat: 1 adult; 1 pieceMandíble piece, I

Connochaetes laurinus, wildebeest: 2 adults; 3 piecesMaxillary piece, 1; left hom, 1; metacarpal, t

Alcelaphus buselaphus, hartebeest: I adult; 2 piecesMandible pieces, 2

Oryx gazella, gemsbok: 2 aduhs, 3 juveniles; 7 piecesMaxiUary pieces, 2; mandible píeces, 2; bom sheath, 2;haof sheaths, 2

Antidorcas marsupialis, springbok: 10 aduhs, 3juveniles; 41pieces

Maxillary pieces, 6; mandible pieces, 7; horn-core pieces,12: calvaría fragment, 1; atlas vertebra, 1; cervical verte­brae, 9; metacarpal, 1; proximal metatarsal, 1; hoofsheaths.B

Raphicerus campestrís, steenbok: 9 adulta, 2juveniles; 15 píecesCalvaria preces, sorne with hom-cores. 10; mandibJepieces, 3; bom sheath, 1: tenninal phalanx, 1

Sylvicapra grimmia, duiker: 1 adult, 1 juvenile; 3 piecesMandible pieces, 3

Feíís caracal, lynx: a aduna, l juvenüe: 6 piecesSkuUs. 4: maxillary pieces, 2

Canis mesomelas, black-backed jackal: 13 adults, 1juvenile; 13pieces

Skulls with articulated mandibles, 3: maxillary pieces, 3;mandible pieces, 4; calvaria pieces, S

Olocyon megalolis, bat-eared fox: 13 OOults; 13 piecesCalvariae in various degrees ofcompleteness, 12; articu·laled forefoot, 1

Prole/es crislalus, aardwolf: I adu1t; 3 piecesCalvaria, 1; mandible pieces, 2

Mellivora capensís, ratel: 2 adults; 2 piecesSkull whh articulated mandible, 1; mandible piece, 1

Orycteropus afer, antbear: 1 adult; 2 piecesCalvaria, 1; innominate piece, 1

Hyslrix afrícaeaustmíis, pcrcupine: 1 adult; 2 piecesQuills, 2; probably not food remains.

Struthio comelus, ostrich: 2 adults, 1eggshell; 8 piecesTibiotarsal píece. 1; shaft piece. 1; claw sheath, 1;eggsheU pieces, 5

Antelope class 1: 18 piecesCervical vertebrae, 3; thoracic vertebra. 1; lumbar verte­brae, 4; specimen of articulated thoracic and lumbar verte­brae, 1; pelvic pieces, 4; femur, 1; proximal tibia, 1; distaltibia, 2; hoof sheath, 1

Antelope c1ass 11: 14 piecesMandibJe piece, 1; alias, 1; axis, 1; cervical vertebra, 1;thoracic vertebra, 1; rib, 1; scapula pieces, 2; innominatepiece, 1; distal humeros, 1; proximal radius, 1; distalradius, 1; distal metacarpal, I: metatarsal.I

Antelope class 111: 24 piecesCervical vertebrae, 4; thoracic vertebra, 1; scapulapieces, 2; pelvis pieces, 5; distal humeros, 1; proximalradius, 1; distal radius, 1; distal femur, 1; proximal tibia,1; distal tibia, 1; proximal metacarpal. 1; metatarsalarticulated with tarsals 1; hoof sheaths, 4

Camivore indet.: 24 piecesCalvaría piece. 1; mandible piece, J; atlas, 1; cervicalvertebrae, 2; thoracic vertebra. 1; lumbar vertebra, 1; aroticulated vertebrae, 13; pelvis piece. 1; proximal humeros,t; tibia, 1; c1aw shea!h, 1

Indeterminate fragments: 6 piecesBone flakes: 21 pieces

296 Appendix: Tables

Table 42. Particulars of Remains frorn Suswa Locality 36E, the "Suswa Lair" ..Papio anubis, anubis baboonCraniums in varying degrees of completeness52. Calvaria of a subadult d. without mandible, complete ex­

cept for the following damage: right orbit extensivelydamaged with tooth marks around ir; left brow ridgedamaged with a punctate mark; puncture in left parietal;depressed fracture in right occipital; three punctatemarks in right auditory region; right zygoma missing andleft damaged; two irregular boles in left orbit and tem­poral.

52 + 54. Cranium of adult d, full dentition except for rightM". Damage to pterygoids; both zygomata missing; twopunctures in right temporal, one in left temporal. Mandi­ble complete except for chewing on right coronoid pro­cess and right angle of the jaw.

55 + 56. Adult o palatal piece, right C. M2-3,left M2-3.Almost complete mandible except for damage tocoronoid processes: canines and 2 incisors missing.

59 + 60. Adult C} calvaria without anterior teeth; pterygoidschewed away; puncture at base of right zygoma and onein midline ofpalate. Mandible with full dentition exceptfor right 12; complete except for right ascending ramus,which has been chewed off; tooth marks present.

63 + 64. Cranium of adult 2. Calvaria without incisors; dam­age to ventral surface of snout; both zygomata missing:porcupine gnawing on borders of orbits. Mandible with­out anterior teeth: ascending rami damaged; weathered.

61 + 62. Cranium of adult <}. Complete calvaria except foranterior teeth; pcrcupme gnawing abo ve orbits; damageto sides of orbits and snout. Mandible without ascendingrami; tooth marks and porcupine gnawing presento

65. Adult C} braincase without snout: extensively damagedand showing porcupine gnawing.

57 + 58. Adult 2 cranium. Calvaria without anterior teeth;zygomata missing; pcrcupine gnawing 00 both orbits andright prernaxilla. Mandible without anterior teeth; dam­age to left coronoid process and right ascending ramus;tooth marks presento

41. Juvenile o mandibular fragmento Canines and premolarsunerupted; incisors and 1 deciduous molar present. CIearchewing damage to both rami and a punctate mark onlingual side of left ramus.

44. Palate from a juvenile ofundetermined sex. Canínes un­erupted; P3_M2 present bilaterally; tbe calvaria hasclearly been chewed away from aboye.

Maxillary pieces39. Right maxilla, P3_M2.40. Left maxilla, C-M3.73,74,75,77,78, and 122. Small maxillary píeces.Mandible pieces48. Mandible with M3 not fuUy erupted; porcupine gnawing

alcng right lower margino49. Adult mandible without anterior teeth or ascending ramio

Sorne possible porcupine gnawing.50.· Old adult mandible without anterior teeth and showing

damage te ascending ramioVarlous cranial pieces42. Ríght parietal with sutured fragments ofoccipital and

frontal. Two hales, apparently caused by canine teeth,spaced 22 mm aparto

Source: Data recorded in Nairobi, June 1970, by C. K. Brain.

43,47,76,92.93. Miscellaneous cranial pieces.Isolated teeth81-86. Isolated canines.87-91. Isolated postcanine teeth.Vertebrae102. Atlas vertebra.99, 101, 104. damaged thoracic vertebrae.100. Damaged lumbar vertebra.103. Damaged sacrum.106. Damaged caudal vertebra.80.82.105,107.119. Undamaged caudal venebrae.Scapulae95, 96, 98. Scapulae showing exrensive damage te the blades.Pelvis pieces94,97,112.113,114.115.116,117. Pelvls pieces extensively

chewed, sorne showing punctate tooth marks.Rib piece150. A single damaged fragment.Forelimb15. Right humeros, complete.16. Complete right ulna. and 17, complete right radius; both

articúlate with 15.IS. Complete left humeros, probably same indívidual as 15.4,5,8,9,24,34, 165. Distal humeri, extensively chewed with

shafts often showing spiral fractures.25,35.36. Humeros shafts, extensively chewed at both ends.19, 166, 167. 168, 169. Radii in varying degrees of complete-

ness, extensively chewed.26,30,170,171. Pieces of ulna, extensively damaged.152, 153, 154, 161. Metacarpals, almost undamaged.155-60, 162. Complete phalanges.Hind limb176. Femur, almost complete but showmg porcupine gnawing.10,46,70, 172, 173, 174. Proximal femur pieces , extensively

damaged and showíng tooth marks.2,6,7,12,13,20.23,32,175. Femur sheft píeces showing

extensive camivore damage.3.29,45, 164. Tibia shaft pieces, extensively damaged.163. Distal tibia piece showing tootb marks.108. Undamaged astragalus.109--11. Damagedcalcanei.t 18. Articulated astragalus and calcaneus. showing tooth

marks.11.14.21.22,28.31,33,37,67. Lcng-bone shaft pieces

showing extensive damage.Panthem pardus, leopardl. Complete braincase. without snout.Oreotragus oreotragus, Klipspringer51. Almost complete calvaria ofan adult O. No mandible.38. Pelvis bone showing chewing around the margins and 3

clearly defined tooth boles.Birds71, 72. Long-bone fragments from large and small unidentified

birds.Unidentified bone fragments: 13 pieces.Bone f1akes: 22 pieces. 1-2 cm, 5; 2-3 cm, 6; 3-4 cm, 2; 4-5 cm,

1; 5-6 cm, 1; 6-7 cm, 6; 8--9 cm, 1.

Appendix: Tables 297

Table 43. Rernains from the Portsmut Leopard Breeding Lair

Papío ursinus. baboon: 1 adult c3; 20 pieces.Almost complete cranium and mandible, l: left femurshaft. 1; metapodials , 9; phalanges. 8

Panthera pardus, !eopard: I adult; 50 piecesPartial cranium. 1; isolated teeth, 9; atlas vertebra, 1;caudal vertebrae, 6; left femur, 1; ulnae, 2; right radíus, 1;left astragalus, 1; right ca1caneus, 1; metapodials, 8;phalanges, 19

Oreotragus oreotragus, klipspringer: 2 juveniles; 26 piecesMandible preces, 2; isolated teeth, 3; cervical vertebrae,4; thoracic vertebrae, 3; lumbar vertebra, 1; left scapula,1; articulated left forefoot, 1; ulna piece. 1; tibia piece, 1;astragalus, 1; metacarpal, 1: tarsal, 1; phalanges,6

Antelope c1ass 11: l adult, 1 juvenile; 10 piecesScapula piece. 1; pelvis piece. 1; femur piece, 1; rnetapo­dial pieces, 4; tarsal, 1; phalanges, 2

Antelope class III: 1 adult ; 2 piecesTibia piece. 1; calcaneus. I

Equus zebra, mountain zebra: 2 adults; 13 piecesRight humen. 2; tibiae, 2; cervical vertebrae, 2; thoracicvertebrae, 3; lumbar vertebrae, 3 articulated; sacrum, 1

Procavia cupensis, dassíe: 1 adult, 3 juveniles; 15 piecesMaxillae. 2; mandibJe pieces, 3; cervical vertebra. 1;scapulae, 2; pelvis pieces, 4; femur shaft, 1; radii, 2

Chelonia indet., tortoise: 1 individual; 1I piecesCarapace and p!astron pieces, 10; limb bone, 1

lndeterminate fragments: 45 piecesCulturalobjects: 17 pieces

Grooved stones, 2; grooved pottery piece, 1; stone pes­tles, 2; potsherd, 1; hematíte pieces, 6; quartz píecessbowing ñake scars, 5

Table 44. Remains from the Hakos River Leopard Breeding Lair

Pupio ursinus, baboon: I adulr; 1 pieceMetacarpal, I

Panthera pardos, leopard: 1 adult. 1 juvenile; 26 piecesAdu!t remains: humerus, 1; calcaneus, 1; astragalus, I

Juvenile remains: cervical vertebrae, 2; thoracic verte­brae, 5; lumbar vertebrae, 8; caudal vertebrae, 2; tibiashaft, 1; distal femoral epiphysis, 1; calcaneus, 1; as­tragalus, 1; innominate pieces. 2

Lvcaon píctus, wild dog: 1 adult; 39 piecesIsolated teeth, 4; atlas vertebra. 1; thoracic vertebrae. 4;caudal vertebra, 1; ribs, 4; inncminate, 1; tarsals, 2;metapodials. 7; phalanges, 15

OvislCapru sp., sheep or goat: I adult; 4 piecesScapulae, 2; tenninal phalanges, 2

80S sp., cattle: 1 adult; 7 piecesArticulated right hind foot, l : metatarsal, 1; terminalphalanges, 3; hoof sheaths, 2

Oreotragus oreotragus, klipspringer: l adult: 4 piecesPalate, 1; isclated incisor, 1; ulna, 1; tibia, 1

Trageíaphus strepsiceros. kudu: 2 juveniles; 8 piecesMaxillary pieces, 3; mandible pieces, 2; tooth fragments, 3

Table 45. Bones Found in the Quartzberg Leopard Lair

A. Probable Porcupine-CoUected Component

Oryx gazetla. gemsbok: 1 subadult; 1 pieceCranium without mandible, 1

Antelope class 11: 1 adult; 7 piecesThoracic vertebra, 1; left humerus, 1; phalanges, 4; hoofsheath, 1

Antelope class Hl: 3 aduhs, 3 juveniles; 96 piecesCalvaria pieces, 3; mandible pieces, 2; tooth fragments, 4;vertebral pleces: atlas, 2; cervical, 5; thoracic, 4; lumbar,6; ribs. 3; scapula, 1; pelvis pieces. 2: humen, 6; radii, 2;ulnae, 2; femur pteces, 6; tibia pieces, 6; metapodialpieces, 3; calcanei, 5; carpals/tarsals, 9; sesamoid, 1;phalaages. 24 pieces

Equus cf. zebra. mountain zebra: 1subadult , 1juvenile; 6 piecesTibia. 1; humerus, J; sesamoid, 1; hoof pieces, 3

Procovia capensís, dassie: 2 subadults: 4 piecesDamaged skull and rnandlble, 1; isolated incisor, 1; ribs, 2

Chelonia Indet., tortoise: 1 individual; 6 piecesPlastrón piece, 1: vertebra, 1; girdle fragmente. 4

Bone f1akes: 13 piecesIndeterminate fragments: 117

B. Probable Leopard-Cotlected Component

Gemsbok (Oryx gazella¡Cranial pieces

Zebra (Equus zebra)Cranial piecesPostcranial pieces

OX (80S sp.)Cranial piece

Baboon (Papio ursinus)Cranial piece (mandible)

Indet. bovid postcranial precesIndet. bone piecesBone ñakes

Total

3

23

11062011

147

Klipspringer (Oreotragus oreotragus)Cranial piecesPostcraniaJ pieces (sorne articulated)

Steenbok (Raphicerus campestris¡Cranial plecesPostcranial pieces

Gemsbok(Oryx gazeíía¡ calfPostcranial pieces (articulated)

Domestic calf(8os sp.JPostcranial piece

Zebra (Equus zebra) coltPostcranial piece

Baboon tPaplo urstnus¡Cranial piecePostcranial pieces

Dassie (Procavia capensís¡Cranial piecesPostcraníal piecesTotalOverall total

632

12

3

12

123

64211

Table 46. Minimum Number of Individual Animals Represented by the Bones in the Quartzberg Leopard Lair

298 Appendix: Tables

Table 47. Bones from Various Zones in the Quartz­berg Leopard Lair

221I1116

15

80.223.4

% Gnawed Bones

14764

Total

2949

11815

Klipspringer maleKlipspringer femaleSteenbok mateGemsbok calfDornestic calfMountain zebra coltBaboon femaleDassie

Total

No. Gnawed No. Ungnawed

B. Probable Leopard-Collected Component

11I1I16

Incidence of pcrcupine gnawing 00 bones

A. Probable Porcupine-Collected Component

Probable pcrcuplne-collected compcnentProbable leopard-ccllected component

Gemsbok adultGemsbok calfDomestic calfMountain zebra adultMountain zebra ccltBaboon

Total

PresumedPorcupine-Collected

Component

PresumedLeopard-Collected

Component

Cave Zone N % N %

Light zone 5 3.4 23 35.9Twilight zone 57 38.8 15 23.4Dark zone 85 51.8 26 40.6

Total 147 100.0 64 99.9

,Appendix: Tables 299

Table 48. Prey of Leopards as Reñected by Observed KilIs

Kafue Naticnal Park Serengetl(Mitchell, Shenron, (Kruuk and Turner Serengeti Matopos Kruger National Park

and Uys I96S) 1967) (Schaller 1972) (Smith 1977) (Pienaar 1969)

Prey Species N % N % N % N % N %

Antelope c1ass 1Gray duiker 1L 115 4 IO.S 148 2.0Grysbok 4 4.2 11 .4Oribi 1 1.1 2Steenbok 2.6 66 .9Klipspringer 2.6 lS .SAntelope c1ass 11Reedbuck 19 19.8 6 10.9 19 11.6 S 11.2 222 1.0Mountain reedbuck S .1Puku IS IS.6Impala 8 8.1 9 16.1 t2 11.6 S.SII 73.8Bushbuck 4 4.2 1 1.8 465 6.2Thomson'sgazelle lS 27.3 104 63.4Grant's gazelle 2 3.6 10 6.1Antelope class IIIHartebeest 9 9.4 2 1.2Nyala 41 .6Kudu 3 3.1 2.6 208 2.8Lechwe 3 3.1Wildebeest I 1.0 S 9.0 11 6.7 3 7.9 8S 1.1Topi 1 1.8 3 1.8Waterbuck I .6 I 2.6 287 3.8Sable 4 IO.S 10 .1Roan 2Tsessebe 2.6 14 .2Antelope class IVEland 2 S.3 9 .1Buffala 4 .1Other mammalsWarthog 2 2.1 I .6 IS3 2.1Zebra 4 7.2 2 1.2 2.6 7S 1.0Hyrax 1 1.8 2.6Ant bear 9 .1Baboon 2 2.1 2 3.6 .6 S8 .8Vervet monkey 3 1.1 3Bat-eared fox 2 1.2Golden jackal 1 .6Black-backedjackal 1.8 1 .6Leopard 2Cheetah 1.8 2Serval I 1.0 2 I.2Genet 1 1.0Civet 1 1.0 4 .1Hare 1 1.0 1Springhare 2 2.1 1.8Porcupine I 1.0 S .1Cane rat 1 1.0 IBirdsEuropean stork 2 1.6 4 2.4Secretary bird 1 1.8Guineafowl 1 1.8Vulture 1 1.8Ostrich 2.6 6 .1

ReptilesPython 1.8FishCatfish 1 1.0

Total 96 99.7 SS 99.S 164 99.8 38 99.8 7,46S 100.0

300 Appendix: Tablea

Table 49. Broad Categories of Leopard Prey as Percentage of Total, Basedon Observed KiIIs

Kafue National Park Serengeti(Mitchell, Sbenton. (Kruuk and Serengeti Matopos Kruger National Park

Prey Species and Uys 1%5) Tumer 1%7) (Schaller 1972) (Smith 1977) (Pienaar 1969) Mean

Antelope c1ass 1 18.8 15.7 3.8 7.7Antelcpe c1ass 11 47.9 59.9 81.1 44.8 83.1 63.4Antelope class JII 16.6 10.8 10.3 26.2 8.6 14.5Antelope class IV 5.3 .2 1.1Primates 5.2 3.6 .6 .8 2.0Camivores 3.0 3.6 3.6 .1 2.1Other mammals 7.2 10.8 1.8 5.2 3.3 5.7Birds, reptiles, físh 1.0 10.8 2.4 2.6 .1 3.4

Total 99.7 99.5 99.8 99.8 100.0 99.9

Table 50. Food Items Represented in Leopard Scatsfrom the Matopos Area, Zimbabwe

Prey Species

DassieKlipspringerHareDuikerRats and miceSableCene ratBirdBaboonImpalaReedbuckSpringhareSteenbokWildebeestElandLeopardGenetSnakeBushpigPorcupineHedgehogScorpionBushbuckMongooseUnidentified

Total

Source: Data from Smith (1971).

Percentage

46.110.28.47.14.84.03.22.81.91.71.71.3.8.s.5.4.4.3.3.6.3.3.2.2

1.899.8

Appendix: Tables 30\

Table 51. Distances over Which Leopards HaveDragged Their Prey in the Matopos National Park,1971-74

Distance Moved(in meters)

Wet Season Dry Season

Prey Species N Mean Range N Mean Range

Impala 6 170 (SO- 300) 3 25 (O- 50)Steenbok I 200Sable antelope 2 650 (300-1,000)Common

duiker 300 5 125 (0-400)Wildeheest 150 1 50Eland 2 200 (0-400)

Reedbuck 4 175 (0-400)Tsessebe I 40Klipspringer 1 50

Source: Data from Smith (1977).

Table 52. Food Remains of Black Eagles from belowNests and Perches in the Matopo Hills, Rhodesia, 1975

Table 53. Food RemainsofBlack Eagles from a FeedingPerch at Wonderbroom, Pretoria, 1969

Precavía capensis Procavia capensís, dassieAdult craniums with articulated mandibles 2 Adult craniums with articulated mandibles 7Adult craniums without mandibles 4 Adult maxillary piece 1Mandibles 6 Adult rnandible 1lsolated upper inciscrs 3 Subadult craniums with articulated mandibles 4Total t5 Subadult cranium without mandible I

Juvenile rnaxilla tHeterohyrax brucei Juvenile rnandible pieces 4

Adult and subadult craniums, without mandibles 23 Calvaría pieces 2Adult and subadult mandibles 26 Isolated upper incisors 4Mandibular rami, left \O Vertebrae 8Mandibular rami, right 14 Sacrum 1Juvenile cranium, without bralncase I Rib piece 1Calvaria pieces 9 Scapula 1Maxillary piece I Femur tIsolated upper incisors 3 Articulated femur and left innominate 1Total 87 Complete pelvises 3

Innominate píeces 3Procavía or Heterohyrax Total 44

Articulated right humeros and radius/ulna 1Articulated right tibia with calcaneus and astragalus 1 Lagornorph, probably Lepus sp.Pelvis pieces 2 Mandible pieces 2Left femur, partly digested from regurgitation I Tibia, right, one with articulated foot 2Total 5 Total 4

Ktnvxís beíílana, hinged tortoise Bird, francolin sizePlastron pieces 3 Tibiotarsus, right 2Loase scale 1 Sacra 2Total 4 Humeros IOverall total 111 Total 5

Testudo pardaJis, mountain tortoiseCarapace pieces 3Overall total 56

302 Appendix: Tables

Table 55. Species Represented in the Bone Accumulations from the "HumanSite." Andrieskraal 1, and the "Porcupine Lair," Andrieskraal2

Table 54. Food Remains of Black Eagles from below aFeeding Percb, Portsmut Farm, South-West África, 22June 1969

s

13

24666

2

2

10

I3

56

Andrieskraal 2

42581393

1131

95130

29

9

I32

211

2124

23

16

20

I6

108

Mínimum Number of Individuals

Andrieskraal I

Procavía capensis, dassieAdult craniums with articulated mandiblesAdult maxil1aeAdult maxillary fragmenteAdult mandible piecesSubadult cranium with articulated mandibleJuvenile mandible piecesPelvis piecesTibia pieces

Limb bcnes and bone fragments, partlydigested, from regurgitatícns

TotalLagomorph, probably Lepus sp ,

Distal humerosOverall total

Taxon

MAMMALIA

PrimatesPaplo ursinus, chacma baboon

CarnívoraGenetta genetta, small spotted genetHerpes/es putverulentis, gray mongoose?Herpestes sp.Mellivora capensis, honey badgerFelis tvbíca, wild catPanthera ?pardus, IeopardCanis mesomelas, black-backed jackal

ProboscideaLoxadonta africana, ACriean elephant

HyraccideaProcavía capensis, dassie

PerissodactylaDlceros bicomís, black rhino

Equus asínus, domestic donkeyArtiodactyía

Hlppopotamus amphibiusPotamochoerus porcusBos sp., domestic OX

Capra sp., domestic goatTragelaphus scriptus, bushbuckTragelaphus strepsíceros, kuduCephalophus monticola, blue duilterSylvicapra grimmia, gray duikerRaphicerus campestris or meíanotis, steenbok

or grysbokRodentia

Hystrix africaeaustmíís, porcupineIndet. small rodents

LagcmorphaLepus sp., hare

REPTIUA

Tortoise or nrrtleAVES

Struthio australis. ostrichIndet. smaller birdsTotal

Source: Data from Hendey and Singer (1965).

" Appendix: Tables 303

Table 56. Analysis of Objects from the Nossob Porcupine Lair

Collection Collectlon

Object 1956 1968 Total Object 1956 1%8 Total

Gemsbok (Oryx gauJla) Lion (Panthera leo)Cranial pieces 7 2 9 Cranial piecesHoms 21 7 28 Hunting dog (Lycaon pictus)

Springbok (Antidorcas marsupíalis¡ Cranial pieces 3 3 6Cranial pieces 12 5 17 Mongoose (cf Cynictis sp.)Hcms 46 26 72 Craniaí pieces 2 2

Wildebeest tConnochaetes taurinus) Skunk (lctonyx striatus¡Horos 5 3 8 Cranial pieces 1 1

Hartcbeest (Alcefaphus buselaphus) Indet. camivore pcstcranial píeces 2 5 7Cranial pieces 2 2 Porcupine (Hystrix ofricaeaostratis¡Homs 11 6 17 Cranial pieces

Steenbok {Raphicerus campestris) Tortoise (Testudo occulifera)Homs 2 4 6 Carapace pieces and scales 56 3 59

Duiker (Sy/vicapra grimmia) Ostrich (Struthio came/us)Cranial pieces 2 2 Postcranial pteces 7 4 11

Goat Eggshell pieces 7 7Horns 2 3 Large bird (gen. et sp. Indet.)

CattJe Postcranial piece 1 1Homs 5 5 Miscellaneous bone fragments 364 31 395

Indet. bovid postcranial pieces 405 243 648 Bone ftakes 277 28 305Equid Pieces oí wood 68 3 71

Cranial pieces 1 1 Metal objects 13 4 17Postcranial pieces 6 4 10 Total 1,328 384 1,712

Table 57. Minimum Numbers of Individual AnimalsRepresented in the Bone Accumulation from the NossobPorcupine Lair

Collection

Animal

SpringbokGemsbokHartebeestWildebeestSteenbokDuikerCatñeGoatHorseHunting dogMongooseSkunkLionPorcupineTortoiseOstrichBird

Total

1956

2511632231212O1161O

67

1968

154323OO123O1OO311

39

Total

4015955232442111921

106

304 Appendix: Tables 1Table 58. Kalahari Park Area Census 1973174 (De Graaff el al.)

Annual PercentageSpecies August November February April Average ofTotal

Springbok 17,560 18,228 24,041 12,894 18,181 40.2Gemsbok 11.183 7,384 16,073 15,915 12,693 27.9Hartebeest 14,178 1,408 5,200 5.011 6,449 14.2Wildebeest 6,378 2,230 4,067 2,522 3,799 8.4Eland 1,433 611 6,569 1,235 2,462 5.4Steenbok 1,402 1,073 1,051 1,645 1,293 2.9Duiker 262 381 710 497 463 1.0

Total 52,396 31,315 57,711 39,719 45,286 100.0

Table 59. Bovid Skeletal Parts, 1968 Collection, Nos- Table 60. Bovid Skeletal Parts in the Bone Accumula-sob Porcupine Lair tion from the Nossob Porcupine Lair

Antelope Size Class Collection

Part Il III Total Part 1956 1968 Total

Hom pieces O 28 17 45 CraniumOther cranial pieces O 7 10 17 Complete skulls O O OVertebrae Maxillary pieces 11 4 15

Atlas O 2 6 8 Mandible pieces 19 5 24Axis O 5 9 14 Isolated teeth 27 O 27Cervical, 3-7 O I7 19 36 Other pieces 16 9 25Thoracic 3 7 6 16 Hom pieces 112 45 157Lumbar 2 20 18 40 VertebraeSacral O 2 2 4 Atlas I7 8 25Caudal O O O O Axis 9 14 23

Rib pieces O 5 O 5 Cervical, J....7 40 36 76Scapula pieces O 3 16 19 Thoracic 40 13 53Pelvis pieces 3 7 20 30 Lumbar 36 40 76Humeros Sacral 3 4 7

Complete O 1 I 2 Caudal 1 O 1Proximal O 1 2 3 Ribs 38 5 43Dis.tal O 2 2 4 Scapula pieces 24 18 42Shaft O 1 O 1 Pelvis pieces 36 30 66

Radius and ulna HumerosComplete O 2 3 5 Complete 2 2 4Proximal O 2 3 5 Proximal 2 3 5Distal O O O O Distal 12 4 16Shaft 1 O O 1 Shaft ) 1 4

Carpal bones O O 1 1 Radius and ulnaMetacarpal pieces O 1 2 3 Complete O 5 5Femur Proximal 13 5 18

Complete O O O O Distal 9 O 9Proximal O 1 10 11 Shaft 1 1 2Distal O O 8 8 Carpal bones 12 5 17Shaft O I 2 3 Metacarpal pieces 7 3 10

Tibia FemurComplete O 1 4 5 Complete 1 O 1PrOximal 1 2 4 7 Proximal 5 tt 16Distal O O 4 4 Distal 6 8 14Shaft 1 O O 1 Shaft 7 3 10

Calcaneus O 1 1 2 TibiaAstragalus 1 1 2 4 Complete ) 5 8Metatarsal pieces O 3 1 4 Proximal 6 7 13Metapodial pieces 2 1 7 10 Distal 9 4 13Phalanges 1 8 ) 12 Shaft 3 1 4

Total 15 132 183 330 Calcaneus 16 2 18Astragalus 15 4 19Other tarsal bones 9 O 9Metatarsal pieces 10 4 14Metapodial pieces 14 JO 24Phalanges 48 12 60Sesamoíds 1 O 1

Total 643 331 974

Appendix: Tables 305

Table 61. Percentage Survival of Bovid Skeletal Parts (Nossob sample: 81bovid individuais)

Survival ofHottentot Goal

Number Original Survival Bone SamplePan Found Number (%) (%)

Hom pieces 157 162 %.9 94.7Pelvis preces 66 162 40.7 9.0Atlas vertebrae 25 81 30.9 6.3Axis vertebrae 23 81 28.4 7.4Scapula pieces 42 162 25.9 9.2Cervical vertebrae, 3-7 76 4Q5 18.8 1.2Maxillae 15 81 18.5 26.3Metatarsal, proximal 26 162 16.0 10.3Metatarsal, distal 26 162 16.0 5.3Lumbar vertebrae 76 486 15.6 2.7Half-mandibles 24 162 14.8 30.7Radius and ulna, proximal 23 162 14.2 17.1Metacarpal, proximal 22 162 13.6 8.4Metacarpal, distal 22 162 13.6 6.0Tibia, proximal 21 162 13.0 3.4Tibia, distal 21 162 13.0 19.0Humerus, distal 20 162 12.3 21.5Astragalus 19 162 11.7 4.2Calcaneus 18 162 11.1 3.7Fernur, proximal 17 162 10.4 4.7Pemur, distal 15 162 9.3 2.4Sacral vertebrae 7 81 8.6 .5Radius and ulna, distal 14 162 8.6 5.8Phalanges 60 972 6.2 .9Humeros. proximal 9 162 5.5 OThoracic vertebrae 53 1,053 5.0 .9Rjbs 43 2,106 2.0 3.4Caudal vertebrae 1 810 .1 O

Table 62. Weights of Bones from the Nossob Porcupine Lair

1956Collection 1968 Collection

Weight N N N N OveraJl(g) Ungnawed Gnawed Total Ungnawed Gnawed Total Total

O- 10 388 283 671 31 30 61 73210- 20 45 100 145 19 34 53 19820- 30 21 70 91 11 19 30 12130- 4Q 16 42 58 4 18 22 8040- 50 16 29 45 3 15 18 63SO- 60 6 20 26 6 15 21 4760-- 70 6 35 41 5 20 25 6670- 80 5 15 20 4 10 14 3480- 90 4 16 20 4 12 16 3690-100 4 13 17 2 12 14 31

100-150 14 59 73 9 41 50 123150-200 13 31 44 4 20 24 68200-250 2 16 18 2 9 11 29250-300 5 13 18 3 8 11 29300-350 3 5 8 4 4 12350-400 5 3 8 2 2 10400-450 1 6 7 2 2 9450-500 1 2 3 3500-550 1 3 4 4550-<>00 1 3 4 56(){H;50 3 3 4650-700 1 2 2700-750 2 2 2

Total 558 770 1,328 107 273 380 1,708

306 Appendix: Thbles

Table 63. Lengths of Bones from the Nossob Porcupine Lair~

Length 1956 Collection 1968 CoJlection

N N N N OveraJl(cm) (inches) Ungnawed Gnawed Total Ungnawed Gnawed Total Total

<l- 2.5 <l-1 59 14 73 2 3 5 78 l2.5- 5.1 1- 2 216 185 401 19 16 35 4365.1- 7.6 2- 3 113 134 247 15 37 52 299

,~¿

7.&- 10.2 3-4 SS 121 176 13 47 60 236 ,10.2- 12.7 4-5 26 71 97 9 41 SO 147

,.<

12.7- 15.2 5-6 21 65 86 15 34 49 135 ....'

15.2- 17.8 &-7 11 34 45 6 20 26 7117.8- 20.3 7- 8 9 30 39 6 14 20 5920.3- 22.9 8-9 13 30 43 3 13 16 5922.9- 25.4 9-10 7 19 26 8 17 25 5125.4- 27.9 I<l-II 3 10 13 4 7 11 2427.9- 30.5 11-12 6 21 27 2 6 8 3530.5- 33.0 12-13 4 4 8 8 8 1633.<l- 35.6 13-14 1 8 9 7 8 1735.&- 38.1 14-15 1 2 3 1 1 438.1- 40.6 15-16 1 3 4 44O.&- 43.2 1&-17 1 4 5 543.2- 45.7 17-18 1 1 2 245.7- 48.3 18-19 2 2 248.3- 50.8 19-20 2 2 2 450.8- 53.3 2<l-2153.3- 55.9 21-22 1 1 155.9- 58.4 22-23 1 1 158.4- 61.0 23-24 2 2 4 2 2 661.<l- 63.5 24-25 1 1 1 1 263.5- 66.0 25-26 1 166.<l- 68.6 2&-27 1 268.&- 71.1 27-2871.1- 73.7 28-2973.7- 76.2 29-3076.2- 78.7 3<l-3178.7- 81.3 31-3281.3- 83.8 32-33 4 4 483.8- 86.4 33-34 1 186.4- 88.9 34-3588.9- 91.4 35-36 1 1 191.4- 94.0 3&-37 1 1 194.0- 96.5 37-38 2 2 296.5- 99.1 38-39 1 1 I99.1-101.6 39-40

Total 558 770 1,328 107 273 380 1,708

Table 64. Percentage Abundance of Gnawed Bones and Other Objects inVarious Weight Classes from the Nossob Porcupine Lair

Woighl N % N % % Bones Gnawed(g) (bonos) ofTotal (gnawed bones) cf Total in This Weight Class

<l- 50 1,194 69.9 640 37.5 53.65<l-100 214 12.5 168 9.8 78.5

10<l-150 123 7.2 100 6.4 81.315<l-200 68 4.0 SI 3.0 75.020<l-250 29 1.7 25 1.4 82.825<l-300 29 1.7 21 1.3 72.430<l-350 12 .7 9 .5 75.0350-400 10 .6 S .3 50.0400-450 9 .5 8 .5 88.945<l-500 3 .2 2 .1 66.750<l-550 4 .2 3 .2 75.0550-600 5 .3 4 .2 80.0600-650 4 .2 4 .2 100.065<l-700 2 .1 1 .1 50.070<l-750 2 .1 2 .1 100.0

Total 1,708 99.9 1,043 61.1

Appendix: Tables 307

Table 65. Percentage Abundanee ofGnawed Bones of Various Lengths frcm theNossob Porcupine Lair

Length N % GnawedN % (gnawed % Dones in This

(cm) (inehes) (bones) ofTota! bones) of Total Size Class

(}- 2.5 (}-l 78 4.6 17 1.0 21.82.5- 5.1 1- 2 436 25.5 201 11.8 46.15.1- 7.6 2- 3 299 17.5 i7I 10.0 57.27.&-10.2 3-4 236 13.8 168 9.8 71.2

10.2-12.7 4-5 147 8.6 112 66 76.212.7-15.2 5-6 135 7.9 99 5.8 73.315.2-17.8 &-7 71 4.1 54 3.2 76.117.S--20.3 7- 8 59 3.5 44 2.6 74.620.3-22.9 8-9 59 3.5 43 2.5 72.922.9-25.4 9-10 51 3.0 36 2.1 70.625.4-27.9 I(}-Il 24 1.4 17 1.0 70.827.9-30.5 11-12 35 2.1 27 1.6 n.130.5-33.0 12-13 16 .9 12 .7 75.033.(}-35.6 13-14 17 1.0 15 .9 88.235.&-38.1 14-15 4 .2 3 .2 75.038.1-40.6 15-16 4 .2 3 .2 75.040.6 a 16 + 37 2.2 21 1.2 56.8

Total 1,708 100.0 1,043 61.1

Table 66. Prey Animals Idenlified from Pellet Collections ofBubo africanus andTyto alha at Swartkrans, between 1972 and 1975

Hubo afrícanus Tvto alba

8 22.2 53 39.09 25.0 37 27.23 8.3 3 2.2

5 13.9 4 2.9

4 2.9

4 11.1 5 3.71 2.8 5 3.7

11 8.1

2 5.6 4 2.9

2 5.6 2 1.536 100.1 136 100.0

Taxon

Order InsectivoraFamily Soricidae

Order RodentiaFamily Muridae

Subfamily MuJinaePraomys nataíensisMus minutoídesAethomys chrysophilus

Subfamily OtomyinaeOtomys irroratus

Family CricetidaeSubfamily Gerbillinae

Tatera sp.Subfamily Dendromurinae

Dendromys sp.Steatomys pratensis

Dendromurine indet.Class Aves

Indet.Class Insecta

Indet.Total

a Individual animals.

N"

2

%

5.6

N

8

%

5.9

rabie 67. Prey of the Cape Eagle Owl, Bubo capensis mackinderí, Determinedfrom Analysis of Food Rernains Collected in the Matopo HilIs, Zimbabwe

Estimated Mass Total Mass Relative %Prey Species (kg) NóI (kg) Mass

Pronolagus crassicaudatus, redrock hare 2.0 416 906.0 63.5

Lepus saxatilis, scrub hare 1.6 SO 88.5 6.2Procavia capensis and Hetcrohyrax

bruceí, dassies .9 275 247.5 17.3Thryonomys gregoríonus, lesser

cane rat 3.6 25 90.0 6.3T. swinderianus, greater cane rat 4.5 9 40.5 2.8Elephantu(us myurus, elepbant shrew .5 21 10.5 .7Otomys angoniensis. vid rat .2 22 4.2 .3üther muríds, rats ano rmce .1 15 .9 .1Pedetes capensis, springhare 2.7 6 16.2 1.1Erinaceus frontaiis. hedgehog .3 4 \.2 .1Paraxerus cepapi, bush squirrel .2 2 .4 < .1Herpestes sanguíneus, slender

mongoose .5 .5 < .\Paracyníctis selousi, juvenile.

Selous's mongoose .5 1 .5 < .1Genetta sp., genet 2.0 4 8.0 .6Viverra cívena. juvenile, civet 4.5 1 4.5 .3Birds 12 7.5 .5Scorpions lO .2 < .1Insects 6 .1 < .1Lizards 2 .1 < .1

Total 925 1,427.3 100.0

Source: Data from Oargeu and Grobler (1976).a Individual animals.

Table 68. Prey Species Most Abundan! in Analyzed Tyto olba Pellet Collections from Southern Africa

,

Order lnsectivoraFamily Soricidae

Crocidora bicolor, Luanda. Ángela (Dean 1974)Crocidura sp. Quibala. Angola (Dean 1974)

Family ChrysccblcridaeEremitalpo granti namíbensis. Urihauchab Mountains,

South-west Africa (Vemon 1972)Order Rodentia

Family MuridaeSubfamily Murinae

Praomys natalensis. Onderstepoort, Transvaal (Kolbe1946); Matope and Lunzu, Matawi (Hanney 1%2);Támara forest, Kingwilliamstown (Skead 1963); SahiePoort, Nwanedzi, and Hape, Kruger National Park(Coetzee 19(3); Pietennaritzburg, Ashburton, Otto'sBluff, and Polly Shorts, Pietermaritzburg arca (Ver­non 1972); Umfolozi, Natal (Yemon 1972); Naboom­sptuit, Settler's, Rus de Winter, Van Rebeek NatureReserve, BJoemhof, Cullinan, and Jack Scott NaturcReserve, Transvaal (Vemon 1972); Wannbaths, Trans­vaal (Deao 19730); Cuanzo Sul, Huambo, Huila, andLucala, AngoIa (Dean 1974): Swartkrans, Bolt's Farm,and Makapansgat Limeworks, Transvaal (Brain, thisstudy)

Mus minutaides, Kingaton, Kimberley area, Cape (Dean1975)

Aethomvs namaquen."is, Valencia Ranch andGamsberg, South·West Africa (Dean 1915)

Subfamily OtomyinaeOtomys irroratus, Sulekama, Qumbu District, Cape

(CIass )944); Kíngwilliamstown, Cape (Bateman 1960);(rene Old Quarry and Irene New Quarry, Tl'3nsvaal(Brain, this study)

Otomy." sp .. Mossel Bay, Cape. and Krugcrsdorp, Trdns­vaal (Vemon 1972)

Family CricetidaeSubfamily Gc:rbiJlinae

Gerbillurus paeba. Twee Rivieren, Ky Ky, Momo,Auchterlonie, and Kamfersboorn, Kalahari NationalPark, Cape (Davis 1958); Samevloeiing, AuchterlonieA and B, Kransbrak, and Kij Kij (or Ky Ky), KalahariNational Park (Nel and NoJte 1965); Upington, Cape,Vanzylsrus, Cape. Auob River, Gochas, Stamprict,Koichab Pan, and Tsumis Estate, South-West Africa(Vemon 1972); Mirabib, South-West Africa (Brain1974b); Gochas and Heuningvlei, South-West Africamean \975)

Desmodilíus auricularis, Craíghlockhart, Kalahari Na­tional Park, Cape (Nel and Nolte 1965); Colesberg,Cape tvemon 1972); Windhoek District, South-westAfrica (Dean 1975)

Tatera leucogaster, Mbangari, Machindudzi, Matahla­Panga, Pretcríus Kop Rest Camp, and Matupa Cave,Ktuger National Park (Coetzee 1963); Namutoni,South-West Africa (Winterbottom 1966)

Tatera sp., Bolt's farro, Transvaal (Davis 1959); Ba­tulama, Kalahari National Park, Cape (Nel and Nolte1965); Marrick, Kimberley District, Cape (Dean 1975)

Subfamily DendromurinaeSteatomys pratensis, Skukuza koppies, Kruger National

Park (Coetzee 1963)Malacomrix typíca, Springbok, Cape (Yemon 1972)

Class AvesVarious, mostly passerine birds, Bryanston, Johannes­

burg (Davis 19.59); Bryanston, Johannesburg (deGraatT1960b): Kingwilliamstown, Cape (Skead 1956); GraaffReinet, Cape and Kub, South-West Africa (Vemon1972)

Class ReptiliaFamity Gekkonidae. Mirabib, South-West Africa (Vernon

1972)

Table 69. Prey Species in a Sample of Tvto alhaPellets from Warmbaths, Transvaal

Table 70. Percentage Contnbuttons Made to the Dietof Barn Owls by Animals of Different Weights

.-_._-

Estimated Weight RangeMean Weight (g) N"per Individual

Species N" (g) O- 20 8120- 40 502

Mammals 40- 60 43Eíephontuius broctiyrvnchus I 46.5 60- 80Suncus lixus 1 6.0 80-100 2Crocídura hirta 6 14.1 100-120 17C. cyoneatsllaceatpitosa 5 8.1 120- 40 Ic. bicolor 9 5.7 Total 646Tatera leucogaster 22 52.0Otomys angoníensis 17 100.5 " Individual animals.Saccostomus campestrís 14 41.7Steatomvs pratensis 48 26.0Mus minutoides 12 5.8Praomys nataíensis 431 38.8Aethomys chrysophilus 2 81.3t.emníscomys griselda 5 59.3Rhabdomys pumilio 1 35.2

BirdsStreptopelia capicota 1 130.0Cisticoía ehiniana 3 13.0Prinia fíavicans 1 9.9Lanius coliaris 1 44.6Passer domesticas 1 20.5P. griseus 1 20.8P. melanurus 1 18.6Sporopipes squamifrons 4 10.9Píoceus velatus 17 23.3Quelea queíea 34 19.4tiupíectes orix 3 23.0E. afer I 15.4Uraegínrhus angotensis 4 9.7

Total 646

Source : Data from Dean (l973a)." Individual animals.

% ofTotal

12.577.76.7

.32.6

.2100.0

Table 71. Faunal Coatent of Bam Owl Pellet Collections frorn Several Localities Referred to in the Text

Irene Old Quarry Irene New Quarry Bolt's FarmMakapansgatLímeworks

41 23.4

1 .65 2.9

89 50.6

.6

9 5.1

1 .68 4.6

4 2.3

12 6.81 .6

175 99.8

42 15.2 15 5.4

4 1.4 10 3.623 8.3 20 7.1

5 1.814 5.1 24 8.610 3.6 2 .7

276 99.8 280 100.0

Taxon

Order InsectívoraFarruly Soticidae

Order RodentiaFamily Muridae

Subfamily MurinaePraomys natalensisAethomys chrysophilusMus mtnutoídesRhabdomys pumilio

Subfamily OtornyinaeOtomys irroratuslangoniensis

Family CricetidaeSubfamily CricetinaeMystromys albicaudatus

Subfamily GerbillinaeTatera sp.

Subfamily DendromurinaeDendromvs sp.Steatomvs pratensis

Family BatbyergidaeCryptomys hottentotus

Rodentia indet.Class A ves, Indet.Class Insecta. indet.

Total

" Individual animals.

N'

6

17

2

32

26

10

7

100

%

6.0

17.0

2.0

32.0

26.0

10.0

7.0

100.0

N

3

%

1.7

N

10

146125

10

%

3.6

52.94.31.8

3.6

N

19

l34136

32

%

6.8

47.94.62.1

11.4

310 Appendix: Tables

Table 72. Vegetation Occurring in Circles of 2 Km Radius around thePositions of Barn Owl Roosts al Bolts Farm, Swartkrans. and MakapansgatLimeworks

Percentage of Total Area

.,Vegetation Type

Cultivated fieldsWooded (other than bushveld)Open grasslandMixcd bushveldEvergreen forest

Total

Bolt·s Farm

8.81.6

89.6

100.0

Swartkrans

12.11.0

86.9

100.0

Makapansgat

8.5

90.51.0

100.0

T.ble 73. Lengths of Bone Flakes Produced by Spotted Hyenas at KilIs inthe Kruger National Park and Deos ín the KaJahari Gemsbok National Park

Length Kruger Park Kalahari Park Combined Samples(cm) N % N % N %

0-11- 2 1 .6 1 .52- 3 15 9.6 1 1.6 16 7.3:1-4 19 12.1 2 3.2 21 9.54-5 16 10.2 4 6.3 20 9.15-6 25 15.9 6 9.5 31 14.16-7 17 10.8 6 9.5 23 10.57- 8 16 10.2 8 12.7 24 10.98-9 12 7.6 8 12.7 20 9.19-10 9 5.7 10 15.9 19 8.6

10-11 9 5.7 8 12.7 17 7.711-12 6 3.8 5 7.9 11 5.012-13 7 4.5 1 1.6 8 3.613-14 1 .6 3 4.8 4 1.814-15 3 1.9 3 1.415-1616-17 1.6 I .5

17t- 1 .6 I .5TotaJ 157 99.8 63 100.0 220 100.1

Table 74. Lengths of Bone F1akes Produced by BrownHyenas al Deos in the Kalahari National Park

Appendix: Tables 311

Length (cm) N %

~I

1-22-3344-55-{;

6-77-88-99-101~11

11-1212-13B-1414-1515-1616-17

17+Total

1 2.91 2.91 2.91 2.92 5.92 5.94 11.86 17.72 5.95 14.7

4 11.83 8.8

2 5.934 100.0

Table 75. Numbers of Genera and Species Represented in the Macrovertebrate Component of the SterkfonteinVa/ley Caves

Genera Species Genera Species

Taxon Touol Extinct Touol Extinct Taxon Total Extinct Touol Extinct

Phylum Chordata Order RodentiaClass Mammalia Family Hystricidae 2Order Primates Class Aves

Famlly Hominidae 3 3 3 3 Order StruthiofonnesFamily Cercopithecidae 6 5 9 9 Family Struthionidae

Order Camivcra OrderFalconifonnesFamily Felidae 6 4 9 7 Family AccipitridaeFamily Hyaenidae 5 2 8 5 OrderGallifonnesFamily Canidae 4 1 6 4 Family NumididaeFamily Mustelidae 1 1 1 Class ReptiliaFamily Viverridae 4 5 2 Order CrocodiJia

Order Artiodactyla Family CrocodylídaeFamily Bovidae 19 5 26 13 Order CheloniaFamily Suidae 3 2 3 3 Family Testudinidae ?1Family Giraffidae 1 1 1 1 Order Squamata

Order Perissodactyla Family VaranidaeFamily Equidae 2 4 2 Family Gerrhosauridae

Order Proboscidea Phylum MolluscaFamily Elephantidae Class Gastropoda

Order Hyracoidea Order Pulmonata 1 ?1Family Procaviidae 3 2 Total 67 24 91 54

Order LagomorphaFamily Leporidae ?2 ?2

312 Appendix: Tables

Table 76. Rernains from Sterkfontein Member 4: Overall Analysís, Taxa Represented, and Mínimum Numberof Individual Animals Involved

Minimum MínimumNumber of Number of Number of Number of

Taxon Specimens Individuals Taxon Specimens Individuals

Phylum Chordata ?HippotraginiClass Mammalia Cranial pieces 23 SOrder Primates Redunca cf arundinum

Family Hominidae Cranial pieceAustralopithecus afríconus Antidorcas cf recki

Cranial pieces 84 ± 45 Cranial pieces 5 3Postcranial pieces S Antidorcas cf bond;

Family Cercopithecidae Cranial pieceParapapio jonesi ?Gazel/a sp.

Cranial pieces 35 27 Cranial pieces 2Postcranial pieces 2 Syncerus sp.

Parapapio broomi CraniaJ pieceCranial pieces 100 91 Tragelaphus sp. aff. angasi

Parapapio whitei Cranial pieces 3Cranial pieces 13 10 Makapania cf. broomiPosteranial pieces 1 Cranial pieces 22 S

Parapapio sp. indet. Antelope c1ass 1Cranial pieees 201 53 Craníal pieces 4 2Posteranial pieces 5 Postcranial pieces S

Cercopithecoides williamsi Antelope c1ass 11Cranial preces 20 17 Cranial pieces SI ±SPostcranial piece 1 Postcranial pieces 30

Cercopithecoid indet. Antelope class [11Cranial pieces 413 :t 100 Cranial pieces 157 :t 15Postcranial pieces 46 Postcranial pieces S9

Order Carnivora Antelope class IVFamily Felidae Cranial pieces 13 2

Panthera pardus Postcrarual pieces 2Cranial pieee Family Suidae

Dinofelis barlowi Cf Pronotochoerus sp.CraniaJ pieces 5 4 Cranial pieces 2 2

Megantereon graciie Order PerissodaetylaCranial piece Family Equidae

Family Hyaenidae Equus capensisEuryboas silberbergi Cranial pieees IS 7

CraniaJ pieces 5 4 Equus sp.Cracuta croeuta Cranial pieces 2

Cranial pieee Order ProboscideaHyaenid indet. Family Elephantidae

Cranial pieces 5 ±3 Elephas el. reckiFamily Canidae Cranial pieee

Canis brevirostris Order HyracoideaCranial piece Family Procaviidae

Canis mesomelas Procavia anticuaCranial píeces 9 ±5 Cranial pieees 17 S

Carnivore indet. Procavia transvaalensísCraniaJ pieces 32 ±S Cranial pieees 9 5Postcranial pieces 16 Order Rodentia

Order Artiodactyla Family HystricidaeFamily Bovidae Hystrix africaeaustratis

Damaliseus el. sp. 2 Cranial pieees 6 5Cranial piece Carnivore eoprolite 1

Damaliscus sp. I or Inseet pupae 3Parmularius sp. Indeterminate fragments 330Cranial pieces 26 7 Bone flakes 12

Medium-sized alcelaphines 2-3 cm 1Cranial pieces 14 7 3-4 cm 1

Cf. Connochaetes sp. 4-5 cm 4Cranial pieces 3 ~cm 4

Cf Megalotragus sp. ~7cm 1Cranial pieees 7-8 cm 1

Hippotragus cf. equinus Total I,S95 470Cranial pieces 2 2

Appendix: Tables 313

Table 77. Skeletal Part Analysís of Bovid Remains from Sterkfontein Member 4

Number of Specirnens

Part Size Class 1 Size Class II Size Class J11 Size Class IV

Calvaría preces I 1Horn-core pieces 3 2 1Maxillary pieces 9 13 1Mandible pieces 3 17 33 1lsolated upper teeth 20 37 5Isclated lower teeth 10 19Tooth fragments 21 52 5Vertebrae

AtlasAxisCervical, 3-7 8Thoracic 4Lumbar 4 14SacralCaudal 2

Rib pieces 13Scapula pieces 3Pelvis piecesHumeros

Proximal pieces 2Distal pieces 3Shaft pieces

Articulated distal humeros!proximal radius and ulna

Radius and ulna. proximalpieces 2

RadiusDistal pieces 2 2Shaft pieces

FemurProximal pieces 4Distal pieces 7Shaft pieces

PatellaTibia

Proximal pieces 2 4Distal piecesShaft pieces

MetacarpalProximal pieces 2Distal pieces 3Shaft pieces 2

MetatarsalProximal pieces 2Distal pieces 2Shaft pieces 2

MetapodialProximal piecesDistal pieces 4 3Shaft pieces ]

Astragalus 2 5Calcaneus 1Tarsal bonesCarpal bones I1st phalanges 3 22d phalanges 1 1Terminal phalanges 1Articulated foot bone pieces

Total 12 111 246 15

314 Appendix: Tables ,Table 78. Allocation lo Age Categories of Fossils Representing 47 Individuals of Australopithecus afrícanus

1-5 Years 6-10 Years 11-15 Years 16-20 Years 21-25 Years 26-30 Years 31-35 Years 36-40 Years

STS 2 STS 57 STS 1 STS 52 TM 1511+ STS 19 STS 5 TM 1514STS 60

TM 1516 STS 23 STS 8 STS 17 STS 35 STS 71 STS 10STS 18 STS 22 TM 1512 STS 61 STS 29 STS 7STS 24 STS 28+37 STS 12 STS 53 STS 42 STS 36STS 56 STS 55 STS 32 STS 38 TM 1532 STS 54STS 9 TM 1523 TM 1561 STS 72 STS 46 TM 1519STS 51 STS 59 STS 21 TM 1520

STS 30STS 31TM 1518STS4

7 2 7 11 7 6 6

Source: Following Mano (1975).

Table 79. Allocation lo Age and Sex Classes of Fossils Representing 27Individuals of Parapapio jonesi

Age Class Male Female Sex Unknown Total

Juvenile 3331-340 2384

Immature adult 306 457 2Young adult 421 287+329 332 7

334 372381+446418+458

Adult 367 565 12485+390 4481925+443 317348+302 313

355307456284

Old adult 250 441 4276368

Total 9 16 2 27

Note: AH specimens are from Sterkfontein Member4 and have numbers prefixed by STS.

Appendix: Tables 315

Table 80. Rernains oí Parapapio broomi from Sterkfontein Member 4:Separation into Age and Sex Classes

Age Class Male FemaJe Sex Unknown Total

Juvenile 558 419 3283

Immature adult 270 354 10279 411 +425

383 328277268266251

Young adult 301 398a-<:+280 438 20414a,b 353 544

371 388a.b390a,b 410

378 261325 256

322+437 382a274 289262 286

Adult 564 {type) 562 271 39272 254+383a 298267 297 326260 385a--e 434346 397+285 406a-<253 3035 445311 530+374b 491314 535 420323 369a.b 511484 338278 557258 426542 409

3800255299335

Old adult 533+534 3960+362 19531+312 264296+351 393

416 331360 356339 363 380a-<337 415a 309

466469539

Total 26 43 22 91

Note: Al! specimen numbers are prefixed STS.

Table 81. Remains oíParapapio whiteí from Sterkfontein Member 4: Separationinlo Age and Sex Classes

Age Class Male Female Sex Unknown Total

Juvenile O O O OIrnmature adult 263+370 1Young adult 342 352+467 2Adult 424+462 259 548 5

336 303Old adult 389 2

359Total 4 4 2 10

Note: AII specimen numbers are prefixed STS.

316 Appendix: Tables

Table 82. Remains of Cercopíthecoídes williamsi from Sterkfontein Member 4:Separation into Age and Sex Classes

Age Class Male Female Sex Unknown Total

Juvenile 300 2290+ 357+435

Imrnature adult 282Young adult 344

518+516 3523

Adults 347 394 279 10350 532 288

366+392 295361252

Qld adult 541 1Total 5 2 10 17

Note: AH specimen numbers are prefixed STS.

Table 83. Fossil Assemblages from Sterkfontein Members 4, S, and 6:Contributions oflndividual Animals ofVarious Taxa to the Preserved Faunas

Member4 Member 5 Member 6

Taxon N" % N % N %

MammaliaPrimates

Hominidae 47 13.4 3 7.3Cercopithecidae 198 56.6 1 2.4

Camivora 28 8.0 4 9.8 3 18.8Artiodactyla 51 14.6 23 56.1 8 50.0Perissodactyla 7 2.0 3 7.3 2 12.5Hyracoidea 13 3.7 1 2.4 1 6.3Proboscidea 1 .3Lagomorpha 1 2.4Rodentia 5 1.4 2 4.9 6.3Aves 1 2.4 6.3Reptilia 2 4.9

Total 350 100.0 41 99.9 16 100.2

a Individual animals.

Table 84. Observed Damage to Bones fromSterkfontein Member 4

Porcupine-gnawed píeceParapapio sp., juvenile mandible, STS 320.

Pieces showing camívore-infíicted damageAustratopithecus africanus, part of ajuvenile mandible, STS 18,showing ragged-edge damage to symphyseal atea; palate, STS53, showing ragged-edge damage around upper margino Exten­sive shattering has also occurred, and diagnosis of carnivore­damage on both these specimens is tentative.Parapapio jonesi, mandible, STS 334, possible camivore dam­age to posterior margin of ramus; Parapapio sp., mandible, STS351. two punctate marks on inner side ofhorizontal ramus.Tragelaphus sp. aff. angasi, mandíble, STS 1865, probable cer­nivcre damage.Procavia antiqua, palate and base ofbraincase, STS 109,showing typical pattem of leopard damage.P. transvaalensis, mandible, STS 101, evidence of carnivorechewing.Antelope class III, distal humeros, STS 1930, extensive toothmarks around distal end; distal humeros, STS 1584, similardamage.

,,,

Appendix: Tables 317

Table 85. Remains from Sterkfontein Member 5: Overall Analysis, Taxa Represented, and Minimum Numbersof Individual AnimaJs Involved

Minimum MinimumNumber of Number oC Number of Number of

Taxon Specimens lndividuals Taxon Specimens Individuals

Phylum Chordata Antelope c1ass JlIClass Mammalia Cranial pieces 32Order Primates Postcranial pieces 108 ±5

Family Hominidae Antelope class IVcf Horno sp. Cranial piece 1

Cranial pieces 6 3 Postcranial pieces 5Family Cercopithecidae Family Suidae

Cercopithecoid, gen. el sp. indet. Suid, gen. et sp. indet.Cranial pieces 2 Cranial pieces 2

Order Carnivora Order PerissodactylaFamily Felidae Family Equidae

Cf. Panthera leo Equus cf. burchelliCranial pieces 2 Cranial pieces 18 3Postcranial pieces I Postcranial píeces 5

cr. Megantereon sp. Order HyracoideaPostcranial pieces 3 Family Proeaviidae

Family Hyaenidae Procavla sp.Proteíes sp. Cranial pieces 2

Cranial pieee Postcranial piece IFamily Canidae Order Lagomorpha

Canis cf. terbíanchei Family LeporidaeCranial piece Gen. et sp. Indet.

Camívore indet. Cranial pieceCranial pieces 7 3 Order RodentiaPostcranial pieces j Famíly Hystricidae

Order Artiodactyla Hvstrix cf. africaeaustralisFamily Bovidae Cranial píeces 6 2

Damaliscus cf. dorcas Postcranial piece ICranial pieces 4 2 Class Aves

Damatiscus cf. sp. 2 Gen. et sp. indet.Craníal pieces 7 4 Cranial piece

Damaliscus sp. Lor Parmularius sp. Postcranial pieceCranial piece Class Reptilia

Medium-sized alcelaphines Order SquamataCranial pieces 7 5 Family Varanidae

Cf. Connochaetes sp. Varanus cf. niloticusCraníal piece Postcranial piece

"Hippotragini Order ChelooiaCranial piece Family Testudínidae

Antidorcas cf. recki Gen. et sp. indet.Cranial pieces 5 J Carapace piece I

Orteotragus majar Indet. fragments 580Cranial pieces 3 Bone flakes

Taurotragus cf. oryx 1-2 cm 12 206Cranial piece 2-3 cm 57

Makapania cf. broomí J..4cm 48Cranial piece 4-5 cm 47

Antelope class 1 5-{, cm 14Cranial piece 1 6-7 cm 15Postcranial pieces 19 ±3 7-8 cm 2

Antelope class [] 8+ cm 11Cranial pieces 29 Total 1,202 58Postcranial pieces 122 ±6

318 Appendix: Tables

Table 86. Skeletal Part Analysis of Bovid Remains from Sterkfontein Member 5

Number of Specimens

Pan Size Class 1 Size Class 1I Size Class 111 Size Class IV

Cal varia pieces 1 7Horn-core pieces 5 6Maxillary pieces 1 1Mandible pieces J 4Isolated upper teeth 6 5Isolated lower teeth 6 2Tooth fragments 7 7Vertebrae

Atlas IAxis 2 1Cervical, 3-7 3 1Thoracic 2 5 6Lumbar 2 2 11Sacral 1Caudal

Rib pieces 18 25Scapula pieces 6 4Pelvis pieces J 2Humerus

Proximal piecesDistal pieces J 7 1Shaft pieces J 1 2

Articulated distal humeros!proximal radius and ulna

Radius and ulna, proximalpieces 2 J

RadiusDistal pieces 4Shaft pieces

FemurProximal pieces 2 JDistal pieces I JShaft pieces 1 I

Patella 1 2Tibia

Proximal pieces I 4Distal pieces 4 1Shaft pieces 2

MetacarpalProximal pieces 1 2Distal pieces 2 JShaft pieces I

MetatarsalProximal pieces J J IDistal pieces 1 I JShaft pieces 1

MetapodialProximal piecesDistal peces 9 4Shaft pieces 2

Astragalus 2 JCalcaneus J 1 2Tarsal bones 5 JCarpal bcnes J 41st phalanges 15 62d phalanges 7 ITerminal phalanges 2 1Articulated Coot bone pieces 3

Total 20 \SI 140 6

-Table 87. Observed Darnage to Bcnes from StcrkfcnteinMember 5

Appendix: Tables 319

Tabie 88. Microfaunal Remains frcm SterkfonteinMember 5

Porcupíne-gnawed otecesAntelope class J: femur shafr, SE 839Antelope class JI: calcaneus, unnurnbercd: sacrum, 1885; prox­imal radiua, 2031Antelope class III: metacarpal, 1458; metatarsal, 127BODe ñakes: 64,127,804.1011. 1458,2061, unnumberedShaft pieces: 1439, unnumbered

Píeces gnawed by smaít rodentsAnttdorcas cf. recki, mandible, 535Antelope ctass 1: scaputa, 657Antelope class JI: phalanges, 826, 681; humeros, 1670; femur1884; metatarsals, 1693, 1401Antelope class III: rib, 645; phalanx, 74'; Iemur, 745Bone ñakes, 117. 657. 745. 815lndet. rib, 653: indet. scapula, 657

Pieces sbowing camivore-infticted damageClear camivore demageAnreíope clase 1: humeros, 1528; humeros, 72 (puactate mark)Antelope c1ass11: rib, 1535; calcaneus, 1707; scapulae, 835, 1361,1265: phalanx, 1263; thoracic vertebrae. 629, 742, 1376; lumbarvertebra, 491; axis, 1264; cervical vertebra, 1321; radii, 1336,unnumberedAntelope class [[1: craníum, 571a,b (punctate marks}; thoracicvertebra, 2162; pelvis, ll; ribs, 1510.625 (punetate mark); femur521 (punctate mark); metacarpal, 1292Bone ñakes: 653,804, 1000,2027; Jndet. pieces: 135,481, 494,620, 745, 825, 902, 1783,2009, 2024Probable camivore damageAntelopc class 11: ribo 1308Suid: mandible, 1069Bone ñake: 402Indet. pieces: 435, 1543, 1647,2164

Pieces showing artificial mariaAntelope class 1: humeros shaft, 1729 (clear chop mark)Antelope class 111; born-ccre, 2031 (hole through horn-core);bom-core piece, 1524 (wom to a pcinf)Bone ñakes: 612 (wom facets, apparently abone tool}; 1000(smoothed edges)Piece showing possible insect boring: shañ piece, 1363

Note: AlI specimen numbers are prefixed SE.

Taxon

Order Chiroptera (indet.)Order Insectívora

Family MacroscelididaeElephantulus sp.

Family SoricidaeSorícid indet.

Order RodentiaFamily Bathyergidae

Cryptomys sp.Family Muridae

Subfamily Murinae, various generaSubfamily Otomyinae

Cf Palaeotomys graciíisFamily Cricetidae

Subfamily CricetinaeMyslromys sp.

Subfamily GerbillinaeTatera sp.

Subfamily DendromurinaeDendromurine indet.

Rodentia (Myomorphal indet.Class ReptihaOrder Squamata

Lacertilia indet.Class Aves.

Indet.Total

MinimumNumber ofIndividuals %

8 1.2

21 3.3

109 16.9

16 2.5

42 6.5

32 5.0

144 22.3

14 2.2

146 22.672 11.2

33 5.1

7 1.1644 99.9

320 Appendix: Tables

TabIe 89. Remains from Sterkfontein Member 6: Overall Analysis, Taxa Representcd, and Minimum Numbersof Individual Animals Involved

Minimum MínimumNumber oí Number of Number oí Number of

Taxon Specimens Individuals Taxon Specimens Indíviduals

Phylum Chordata Order HyracoídeaCtass Mammalia Family ProcaviidaeOrder Carnivora Procavla cr. capensis

Gen. el sp. índet. Cranial pieces 3Cranial pieces 3 =3 Postcranial pieces 2Postcranial pieces 6 Order Rodentia

Order Artiodactyla Family HystricidaeFamily Bovidae Hystríx ef. afrtcaeaustratis

Antídorcas bond; Cranial pieces 2Cranial pieces 3 2 Class Aves

Damaliscus cf. dorcas Order FalconiformesCranial piece Family Accipitridae

Antelope class 1 Postcranial pieces 2Postcranial pieces 10 1-2 Indet. fragments 70

Antelope class [[ Bone ñakes 290Cranial pieces 3 =4 )-2 cm 17Postcranial pieces 34 2-3 cm 112

Antelope class Hl 3-4 cm 85Cranial pieces 4 =2 4-5 cm 41Posteranial pieces 15 5-{;cm 17

Order Perissodactyla ~7cm 8Family Equidae 7--8 em 7

Equus ef. burcñeíí¡ 8 + cm 3Cranial pieces 6 2 Total 454 19

Appendix: Tables 321

Table 90. Skeletal Part Analysis of Bovid Remains from Sterkfontein Member 6

Number of Specimens

Part Size Class 1 Size Class 11 Síze C1ass III

Cal varia piecesHom-core piecesMaxillary piecesMandible piecesIsolated upper teethIsolated lower teethTooth fragments 2 2Vertebrae

AtlasAxisCervical, >-7ThoracicLumbarSacralCaudal

Rib pieces 2 11 6

Scapula pieces 1 1Pelvis pieces 1 1

HumerosProximal piecesDistal preces 3 2Shaft pieces

Articulated distal humerus/proximal radius and ulna

Radius and ulna, proximalpieces

RadiusDistal piecesShaft pieces

FemurProximal piecesDistal piecesShaft pieces

PatellaTibia

Proximal piecesDistal piecesShaft pieces 2

MetacarpalProximal piecesDistal piecesShaft pieces

MetatarsalProximal preces 1 2

Distal pieces 1

Sbañ pieces 2Metapodial

Proximal piecesDistal pieces 3Shaft pieces 2

AstragalusCalcaneus 2Tarsal bonesCarpal bones1st phalanges ) 22d phalanges 2

Tenninal phalangesArticulated foot bones

Total 10 37 19

322 Appendix: Tables 1Table 91. Observed Damage to Bones from Sterkfontein Member 6

Porcupine-gnawed piecesShaft pieces: SE 702, 1526Bone flakes: 737, 867, 907, 1006,2 unnumbered pieces

Pieces gnawed by small rodemsAntelope c1ass 1: phalanx, 984Antelope class 11: phalanx, 878Shaft piece, unnumbered, and bone ftake, 719

Pieces showing camivore-ínfíicted damageClear camivore damage

Note: AII specimen numbers are prefixed SE.

Antelope c1ass 11: pelvis. 697; phalanx. 1890Antelope class III: rib, 843Indet. pieces: 913, unnumberedBone ñakes: 725.932,985Probable camivcre damageIndet. pelvis piece: 2084Bone ñake, 848

Pieces showing artificial marksAntidorcas bondi, mandible, 690 (cut marks)Bone ftakes, 1054, 2384, unnumbered (rounded and wom)

Table 92. Remains from Swartkrans Member 1: Overall Analysis, Taxa Represented, and Minimum Numbersof Individual Animals Involved

Minimum MinimumNumber of Number of Number of Number of

Taxoo Specimens Individuals Taxon Specimens Individuals

Phylum Chordata Family HyaenidaeClass Mammalia Hvaena brunneo disparOrder Primates Cranial pieces 7 3

Family Hominidae Crocuta crocuta venostulaHorno sp. Cranial pieces 3 2

Cranial pieces 4 3 Hvaenictis forfexAustralopithecus robustus Cranial pieces

Cranial pieces 218 87 Euryboas nítidula (type B, advancedlPostcranial pieces 11 Cranial pieces 9 5

Family Cerccpithecidae Enryboas nitidula (type A. primitive)Parapapio jonesi Cranial pfeces 5 2

CraniaJ pleces 29 8 Eurvboas nítidula (type uncertain)Papio robinsoni Cranial pieces 4

Cranial pieces 121 38 Proteles transvaalensisTheropithecus danieli Cranial pieces 2

Cranial pieces 31 17 Hyaenid indet.Dinopithecus íngens Cranial pieces 16 ±4

Cranial pieces 57 26 Family CanidaeCercopithecoid Indet. Canis mesomeías pappos

Cranial pieces 134 ±28 Cranial pieces 9 4Postcranial pieces 31 Vu/pes pulcher

Order Carnivora Cranial pieces 2 2Family Felidae Carnivore indet.

Panthera pardus Cranial pieces 28 ±15Cranial pieces 29 12 Postcranial pieces 28Postcranial pieces 3 Order Artiodactyla

Dinofelis sp. Family BovidaeCranial pieces 2 Damalíscus sp. 1 or Parmularius sp.Postcranial pieces 2 Cranial pieces 8 4

Megantereon sp. Rabatíceras poerocomutusCranial piece Cranial pieces 3 2

Continued 0'1 following page

Appendix: Tables 323

Table 92. Continued

Mínimum MinimumNumber of Number of Number of Number of

Taxon Specunens Individuals Taxon Specimens Individuals

Medium-sized alcelaphines Order PerissodactylaCranial pieces 97 22 Family Equidae

Cf. Connochaetes sp. Equus capensisCranial pieces 70 19 Cranial pieces 9 6Postcranial pieces 4 Hlpparíon stevtíeri

Cf. Megalotragus sp. Cranial piecesCraniaJ pieces 8 3 Equid indet.

"Hippotragini Craníal pieces 14 ±4Cranial pieces 2 Postcranial pieces 3

Redunca cf. arundinum Order HyracoideaCranial pieces Family Procaviidae

Pelea cf. capreolus Procavia antiquaCranial pieces 2 2 Cranial pieces 35 16

Antídorcas cf. recki Procavia transvaalensisCranial pieces 12 6 Cranial pieces 20 8

rcaeeuo sp. Postcraníal pieces 1Cranial pieces 8 6 Hyracoid indet.

Antilopine or Neotragine indet. Cranial pieees 6 2Cranial pieees 3 3 Order Rodentia

Oreotragus cf. major Family HystricídaeCranial pieces Hystrix afdcaeaustralis

Syncerus sp. Cranial pieces 5 2Cranial pieces 13 3 Hystrix ?makapanensis

Trageíaphus cf. strepsiceros Cranial píecesCranial pie ces 12 4 Class Aves

• CC. Makapania sp. lndet.: eranial pieceCranial pieces 10 3 Phylum Mollusca

Antelope c1ass I C1ass GastropodaCranial pieces 25 ±5 Order PulmonataPostcranial preces 11 Cf. Achatino sp.

Antelope c1ass 11 Shells 15 15Cranial pieces 117 ±20 Indeterminate fragments 500Postcranial pieces 113 Bone ñakes 49

Antelope class III 2-3 cm 7Cranial pieees 199 ±25 3-4 cm 11Postcranial pieces 162 4-5 cm 15

Antelope class lV 5-{)cm 10Cranial pieces 15 2 6-7 cm 4Postcranial pieces 2 7-8 cm I

Family Suidae 8+ cm 1Tapinochoerus meadowsi Total 2,381 463

Cranial pieces 8 7Suid indet.

Cranial pieces 9 ±3

Note: Corrected minimum number of individuals, 339.

324 Appendix: Tables ~Table 93. Skeletal Part Analysis of Bovid Remains from Swartkrans Member J

Number of Specimens

Part Size Class I Size Class Il Size Class III Size Class IV

Calvaria pieces 2 I 4Horn-ccre pieces 1 1 6Maxillary pieces 2 14 21 2Mandible pieces 9 42 48 2Isolated upper teeth 4 23 53 9Isolated lower teeth 3 18 39 1Tooth fragments 4 18 28Vertebrae

Atlas IAxis 2Cervical, 3-7 4 3Thoracic 13 ILumbar 7 2Sacral 2Caudal

Rib pieces 4 IScapula pieces 5 5Pelvis pieces 7 4Humeros

Proximal pieces 9Distal pieces 3 9Shaft pieces 1 3

Articulated distal humerus!proximal radias. and ulna

Radius and ulna, proximalpieces 4

RadíusDistal piecesShaft pieces

FemurProximal pieces 2 6Distal pieces 5 9Shaft pieces I

PatellaTibia

Proximal pieces I IDistal pieces 2 4Shaft pieces I

MetacarpalProximal pieces 2Distal pieces 3Shaft pieces 3

MetatarsalProximal pieces 2 2Distal pieces 1 IShaft pieces 2 3

MetapodialProximal piecesDistal pieces 5 6Shaft pieces 7 5

Astragalus 4 9Calcaneus 3 9Tarsal bones 3 4Carpal bones 1 9lst phaJanges 12 282d phaJanges 5 9Terminal phalanges 2 5 2Articulated foot bone pieces 1 2

Total 36 230 361 17

Appendix: Tables 325

Table 94. Allocation of 113 AustralopithecineSpecirnens from Swartkrans Member I to Age-at-DeathCategories

Age Category(years) N %

1- 5 11 10.06-10 22 20.0

11-15 21 19.016-20 21 19.021-25 12 11.026-30 14 12.031-35 11 10.036-40 1 .1

Total 113 101.1

Source: Data frorn Mano (1975).

Table 95. Parapapío jonesi from Swartkrans Member 1: Allocation of Specimens to age Classes

Maxillae Mandibles

Sex Sex Mínimum NumberAge Class Male Female Unknown Male Female Unknown of Individuals

Young juvenileJuvenile 462 418,433 2Immature adultYoung aduh 588b IAdult 612 1835 1Old adult 588 573b, 543 14161 573a, 414 4Very old adult

Note: AII specimens are prefixed SK.

Table 96, Papio robinsoní from Swartkrans Member 1: Allocation of Specimens to Age Classes

Maxillae MandiblesMinimum

Se. Sex Number oíAge Class Male Female Unknown Male Female Unknown Individuals

Young juveniJeJuvenile 447.610 453.446.

420, 463 4Immature adult 536.436, 14083. 5

458.456. 445.4531497

Young adult 602.631.544. 558.562. 408. 570. 407,409, 11555 (lype) 565,476 429,419 410.457.

540,2319,427

Adult 14006. 550 557.549. 417,435, 321b 721SO.537 1468

Old adult 590,629. 571b 572,423. 421,5710, 714163 431,538 459

very old adult 560 566 406.430. 416. 14156 62161.12171

Note: AH specimen numbers are preñxed SK.

326 Appendix: Thbles

Table 97. Theropithecus danieli Remains from Swartkrans Member 1: AUocation of Specimens to Age Classes

Age Class

Young juveniteJuvenileImmature adult

Young adult

Adult

Old adultVery old adult

Maxillae MandiblesMínimum

Sex Sex Number ofMaJe Female Unknown Male Female Unknown Individuals

564,593, 426,403 6448,507567,597, 569 411 4561479,563 581 439, 575a, 4(goes with 405 + 402405/402)575b,< 1461 491 2

Note: AU specimen numbers are prefixed SK.

Table 98. Dinopíthecus íngens Remains from Swartkrans Mernber 1: AUocation of Specimens to Age Classes

Maxillae MandiblesMinimum

Sex Sex Number ofAge Class Maje Female Unknown Male Female Unknown Individuals

Young juvenile 415,413 2Juvenile 554 1518 589 2Immature adult 447 574,548, 532, 1407 6

440,443YOllO¡ adult 546 600,603, 404,428 7

542a 492,455AduIt 578a,b 604,571b, 474 5

2625 5760Old aduh 545 553,473 422 3Very old adult 401 I

Note: AU specimen numbers are prefixed SK.

Table 99. Fossil Assemblages from Swartkrans Members 1 and 2 and ChannelFi1l: Contributions of IndividualAnimals of Various Taxa to the PreservedFaunas

Member I Member2 Channel FiII

Taxon % N % N % N

MammaliaPrimates

Hominidae 90 26.6 I .4Cercopithecidae 89 26.3 8 3.1

Carnívora 36 10.6 21 8.1 3 10.0Artiodactyla 89 26.3 166 64.3 17 56.7Perissodactyla 7 2.1 13 5.0 I 3.3

Hyracoidea 24 7.1 37 14.3 5 16.7Lagomorpha 4 1.5 2 6.7Rodentia 3 .9 5 1.9

Aves 1 .3 1 .4 I 3.3Reptilia 2 .8 I 3.3

To'a1 339 100.2 258 99.8 30 100.0

Table 100. Observed Damage to Bones fram SwartkransMember I

Porcupine-gnawed pieceAntelope class IlI: metapodlal, SK 11819

Pieces gnawed by small rodentsAntelope class 1: phalanx, 14020

Pieces showing camivore-infíicted damageClear camivore damageAustralopimecus robustas: juvenile calveria, 54 (two punctatemarks, one in each parietal); juvenile mandible, 3978 (regged­edge damage to right horizontal ramus); innominate, 3155bItooth marks around severa! margins)Antelope class T: humeros, 14074Antelope class fI: mandible, 4209; femurs, 140760, 6294; tibia,1803; metacarpal. 2646; rib, 3093Antelope class UI: scapula, 4203; calcaneus, 6694Indet. shaft piece, 8165Probable camivcre damageHorno sp.: mandible, 45 (ragged-edge break through horizontalramus)Australopithecus robustus: endocranial cast, 1585 (punctate de­pression in right frontal, probably caused by a tooth); adultmandibles,74a. 1587, 1588 (damage to lower margin oframus);juvenile mandibles, 61, 62, 64, 438, 869 (damage te horizontal orascending rami); axis vertebra, 854 (edge damage): adult 500ut,1J (damage to right side of maxilla): adolescent snout, 13 (ir­regular damage all around); innominate, 50 (damage lo iliaemargin); distal humeros, 860 (damage to shaft)Cercopíthecoíd indet.: calvaría, 2132 (three punctate marks)

Note: Al! specimen numbers are prefixed SK.

Appendix: Tables 327

. IIIIII

328 Appendix: Tables

Table 102. Remains from Swartkrans Member 2: Overall Analysis, Taxa Represented, and Minimum Numbersof Individual Anírnals Involved

Minimum MínimumNumber of Number of Number oC Number of

Taxon Specimens Individuals Taxon Specimens Individuals

Phylurn Chordata Family MustelidaeClass Mammalia Mellivora aff. stvalensisOrder Primates Cranial pieces 3

Family Hominidae Family VivenidaeHorno sp. Herpestes sanguineus

Cranial pieces Cranial píecePostcranial pieces Cyníctis peníciltata

Family Cercopithecidae Cranial piecePapío sp. Camívore indet.

Cranial pieces 7 4 Cranial peces 12 8Cercopithecoídes williamsi Postcranial pieces 44

Cranial pieces 5 4 Order AniodactylaCercopuhecoíd indet. Family Bovidae

Cranial pieces 10 4 Damaliscus cf. dorcasPostcraniaJ pieces 4 Cranial pieces 21 9

Order Carnivora Damaiiscus sp. 2 (niro?)Family Felidae CraniaJ pieces 57 9

Panthera aff. leo Medium-sized alcelaphínes (includingCranial pieces 4 3 ef. Beatmgus sp.)Postcranial píeces 6 Cranial pieees 30 9

Panthera pardus Cf Connochaetes sp.Cranial pieces 2 2 Cranial pieces 2Postcreniel pieces 1 Cf. Megaíotragus sp.

Felid indet. Cranial pieces 3 2Posteranial pieces 2 Hippotragus ct. niger

Family Hyaenidae Cranial pieces 25 9Hyaena brunnea Cf. Kobus ellipsiprymnus

Cranial pieces 3 2 Cranial pieces 2 2Ptoteles crístatus Pelea capreoius

Craniel pieces 2 Cranial pieces 33 10Hyaenid indet. Antidorcas australis andJor

Cranial pieces marsupialisPostcranial pieces Cranial pieces 64 16

Family Canidae Antidorcas bondíCanis mesomeías Cranial pieces 30S 70

Cranial pieces 27 8 PostcraniaJ pieces 10Postcranial pleces 9 Oreotragus cf. majar

Cf. Lycaon sp. Cranial pieceCranlal piece Oreotragus cf. oreotragus

Otocyon recki Cranial pieces 3 2Cranial pieces 2

Contínued on following page

Appendix: Tables 329

Table 102. Continued

Mínimum MínimumNumber of Number of Number of Number of

Taxoo Specimens Individuals Taxon Specirnens Individuals

Raphicerus cf. campestrís Order HyracoideaCranial pieces 10 5 Family Procaviidae

Cf. Raphicerus sp. Procavia cf. antiquaCraniai piece Cranial pieces 103 21

Ourebia cf. ourebia Postcranial pieces 8Cranial pieces 5 3 Procavia transvaatensts

Tragelaphus ef. scriptus Cranial pieces 33Cranial pieces 5 4 Postcranial pieces 4 16

Tragetaphus d. strepsiceros Hyracoid indet.Cranial pieces 11 5 Cranial pieces 24

Tragelaphus sp. aff. angasi Postcranial pieces 2 10Cranial pieces 2 Order Lagcmorpha

Taurotragus ef. oryx Family LeporidaeCranial piece Lagomorph, gen. el sp. indet.

Antelope class 1 Cranial pieces IJ 4Cranial pieces 15 Order RodentiaPostcranial píeces 224 4 Family Hystricidae

Antelope class Ha Hystrix africaeaustralisCraniaJ pieces 152 Cranial pieces 8 ±5Postcranial pieces 2,708 75 Class Aves

Antelope class JIb Order StruthioforrnesCranial pieces 39 Family StruthionidaePostcranial pieces 419 ± 10 Struthio sp.

Antelope class JIJ Postcranial pieces 2Cranial pieces 42 Eggsbell pieces 2Postcranial pieces 203 7 Class Reptilia

Antelope c1ass IV Order ChelcniaCranial pieces 1 Family TestudinidaePostcraníal pieces 8 2 Chelonian indet.

Famíly Suidae Postcranial pieces 2 2Phacochoerus antiquus Carapace pieces 2

Crania! preces 8 4 Phylum Mol1uscaSuid indet. Class Gastropoda

Cranial pieces 2 Order PulmonataFamily Giraffidae Land snail, cf. Achatina sp.

Sivatherium maurusium SheUs 10 10Cranial piece Indet. fragments 723

Order Perissodactyla Bone ftakes 330Family Equidae 0-1 cm I

Equus quagga 1-2 cm 6Cranial preces 16 9 2-3 cm 38Postcranial pieces 5 l-4 cm 104

Equus capensis 4-5 cm 94Cranial pieces 8 5--<; cm 45Postcranial pieces 4 4 6-7 cm 19

Equid indet. 7--8 cm 12Cranial pieces 21 8+ cm 11Postcranial pieces 6 3 Total 5,894 392

Note: Corrected mínimum number of individuals, 258.

330 Appendix: Tables ~Table 103. Skeletal Part Analysis of Bovid Remains from Swartkrans Member 2 ,

Number of Specimens

Par! Size Class 1 Slze Class Il Size Class III Size Class IV

Calvaria pieces O I 1 OHom-core pieces 3 4 1 OMaxillary pieces O 1 2 OMandible pieces 10 3 14 IIsolated upper teeth O 12 11 Ilsolated lower teeth 2 17 7 OTooth fragments O 1 6 OVertebrae

Atlas 3 2 O OAxis 2 2 2 OCervical, ~7 7 9 7 OThoracic lO 10 6 OLumbar 4 24 5 OSacral O 3 1 OCaudal O O 2 O

Rib pieces 16 26 18 OScapula pieces 5 10 10 OPelvis píeces 6 4 3 IHumeros

Proximal pieces 1 2 6 ODistal pieces 6 7 14 OShaft pieces 17 5 2 O

Articulated distal humeros!proximal radius, and ulna O O O

Radius and ulna, proximalpieces 4 21 8

RadiusDistal pieces 2 4 4 OShaft pieces 23 15 6 O

FemurProximal pieces 3 8 7 IDistal pieces 3 4 5 2Shaft pieces 16 9 1 O

Patella O O 2 OTibia

Proximal pieces 6 5 1 ODistal pieces 4 17 5 OShaft pieces 5 8 2 O

MetacarpaJProximal pieces 5 16 5 ODistal pieces 5 18 3 OShaft pieces 19 18 1 O

MetatarsalProximal pieces 5 18 7 ODistal pieces 2 12 1 OShaft pieces 23 19 1 O

MetapodialProximal pieces O I O ODistal pieces 1 16 15 1Shaft píeces 3 6 O O

Astragalus 1 6 5 OCalcaneus 4 13 11 ITarsal bones 1 5 1 OCarpa! bones O 8 1 O1st phalanges 8 39 21 O2d phalanges 2 21 9 ITerminal pha1anges 2 7 3 O

Total 239 458 245 9

Appendix: Tables 331

Table 104. Skeletal Part Analysis nfBovid Remains (Class Ha) from Swartkrans Member 2

Number ofPart Specimens

Calvaria pieces 7Hom-core pieces 6Maxillary pieces 21Mandible pieces 41Isolated upper teeth 42Isolated lower teeth 35Vertebrae

Atlas 13Axis 10Cervical, 3-7 47Thoracic 68Lumbar 84Sacral 4Indet. vertebral pieces 5

Rib pieces 66Scapula pieces 112

Left 41Right 45? sirle 26

Pelvis pieces 45Lef! 16Right 18? side 11

Humerus. proximal píeces 26Left 9Right 9? sirle 8

Humerus. distal pieces 136Left 71Right 56? side 9

Humeros, shaft pieces. ? sirle 70Humeros, distal, artículated to radius and ulna,

proximal, left 1Radius and ulna, proximal 106

Left 35Right 62? side 9

Radius, distal pieces 69Lef! 33Right 30? side 6

Radius, shaft pieces 115Leñ 37Right 23? side 55

Fémur, proximal pieces 64Lef! 22Right 24? side 18

Part

Fémur. distal piecesLef! 17Rigbt 16? side 6

Femur, shaft piecesLeft 31Right 27? side 81

Tibia, proximal piecesLeft 20Right 20? side 6

Tibia, distal piecesLeñ 44Right 52? side 8

Tibia, shaft piecesLeft t5Right 14'] side 134

Metacarpals, completeLeft 2Right 3

Metacarpals, proximal piecesLef! 32Right 32? side 4

Metacarpals, distal pieces, ? sirleMetacarpals, shaft pieces, ? sideMetatarsals, proximal pieces

Left 53Right 6t'] side 21

Metatarsals, distal pieces, ? sideMetatarsals, shaft pieces, ? sideMetapodial, proximal pieces, '] sideMetapodial, distal pieces, ? sideMetapodial, shaft pieces, '] sideAstragalus

Lef! 40Right 22? sirle 4

CalcaneusLeft 37Right 46'] side 2

Tarsal bonesCarpal bones1st phalanges2d phalangesTerminal phalangesSesamoid bones

Total

Number ofSpecimens

39

139

46

104

163

5

68

77161135

58%2

827

66

85

1515

25040122

2,860

Note: Cranial parts represent only those specimens not assigned to specific taxa.

332 Appendix: Tables

Table 105. Percentage Survival oí Skeletal Parts Attributed lo Antídorcas bondi from Swartkrans Member 2

Number Original % Number Original %Part Found Number Survival Part Found Number Survival

Atlas vertebrae 13 70 18.6 Distal radius/ulna 69 140 49.3Axis vertebrae 10 70 14.3 Proximal femur 64 140 45.7Cervicals, 3-7 47 350 13.4 Distal femur 39 140 27.9Thoracic vertebrae 68 884 7.6 Proximal tibia 46 140 32.9Lumbar vertebrae 84 672 12.5 Distal tibia 104 140 74.3Sacra! vertebrae 4 70 5.7 Proximal metacarpal 74 140 52.9Caudal vertebrae O 300 O Distal metacarpal 123 140 87.9Ribs 66 1,820 3.6 Proximal metatarsal 136 140 97.1Scapula 112 140 80.0 Distal metatarsal 99 140 70.7Pelvic bone 45 140 32.1 Astragalus 66 140 47.1Proximal humeros 26 140 18.6 Calcanens 85 140 60.7Distal humeros 137 140 97.9 Phalanges 302 1,680 18.0Proximal radius/ulna 107 140 76.4

Note: A minimum of 70 individuals are represented by cranial remains.

Table 106. Observed Damage lo Bones from Swartkrans Member 2

Porcupíne-gnawed píecesAntidorcas bondi: mandlble, SK 7698: frontal. 2880; hom-core,

14215Hippotragus cf. níger: juvenile mandible, 1992Antelope class Ha: mandible, 2034; tibia, 6229: indet. shaft

píece.8232Antelope c1ass 111: hom-core, 9282; humeros, 2623; indet. shaft

piece,9267Antelope class IV, ealcaneus, 14199Indet. fragment, 3753

Pieces gnawed by small rodentsHippotragus cf. niger: juvenile mandible, 14243Procavia anticua: mandible, 202Large camivore: metapodial, 463Antelope c1ass 1: pelvis. 9604; radias, 11611; tibia, 7788;

metapodials, 4502, 5076, 5189, 5400, 5662, 7100, 7752, 8957;calcaneus, 6754; phalanx, 11672

Antelope class Ha: mandibles, 3124, 3116; lumbar vertebra,12738; scapula, 9082; humeri, 4343, 6308, 6339, 6345,7689,9351,9432; radii, 2653,10788,10811,12009,12087,12552;femurs, 1901,2618,2642,3629,5125,5299,6281,8730,9235,9284,10854,11787,14037; tibias, 1906,5027,5794,6196,6521,7390,8173,9305,9564; calcanei, 4130, 5271, 5276, 6735,8079, 10922, 11788, 12151, 12497: metapodials, 4594, 4709,5081,5245,5252,5309,5394,5600,5625,5635,5696,5702,5710,5714,5715,5763,5787,5799,5840,5895,7108,7695,7894,7924,8077,8186,8284,8648,9256,9304,9320,9596,9804, 10550, 10767, 10938, 10990, 11240, 11293, 11719, 11728,11978,12001,12036,12049,12058,12086,12484,2 un­numbered; phalanges, 1530, 1576,4376,4401,4493,5146,5264,5391,5876,6742,6746,6820,6830,6836,10079,10738,11593,11991: indet. shaft pieces, 5257, 7287, 7850, 8728,12406, 12492

Antelope c1ass IIb; Radius, 7384; femur, 14027; tibia. 10341;calcaneus, 11058; metapodial, 5825, 9210, 14177; phalanx,10434, unnumbered

Antelcpe c1ass IIl: humen, 4406,14031; femur, 1597; phalanx,

6732Bone flakes: 4182, 5413, 9877,12708Indet. pieces: 6016, 7142, 7415, 8289, 8383, 10049, 10470, 12083

Continued on following page

Pieces showíng camívore-infíicted domageIndeterminate piecesClear camivore damage: 4804, 5088, 6850, 6917, 7065, 7589,

9253,9332,9377,9537,10057,10430,10461,11460,11776Probable camivcre damage: 4601, 5761, 6366, 7142. 9168,

10641,11290,11972,12766Bone ñakes

Clear camivore damage: 5381,9068, 10208, 10937, 11838, 12124,12398, unnumbered

Carnivore pieces (sp. indet.)Clear carnivore damage: 6291, 6785, unnumberedAntelope c1ass 1Clear camivore damage: thoracie vertebra. 5364; humerii, 6391,

7797,10547,10567,11350; radii, 10366, 10888; femurs, 7539,7789,9770, 11211. unnumbered; tibia. 4223; metacarpals,8078,8482,9974, 11023; metatarsals, 9536, 11024, 11633

Probable carnivore damage: hcm-core. 9136; mandible, 11458;femur, 10346; metacarpals, 9905, 11594; metatarsals, 6676,8914,9232

AnteJope class JlaClear carnivore damage: mandible, 10410: atlas vertebra, 10234;

axis vertebrae, 3844, 7604, 10783; other cervical vertebrae,5451,9734; thoraeic vertebrae, 5062, 5564, 9651. 11572; lum­bar vertebrae, 5464, 10164, 10589, 10833, 10988, 11892: saocrum, 11382; scapulae, 9924, 10118, 10716, 11812, 12576,12580; pelvises, 6604, 6622, 7587, 9313, 9352, 9529,10884,10928, 11280; humerii, 2885, 4340, 4343, 4388, 4422, 4774,4826,5016,5296,5338,5593,6247,6306,6311,6313,6325,6353,6358,6370,6399,6404,6922,7174,7546,7772,8140,8475,9351,9370,9552,9588,9741,10034,10173,10495,10508,10587,10606,10747,10752,10813,11103,11180,11736, 11765, 12080, 12103, 12374, 12428, 12509, 12634,14018.14032. unnumbered; radh and ufnae, 4331, 4359, 5431,6140,6234,6409,7775,7891,9461,9684,10015,10016,10423,10846, 10975; femurs, 4116, 4122, 4246, 4278, 4358, 4710,4855,5053,5104,5117,5345,5361,5675,6251,6255,6263,6269,6280,6290,7147,7582,7698,8164,8185,8204,8228,8722,8857,9233,9327,9391,9395,9523,9549,9559,9685,10540,10581,10618,10854,10989,10999,11217,12028,12420,12488; tibiae, 4658, 5010, 5199, 5319, 6154, 6197, 6249,6965,7296,7351,7624,7755,8040,8950,9216,9375,9413,

Table 106. Continued

10673. 12202; calcanei, 5212, 6112,6780; metacarpals, 5216,5429,5616,5757,5811,9110,9788, 10219; metatarsals ; 4686,5215,5594,5596,5598,5669,5800,5847,5845,5859,7723,8205,9288,9441,10037,10196,10571,10732,10945,10978,11528,11810,11881,11978,12070,12111,12413, un­

numbered; metapodials, 5653, 8043, unnumbered; indet. shaftpieces, 1579,5970,6985,7006,8225,9826,9832, 10156,10247, 10488, 10893, 1t325

Probable camivore damage: mandible, 7724; thoracic vertebra,11361; lumbar vertebra, 14011; scapulae, 9082, 9430, 9848,10645,10886,11427,11797,12594; pelvises, 5089, 6006, 6638,11466; humeri, 5337, 6317, 6336, 9567,10222,10250,10264,10757,11909,12574,12587; radii, 6191, 6393, 8794, 9474;femurs, 1913,4472,4676,4725,5150,5198,6646,7479,80\2,8522,8668, 10011, 10387; tibiae, 4398, 4617, 4880, 5171, 7000,7148,7849,9266,9824,9868,11139,11421,12070,12150; as­tragalus, 11770; calcaneus, 6771; metacarpals, 5807, 5814,5869,9518,9692, 10086, 10781, 12358; metatarsals, 4143,4500,4627,4720,4903,5230,5612,5745,5752,5852,8099,8233,9438,9486,9643, 10102, 10475, 11333, 11535, 12104,unnumbered; indet. shaft pieces, 4872, 5242. 5772, 7004,7287,7405,9101,11150,12102

Note: AH specimen numbers are prefixed SK.

Antelope c1ass IIbC1ear camivore damage: scapula. 10227; ríbs , 5661. 7972;

radius, 7687; femurs, 6307, 9315,10996; calcaneus. 8854;metacarpals, 3368, 7710, 11337; metatarsals, 5648. 5677, 5698,10642

Probable eamivore damage: pelvis. 12040; radii and ulnae, 6163,12383; femur, 9611; tibia, 5025; metacarpals, 4567,9058;metatarsal, 10351; indet. shaft piece. 7074

Antelope c1ass 111Clear camívore damage: mandibles, 6063, 7476; axis vertebra.

11066; humeri, 2884. 6641,14127; radius. 10955; calcaneus,10151; metacarpal, 5857; metatarsal, 5753

Artificial cut marksOurebia cf. ourebí: mandible, 14168 (two cut marks 00 lower

margin of ramus)Antelope class IIb: proximal radius and ulna, 6136 (euts and

scratches)Indet. fragment: 4729 (cut marks)

Artificial chop marksIndet. shaft piece: 8289 (ehop marks)

Table 107. Remains from the Swartkrans Channel Fill: Oyeran Analysis, Taxa Represented, and MinimumNumbers of Individual Animals Involved

Minimum MinimumNumber of Numberof Number of Number of

Taxon Specimens Individuals Taxon Specimens Individuals

Phylum Chordata Order PerissodactylaClass Mammalia Family EquidaeOrder Primates Equus cf burchelli

Family Hominidae Cranial pieces 4Austrafopithecus robustus Postcranial pieces 3

Cranial piece Order HyracoideaOrder Carnívora Family Procaviidae

Family Felidae Procavia ef. capensisCf. Panthera pardus Cranial pieces 11

Cranial piece Postcranial pieces 16 5Postcranial piece Order Lagcmorpha

Family Canidae Family LeporidaeCanis mesomelas Gen. et sp. indet.

CraniaJ pieces 3 2 Cranial pieces 3 2Camivore Indet. Class Aves

Posteranial pieces 12 ±3 Order GalliformesOrder Artiodactyla Family Numididae

Farmly Bovidae Cf Numida sp.Medium alcelaphine, including Postcranial pieces 5Damaliscus sp. and Conno- Class Reptiliachaetes sp. Order Chelonia

Cranial pieces 17 5 Famíly TestudinidaeAntidorcas bondi Gen. et sp. indet.

Cranial pieces 6 3 Carapace pieces 3Antidorcas cf marsupialis Indeterminate fragmenta 193

Cranial pieces 5 2 Bone flakes 375Antelope class l 0-1 cm 2

Cranial pieces 4 1-2 cm 2~

Postcranial pieces 37 ±3 2-3 cm 97Antelope class 11 3-4 cm 106

Cranial pieces 33 4-5 cm 60Postcranial píeces 73 ±IO 5-6 cm 63

Antelope class 111 6-7 cm 9Cranial pieces 6 7-8 cm 4Postcranial pieces 24 ±3 8 + cm 5

Antelope class IV Total 837 44CraniaJ pieces OPostcranial pieces 1

334 Appendix: Tables 1- !~

Table 108. Skeletal Part AnaIysis of Bovid Remains from SwartkransChannel FiII

Number of Specimens

Part Size Class 1 Size Class II Size Class III Size Class IV

Calvaría pieces 2Horn-core pieces 6MaxiUary piecesMandible pieces 17 2lsolated upper teethIsclated lower teeth 3 5 1Tooth fragmenta 3 2Vertebrae

Atlas 3Axis 1Cervical. 3-7 3Thoracic 5 2LumbarSacra!Caudal

Rib pieces 1Scapula píeces 2 1Pelvis pieces 5 2Humerus

Proximal piecesDistal pieces 2 3 2Shaft pieces 2 1

Articulated distal humeros!proximal radius, and uIna

Radius and ulna, proximalpieces 4 3 2

RadiusDistal pieces 2Shaft pieces 1

FemurProximal pieces 1Distal pieces 2Shaft pieces

PatellaTibia

Proximal piecesDistal pieces 2 4Shaft pieces

MetaearpalProximal pieces 3 2Distal pieces 1Shaft pieces 1

Mc:tatarsalProximal piecesDistal piecesShaft pieces

MetapndialProximal pieces 2Distal píeces 4 9 1Shaft píeces 4

Astragalus I 3Calcaneus 1 5Tarsal bones 2Carpal bones 4 11st phalanges 10 22d phalanges 4 3 ITerminal phalanges 2 IArticulated foot bone pieces

Total 41 106 30

Appendix: Tables 335

Table 109. Observed Damage to Bonos fromSwartkrans ChanneJ FilI

Porcuptne-gnawed pieceShaft piece , SKW 2681

Pieces gnawed by small rodentsAntelope class 11, phalanx, 1390; indet. rib, 2424

Pieces showing camivore-infiicted damageClear camivore damageAntelope class 1: phaJanx 2381Antelope class Il: pelvis, 2352; phalanx, 2376; bone ñake,

3042Possible carnivore damageBone flakes, 3013, and 3088

Píeces showing artificial marksBone ftake: 3220 (wom edges)Antelope class 11: hom-core, 2628 (probable chop mark)

Note: Al! specimen numbers are prefíxed SKW.

Table 110. Remains from Kromdraai A: Overall AnaJysís, Taxa Represented, and Mínimum Numbers oí Indi­vidual Animals Involved

Taxon

Phylurn ChordataC1ass MarnrnaliaOrder Primates

Family CercopithecidaeParapapio jonesi

Cranial piecesPapio robinsoni

Cranial piecePapio angusticeps

Cranial piecesPostcranial pieces

Gorgopithecus majarCranial pieces

Cercopithecoid indet.Cranial piecesPostcranial pieces

Order CamivoraFamily Felidae

Panthera leoCranial pieces

Fetis crassidensCranial piece

Felís sp.Cranial piece

Homotherium sp.Cranial piece

Megantereon eurynodonCranial piecesPostcranial pieces

Continued on following page

NumberofSpecirnens

2

125

17

508

4

16

MínimumNumberofIndividuals

2

II

8

± 12

Taxon

Dínofetis píveteaui

Cranial piecesFamily Hyaenidae

Crocuta crocutaCranial piecesPostcranial pieces

Hvaena beííaxCranial piece

Hyaenid indet.Cranial pieces

Family CanidaeLycaon sp.

Cranial piecesCanis mesometas

Cranial piecesPostcranial piece

Canls terblancheiCranial piece

Vulpes pulcherCranial piece

Family ViverridaeHerpestes mesotes

Cranial piecesPostcraniaJ pieces

Crossarchus transvaalensisCraniaJ piece

Viverrid indet.Cranial pieces

NumberofSpecimens

3

753

9

2

17I

212

2

MinimurnNumber ofIndividuals

3

3

8

2

336 Appendix: Tables ,Table 110. Continued

Minimum MinimumNumber of Number of Number of Number of

Taxon Specimens Individuals Taxon Specimens Individuals

Carnivore indet. Order PerissodactylaCranial pieces 8 Family EquidaePostcranial pieces 39 ::!:IO Hipparion steytieri

Camivore coprolites 5 Cranial pieceOrder Articdactyla Equus burchelli

Family Bovidae Cranial pleces 5 5Damaliscus sp. 1 oc Parmularius Equus capensís

Cranlal píeces 166 32 Cranial pieces 54 23Mediurn-sized alcelaphines Postcranial pieces 5

Cranial pieces 35 12 Equid indet.Cf Connochaetes sp. Cranial pieces 17 ±4

Cranial pieces 17 5 Postcranial píeces 5Cf. Megalotragus sp. Order Hyracoidea

Cranial pieces 2 2 Family ProcaviidaeHippotragus cf. equinus Procavia antiqua

Cranial piece Cranial pieces 29 10?Hippotragíni Procavia transvaalensis

Cranial piece Cranial píeces 19 6Redunca cf arundinum Postcranial piece 1

Cranial piece Hyracoid mdet.Pelea cf. capreolus Craníal pieces 3 2

Cranial pieces 6 3 Order LagomorphaAntidorcas recki Family Leporidae

Cranial pieces 43 13 Lagomorph indet.Postcranial pieces 1 Cranial pieces 4 2

Antidorcas bond; Order RodentiaCranial pieces 12 6 Family Hystricidae

Rapíücerus sp. Hystrix africaeaustralisCranial piece Cranial pieces 6 3

Syncerus sp. Hystrix cf makapanensisCranial pieces 4 2 Cranial piece

Tragelaphus ce. scríptus Class AvesCranial piece CC. Columba sp.

Tragelaphus cf strepsiceros Eggshel1pieceCrarual pieces 8 5 Class Reptilia

Taurotragus cC.oryx Order CheloniaCranial pieces 2 2 Family Testudinidae

Antelope class [ Cf Testudo sp.Cranial pieces 7 Carapace piecePostcranial pieces 56 ±5 Phylum MoUusea

Antelope class 11 Class GastropodaCranial pieces 214 ± 12 Order PulmonataPosteranial pieces 299 CC. Achatina sp.

Antelope class [11 Shell pieces 6 6CraniaJ píeces 49 ±8 Indet. fragments 396Postcranial píeces 73 Bone ftakes 14

Antelope class IV 2-3 cm 1CraniaJ pieces 2 ~cm 1Postcranial pieces 2 4-5 cm 7

Family Suidae 5--6 cm 1Phacochoerus antiquus 6-7 cm 2

Cranial piece 7--8cm 1Tapinochoerus meadowsi ~9cm I

Cranial pieces 2 2 Total 1,847 253

Note: Corrected minimum numberoCindividuals, 194.

Appendix: Tables 337

Table 111. Skeletal Par! Analysis of Bovid Rernaíns from Kromdraai A

Number of Specimens

Part Size Class 1 Size Class II Size Class In Size Class IV

Calvaria pieces 4 2Hom-core pieces 2Maxillary pieces 10Mandible pieces 25 4lsolated upper teeth 1 62 12Isolated lower teeth 2 53 18Tocth fragments 3 60 11Vertebrae

Atlas 6Axis 4 2Cervical,3-7 1 3 1Thoracic 7 [4 2Lumbar 5 9Sacral 4 2Caudal 1

Rib pieces 16 10Scapula pieces 17 1Pelvis pieces 2 2Humeros

Complete rones 3Proximal pieces 2 17 7Distal pieces 2 12 1Shaft pieces 2 4

Articulated distal humeros/proximal radius, and ulna

Radius and ulna, proximalpieces 3 16 3

RadiusDistal pieces 2 4 3Shaft pieces 4 10 3

FémurComplete bones 1Proximal pieces 9 1Distal pieces 13 5Shaft pieces 2

TibiaProximal pieces 1 10Distal pieces 2 4Shaft pieces 4

MetacarpalComplete bones 1Proximal pieces 10Distal pieces 4 2Shaft pieces 11 1

MetatarsalProximal pieces 6 8Distal pieces I 6Shaftpieces 3 9

MetapodialProximal piecesDistal pieces I 6

Astragalus 1 10 2

Calcaneus 2 8 4

Tarsal bones 3Carpal bones 3 3lst phalanges 2 20 102d phalanges 1 12 3 2

Terminal phalanges 1 1

Articulated foot bone pieces 3 2Total 63 513 122 4

338 Appendix: Tables

Table 112. Fossil Assemblages from Kromdraai A andB: Contributions oflndividuaJ Animals ofVarious Taxato the Preserved Faunas

Krorndraai A Kromdraai B

Taxon " % " %

MarnmaliaPrimates

Hominidae 6 6.4Cercopithecidae 22 11.3 37 39.4

Carmvora 25 12.9 15 16.0Artiodactyla 94 48.5 26 27.7Perissodactyla 29 15.0Hyracoidea 16 8.2 1 l.lLagomorpha 2 1.0 2 2.1Rodentia 4 2.1

Aves 1 .5 2 2.1Reptilia 1 .5 5 5.3

Total 194 100.0 94 100.1

Table 113. Observed Damage lo Bones fromKromdraai A

Porcupíne-gnawed piecesProcavia transvaalensis: mandible, KA 4Antelope ctass 111: ulna, 1767Bone flake: 1305Indet. shaft preces; 1312, 1594

Pieces gnawed by small rodentsProcavia antíqua: maxiíla, 11

Antelope class Il: tibia, unnumbered

Pieces showing camivore-infíícted damageGorgapithecus major: mandible, 198 (punctate tooth marks)Antidorcas recki : metacarpal, 1J23b (clear tooth marks)Damaliscus sp. or Parmutarius sp.Clear carnivore damage: mandible, 566, 929; cranium with ar-

ticulated vertebrae, 731Probable camivore damage: mandible pieces, 1101, t739, 18270

Procovia transvaalensisClear camivore damage: mandible pieces, 1, 5Probable camivore damage: partial snout, 53; mandible, 4

Procavia antiquaClear carnivore damage: mandible pieces. 8, 17, 19Probable carnivore damage: mandible píeces. 6, 7Antelope class [Clear carnivcre damage: humeros, 646d; calcaneus, t619cAntelope class 11Clear carnivore damage: humeros, 724; metacarpals, 1343,

1619; tibiae, 591, 911.1135Probable camivore damage: humeri,S3l, 961; fémur, 513Indet. fragmentsClear carnivore damage: 1584bProbable camivore damage: 535, unnumbered

Note: AH specimen numbers are prefixed KA.Damage apparently caused by insects that bored into the

bone was observed on six specimens: 646, 883, 1069, 1204.1392. 1573.

Table 114. Microvertebrate Remains from Kromdraai A

MinimumNumber of

Taxon Individuals %

Order Chiroptera Iindet.) 3 1.1Order Insectívora

Family Chrysochloridae (indet.) .4Family Macroscetidae

Eleptumtulus sp. 9 3.3Family Soricidae (indet.) 22 8.1

Order RodentiaFamily Bathyergidee

Cryptomys sp. 2 .7Family Muridae

Subfamily Murinae, various genera 14 5.1Subfamily Otomyinae

Cf. Palaeotomys gracilis 21 8.0Family Cricetidae

Subfamily CrícetinaeMystromys sp. 69 25.3

Subfamily GerbíllinaeTatera sp. 2 .7

Subfamily Dendromurinae (indet.) 81 29.7Rodentia (Myomorpha) indet. 44 16.1

Class ReptiliaOrder Squamata

Lacertilia Indet. 2 .7Class Aves (indet.) 3 1.1

Total 273 100.3

,,

Appendix: Tables 339

Table 115. Remains frorn Kromdraai B: Overall Analysis, Taxa Represented, and Minimum Numbers ofIndividual Animals lnvolved

Layer 1 Layer 2 Layer 3 Total

Minimum Minimum Minimum MínimumNumber of Number of Number of Number uf Nurnber of Number of Number of Number of

Taxon Specimens Individuals Specimens lndividuals Specimens Individuals Specimens Individuals

Phylum ChordataClass MammaliaOrder Primates

Family HominidacAustra/opithecus robustus

Cranial pieces 6 6 6 6Postcranial pieces 2 2

Family CercopithecidaePapio robinsoní

Cranial pieces 75 11 7 4 7 3 89 18Papto angusticeps

Cranial pieces 17 9 2 7 4 26 14Cercopithecoídes wiíliamsi

Crenial pieces 7 2 6 3 13 5Cercopithecoid indet.

Cranial pieces 71 indet. 12 indet. 19 indet. 102 Indet.PoslcraniaJ preces 194 37 34 265

Order CamivoraFamily Felidae

Panthera pardusCraníal piecesPostcranial pieces 3 3 6 2

"Megantereon sp.Postcranial pieces 9 9

YDinofelis sp.Pcstcranial pieces

Family HyaenidaeHyaena cf brunnea

Cranial pieces 1 3Postcranial pieces 2

Proteíes sp.Cranlal pieces

Family CanídaeCanis sp.

Cranial pieces 2 4 2 7 4Postcranial pieces 2 2

Family ViverridaeHerpestes sp.

Cranial pleces 2 2Viverra sp.

Postcranial pieceCamivore indet.

Craníal pieces 14 indet. 2 Indet. indet 16 mdet.Postcranial pieces 24 2 26

Order ArtiodactylaFamily Bovidae

Connochaeres sp.Cranial pieces 2 2 4 3

Antídorcas cf. reckiCraníal píeces 3 2 6 2 or 3

Continued on [ollowing page

340 Appendix: Tables ,TabJe 115. Continued

'~

Layer I Layer 2 Layer 3 Total

Minimum Minimum Minimum MinimumNumber of Number of Number of Nurnber of Number of Number of Numbcr oCNumber of

Taxon Specimens Individuals Specimens Individuals Specimens Individuals Specirnens Individuals

Cf Antidorcas bondiCranial pieces 3 4 2

Gazella sp.Cranial pieces

lncertae sedisCranial pieces 3 3

Antelope class 1Cranial pieces 6 2 2 3 3 11 6Pcstcraeial pieces 7 2 7 26

Antelope c1ass 11Cranial picces 20 ~5 4 ~3 2 2 26 10Postcranial pieces 42 15 15 72

Antelope class IIICranial pieces 2 3 12 3 I 2 15 8Postcranial pieces 20 15 10 45

Antelope c1ass IVCranial piecesPostcranial pieces 2 2

Family SuidaePhocochoerus antiquus

Cranial piecesOrder Hyracoidea

Family ProcaviidaeProcavia sp.

Postcranial piecesOrder Lagomorpha

Family LeporidaeCf. Lepus sp.

Cranial pieces I 2 2Postcranial pieces 4 4

Class AvesStruthio sp.

Eggshell pieceBird: indet. large raptor

Postcranial pieces 2 2Class RepulíaOrder Squamata

Cf. Cordylus giganteusPostcranial pieces 8 8

Order CrocodiliaCf. Crocodylus nitoticus

Postcranlal píecesOrder Chelonia

Cf. Testudo sp.Pcstcrantal pieces 6 2 I 7 3

Coprolites 4 indet. indet. índet. 4 indet.Bone ñakes 1,512 indet. 627 indet. 748 indet. 2,887 indet.Indet. fragments 838 lndet. 212 indet. 224 indel. 1,275 indet ,Total 2,897 52 972 20 1,116 32 4,985 104

Tab

le11

6.Sk

elet

alPa

rtA

naly

sis

ofB

ovid

Rem

ains

from

Kro

mdr

aai

B

Num

ber

01Sp

ecim

ens

La

yer

1L

ayer

2L

ayer

3

Cla

ss1

Cla

ss11

Cla

ss11

1C

lass

1C

lass

IIC

lass

IUC

lass

1C

lass

11C

lass

111

Cla

ssIV

Tot

al I5

11

31

11

33

22

612

12

12I

135

Pan

Cal

vari

api

eces

Hom

-cor

epi

eces

Max

iUar

ypi

eces

Man

dibl

epi

eces

Isol

ated

uppe

rte

eth

Isol

ated

Iow

erte

eth

Too

thfr

agm

ents

Ver

tebr

aeA

tlas

Axi

sC

ervi

cal,

3-7

Tho

raci

cL

umba

rSa

cral

Cau

dal

Rib

piec

esS

capu

lapi

eces

Pelv

ispi

eces

Hum

eros

Prox

imal

piec

esD

ista

lpi

eces

Shaf

tpi

eces

Art

icuJ

ated

dist

alhu

mer

os/p

roxi

mal

radi

usan

dul

naR

adiu

san

dul

na,

prox

imal

piec

esR

adiu

sD

ista

lpie

ces

Shaf

tpi

eces

Con

tinu

edon

foll

owin

gpa

ge

2 6 1 10 2

9 2

2

2 I 2

2 11 3 2 21 2 4 4 5

Tab

le11

6.C

ontí

nued

Num

bero

fS

peci

men

s

Lay

er1

Lay

er2

Lay

er3

Pan

Cla

ss1

Cla

ssII

Cla

ssII

IC

lass

1C

lass

11C

lass

III

Cla

ss1

Cla

ss11

Cla

ss11

IC

lass

IVT

otal

Fem

urPr

oxim

alpi

eces

Dist

alpi

eces

Shaf

tpi

eces

Pat

eHa

Tib

ia Prox

imal

piec

esD

ista

lpie

ces

Shaf

tpi

eces

II

2I

5M

etac

arpa

lPr

oxim

alpi

eces

Dis

talp

iece

sS

haf

tpíe

ces

Met

atar

sal

Prox

imal

piec

esI

12

Dis

talp

iece

sSh

aft

piec

esM

etap

odia

lPr

oxim

alpi

eces

Dis

talp

iece

s2

II

II

6Sh

aft

piec

esA

stra

galu

s1

2I

4C

alca

neus

14

II

II

9T

arsa

lbon

esI

J2

Car

pal

bone

sI

2I

15

1stp

hala

nges

72

I4

44

5I

282d

phal

ange

s1

2I

22

8T

enni

naJ

phal

ange

s2

32

32

113

Sesa

moi

dJ

I1

3A

rtic

ulat

edfo

otbo

nepi

eces

Tot

al13

6222

419

2720

17I1

219

7

..

Table 118. Cercopithecoid Remains from Krorndraai B: Minimum Numberof Individual, per Age Class

Total NumberAge Ctass Layer t Layer 2 Layer 3 oí Individuals

Papio rohinsoniJuveniles 7 2 2 11Immature adultsYoung adulta 1 2Adults 2 4üld adults \ l

Total 11 4 3 18

Papio angusticepsJuveniles 1 3Immature adults 1 \Young adults 1 2 4Adults 4 I 5Qld adults 1 1

Total 8 2 4 14

Cercopíthecoídes wiílíamsíJuveniles 2Immature adultsYcung adults 2AdultsOíd adults \ 1

Total 2 O 3 5

344 Appendix: Tables "s

Table 119. List of Skeletal Parts from Cercopithecoids from Kromdraai B ThatHave Not Been Specífically Identified

Par! Layer I Layer 2 Layer 3 Total

Fragment ofCalvaria 40 14 54MaxillaMandible 1 tTooth 30 12 5 47Vertebrae

Cervical 7 7Thoracic 2 3Lumbar 1 2 4Sacra!Caudal 33 2 6 41

Ríb 52 S2Scapufa 1 2Pelvis 2 3 6Humeros

ProximalDistal 2 3Shaft 1

RadiusProximal 4 2 6Distal 1 1Shaft 3 4

UlnaProximal 4 2 6DistalShaft

FemurProximal 6 8Distal 2 4Shaft I 3

TibiaProximal 1 1Distal 3 3Shaft

Metapodial 10 S 3 18Calcaneus S 2 2 9Astragalus 2 I 3Phalanges SO 11 6 67Miscellaneous 9 4 13

Total 265 49 53 367

Tab1e 120. Skeletal Parts from Unidentified Carnivores from Kromdraai B

Part Layer I Layer 2 Layer 3 Total

Fragment 01Tooth 10 2 12Sku\l 4 4Pelvis 1Proximal ufna 1Metapodial 9 9Phalanges IS IS

Total 38 4 42

Table 121. Observed Damage to Bones fromKrorndraai B

Porcupine-gnawed pieceCarnivore mandible piece, KB 2884

Pieces showing camivore tooth marksAustralopíthecus rohustus: pelvis piece , TM 1605(the ragged­

edged margins probably represent camivore damage)Antelope c1ass II: tibia, 3283; shaft pieces, 2275, unnumbered

Pieces showing rounding and wearProte!es sp.: mandible, 2945.2Cercopithecoid: isolated teeth, 247, 996, 1396, unnumbered;

phalanx, 1196; caudal vertebrae, 429, 450Bovíd hom-cores, 127,378; miscellaneous pieces, 20, 137, 243,

403,592,647,872,942,1342,1377,1459,1709,1907,2016.2049.2123.2224,2329,2675,2765, unnumbered

Bone flakes. 60, 838, 930,1024,1056,1367,1397,1491,1524.1585,1601,1693,1758,1867,1892,1983,2116,2131,2294,2414.2565,2592,2653,2670,2677,2690,2705,2867, II un­numbered pieces

Lengths 01 bone píeces showing rounding and wearLength of piece

(cm) N

o-r I1-2 282-3 263-4 84--5 65-6 ITotal 70

Note: AH specimen numbers are preñxed KB unless otherwisenoted.

Appendix: Tables 345

References

Acocks, l. P. H. 1953. Veld types of South Afriea. Bo­tanical Survey S. Afr. Mem, 28:1-192.

Alexander, A. J. 1956. Bone carrying by a porcupine. S.Afr. J. Scí. 52 (11): 257-58.

Alexander, J. E. 1838, An expeditíon 01 díscovery into theinterior 01 Africa, through the hitherto undescribedcountries al/he Great Namaquas. Boschmans and HillDamaras. London: Henry Colbum.

Ansell, W. F. H. 1971. Order Artiodaetyla. In The mam­mals of Africa: An identificarían manual, ed. J. Mees­ter and H. W. Setzer, part 15. Washington, O.e.:Smithsonian Institution Press.

Arambourg, e. 1947. Contribution á l'étude géologique etpaléontologique du bassin du Lae Rudolphe et de labasse vallée de l'Omo: Deuxíéme portie, paléontologie.Mission Sci. Omo 1932-1933, T. Géol. Anth, 1 (3):231-562.

Arkell, A. J. 1957. A possibly palaeolithie bone spatulafrom Egypt. Proc. Prehist, Soco 23:234-36.

Armstrong, A. L. 1936. A bull-roarer of Mousterian agefrom Pin Hole Cave, Cresswell Crags. Antiqua. J.(London) 16:322-23.

Astiey Maberly, C. T. 1953. A plea for the leopard. Afr.Wild Life 7 (1): 19-27.

Bakr, A. 1959. Ph.D. diss. Biology Dept., Harvard Uni­versity.

Balestra, F. A. 1962. The rnan-eating hyaenas of Mlanje.Afr. Wild Life 16:25-27.

Banks, R. C. 1965. Some information from barn owl pel­lets. Auk 82:506.

Bateman, J. A. 1960. Owl habits from pellets. Bull. Zool.Soco S. Afr. 1 (3): 2(}...21.

Bates, H. J. 1971. Baboons. Afr. Wild Life 25 (4): 154.

t' /i Bearder, S. K. 1977. Feeding habits of spotted hyaenas ina woodland habitat. E. Afr, Wildl. J. 15:263-80.

Beaumont, P. B. 1973. Border Cave: A progress report.S. Afr, J. Sci. 69:41-46, ,,_

"'! Behrensmeyer, A. K. 1975;>,. The taphonomy andpaleoecology of Plio-Pleistocene vertebrate as­semblages east of Lake Rudolf, Kenya. Bull. Mus.Comp-,ZDQI., Harv. 146:473-578.

---o '1975h. Taphonomy and paleoecology in thehominid fossil record. Yearb. Phys. Anthrop . 19:36-50.

Benson, e. W. 1962. The food of the spotled eagle-owlBuba africanus. Ostrich 33:35.

Benson, C. W., and Irwin, M. P. S. 1967. The distributionand systematics ofBubo capensis Smith (Aves). Arnol­dio (Rhodesia) 3 (19): 1-19.

Bere, R. 1966. The African elephant. New York: GoldenPress.

Best, A. A., ed. 1973. Rowland Ward's records of biggame, 15th ed. (Africa). London: Rowland Ward, Ltd.

Bigalke, E. 1973. The exploitalion of shel1fish by eoastaltribesmen of the Transkei. Ann. Cape Prov. Mus.(Nal. nt«¡ 9 (9): 159-75.

Binford, L. R., and Bertram, J. B. 1977. Bone!Ufrequencies-and attritional processes. In Por theorybuilding in archaeology: Essays on fauna! remains,aquatic resaurces, spatial anaiysis and systematicmodeling, ed. L. R. Binford, pp. 77-153. New York:Academic Press.

Boaz, N. T., and Behrensmeyer, A. K. 1976. Hominid -:-,taphonorny: Transport of human skeletal parts in anartificial fluviatile environment. Am. J. Phys. Anthrop.45 (1): 53-60.

Boaz, N. T. and Howel1, F. C. 1977. A gracile hominidcranium from upper Member G of the Shungura For­mation, Ethiopia. Am. J. Phys. Anthrop. 46: 93-108.

Bodenheimer, F. S. 1949. Prohlems ofvole populatíons inthe Middle East, Report on the population dynamies ofthe Levant vole. Jerusalern: Research Council ofIsrael.

Bond, G., and Summers, R. 1951. The Quartemary suc­cession and archaeology at Chelmer near Bulawayo,Southern Rhodesia. S. Afr. J. Sci. 47:20(}"'204.

Boné, E. L., and Sínger, R. 1965. Hipparion fromLangebaanweg, Cape Province, and a revisión of thegenus in Africa. Ann. S. Afr. Mus. 48:273-397.

Bonnichsen, R. N .d. Sorne operational aspects of humanand animal bone alteration. In Mammalian osteo­archaeology: North America, ed. B. M. Gilbert. Co­lumbia, Mo.: Missouri Archaeologieal Society.

Bothma, J. du P. 1971. Order Hyracoidea, In The mam­mals 01 Afríca: An ídentifícation manual, ed. J. Mees­ter and H. W. Setzer, pan 12. Washington O.e.:Smithsonian Instítution Press.

Brain, C.K. 1958. The Transvaal ape-man-bearing cavedeposíts, TransvaaJ Museum Memoirs, no. 11.

---o 19670. The Transvaal Museum's fossil project atSwartkrans. S. Afr, J. Scí. 63:37S-84.

---o 1967b. Hottentot food remains and their bearingon the interpretation of fossil bone assemblages. Sct.Papers Namih Desert Res. Statton 32:1-11. .)

---o 1967c. Bone weathe~ng and lhe problem of bone ()/5pseudo-tools, S. Afr, J. SC•. 63 (3): 97-99. (

---o t%9a. New evídence for climatic change duringMiddle and Late Stone Age times in Rhodesia. S. Afr.

347

348 References

Arch. Bull. 24:127-43.--~. ] 969h. The contribution of Namib Desert Hot­

tentots to an understanding of australopithecine boneaccumulations. Sci. Papers Namih Desert Res. Sta/ion39:13-22.

---o 1969c. Faunal remains from the Bushman RockShelter, Eastem Traosvaal. S. Afr. Archaeol. Bull. 24(94): 52-55.

---o 1970. New fiods at the Swartkraos australopíthe­cine site. Nature (London) 225 (5238):1112-19.

---o 1972. Ao attempt to recoostruet the hehaviour ofaustraJopithecines: The evidence for interpersonalviolence. Zool. Afr. 7 (1): 379-401.

---o 1973. Seven years' hard labour at Swartkrans.Bull. Transv. Mus. 14:5-6.

---o ]9740. Sorne suggested procedures in theanalysis of bone accumulations from southem MricanQuatemary sites. Ann. Transv. Mus. 29 (1): 1--1l.

---o 1974h. The use of microfaunal rernains as habitatindicators in the Namib. S. Afr. Archaeol. Soc., Good­win ser. no. 2, pp. 55-60.

---o 19740. Human food remaios from the Iroo Age atZimbabwe. S. Afr. J. Sct, 70:30J..-9.

---o 1975a. An introduction to the South African aus­tralopithecine bone accumulations. In Archaeologicalstudies, ed. A. T. Clason, pp, 109-19. Amsterdam:North Holland.

---o 1975h. Ao interpretation of the booe assemblagefrom the Kromdraai australopithecine site, South Af­rica. In Palaeoanthropoíogv, morphology andpaleoecology, ed. R. H. Tuttle, pp. 225-43. The Hágue:Mouton.

---o ]976a. A re-interpretatíon of the Swartkrans siteaod its remains. S. Afr. J. Sci. 72:141-46.

---o 1976h. Some priociples in the ioterpretatioo ofbone accurnulations associated with mano In Humanorígins: Louis Leakey and tñe East Afrícan evidence,ed. G. L1. Isaac and E. R. MeCown, pp. 96-116. MenloPark, Calif.: Staples Press.

---o 1978. Some aspects of the South Afríean austra­lopithecine sites and their bone accumulations. In Afri­can Homlnídae of the Plio-Píeistocene, ed. C. Jolly.Londoo: Duckworth.

Brain, C. K., and Brain, V. 1977. Microfaunal remainsfrom Mirabib: Some evidenee of palaeoecologícalehanges in the Namib. Madoqua lO (4): 285--93.

Brain, C. K., and Cooke, C. K. 1967. A prelimioary ac­eount of the Rede1iff Stooe Age cave site io Rhodesia.Bull. S. Afr, Archaeol. Soco 21 (84): 171--1l2.

Braio, C. K., and Meester, 1.1964. Past elimatie changesas bíological isolating mechanisrns in southem Africa.In Ecological studies in Southern Afdea, ed. D. H. S.Davis, pp. 332-40. Deo Haag: lunk Publíshers.

Brelsford, V. 1950. Unusual events in aoimal life-IV.Afr. Wild Life 4 (1): 67.

Brock, A.; MeFadden, P. L.; and Partridge, T. C. 1977.Pre1iminary palaeomagnetie results from Makapaosgatand Swartkrans. Nature (London) 266:249-50.

Brooke, R. K. 1967. On the foodofthe woodowl. Ostrich38 (1):55.

---o 1973. Notes 00 the distribution aod food of theCape eagle owl in Rhodesia. Ostrich 44:137-39.

Broom, R. 1907.00 sorne new species ofChrysoehloris.Ann. Mag. Nat. Hist. 19:262~.

---o 1909a.On a large extinct species ofBubalís. Ann.S. Afr. Mus. 7:279--1l0.

---o 1909b. On evidence of a large horse recently ex­tinct io South África. Ann. S. Afr. Mus. 7:281-82.

---o ]9130. Man contemporaneous with extinct ani­mals in South África. Ann. S. Afr. Mus. 12:13-16.

---o 1913b. Note on Equus capensis, Bull. Am. Mus.Nat. Hist. 32:437-39.

---o 1925. 00 evidence of a giant pig from the LaterTertiaries of South Afríea. Rec. A/bany Mus. 3:307--1l.

---o 1928. 00 sorne new marnmals from the diamondgravels of the Kimberley district, Ann. S. Afr, Mus.22:439-44.

---o 1931. A new extinct giant pig from the diamondgravels oí Windsorton, South Africa. Rec. Albany Mus.4:167~.

---o 1934. Do the fossil rernains associated with Aus­traiopithecus afrícanus, S. Afr. J. Sci. 31:471--1l0.

---o 1936. A oew fossil anthropoid skull from SouthAfrica, Nature (Londoo) 138 (3490): 486--1l8.

---o 19370. Díscovery of a lower molar of Australo­pithecus, Nature (London) 140:681--1l2.

---o 1937b. Notes of a few more new fossil mammalsfrom the caves ofthe Transvaal. Ann. Mag. Nat. Hist.20 (10): 509-14.

---o 1937c. On sorne new Pleistocene rnarnmaIs fromlimestone caves in the Transvaal. S. Afr. J. Sct,33:75()-68.

---o 1938a. More discoveries of Austraíopíthecus,Nature (London) 141:82S-29.

---o 1938h. Tbe Pleistoeeoe aothropoid apes of SouthAfrica, Nature (London) 142 (3591): 377-79.

---o 193&. Further evidence of the strueture of theSouth African Pleistocene anthropoids. Nature (Lon­doo) 142 (3603): 897-99.

---o 1939a. A pretiminary account of the Pleistocenecarnivores of the Transvaal caves, Ann. Transv: Mus.19:331-38.

---o 1939b. A restoratioo of the Kromdraai skull.Ann. Transv, Mus. 19 (3): 327-29.

---o 1940. The South Mrican Pteistocene cer­copíthecoíd apes. Ann. Transv. Mus. 20:89-100.

---o 19410. Structure ofthe Sterkfontein apeo Nature(London) 147:86.

---o 1941h. Mandible of a young Purunthropus ehild.Nature (Loodoo) 147:607--1l.

---o 1941e. 00 two Pleistocene golden moles. Ann.Transv. Mus. 20:215--16.

---o 1942. Tbe hand of the ape-rnan, Paranthropusrobustus, Nature (London) 149:51J..-14.

---o 1943. Ao ankle-bone of the ape-rnan, Paran­thropus rohustus. Nature (London) 152:689-90.

---o 1947. Diseovery of a new skull of!he South Afri­can ape-rnan, Ptesianthropus. Nature (London)159:672.

---o 19480. Some South Afríean Plioceoe and Pleis­tocene marnmals. Ann. Transv. Mus. 21:1-38.

---o 1948b. The gianl rodeot mole, Gypsornvchus.Ann. Transv. Mus. 21:47-49.

---o 1949. Another oew type of fossil ape-rnan. Na­ture (London) 163:57.

---o 1950. Finding the missíng link. London: Watts.Broom, R., and leosen, 1. S. 1946. A new fossil baboon

from the caves at Potgietersrust. Ann. Transv. Mus.20:337-40.

Broom, R., and Le Riehe, H. 1937. The deotitioo ofEquus capensis Broom. S. Afr. J. Sci. 33:769-70.

Broom, R., aod Robioson, J. T. 19470. Two fealures of

thePlesianthropas skull. Natare (London) 159:809-10.---, 1947h. Further remains of the Sterkfontein ape­

rnan, Plesianthropus, Nature (London) 160:43Ci-31.---o 1949a. A new type of fossil mano Nature (Loo­

don) 164:322.---o 1949b. Thumb of the Swartkrans ape-rnan. Na­

tare (London) 164:841-42.---o 1949c. A new type of fossil babeen, Gor­

gopíthecus major. Proc. Zool. Soco Lond. 119:379-1l6.---o 1950. Man contemporaneous with the Swartkrans

ape-rnan. Am. J. Phys. Anthrop., n.s., 8:151-56.---o 1952. Swartkrans ape-man Paranthropus crassi­

denso Transvaal Museum Memoirs, no. 6.Broorn, R.; Robinson, J. T.; and Schepers, G. W. H.

1950. Part 1: Furtñer evidence of the structure of theSterkfontein ape-rnan Plesianthropus. Part 2: The braincasts 01 the recently discovered Plesianthropus skulls.Transvaal Museum Memoirs, no. 4.

Broom, R., and Schepers, G. W. H. 1946. The SoathAfricanfossil ape-rnen: The Australopithecinae. Trans­vaal Museum Memoirs, no. 2.

Brothwell, D. R. 1963. Digging up bones. London:British Museum (Natural Hístory).

Brown, L. H. 1952. On the biology of the large birds ofprey of Embu District, Kenya Colony. Ibis 94:577~20.

---o 1955. Supplementary notes on the biology of thelarge birds of prey of Embu District, Kenya Colony.Ibis 97:3~.

---, 1965. Observations on Verreaux's eagle owI inKenya. J. E. Afr. Nal. Hist. Soc. 25 (2): 101-7.

---o 1966. Observations 00 sorne Kenya eagles. Ibis\08:531-72.

---o 1970. Afrícon bírds ofprey. London: Collins.Buckland, W. 1822. Account of an assemblage of fossil

teeth and bones of elephant, rhinoceros, hip­popotarnus, bear, tiger, and hyaena and sixteen otheranimals: Discovered in a cave al Kirkdale, Yorkshire,in the year 1821; with a comparatíve view offive similarcavems in various parts of England and others 00 theContinent. Phi/o Trans. Roy, Soco Lond. 112:171-237.

---o 1823. Reliquae diluvianae; or Observations onthe organic remains contained in cave físsures, andother diluvial gravel, and on other geologicalphenomena attesting the action 01 a universal deluge.2d ed. London: Murray.

Buettner-Janusch, J. 1966. A problem in evolutionarysysternatics: Nomenclature and classification of ba­boons, genus Papio. Folia Primal. 4:288-308.

Butler, P. M., and Greenwood, M. 1973. The early Pleis­tocene hedgehog from Olduvaí, Tanzania. Foss. Vert.Afr. 3:7-42.

---o 1976. Elephant shrews (Macroscelididae) fromOlduvaí and Makapansgat. Foss. Verl. Afr. 4:1-56.

Butzer, K. W. 1974a. Geology of lbe Comelia Beds,northeastem Orange Free State. In The geology, ar­chaeology and fossil mammals 01 the Comelia Beds,O.F.S., pp. 7-32. Mem. National Mus., Bloemfontein,no. 9.

---o 1974b. Paleoecology of South African australo­pithecines. Taung revisited. Curro Anthrop. 15 (4):367-112.

--o 1976. Lithostratigraphy of the Swartkrans For­mation. S. Afr. J. Sci. 72:136-41.

Campbell, A. M. 1959. Apple-scoops. Nalare (London)183:1542.

Camegie, A. J. M. 1961. The stomach contents of a spot-

References 349

ted eagle owl (Baba africanas). Ostrich 32:97.Cárter, P. L. 1970. Late Stone Age exploitation patterns

in southem Natal. S. Afr, Archaeol. Bull, 25 (98):55-58.Chaplin, R. E. 1971. A study of animal bones from ar­

chaeological sites. London and New York: SeminarPress.

Chave, K. E. 1964. Skeletal durability and preservation.In Approaches lo paleoecology. ed. J. Imbrie and N.Newel, pp. 377-117. New York: John Wiley.

Chitty, D. 1938. A laboratory study ofpellet formatlcn inthe short-eared owl. Proc. Zool. Soco Lond., ser. A,108:2.

Churcher, C. S. 1956. The fossil Hyracoidea of theTransvaal and Taungs deposits. Ann. Transv, Mus. 22(4): 477-501.

---o 1970. The fossil Equidae from the KrugersdorpCaves. Ann. Transv. Mas. 26 (6): 14~.

---o 1974. Sivatherium maurusium (Pornel) from theSwartkrans australopithecine site, TransvaaJ (Mam­malia: Girallidae). Ann. Transv. Mas. 29 (6): 65~9.

---o 1978. Giraffidae. In Evolution 01 Afrícan mam­maís, ed, V. J. Maglio and H. B. S. Cooke, pp. 509-35.Cambridge: Harvard University Press.

Churcher, C. S., and Richardson, M. L. 1978. Equidae.lnEvolulion 01 African mammals, ed. V. J. Maglio andH. B. S. Cooke, pp. 379-422. Cambridge: HarvardUniversity Press.

Clancey, P. A. 1964. The birds of Natal and Zala/and.Edinburgh: Oliver and Boyd.

Clark, J. D. 1972. Palaeolithic butchery practices, InMan, seuíemeru and urbanism, P. J. Ucko, R. Tring­ham, and G. W. Dimbleby, pp. 149-56. London:Duckworth.

---o 1974. The stone artefacts from Comelia, O.F.S.,South Africa. In The geolagy, archaeology and fossíímammats of the Comelia Beds, O.F's.. pp. 3~1.

Mem. NationaJ Mus., Bloemfontein, no. 9.Clark, J. D., and Haynes, C. V. 1970. An elephant butch­

ery site at Mwanganda's Village, Karonga, Malawi, andits relevance for palaeolithic archaeology. World Ar­chaeology 1 (3): 39Ci-41 J.

Clark, R. J. 1972. Pellets ofthe short-eared owl and marshhawk compared. J. Wildl. Mgml. 36:962--64.

Clark, W. E. Le Gros. 1978. The fossil evidence forhaman evolutíon, 3d. ed .. ed. B. Campbell, Chicago:University of Chicago Press.

Clarke, R. J. 1977. A juvenile cranium and sorne adultteeth of early Homo from Swartkrans, Transvaal. S.Afr. J. su. 73:46-49,

Clarke, R. J.; Howell, F. C.; and Brain, C. K. 1970. Moreevidence of an advanced hominid al Swartkrans. Na­lare (London) 225:1217-20.

Clason, A. T. 1972. Sorne remarks on the use andpresentation of archaeozoological data. Helinium12:139-53.

CoetlOe, C. G. 1963. The prey of owls in the KrugerNational Park as indicated by owl pellets collectedduring 196Ci-6J. Koedoe 6:115-25.

---o 197J. Camivora (excluding the family Felidae).In The mammals of Aldca: An identification manual,ed. J. Meester and H. W. SetlOr. Washington, D.C.:Smithsonian Institution Press.

---o 1972. The identification of southem African smallmammal remains in owl pellets. Cimbebasia, ser. A.,2:5~.

Colbert, E. H. 1935. Siwalik mammals in the American

350 References

Museum of Natural History. Trans. Am. Phi/o Soc ..n.s.. 26:1-401.

Collings, G. E. 1972. A new species of machaerodontfrom Makapansgat. Palaeont, Afr. 14:87-92.

Collings, G. E.; Cruíckshank, A. R. 1.; Maguire, J. M.;and Randall, R. M. 1976. Reeenl faunal studies atMakapansgat Limeworks. Transvaaí, South Africa.Ann. S. Afr. Mus. 71:153-{j5.

Cooke, C. K. 1963. Report 00 excavations at Pomongweand TshanguJa caves, Matopo HilJs, SouthemRhodesia. S. Afr. Archoeol. Bull. 18 (71); 73--151.

---o 1966. Reappraisal of the industry hitherto namedProto-Stillbay. Amoldia (Rhodesia) 2 (22): 1-14.

---o 1968. The Early Stone Age in Rhodesia. Amoldia(Rhodesía) 3 (39): 1-12.

---o 1%9. A re-examination of the "Middle StoneAge' industries of Rhodesia. Amoldia (Rhodesia) 4 (7):1-20.

Cooke, H. B. S. 1938. The Sterkfontein bone breccia: Ageological note. S. Afr. J. Sci. 35:2~.

---o 1941. A preliminary aceount of the WonderwerkCave, Kuruman District. Section II: The fossil remains.S. Afr. J. Sci. 37:300--12.

---o 1943. Cranial and dental characters of the recentSouth Afrieao Equidae. S. Afr. J. Sci. 40:254--57.

---o 1947. Sorne fossil hippotragine antelopes fromSouth Afriea. S. Afr. J. Sci, 63:226-31.

---o 19490. Fossil mammals of the Vaal River de­posits. Geol. Surv. S. Afr, Mem. 35 (3): 1-117.

---o 1949b. The fossil Suina of South Africa, Trans,Roy. Soc. S. Afr. 32: 1-44.

---o 1950. A critical revisión of the Quaternary Peris­sodactyla of southem Afriea. Ann. S. Afr. Mus.31:393-479.

---o 1%0. Further revision of the fossil Elephantidaeof southern Afriea. Palaeont. Afr. 7:46-58.

---o 1962. Notes on (he faunal material from the Caveof Hearths and Kalkbank. In Prehistory of tñe Trans­vaaí, ed. R. 1. Mason, pp. 447-53. Johannesburg: Wit­watersrand University Press.

---o 1963. Pleistocene mammal faunas of Afriea, withparticular reference to southern Afríca. In Africanecology and human evolutton, ed. F. C. Howell and F.Bourliere, pp- 65-116. Viking Fund Pubhcations in An­thropology. no. 36. New York: Wenner-Gren Founda­tion.

--o 1974. The fossíl maromals of Camelia, O.F.S.,South Afriea. In The geoiogy, archaeology and fossilmamma/s o[ the Cornelia Beds, O.F.S., pp. 63-84.Mem. National Mus., Bloemfontein, no. 9.

---o 1976. Suidae from Plio-P1eistocene strata of theRudolf basin. In Earliest man and env;ranments in theLake Rudo/{ basin, ed. Y. Coppens, F. C. Howel!,G. Ll. Isaac, and Richard E. F. Leakey, pp. 251-{j3.Chicago: University of Chicago Press.

---o N .d. South Mocan Pleistocene mammaJs in theUniversity of California collections. Typescript.

Cooke, H. B. S., and MagUo, V.]. 1972. Plio-Pleisloeenestratigraphy in East Africa in relation to proboseideanand suid evolution. In Calibratían o/ hominoid evo/u­/ion, ed. W. W. Bishop, and J. A. Miller, pp. 303-29.Edinburgh: Seoltish Aeademie Press.

Cooke, H. B. S., and Wells, L. H. 1947. Fossil mammalsfrom the Makapan valley, Polgietersrust. IIl.Giraffidae. S. Afr. J. Sci. 43:232-35.

---o 1951. FossiJ remains from Chelmer, nearBulawayo, Southem Rhodesia. S. Afr. J. Sci. 47:205--9.

Cooke. H. B. S., and Wilkinson, A. F. 1978. Suidae andTayassuidae. In Evolution 01 African mammuls, ed,V. r. Maglio and H. B. S. Cooke, pp. 435-82. Cam­bridge: Harvard University Press.

Coppens. Y., and Howell, F. e. 1976. Mammalian faunasof the Omo group: Distributional and biostratigraphicalaspects. InEarliest man and environments in the Laí:eRudo/{ basin, ed. Y. Coppens, F. C. Howel!, G. Ll.Isaac, and R. E. F. Leakey, pp. 177-92. Chicago: Uni­versity of Chicago Press.

Coppens, Y.; Maglio, V. J.; Madden, C. r.; and Beden,M. Proboseidea. InEvoíution ofAfrican mammals, ed.V. 1. Maglio and H. B. S. Cooke, pp. 33W7. Cam­bridge: Harvard University Press.

Corbet, G. B. 1971. Family Macroseelididae. In Themammals 01 Africo: An ídesuifícation manual, ed. J.Meester and H. W. Setzer. Washington, D.e.: Smith­sonian Institution Press.

Corbet, G. B., and Hanks, J. 1968. A revision of theelephant shrews, family Macroscelididae. Bull. Brit,Mus. Nat, Hist. (Zool.) 16:47-111.

Corbett, J. 1954. The man-eating leopard of Rudro­prayag. London: Oxford Uníversity Press.

Comwall, 1. W. 1956. Bones for the Archaeologist. Lon­don: Phoenix House.

Coryndon, S. C. 1964. Bone remains in the caves. In Thelava caves ofMount Suswa, Kenya, ed. P. E. Glover etal. Ssudíes in Speleology 1 (1): 6(}.63.

Crass, R. S. 1944. The birds of Sulenkama, Qumbu Dis­triet, Cape Province. Ostrich 15 (1): 10-20.

Cullen, A. 1969. Window 01110 ....iíderness, Nairobi: EastAfrican Publishing House.

Cummings, J. H.; Duke, G. E.; and Jegers, A. A. 1976.Corrosion of bone by solutions simuJating raptor gastricjuice. Raptor Researeh 10 (2): 55--57.

Dale, M. M. 1948. New fossil Suidae from the LimeworksQuarry, Makapansgat, Potgietersrust. S. AIr. J. Sei.2:114--16.

Dart, R. A. 1925a. Australopíthecus africanus: Theman-ape of South Afriea. Nature (London) 115 (2884):195--99.

---o 1925b. A note 00 Makapansgat: A site of earlyhuman occupation. S. Afr. J. Sci. 22:454.

--o 1929. A note on the Taungs skull. S. Afr. J. Scí,26:648--58.

---o 19480. An adolescent promethean austra1opithe­cine mandible from Makapansgat. S. Afr. J. Sci.45:73--75.

---o 1948h. The Makapansgat proto-human Australo­pithecus prometheus. Am. J. Phys. Anthrop., n.s.,6:259-84.

--~. 194&. The adolescent mandible ofAu:i'tralopilh­ecus prometheus. Am. J. Ph)'s. Anthrop., n.s., 6:391­412.

---o 1949a. The predatory implementa! teehnique ofAustralopitheeus. Am. J. Phys. Anthrop., n.s., 7: 1-38.

---o 1949b. The bone-blud8eon hunting teehniqne ofAustralopitheeus. S. Afr. J. Sci. 2:150-52.

--o 19530. The predatory transition from ape to manoIn/o Anthrop. Llng. Rev. 1 (4): 201-18.

---o 1953b. The proto·human inhabitants of southemafriea. InAIriea's place in /he human s/ory, pp. 19--22.Johannesburg: South Afriean Broadcasting Corpora­tion.

---o 1954. The adult female lower jaw fromMakapansgat. No/u'e (London) 173:286.

---o 1955. The first auslra10pithecine fragment from

1

the Makapansgat pebble culture stratum. Nature (Lon­don) 176;170.

---o 19560. The myth of the bone-accumulatinghyena. Am. Anthrop. 58 (1); 4<k>2.

---o 1956h. Cultural status of tbe Scuth African man­apes. Smithsonian Report, n. 4240, pp. 317-38.

---o 1957a. The Makapansgat australopithecine osteo­dontokeratic culture. Proceedings of the Third Pan­African Congress on Prehistory (Livingstone, 1955).London: Chalto and Windus.

---o 1957b. The osteodontokeratic culture of Austra­lopithecus prometheus. Transvaal Museum Memoirs,no. 10.

---o ]957c. Ao austraJopithecine object fromMakapansgat, Nature (London) 179;69J-95.

---o 1958a. The minimal bone-breccia content ofMakapansgat and the austraIopithecine predatoryhabit. Am. Anthrop. 6O;92J-31.

---o 1958b. Bone tools and porcupine gnawing. Am.Anthrop. 60 (4); 715-24.

---o 1959a. Africo:s place in the emergence 01 civíii­sation, Johannesburg: South African BroadcastingCorporation.

---o 1959b. The ape-man tool-makers of a millionyears ago: South African Australopithecus-His life,habits and skills. Il/us. Lond. News 234;79lh'l01.

---o 1959c. An australopithecine scoop fromHerefordshire. Nature (London) 183;844.

---o 1959d. The first Australopithecus craníum fromthe pink breccia at Makapansgat. Am. J. Phys. An­throp . 17 (1); 77-4l2.

---o 195ge. Cannon-bone scoops and daggers, S. Afr.J. Sci. 55 (3); 79-4l2.

---o 1959[. Osteodontokeratic ripping tools and pulpscoops for teething and edentulous australopithecines.J. Dent. Ass. S. Afr. 14 (5); 164-78.

---o 1959g. Further light on australopithecine humeraland femoral weapons. Am. J. Phys. Anthrop., n.s., 17(2): 87-94.

---o 196Oa. Píthecanthropus andAustra/opithecus. Z.Morph. Anthrop. 50 (3): 261-74.

---o 1960b. The bone tool-manufacturing ability ofAustraloplthecus prometheus. Am. Anthrop. 62 (1):134-43.

---o 196Oc. Africa's place in the evolution of man.Ann. Proc., 1959-60, Associated Scientiñc and Techni­cal Societies, South Africa, pp. 21-41.

---o 196Od. The place of antelope cannon-bones (ormetapodials) in australopithecine economy. Z. Wiss.Zool. 68:1-15.

---o 196Oe. The persistence of sorne tools and utensilsfound tirst in the Makapansgat grey breccia. S. Afr, J.Sci. 56 (3): 71-74.

---. 19610. An australopithecine scoop rnade from aright australopithecine upper arm bone, Nature (Lon­don) 191:372-73.

---o 1961b. Further information about how Australo­pithecus made bone tools and utensils, S. Afr, J. Sci. 57(5): 127-34.

---o 19620. The continuity and originality of australo­pithecine osteodontokeratic culture. Actes IV CongrisPanafrican de Prehistoire et de l' étude du Quatemaire,pp. 27-41.

---o 1962b. Substitution of stone tools for bone toolsat Makapansgat. Noture (London) 196;314-16.

---o I962c . Stalactites as tool material for the austra­lopithecines: A missing cultural link between skeletal

References 351

and stone tool-making from the Makapansgat stalactiticcavern. II/us. Lond, News. 242:1052-55.

---o 1962d. Siegfried and Australopithecus. Probe: J.Scí, Stud. Council Univ, Witwatersrand 1:14-21.

---o I962e. The gradual appraisal of Australopithe­cus, InEvolutíon und Horninísation, ed. G. Kurth, pp.141-56. Stuttgart: Gustav Fischer Verlag.

---o 1962f From cannon-bone scoops to skull bowlsat Makapansgat. Am. J. Phys, Anthrop. 20 (3): 287-96.

---o 19640. A brief review of the Makapansgat in­vestigations 1925-1963. Academia das Ciencias de Lis­boa. Memorias 9:1-17.

---o 1964b. The Abbé Breuil and the osteodonto­keratic culture. Instituto de Prehistoria y Arqueotogia,Monografías 1l;347-70.

---o 19640. The Ecology of the South African man­apes, In Ecological studíes in Southem Africa, ed,D. H. S. Davis, pp. 49-<>6. Den Haag: Junk Publishers.

---o ]965a. The unavoidabLe osteodontokeratic cul­ture. In Dr. D. N. Wadia Commemorattve Volume, pp.231-54. Lucknow: Mining and Metallurgical1nstitute ofIndia.

---o 1965b. Australopithecine cordage and thongs. InHomenaje a Juan Comas en su 65 aniversario, 2:43-61.Mexico.

---o 1965c. Pounding as a process and the producer ofother artefacts. S. Afr. Arcbaeoí, Bull. 20 (79); 141-47.

---o 1965d. Tree chopping with an elephant rih. S.Afr, J. sa. 61 (11); 395.

Dart, R. A., and Craig, D. ]959. Adventures with themissíng link. New York: Harper; London: HamishHamilton.

Dart, R. A., and Kitching, J. W. 1958. Bone tools at theKalbank Middle Stone Age site and the Makapansgataustralopithecine locality, Central Transvaa!. Part 2.The osteodontokeratic contributíon. S. Afr. Archaeol.Bull. 13 (5I); 94-116.

Davis, D. H. S. 1955. Taxonomic reporto Swartkrans fos­sil microfauna. Unpublished reporto

---o 1958. Notes 00 the small mammals in theKalahari Gemsbok National Park, with speciaJ refer­ence to those preyed upon by bam owls. Koedoe1:184-4l8.

---o 1959. The bam-owl's contributicn to ecology andpalaeoecology.Ostrich, suppl. 3: 144-53.

---o 1962. Oistribution patterns of southern AfricanMuridae, with notes on sorne of their fossil anteced­ants. Ann. Cape Prov. Mus. 2:56-76.

---o 197i. Genera Tatera and Gerbillurus. In Themammals of Africo: An identifícation manual, ed. J.Meester and H. W. Setzer. Washington, O.e.: Srnith­sonian Institution Press.

---o 1974. The distribution of sorne smal! southern Af­rican mammals (MarnmaIia: Insectivora, Rodentia).Ann. Transv, Mus. 29;135-84.

Dawkins, W. B. 1874. Cave hunting. London.---o 1877. On the mammalian fauna of the cave of

Cresswel! Crags. Q. J. Geol. Soco Lond. 33:606-7.Deacon, H. J. 1969. Melkhoutboom Cave, Alexandria

Oistrict, Cape Province: A report on the 1967 in­vestigation. Ann. Cape Prov. Mus. 6 (13); 141-69.

---o 1970. Plant remains fmm Melkhoutboom Cave,South Africa. Proc. Transkei and Císket Res. Soco1:1J-15.

---o 1972. A review of the post-Pleistocene in SouthAfrica. S. Afr. Archaeol. Soc., Goodwin ser. 1;26-45.

---o 1975. Demography, subsistence, and culture

352 References

during the Acheulean in southem Africa. In Alter theaustraíopithecines, ed. K. W. Butzer and G. Ll. Isaac,pp. 54J-{;9. The Hague: Mouton.

Deacon, J. 1965. Part 1. Cultural material from theGamtoos valley shelters (Andrieskraal 1). S. Afr. Arcñ.Bull. 20 (80): 193-200.

---o 1972. Wilton: An assessment after fifty years. S.Afr. Archaeol. Bull. 27: 10-48.

Dean, W. R. J. 1973a. Analysis of a collection ofbam owlTyto alba pellets from Warmbaths, Transvaal, Zool.Afr. 8 (1): 75-82.

---o 1973b. Age distribution of Praomys natalensisprey in Tyto alba pellets. Zooí. Afr. 8 (1): 140.

---o 1974. Analysis of Tvto alba pellets from Angola,Zool. Afr, 9 (1): 89-90.

---o 1975. Tvto alba prey in South West Africa andthe northem Cape. Zool. Afr, 10:217-19.

Deane, N. N. 1962. The spotted hyaena, Crocuta c.crocuta. Lammergeyer 2:26--44.

de Graaff, G. 1957. A new chrysochlorid fromMakapansgat. Palaeont. Afr. 5:21-27.

---o 1960a. A preliminary investigation of the maro­malian microfauna in Pleistocene deposíts of caves inthe Transvaal Systern. Paíaeom, Afr. 7:59-118.

---. 1%Oh. '0 Ontleding van uilklonte van. die Non­netjiesuil, Tyto alba. Ostrich 31 (1): 1-5.

---o 1961. 00 the fossil mammalian microfauna col­lected at Kromdraal by Draper in 1895.S. Afr, J. Sci,56:259-ffl.

---o 1971. Family Bathyergiuee. In The mammals ofAfrica: An identification manual, ed. J. Meester andH. W. Setzer. Washington, D.C.: Smithsonian Instt­tution Press.

Delson, E. 1973. Fossü colobine makeup of the circum­Mediterranean region and the evolutionary history ofthe Cercopithecidae (Primates, Marnmalia). Ph.D.diss., Columbia University.

---o 1975. Evo1utionary history of the Cer­copithecidae. Con/ribo Prímatol. 5:167-217.

Díetrich, W. O. 1942. Altestquartáre Sáugetíere aus dersüdlichen Serengeti, Deutsch-Ostafríka. Palóeonto­graphica 94:43-133.

Dorst, J., and Dandelot, P. 1972. A fieíd guide 10 filelarger mammals o/ Africa. London: Collins.

Draper, D. 1896. Report of meeting, 8lh Apri1, 1895,Geological Society of South África. Trans, Ceo/. SOCo

S.Afr. 1(1): 11.--o 1898. Meeting of lhe Geological Society of South

Africa, 12 July 1897. Trans. Geol. Soco S. A,{r. 3:63.Duke, G. E.; Jegers, A. A.; Loff, G.; andEvanson, O. A.

1975. Gastric digestion in sorne raptors. Comp.Bioehem. Physiol. 50A:649-56.

Eck, G. 1976. Cercopithecoidea from the Dmo Group de­posits. In Earlíest man and environments in the LakeRudolf bas;n, ed. Y. Coppens, F. C. Howell, G. LI.Isaac, and R. E. F. Leakey, pp. 332-44. Chicago: Uni­versity of Chicago Press.

Edwards, D. 1974. Survey to determine the adequacy ofexisting conserved areas in relation to vegetation types:A preliminary reporto Koedoe 17:1~38.

Efremov, 1. A. 1940. Taphonomy: A new bmnch ofpalaeontology. Pan-Amo Geol. 74:81-93.

EisenOOrg, J. F. 1970. A splendid predator docs its ownthing untroub1ed by mano Sm;thsonian 1 (6): 48-53.

Eisenberg, J. F., and Lockhart, M. 1972. An ecologicalreconnaissance of Wilpattu National Park, Ceylon.

Smíthsonian Contrib, Zool., no. 101, pp. 1-118.Eisenhart, W. L. 1974. The fossil Cercopithecoids of

Makapansgat and Sterkfontein. B.A. thesis, Oept. of··Anthropology, Harvard College.

Eloff, F. C. 1964. On the predatory habits of lions andhyaenas,--Koedoe 7:1O~12.

---.·1975) The spotted hyaena Crocuta crocuta (Erx- .,._.Ieben) in arid regions of southem Africa. Pub. Univ.ÓPretoria. n.s., 97:35-39.

Eloff, J. F. 1969. Bushman Rock Shelter, eastem Trans­vaal: Excavations, 1967--{;8. S. Afr. Arcllaeo/. Bull. 24(94): 60.

Eng/ish Mechanic and the Wor/d of Science. 1897. No.1692 (27 August), pp. 36.

Ennouchi, E. 1953. Un nouveau genre d'Ovicapriné dansun gisement pléistocene de Rabat. C. R. Seanc. SocoGéol. Fr. 1953:126-28.

Éstes, R. D. 1967. Predators and scavengers. Nat. Híst.N. Y. 76 (2): 20-29; 76 (3): 38-47.

Ewer, R. F. 1954u. Sabre-toothed tigers. New Biology17:27-40.

---o 1954b. The Hyaenidae of Kromdraai, Proc. Zool.Soe. Lond. 124:565-85.

---o 1955a. Hyaenidae, olher than Lycvaena ofSwankrans and Sterkfontein. Proc, Zool. Soco Lond.124:815-37.

--o 1955b. The Lycyaenas of Sterkfontein andSwartkrans, together with sorne general considerationsof the Transvaal fossil hyaenids. Proc. Zoo/. SOCo

Lond. 124:839-57.--o 1955c. The fossil camivores of the Transvaal

caves: Machairodontinae. Proc. Zool. Soco Lond. 125(3-4): 587-615.

--o 19560. The fossil camívores of the Transvaalcaves: Felinae, Proc, Zool. Soco Lond. 126 (1): 83-95.

---o 1956b. The fossil camivores of the Transvaalcaves: Canidae. Proe. Zool. Soco Lond, 126(1): 97-119.

--o 19560. Too fossi1 carnivores of lhe Transvaalcaves: Two new viverrids, together with sorne generalconsiderations, Proe. Zool. Soe. Lond. 126:259-74.

--o 1956d. Sorne fossil camivores from theMakapansgat valley. Palaeont. Afr. 4:57-67.

--o 1956e. The fossil suids of the Transvaal caves.Proc. 2001. Soco Lond. 127:527-44.

--o 1958a. The fossil Suidae of Makapansgat. Proc,Zool. Soco Lond. 130:329-72.

--o 1958b. Adaptive features in lhe skulls of AfricanSuidae. Proc. 2001. SOCo Lond. 131 (1): 135-55.

--o 1965. Large Camivora. In Olduvai Gorge 1951­196I. ed. L. S. B. Leakey. Cambridge: UniversityPress.

--o 1967. The fossil hyaenids of Africa-A ce­appraisal. In Background lo evolution in Africa, ed.W. W. Bishop and J. D. Clark. Chicago: University ofChicago Press.

--o 1973. Tlle cam;vores. The World Naturalis!.London: Weidenfeld and Nicolson.

Exton, H. 1899. Presidential address, Geological SocietyofSoulh Africa, 22d February, 1898. Trans. Geo/. SocoS. Afr. 4:6-10.

Frames, M. E. 1898. Rem!U"ks on Peof. Pristec's paper onGlacial phenomena al Pretoria and on the Rand. Trans.Geol. Soc. S. A,{r. 3:91-95.

Freedman, L. 1957.The fossil Cercopithecoidea ofSouthAñica. Ann. Transv. Mus. 23:121-262.

--o 1965. Fossil and subfossil primates from Ihe

limestone deposits at Taung, Bolt's Farm and Wit­krans, South Afriea. Palaeont. Afr. 9:19-48.

---o 1970. A new checklist of fossil Cercopithecoideaof South Afriea. Palaeont, Afr, 13:1~10.

Freedman, L., and Brain, C. K. 1972. Fossil cer­copithecoid remains from the Kromdraai australopithe­cine síte (Mammalia: Primates). Ann. Transv, Mus. 28(1): 1-16.

---o 1977. A re-examination ofthe cercopithecoid fos­sils from Swartkrans (Mammalia: Cercopithecidae),Ann. Transv. Mus. 30 (18): 211-18.

Freedman, L., and Stenhouse, N. S. 1972. TheParapapiospecies of Sterkfontein, Transvaal, South Africa.Paíaeont, Afr. 14:93--111.

Freeman, L. G., and Butzer, K. W. 1966. The Acheuleanstation at Torralba (Spain): A progress reporto Quater­noria 8:9-21.

Un frere Mariste de Johannesburg. 1898. Les grottes deSterkfontein. Cosmos 679: 133--35.

Frison, G. C., ed. 1974. The Casper síte, New York:Academic Press.

Galton, F. 1889. Narrative of on explorer in tropicalSouth Africa, being an account 01 a visit to Damara­land in /85/. London: Henry Frowde.

Gargett, Y. 1965. Death of a black eagle. Honeyguide50:9-10.

---o 1967. B1ack eagle experiment. Bokmakarie19:88-90.

---o 1970a. B1ack eagle experiment, no. 2. Bok­makaríe 22:32-35.

---o 1970b. Black eagle survey, Rhodes Matopos Na­tional Park: A population study 1964-1968. Oslrich,suppl. 8:397-414.

---o 1971. Some observations on the black eagles inthe Malopos, Rhodesia. Ostricñ. suppl. 9:91-124.

---o 1972a. Observatíons at a black eagle nest in theMatopos, Rhodesia. Ostricñ, 43 (2): 77-108.

---o 1972b. Black eagle Aquila verreauxí populationdynamies. Ostricñ 43 (3): 177-78.

---o 1975. The spacing ofblack eagles in the Matopos,Rhodesia. Ostrícñ 46 (1): 1-44.

--o 1976. Dead or Alive? Afr. Wild Life 30 (2): 40-41.---o 1977. A thirteen-year population study of the

black eagles in the Matopos, Rhodesia, 1964-1976. Os­Irich 48: 17-27.

---o 1978. Mackinder's eagle owl, Bubo capensismockinderi, in the Matopos, Rhodesia. In Proceedings01 the Svmposium on A/rican Predatory Birds, pp.46-61. Pretoria: Nortbern Transvaal Ornithological So­eiety.

Gargett, Y., and Grobler, J. H. 1976. Prey of the Capeeagle owl, Bubo capensis mackinderi Sharpe 1899, inthe Matopos, Rhodesia.Arnoldia (Rhodesia) 8 (7): 1-7.

Gentry, A. W. 1965. New evidence on the systematicposition of Hippotragus niro Hopwood 1936 (Mam­malia). Ann. Mag. Nat, Hist, ser. 13, 8:335-38.

---o 1966. Fossil Antilopini of East Afriea. Bull. Brit,Mus. Na/. Hisl. (Geol.) 12:45-106.

---o 1970a. The Langebaanweg Bovidae. Appendix inA review of the geology and palaeontology of the Plio/Pleistocene deposits at Langebaanweg, Cape Province,ed. Q. B. Hendey. Ann. S. Afr, Mus. 56 (2): 114-17.

---o 1970b. Revised c1assifieation for Makapaniabroomi Wells and Cooke (Bovidae, Mammalia).Palaeont, Afr. 13:63--67.

---o 1971a. Genus Gazella. In The mammals of Af-

References 353

rica: An identification manual, ed. J. Meester andH. W. Setzer. Washington D.C.: Smithsonian Insti­tution Press.

---o 1971b. The earliest goats and otber antelopesfrom the SamosHipparion fauna. Bull. Brit. Mus. Nat.Hist. (Geol.) 20 (6): 231-96.

---o 1974. A new genus and species of Plioceneboselaphine (Bovídae, Mammalia) from South Afriea.Ann. S. Afr, Mus. 65 (5): 145-88.

---o 1976. Bovidae of the Omo Group deposits. InEarliest man and environments in the Lake Rudolfbasin, ed. Y. Coppens, F. C. Howell, G. Ll. Isaac, andR. E. F. Leakey, pp. 275-92. Chicago: Unlversity ofChicago Press.

Gingerich, P. D. 1974. Proteles crlstatus Sparrman fromthe Pleistocene of South Africa, with a note on toothreplacemenl in the aardwolf (Mammalia: Hyaenidae).Ann. Transv. Mus. 29 (4): 49-54.

Glover, P. E.; Glover, E. C.; Trump, E. C.; andWateridge, L. E. D. 1964. The lava caves of MountSuswa, Kenya, with particular reference to theireeologieal role. Studies in Speleology 1 (1): 51--M.

Gow, C. E. 1973. Habitual sbeltering in an extensive cavesystem by baboons near Bredasdorp, South Afriea. S.Afr. J. Sci. 69:182.

Grabam, C. 1953. Bold leopards. Afr, Wild Life 7 (3):244-45.

Greenwood, M. 1955. Fossil Hystricoidea from theMakapan Yalley, Transvaal, Pataeont. Afr. 3:77--1l5.

---o 1958. Fosstl Hystrieoidea from the MakapanValley, Transvaal: Hystrix makapanensis nomo nov.for Hystrix major Greenwood. Ann. Mag. Nat. Hist.,ser. 13 (5): 365.

Grindley, J. R., and Nel, E. 1970. Red water and musselpoisoning at Elands Bay, December 1966. FísheriesBull. S. Afr. 6:36-55.

Grindley, J. R.; Speed, E.; and Maggs, T. 1970. The ageof the Bonteherg shelter deposits, Cape Peninsula. S.Afr. Archaeol. Bull. 25 (97): 24.

Guérin, G. 1928. La vle des chouenes: Régime et crois­sanee-de l' effraye commune Tyto alba alba (L.) en ven­dée. Paris: P. Lechevalier.

Hanney, P. 1962. Observations on the food of the barnowl (Tyto alba) in Southern Nyasaland. Ann. Mag.Na/. m«., ser. 13 (6): 305-13.

Harris, J. M. 1974. Orientation and variability in the os­sicones of African Sivatheriinae (Mammalia:Giraffidae). Ann. S. Afr. Mus. 65 (6): 189-98.

---o 19760. Bovidae from the East Rudolf sueeession.In Earliest man and environments in the Lake Rudoífbasín, ed. Y. Coppens, F. C. Howell, G. Ll. Isaac, andR. E. F. Leakey, pp. 293--301. Chicago: University ofChicago Press.

---o 1976b. Pleistocene Giraffidae (Mammalia, Ar­tiodactyla) from East Rudolf, Kenya. In Fossil verte­bra/es ofAfrica, ed. R. J. G. Savage and S. C. Coryn­don, 4:283--332. London: Aeademic Press.

---o 1976c. Pliocene Giraffoidea (Mammalia, Ar­tiodactyla) from the Cape Provinee. Ann. S. Afr, Mus.69 (12): 325-53.

Haughton, S. H. 1922. A note on some fossils from theYaaI River gravels. Trans. Geol. Soco S. Afr. 24:11-16.

--.1932. The fossil Equidae ofSouth Afriea.Ann. S.Afr. Mus. 28:407-27.

Hemmer, H. 1965. Zur Nomenklatur und Verbreitung desGenus Dinofelis Zdansky, 1924 (Therailurus Piveteau,

354 References

1948). Palaeont, Afr. 9:75-89.Hendey, Q. B. 1%8. The Melkbcs slte: An upper Pleis­

tocene fossil oecurrence in the south-weslern CapeProvince. Ann. S. Afr. Mus. 52:8~119.

---o 1973. Carnivore rernains from the Kromdraaiaustralopithecine site (Marnmalia: Carnivora). Ann.Transv, Mus. 28:~1I2.

---o 1974a. The Late Cenozoic Carnivora of thesouth-sestem Cape Province. Ann. S. Afr. Mus. 63:1­369.

---o 1974b. New fossil camivores from the Swart­krans australopithecine site (Marnmalia: carnivora).Ann. Transv, Mus. 29 (3): 27-48.

---o 1976. The Pliocene fossil occurrences in "E"Quarry, Langebaanweg, South Africa. Ann. S. Afr.Mus. 69 (9): 215-47.

Hendey, Q. B., and Hendey, H. 1968. New Quaternaryfossil sites near Swertklip, Cape Province. Ann. S. Afr.Mus. 52:43-73.

Hendey, Q. B., and Singer, R. 1%5. Part 111. The faunalassemblages from the Gamtoos valley shelters. S. Afr,Areh. Bull. 20 (80): 206-13.

Hewitt , J. 1921. On several implements and omamentsfrom Strandloper sttes in the Eastem Province. S. Afr.J. Sei. 18:454-{;7.

---o 1931. Artefacts from Melkhoutboom. S. Afr, J.Sci. 28: 540-48.

HiIl, . A: 1915. Taphonomy of contemporary and lateCenozoic East Afriean vertebrates. Ph.D. diss. Univer­sity of London.

---o 1978. Hyaenas, bones and fossil mano Kenya,PaSI and Present 9:9-14.

Hill, A., and Walker, A. 1972. Procedures in vertebratetaphonomy: Notes en a Uganda Miocene fossil locality.J. Geol. Soco Lond, 128:3~.

Hoernlé, A. 1923. South West Africa as a primitive cul­tural area. S. Afr, Geogr. J. 6:14-28.

Hoesch, W. 1936. Ornithologische Monatsberichte 44:6.Hoffman, A. C. 1953. The fossil alcelaphines of South

Africa--genera Peíoroceras. Lunatoceras andAlcelaphus. Navors. Nas. Mus. Bíoemfontein, 1 (3):41-56.

Holloway, R. L. 19700. Australopithecine endocast(Taung specimen, 1924): A new volume determination,Scíence 1%8:961>-68.

---o 1970b. New endocranial values for the australo­pithecines. Nature (London) 227:1~200.

---o 1972. New australopithecine endocast, SK 1585,fromSwartkrans, South Africa. Am. J. Phys. Anthrop.37 (2): 173-86.

Honer, M. R. 1%3. Observations on the barn owl (Tytoalba guttata} in the Netherlands in relation te ítsecologieal population fluctuations. Ardea 51:158-95.

Hooijer, O. A. 1975. Miocene lo Pleistocene hipparions oíKenya, Tanzania and Ethiopia. Zool. Verh. 142:1-80.

---o 1976. The late Pliocene Equidae of Langebaan­weg, Cape Province, South Africa. Zool. Verh.148:1-39.

Hopwood, A. T. 1936. New and little-known fossil mam­mals from the Pleistocene of Kenya Colony and Tan­ganyika Terrítory. Ann. Mag. Nat. Hist., ser. !O,17:636-41.

Howell, F. C. 1%1. Ismila: A palaeolithic site in Africa.Sci. Am. 205 (4): 118-29.

---o 1%6. Observations on the earlier phases of the

European Lower Palaeolithic. Am. Anthrop, 68 (2):88-201.

Howell, F. C. J978. Hominidae. In Evoíution 01 Africanmammals, ed. V. J. Maglio and H. B. S. Cooke, pp.154-248. Cambridge: Harvard University Press.

Howell, F. c.: Cole, G. H.; and Kleindienst, M. R. 1%2.Isimila: An Acheulian occupation site ín the IringaHighlands, Southern Highlands Province, Tanganyika.Ann. Mus. Roy. Afr, Cent. 40:43-80.

Howell, F. C., and Coppens, Y. 1974. Inventory of re­mains of Hominidae from PliocenelPleistocene forma­tions of the lower Omo basin, Ethiopia (1%7-1972).Am. J. Phys. Anthrop, 40: 1-16.

--o 1976. An overview of Hominidae from the Omosuccession, Ethiopia. In Earliest man and enviran­ments in the Lake Rudolfbasin, ed. Y. Coppens, F. C.Howell, G. L1. Isaac, and R. E. F. Leakey, pp. 522-32.Chicago: University of Chicago Press,

Howell, F. C., and Petter, G. 1976. Camivora frorn OmoGroup formaticns, Southem Ethiopia, lo Earliest manand environmerüs in the Lake Rudo/f basin, ed. Y.Coppens, F. C. Howell, G. L1. Isaac, and R. E. F.Leakey, pp. 314-31. Chicago: University of ChicagoPress.

Howell, F. C.; Fichter, L. S.; and Eck, G. 1%9. Verte­brate assembleges from the USRO Formetíon, WhiteSands and Brown Sands localities, lower Omo basin,Ethiopia, Qualernaria 2:65-88.

.Hughes, A. R:1954a. Hyaenas versus australopithecinesas-ageots.ef bone accumulations. Am. J. Phys. An­mrop., n.s., 12 (4): 467-86.

--o 1954b. Habits of hyaenas. S. Afr. J. Sci,51:156-58.

--o 1958. Sorne ancient and recent observations onhyaenas. Koedoe 1:105-14.

--o 1%1. Further notes on the habits of hyaenas andbone gathering by porcupines. Zool. Soco S. Afr. NewsBull. 3 (1): 35-37.

Hughes, A. R., and Tobias, P. V. 1977. A fossil skullprobably of the genus Horno from Sterkfontein, Trans­vaal, Nature (London) 265:310-12.

llani, G. 1975. Hyenas in Israel. lsrael-Lond and Na­ture, October 1975, pp. 10-18. From Teva Va'aretz 16(4).

Isaac, G. Ll. 1%70. The stratigraphy ofthe Peninj Group:EarIy Middle Pleistocene formations west of Lake Nat­ron, Tanzania. In Background lo evotution in Afríca,ed. W. W. Bishop and J. D. Clark, pp. 22~58. Chicago:University of Chicago Press.

--o 1%7b. Towards the interpretation of occupa­tional debris: Sorne experiments and observations.Kroeber Anthrop. Soco Papers 37:31-55.

---o 1968. Traces of Pleistocene hunters: An East Af­rican example. In Man the hunter, ed. R. B. Lee and 1.De Vore., pp. 253-61. Chicago: Aldine.

---o 1%9. Studies of early culture in East Africa.World Arehaeology 1:1-28.

---o 1971. The diet of early mano Aspects of arochaeological evidence from Lower and Middle Pleis­tocene sites in Africa. World Archaeology 2 (3): 278-99.

---o 1976. The activities of early African hominids: Areview oí the archaeological evidence for the time spantwo and a half to one miHion years ago. In Humanorigins: Louis Leakey and the East African evidence,ed. G. Ll. Isaac and E. R. McCown, pp. 482-514.

Menlo Park, Calif.: Staples Press.Isaac, G. L1.; Leakey, R. E. F.: and Behrensmeyer, A.

K. 1971. Archaeological traces of early hominid ac­tivities, east oí Lake Rudolf', Kenya. Science173:1129-34.

Jackson, H. D. 1973a. The Cape eagle owl, Bubo capen­sís in Mozambique. Bull. Brít, Orn. Club 93:10.

---o 1973b. Records of sorne birds and mammals inthe central Chimanimani Mountains of Mozambiqueand Rhodesia. Durban Mus. Novit. 9 (20): 291-305.

Jenkins, T., and Brain, C. K. 1967. The peoples of thelower Kuiseb valley, South West Afriea. Sci. PapersNamib Desert Res. Station 35:1-24.

Johanson, D. C. 1976. Ethiopia yields tirst "fami1y" ofear1y mano Natl. Geogr. Mag. 150 (6): 791-411 I.

Johanson, D. C.; and Taieb, M. 1976. Plio-Pleistocenehominíd discoveries in Hadar. Ethiopia. Nature (1.00­don) 260:293-97.

Johanson, D. C., and White, T. D. 1979. A systematicassessment of early African hominids. Science 203(4378): 321-30.

Johanson, D. C.; White, T. D.; and Coppens, Y. 1978. Anew species oí the genus Australopíthecus (Primates:Hominidae) from the Pliocene of Eastem Afriea.Kirtlandia 28:1-14.

Jolly, C. J. 1966. The evo1ution of the baboons. In Thebaboon in medical research, ed. H. Vogtborg, 2:23-50.Austin: University of Texas Press.

---o 1970. The large African monkeys as an adaptivearray. In Old Wor/d monkevs, ed. J. Napier and P.Napier, pp. 139-74. London: Academic Press.

---o 1972. The c1assitication and natural history ofTheropithecus (Símopitñecus] (Andrews), 1916 ba­boons ofthe African Pleistocene. Bul/. Brit, Mus. Nat.Hisl. (Geol.) 22 (1): 1-123.

Jones, T. R. 1937. A new fossil primate from Sterkfon­tein, Krugersdorp, Transvaal. S. Afr. J. Scí, 33:709-28.

---o 1978. The skul/ and mandible of the South Afri­can haboon: A morphological study. Johannesburg:Witwatersrand Uníversity Press.

Kirby, P. R. 1940. Robert Knox and his South Africanresearch. S. Afr. Med. J. 14:13.

Kitehing, J. W. 1953. A new speciesoffossil baboon fromPotgietersrust. S. Afr. J. Sci, 50:~9.

---o 1963. Bone, tooth and horn tools of Palaeolithicmano Manchester: Manchester University Press.

---o 1965. A new giant hyracoid from the LimeworksQuarry, Makapansgat, Potgietersrust. Palaeonl. Afr,9:91-96.

Klein, R. G. I972a. Preliminary report on the Julythrough September 1970 excavation al Ne1son BayCave, Plettenberg Bay (Cape Prov., South Africa). InPa/aeoecology of Africa, 1969-7/, ed. E. M. vanZinderen Bakker, pp. 177-208. Cape Town: Balkema.

---o 1972b. The late Quatemary mammalian fauna ofNe1son Bay Cave (Cape Province, South Africa): Itsirnplications for megafaunal extinctíons and environ­mental and culture change. Qualernary Res. 2 (2):135-42.

---o 1974a. A provisional statement 00 terminalPleistocene mammalian extinetions in the Cape bioticzone (Southem Cape Province, South Africa). S. Afr.Arch. Soc., Goodwin ser., no. 2, pp. 39-45.

---o 1974b. Environmenl and subsistence of pre­historie man in the southern Cape Province, South Af-

References 355

rica. World Archueol. 5 (3): 249-84.---o 1975a. Palaeoanthropological implications of the

nonarchaeological bone assemblage from Swartklip 1,south-western Cape Province, South Africa. Quater­nary Res. 5:275-418.

---o 1975h. Middle Stone Age mari-animal re­lationships in southem Africa: Evidence from Die Kel­ders and Klasies River Mouth. Scíence 190:265-67.

---o 1975c. Ecology of Stone Age man at the southerntip of Africa. Archaeology 28 (4): 238-47.

---o 19760. The marnmalian fauna of the KlasiesRiver Mouth sites, southern Cape Province, South Af­rica. S. Afr. Archaeol. Bull. 31:75-98.

---o 1976b. A preliminary report on the "MiddleStone Age" open-air site of Duinefontein 2(Melkbosstrand, south-westem Cape Province, SouthAfrica). S. Afr, Archaeol. Bull. 31:12-20.

---o 1976c. The fossil history ofRaphicerus H. Smith,1827 (Bovidae, Mammalia) in the Cape biotic zone.Ann. S. Afr, Mus. 71:169-91.

---o 1977. The ecology of early man in southem Af­rica. Science 197:115-26.

---o N .d. Results of the preliminary analysis of themamrnalian fauna from the Redcliff Stone Age site,Rhodesia. Typescript.

Klein, R. G., and Scott, K. 1974. The fauna of Sccn'sCave, Garntoos valley, south-eastern Cape Province.S. Afr, J. sa. 70:186-417.

Knox, R. 1822. Notice relative to the habits of thehyaenas in southem Africa. Trans, Wernerian Nat.Hist. Soc., Edinburgh 4:383.

Kolbe, F. F. 1946. The case for the barn owl. Afr. WildLije 1:69-73.

Kruuk, H. 1966. Clan system and feeding habíts of spot­ted hyaenas (Croeuta croeuta Erxleben). Nature (Len­don) 209:1257-58.

---o 1968. Hyaenas: The hunters nobody knows.Naü. Geogr. Mag. 134:44-57.

---o 1972. The spotted hyaena: A study of predationand social behavíor, Chicago: University of ChicagoPress.

---o 1976. Feeding and social behaviour of the stripedhyaena (Hyaena vulgaris Desmarest). E. Afr, Wildl. J.14:91-111.

Kruuk, H., and Tumer, M. 1967. Comparative notes onpredation by lion, leopard, cheetah and wild dog in theSerengeti area, E. Africa. Mammalia '6731:1-27.

Kuhn, H. J. 1967. Zur Systematik des Cercopithecidae. InProgress in primatology, ed. D. Starck, R. Scbneider,and H. J. Kuhn, pp. 25-46. Stuttgart: Fisher.

Kurtén, B. 1956. The status and affinities of Hyaenasinensís Owen and Hyaena ultima Matsumoto. Amer.Mus. Novit., no. 1764, pp. 1-48.

---o 1957a. The bears and byaenas ofthe interglacials.Quaternaria 4:1-13.

---o 1957b. Mammal migrations, Cenozoic stratig­raphy and the age of Peking man and the australopithe­cines. J. Paíaeont, 31:215-27.

---o 1968. Pleistocene mammals of Europe. London:Weidenfeld and Nicolson.

Laubscher, N. F.; Steffens, F. E.; and Vrba, E. S. 1972.Statistieal evaluation of the taxonomic status of a fcssilmember of the Bovidae (Mammalia: Artiodactyla).Ann. Transv. Mus. 28 (2): 17-26.

Lavocat, R. 1956. La Caune de rongeurs des grottes a

356 References

australopithéques. Palaeont, Afr. 4:6~75.

Leakey, L. S. B. 1943. New fossil Suidae from Shungura,Omo. J. East Afr. Nat. Hist. Soco 17:45-61.

--o 1965. Olduvai Gorge, 1951-6/. Cambridge: Cam­bridge University Press.

Leakey, L. S. B., and Leakey, M. D. 1964. Recent dis­coveries of fossil hominids in Tanganyika: At Olduvaiand near Lake Natron. Nature (London) 202:5-7.

Leakey, L. S. B.; Tobias, P. V.; and Napier, J. R. 1964. Anew species of the genus Horno from OJduvai Gorge.Nature (London) 202:7-9.

Leakey, M. D. 1970. Stone artefacts from Swartkrans.Nature (London) 225:1222-25.

---o 1971.Oldavai Gorge. Vol. 3. Excavatíons in Beds1 and Il, 1960-/963. Cambridge: Cambridge UniversityPress.

Leakey, M. G. 1976. Carnivora of Ibe East Rudolf suc­cession. In Earliest man and environments in the LakeRudo/f basin, ed. Y. Coppens, F. C. Howell, G. Ll.Isaac, and R. E. F. Leakey, pp. 302-13. Chicago: Uni­versity of Chicago Press.

Leakey, R. E. F. 1973. Further evidence of Lower Pleis­tocene hominids from East Rudolf, North Kenya, 1972.Nature (London) 242:170--73.

---o 1974. Further evidence of Lower Pleistocenehominids from East Rudolf, North Kenya, 1973. Na­ture (London) 248:653--56.

---o 19760. Hominids in Africa. Am. Scient, 64 (2):174-78.

---o 1976h. An overview of Ibe Hominidae frorn EastRudolf, Kenya. In Earliest man and environments inthe Lake Rudo/fhasin, ed. Y. Coppens, F. C. Howell,G. Ll. Isaac, and R. E. F. Leakey, pp. 47lHl3.Chicago: University of Chicago Press.

Leakey, R. E. F., and Walker, A. C. 1976.Australopithe­CUS, Horno erectus and the single species hypothesis.Nature (London) 261:572-74.

Lee, R. B. 1968. What hunters do for a living; or, How tomake out on scarce resources. In Man the hunter, ed.R. B. Lee and I. Devore, pp. 30-48. Chicago: Aldine.

---o 1972. The !Kung bushmen of Botswana. In Hun­ters and gatherers today, ed. M. G. Bicchierí, pp.327-68. New York: Holt, Rinehart and Winston.

Lee, R. B., and Devore, I. 1968. Problems in the study ofhunters and galberers. In Man the hunter, ed. R. B.Lee and I. Devore, pp, 3--12. Chicago: Aldine.

Leechman, D. 1951. Bone grease. Am. Antlq, 16 (4):355-56.

Louw, A. W. 1969. Bushman Rock She1ter, Ohrigstad,eastem Transvaal: A preliminary ínvestigation, 1965.S. Afr. Archaeol. Bull. 24 (94): 3~51.

Lubbock, J. 1865. Prehistoríc times. London: Williamsand Norgate.

MacArthur, R. H., and Pianka, E. R. 1966. An optimaluse ofa patchy environment.Am. Nat. 100(916):603--9.

MacCalman, H. R. 1967. The zoo park elephant síte,Windhoek. In Palaeoecology of Africa, /964-65, ed.E. M. van Zinderen Bakker, pp. 102-3. Cape Town:Balkema.

McHenry, H. 1975a. Fossil hominid body weight andbrain size. Nature (London) 254:68lHl8.

---o 1975h. Fossils and the mosaic nature of humanevolution. Scíence 190:425-31.

McLachlan, G. R., and Liversidge, R. 1970. Rohertshirds of South Afríca. Cape Town: Central NewsAgency.

Maggs, T. M. O'C. 1971. Some obscrvatíons on the sizeof human groups during the Late Stone Age. In Rockpaintings of Soutb Afríca, ed. M. Schoonraad, pp.4~53. S. Afr, J. Sci., special issue no. 2.

Maggs, T., and Speed, E. 1967. Bonteberg Shelter. S.Afr. Archaeol. Bull. 22:80--93.

Maglio, V. J. 1970. Early Elephantidae of Africa and atentative correlation of African Plio-Pleistocene de­posíts. Nature (London) 225:328-32.

---o 1973. Origin and evolution of the Elephantidae.Trans. Amer. Phi/o Soco 63:1-149.

Maguire, J. 1976. A taxonomic and ecological study oftheliving and fossil Hystricidae with particular reference 10

southern Afríca. Ph.D. diss., Department of Geology,University of the Witwatersrand.

Maier, W. 19700. Neue Ergebnisse der Systematik undder Stammesgeschichte der Cercopithecoidea. Z.Sdugetierk. 35:193--214.

---o 1970h. New fossil Cercopithecoidea from theLower Pleistocene cave deposito of the MakapansgatLimeworks, South Mrica. Palaeont. Afr. 3:6~107.

---o 1973.Palaookologie und zeitliche Einordnung dersudafrikanischen Australopithecinen. Z. Morph. An­throp. 65:70--105.

Malan, B. D. 1959. Early references to the Sterkfonteincaves. S. Afr. J. Scí, 55: 321-24.

---o 1960. A history of Sterkfontein with a commentby B. D. Malan. S. Afr, J. sa. 56 (5): 110.

Mann, A. E. 1968. The palaeodemography of Australo­pithecus. Ph.D. diss., University of California, Berke­Ley.

---o 1975.Paleodemographic aspects ofthe South Af­rícan australopitñecines. University of PennsylvaniaPublications in Anthropology, no. 1. Philadelphia: Uni­versity of Pennsylvania.

Martyn, J., and Tobias, P. V. 1967. Pleistocene depositsand new fossil localities in Kenya. Nature (London)215 (5100): 47lHlO.

Mason, R. J. 1958. Bone tools at the Kalkbank MiddleStone Age site and Ibe Makapansgat australopithecinelocality, central Transvaal, Par! 1. The KaJkbank site.S. Afr. Archaeol. Bull. 13 (51): 94-116.

---o 1962a. Prehistory of the Transvaal: A record ofhuman activity, Johannesburg: Witwatersrand Univer­sity Press.

---o 1962h. The Sterkfontein stone artefacts and Ibeirmaker. S. Afr. Archaeol. Bull. 17 (66): 109-25.

---o 1964. Iron Age bone artefacts. S. Afr. Archaeol.Bull. 19 (74): 38.

---o 1969. Tentative interpretations of new radiocar­bon dates for stone artefact assemblages from RoseCottage Cave, O.F.S. and Bushman Rock Shelter, Tvl.S. Afr. Archaeol. Bull. 24 (94): 57-59.

Mason, R. J.; Friede, H.; and Pienaar, J. N. 1974. KrugerCave. S. Afr. J. Sci. 70:375.

Matlbews, L. H. 1939. The bionomics of Ibe spottedhyaena, Crocuta crocuta Erxl. Proc. Zool. Soc, Lond.,ser. A, 109:43--56.

Mayhew, D. F. 19n. Avian predators as accumulators offossil mammal material. Boreas 6:25-31.

Meester, J. 1955. Fossil shrews of South Africa. Ann.Transv, Mus. 22:271-78.

---o 1958. Variation in the shrew genus Myosorex insouthern Mrica. J. Mammal. 39:325-39.

---o 1963. A systematic revision of the shrew genusCrocidura in southern Africa. Transvaal Museum

Memoirs, no. 13.---o 1971. Family Chrysochloridae. In The mammaís

01 Africo: An idetuification manual, ed. J. Meester andH. W. Setzer. Washington, D.e.: Smithsonian Institu­tion Press.

Meester, J.; Davis, D. H. S.; and Coetzee, C. F. 1964.Anínterírn classíficatíon of southern Afrícan mammals:Cyclostvled, King Williams Town: Zoological Societyof South África and South African Council for Scientificand Industrial Research,

Meester, J., and Dippenaar, N. J. 1978. A new species ofMyosorex from Knysna. Ann. Transv. Mus. 3J (4):29-42.

Meester,J.; Davis, D. H. S.; and Coetzee, C. G. 1964.Anmalia: Soricidae) from Southern Mrica. Ann. Transv,Mus. 27;269-77.

Meyer, G. E. 1978. Hyracoídea. In Evolution 01Afrícanmammals, ed. V. J. Maglio and H. B. S. Ccoke, pp.284--314. Cambridge; Harvard Uníversity Press.

Milis, M. G. L. 1973. Tbe brown hyaena. Afr, Wildl. 27(4):150-53.

---o 1974. Brown hyaena in the Kalahari GemsbokNationaí Park: Ecology. Wld. Wildl. Yb., 1973-4, pp.140-48.

---o 1976. Ecology and behaviour of the brownhyaena in the Kalahari, with sorne suggestions for rnan­agement. Proceedings o[ the Symposium on the En­dangered Wildlife Trust, Pretoria, July 1976, pp. 36-42.

Milis, M. G. L., and Milis, M. E. J. 1977. An analysis ofbones collected at hyaena breeding dens in theGemsbok National Parks (Mamrnalia: Carnívora. Ann.Transv. Mus. 30 (14); 145-55.

Misonne, X. 1969. African and Indo-Australian Muridae:Evolutionary trends. Ann. Mus. Afr, Cent., ser. 8,172:1-129.

---o 1971. Order Rodentía, In The mammals of Af­rica: An identification manual, ed. J Meester andH. W. Setzer, Part 6. Washington, D.C.: SmithsonianInstitution.

Mitchel\, B. L.; Shenton, J. B.; and Uys, J. C. M. 1965.Predation on large marnmals in the Kafue NatíonalPark, Zambia. Zoologica Africana 1 (1): 297-318.

Mogg, A. O. D., 1975. Important plants of Sterkfontein:An iIlustrated guide. Johannesburg: University of theWitwatersrand.

Mohr, E. 1967. Der Blaubock, Hippotragus Ieucophaeus(Pallas, 1766): Eine Dokumentaüon. Mamrnalia De­picta. Hamburg: Paul Parey.

Mollett, O. 1947. Fossil mammals from the MakapanValley, Potgietersrust. I. Primates. S. Afr. J. Sci.43:295-303.

Montgomery, T. H. 1899. Observations on owls, withparticular reference to their feeding habits. Am. Nat,33:56J.-72.

Mouritz, L. B. 1915. Notes on the ornilhology of theMatopos district, Soutbern Rhodesia. Ibis, 10th ser. 3(2): 185-216.

Mundy, P. J., and Ledger, J. A. 1976. Griffon vultures,carnivores and bones. S. Afr. J. Sci. 72:106-10.

Napier, J., and Napier, P. 1967. A handbook of livingprimates. London: Academic Press.

Nel, J. A. J. 1969. Tbe prey ofowls in the Namib Desert.1. The spotted eagle owI, Bubo africanus, at SossusVlei. Sci. Papo Namib Oeserl Res. Sla. 4 (37­53):55-58.

Nel, J. A. J., and Nolte, H. 1965. Notes on the prey of

References 357

owls in the Kalahari Gemsbok National Park. Koedoe8:75-81.

Noddle, B. 1974. Ages of epiphyseal closure in feral anddomestic goats and ages of dental eruption. J. Ar­chaeol, Sci. 1:195-204.

Oakley, K. P. 19540. The dating nf the Australopithe­cinae of África. Am J. Phys. Anthrop., n.s., 12 (1):9-28.

---o 1954b. Dawn ofman in South Africa. Proc, Roy.lnst, GI. Brit. 35 (3):641-56.

---o 1956. The earliest ñre-makers. Antíquity30:102-7.

---o 1960. The hístory ofSterkfontein with a commentby B. D. Malan. S. Afr, J. Sci, 56:1\0.

Olson, T. R. 1974. Taxonomy of the Taung skul\. Nature(London) 252:85-86.

Pager, H. 1971. Tbe rock art oflhe Ndedema Gorge andneighbouring valleys, Natal Drakensberg. In Rackpaintings of Soutñ Afríca, ed, M. Schoonraad, pp.27-33. S. Afr. J. Sci., Special issue no. 2.

Parkington, J. E. 1972. Seasonal mobi1ity in the LateStone Age. Afr. Slud. 31 (4): 223-43.

---o 1976. Coastal settlement between the mouths ofthe Berg and Olifants rivers, Cape Provínce. S. Afr.Archaeol, Búll. 31:127-40.

Parkington, J. E., and Poggenpoel, C. 1971. Excavationsat De Hangen, 1968. S. Afr, Archaeoí, Bull. 26:3-36.

Partridge, T. C. 1973. Geomorphological dating of caveopening at Makapansgat, Sterkfontein, Swartkrans andTaung. Nature (London) 246:75-79.

---o 1975. Stratigraphic, geornorphological andpalaeoenvironmental studies of the MakapansgatLimeworks and Sterkfontein hominid sites: A progressreport on research carried out between 1965 and 1975.Manuscript.

Partridge, T. C., aand Talma, A. S. N.d. Carbon andoxygen ísotope measurements at Makapansgat andSterkfonteín, and their palaeoenvironmental signifí­canee. In Press.

Patterson, B. J968. The extinct baboon Parapapio jonesíin the early Pleistocene of north western Kenya. Bre­vioro 282:1-4.

Patterson, B., and Howells, W. W. 1967. Hominid hum­eral fragment from early Pleistocene of NorthwesternKenya. Science 156 (3771): 64-M.

Patterson, B.; Behrensmeyer, A. K.; and SiII. W. D.1970. Geology and fauna of a new Pliocene locality innorth western Kenya. Nature (London) 226:91S-21.

Peabody, F. E. 1954.Travertines and cave deposits oflheKaap Escarpment of South Africa, and the type localityof Auslralopithecus africanus Oart. Bull. Geol. Soc.Am. 65:671-706.

Pearson, K. 1914.The lije. letters and labours ofFrancisGalton, Vol. 1. Cambridge: Cambridge UniversityPress.

Percival, A. B. 1924. A gorne ranger' s note book. Lon­don: Nisbel.

Peringuey, L. 1911. Tbe Stone Ages of South Mrica asrepresenled in the colleclion of the South AfricanMuseum. Ann. S. Afr. Mus. 8:1-218.

Perkins, D. 1973. A critique on the methods of quantify­ing faunal remains from archaeological sites. InDomestikations/orschung und Geschichte der Haus­tiere, ed. J. Matolcsi, pp. 367~9. Budapest: AkadémiaiKiadó.

Perkins, D., and Daly, P. 1968. A hunters' village in

358 References 1Neolithic Turkey. Scl, Am. 2t9 (5): 96-106.

Petter, F. t971a. Order Lagomorpha. In The marnmals ofAfrica: An identification manual, ed. J. Meester andH. W. Setzer, part 5. Washington, D.C.: SmithsonianInstitution Press.

---o 197tb. Subfamily Gerbiliinae. In The mamrnatsofAfríca: An identifícation manual, ed. J. Meester andH. W. Setzer, part 6.4. Washington, D.C.: SmithsonianInstitution Press.

Petter, G. 1973. Carnivores Pleistocenes du Ravin d'OI­duvai (Tanzanie), In Fossil vertebrales al Afríca, ed.L. S. B. Leakey, R. J. Savage, and S. C. Coryndon,3:43-100. London: Aeademic Press.

Pienaar, U. de V. 1969. Predator-prey re1ationshipsamongst the larger marnmals in the Kruger NationalPark, Koedoe 12:108-76.

Plug, 1. 1978. Collecting patterns of six specíes of vul­tures, Ann. Transv. Mus. 31 (6): 51---{j3.

Poeoek, R. 1. 1932. The leopards of África. Proc, Zoo/.Soco Lond. 1:543-91.

---o 1941. The fauna of Britisb India. includíngCeylon and Burma. Vol. 2. Mammalia. London: Taylorand Franc:is.

Pocock, T. N. 1969. Micro-fauna provisionalIy identifiedfrom sieved material recovered thus far from Dump 8 atSterkfontein, 1967-1969. Appendix. S. Afr, Arch. Bull.24:168-69.

---o 1970. Pleistocene bird fossils from Kromdraaiand Sterkfontein. Ostrich, suppl. 8, pp. 1---{j.

---o 1976. Pliocene mammalian microfauna fromLangebaanweg: A new fossil genus linking theOtornyinae with the Murinae. S. Afr. J. Sci, 72:58---{j0.

Prister, A. 1898. Preliminary notes on glacial phenomenaat Pretoria and on the Rand. Trans. Geol. Soco S. Afr,3:7~2.

Raczynski, J., and Ruprecht, A. L. 1974. The effeel ofdigestion 00 the osteologicaI cornpositíon of owI pel­lets. Acta Om., Warsz. 14 (2): 1-13.

Rau, C. 1876. Early man in Europe, New York: HarperBros.

Rau, R. E. 1974. Revised Iist of the preserved material ofthe extinct Cape Colony quagga, Equus quagga quagga(Gmelin). Ann. S. Afr, Mus. 65 (2): 41~7.

Reed, C. r., and Reed, B. P. 1928. The meehanism ofpellet formatíon in the great homed owI (Bubo vír­giniatus). Sclence 68:359-60.

Repenning, C. A. 1965. An extinct shrew from the earlyPleistoeene of South Africa. J. Mamma/. 46 (2):189-96.

Roberts, A. 1951. The mammals ofSoutb Africa. Johan­nesburg: Centra! News Ageney, South Afríca.

Robinson, J. T. 1953a. Meganthropus, auslralopithecinesand hominids. Am. J. Phys. Anthrop., n.s., 11 (1): 1-38.

---o 1953b. Te/antbropus and its phylogeneticsignifieance. Am. J. Phys. Anthrop., n.s., 11:44>'-502.

---o 1954. The genera and speeies of the Australopith­ecinac. Am. J. Phys. Anthrop., n.s., 12:181-200.

---o 1956. The dentition 01 the Australopithecinae.Transvaa] Museum Memoirs, no. 9.

---o 1957. Occurrence of stone artefacts wilh Austra-/opilbecus al Sterkfontein. Nature (London)180:521-24.

---o 1958. The Sterkfonlein lool-maker. Leech28:94--100.

---o 1959. Abone imp1ement from Sterkfonlein. Na­lure (London) 184:583-85.

---o 1961. The austraJopithecines and their bearing onthe origin of man and of stone tool-making. s. Afr. J.Sci. 57:3-13.

---o 1962. Sterkfontein stratigraphy and tbesignificance of the extensión site. S. Afr. Arcnaeol,Bull. 17:87-107.

---o 1963. Adaptive radiation in the australopithecinesand the origin of mano In Afrícan ecology and humanevo/ution, ed. F. C. Howetl and F. Bourliere, pp. 38>'­416. Viking Fund Publications in Anthropology, no. 36.New York: Wenner-Gren Foundatíon.

---o ]966. The distinctiveness of Homo habitis, Na­IUre (London) 209:957---{j().

---o 1972. Ear/y homíníd posture and locomotion,Chicago: Universíty of Chícago Press.

Robinson, J. T., and Mason, R. J. 1957. Occurrence ofstone artefacts with Austraíopithecus at Sterkfontein.Nature (London) 180:521-24.

Rowe, E. G. 1947. The breeding biology of Aquila ver­reauxi Lesson. Part l. [bis 89:387--410; par! 2, Ibis89:57&--606.

Ryan, M. 1961. Man-eating leopard. Afr. Wild Lije 15 (1):67-71.

SaLe, J. B. 1965. The feeding behaviour of rack hyraxes.(genera Procavia and Heterohyrax) in Kenya. E. Afr.Wildl. J. 3:1-18.

---o 1966. Daily food consumprion and mode of in­gestion in the hyrax. J. E. Afr. Nat. Hist. Soco 25 (3):214-24.

Sandelowsky, B. H. 1974. Archaeological investigationsal Mirabib Hill Rock Shelter, S. Afr, Archaeol. Soc.,Goodwin ser. 2:6>.-72.

Schaller, G. B. 1967. The deer and the tiger: A study ofwildlife in India. Chicago: University of Chicago Press.

---o 1972. The Serengeti lion : A study in predotor­prev relations, Chicago: University of Chicago Press.

Schütt, G. 1974. Die Camivoren von Würzburg­Sehalksberg: Mil einem Beitrag zur biostratigraphis­chen und zoogeographischen Stellung deraltpleistozánen Wirbeltier-faunen vom Mittelmain(unterfranken). N. Jb, Geo/. Palüom, Abh. 147 (1):61-90.

Selous, F. C. 1908. African nature notes and reminis­cences, London: Macmillan.

Sessions, P. H. B. 1966. Notes on rhe birds of Lengetiafarm, Mau Narok. J. E. Afr. Nat, Bist. Soco 26 (1):18-48.

---o 1972. Observations 00 Mackinder's eagle owlBubo capensis mockinderl Sharpe on a Kenya fann. J.E. Afr. Nat. Hist, Soc.• Nat. Mus. 138:1-19.

Shackleton, N. J. 1973. Oxygen isotope analysis as ameans of determining season of occupation of pre­historie mídden sites. Archaeometry 15:133-41.

Shapiro, M. M. J. 1943. Fossil mammalian remains fromBankies, Kroonstad districl, O.F.S. S. Afr. J. Sci.39:176-81.

Shaw, E. M.; Woolley, P. L.; and Rae, F. A. 1963.Bushman arrow poisons. Cimbebasia 7:2-41.

Shaw, J. C. M. 1937. Evidenee conceming a large fossilhyrax. J. Dent. Res. no. 1 (16): 37.

---o 1938. The leeth of lbe Soulh Mrican fossil pig(Notochoerus capensis syn. meadowsi) and theirgeological signifieance. Trans. R. Soco S. Afr. 26:2>.-37.

Shaw, J. C. Middlelon, and Cooke, H. B. S. 1941. Newfossil pig remaios from the Vaal Ríver graveIs. Trans.R. Soco S. Afr. 28:293-99.

Shipman, P., and Phillips, J. E. 1976. On scavenging byhominids and other carnivores. Current Anthrop, 17(1): 17{}-72.

Shortridge, E. C. 1934. The mammals 01 south west Af­rica. London: William Heineman.

Siegfried, W. R. 1965. On the food habits of the spottedeagle ow1. Ostrícñ 36:146.

---o 1968. Breeding season, clutch and brood sizes inVerreaux's eagle. Ostrích 39:139-45.

Silberbauer, G. B. 1972. The G/wi Bushmen. In Huntersand gatherers today, ed. M. G. Bicchierí. New York:Hoh, Rinehart and Winston.

Silver, I. A. 1969. The ageing of domestic animals. InScience in Archaeology, ed. D. Brothwell and E.Higgs, pp. 283-302. Leipzig: Thames and Hudson.

Simons, E. 1972. Primate evolution: An introduction 10

man's place in nature. New York: Macmillan.Simons, E. L., and Delson, E. 1978. Cercopithecidae and

Parapithecidae, In Evo/ution 01 African mammaís, ed.V. J. Maglio and H. B. S. Cooke, pp. 10{}-1I9. Cam­bridge: Harvard University Press.

Simons, J. W. 1966. The presence ofleopard and a studyof the food debris in the leopard lairs of the MountSuswa Caves. Kenya. Bull. Cave Exploration Group E.Afr. 1:51-69.

Singer, R., and Boné, E. L. 1960. Modem giraffes and thefossil giraffids of Africa. Ann. S. Afr. Mus. 45 (4): 375­548.

---o 1966. Hipparion in Afriea. Quaternaria 8:187-91.Skead, C. J. 1956. Prey of the bam ow1. Ostrícñ 27

(3): 148.---o 1963. Contents of pellets of bam owl, Tyto alba.

at a rural roost. Oslrieh 34 (3): 171-72.Skinner, J. D. 1976. Ecology of the brown hyaena,

Hyaena brunnea in the Transvaal, with a distributionmap for southem Africa. S. Afr. J. Sci, 72:262-69.

---o N .d. The striped Hyaena hyaena of the Judeanand Negev deserts and a comparison wilb the brownhyaena, H. brunnea. Israel J. Zool. In press.

Smart, C. 1976. The Lothagam 1 fauna: Its phylogenetic,ecological, and biogeographic signíficance. In Earliestman and envíronments in (he Lake Rudolfbosín, ed. Y.Coppens, F. C. Howell, G. Ll. Isaac, and R. E. F.Leakey, pp. 361-69. Chicago: University of ChicagoPress.

Smith, R. M. 1977. Movement pattems and feeding 00­haviour of leopard in the Rhodes Matopos NationalPark, Rhodesia. Arnoldia (Rhodesia) 8 (13): 1-16.

Smithers, R. H. N. 19710. The mammals of Bolswana.National Museum of Rhodesia Memoirs, no. 4.

--o 1971b. Family Felidae. In The mammals of Af­rica: An ídentificatíon manual, ed. J. Meester andH. W. Setzer, part 8:1, pp. 1-10. Washington, D.C.:Smithsonian Institution Press.

Southern, H. N. 1954. Tawny owls and their prey. Ibis96:348-410.

Speth, J. D., and Davis, D. D. 1975. Seasonal variabilityin early hominid predation. Paper presented at thesymposium Archeology in Anthropology: BroadeningSubject Matter, Seventy-fourth annual meeting of theAmerican Anthropological Association, 1975.

Stevenson-Hamilton, J. 1934. The low-veld: Its wildlifeand its people. London: Cassell.

---o 1947. Wild Iife in Soulh Afríca. London: Cassell.Stiles, D. N.; Hay, R. L.; and O'Neil, J. R. 1974. The

MNK chert factory site, Olduvai Gorge, Tanzania.

References 359

World Arehaeology 5:28>-308.Stoltz, L. P. 1977. The population dynamics of baboons

Papio ursinas Kerr 1792 in the Transvaal. D.Se.(wildlife management) thes¡s. University of Pretoria.

Sutcliffe, A. J. 1969. Adaptations of spotted hyaenas 10I)kliving in the British Isles. Mamm. Soco Bull. 31:1-4. -

---o 1970. Spotted hyaena: Crusher, gnawer, digestand collector of bones. Nature (London) 246 (5433d 5'f',1-211I{}-13. r

---o 19730. Caves of the East African Rift Valley.Trans. Cave Res. Group ofGr. Brít, 15 (1): 41-65.

---o 1973h. Similarity ofbones and antlers gnawed bydeer to human artefacts. Nature (London) 246 (5433):­428-30.

---o 1977. Further notes on borres and antlers chewedby deer and other ungulates, Deer 4 (2): 73-82.

Swayne, H. G. C. 1899. The leopard in Somaliland. InGreat and small game ofAfrica. ed. H. A. Bryden, pp.57>-79. London: Roland Ward.

Sydow, W. 1961. Fund eines Elefantenskeletts:Windhoek-Zoo (fossil). Mili. S. W. Afrika Wiss. Ges. 2(1): 2-3.

---o 1963. Elefanten in Zoo. Mili. S. W. Afrika Wiss.Ges. 4 (9): 5-6. 'ji

Tappen, N. C. 1969. The relationship of weathering _~ '10cracks to split-line orientation in bone. Am. J. Phys. /Anthrop., n.s., 31 (2): 191-97.

---o 1970. Main pattems and individual differences inbaboon skuIl split-lines and theories of causes of split­line orientation in bone. Am. J. Phys. Anthrop., n.s., 33(1): 61-71.

Thenius, E. 1966. Zur Stammesgeschichte der Hyánen(Carnívora, Marnmalia). Z. Sóugetíerk. 31:293-300.

Tobias, P. V. 1965. Australopitheeus, Homo hahiUs,tool-using and tool-making. S. Afr. Archaeol. Bull.20:167-92.

---o 1967. Olduvai Gorge, Vol 2. The eranium ofAustralopithecus (Zinjanthropus) boisei. Cambridge:Cambridge University Press.

---o 19680. The taxonomy and phylogeny of the aus­tralopithecines. In Taxonomy and phylogeny of OldWorld primates witñ reference 10 the orígín o/man, ed.B. Chiarelli. Turin: Rosenberg and Sellier.

---o 1968b. The age of death among the australopithe­cines. Anthropologist (Delhi), special volume, pp.23-28.

---o 1971.The brain in hominid evolutíon, New York:Columbia University Press.

---o 1973a. Implications of the new age estimates ofthe early South African hominids. Nature (London)246:79-83.

---o 1973b. A new chapter in the history of theSterkfontein early hominid síte. J. S. Afr. Biol. Soe.14:3{}-44.

--o 1974. Aspects of pathology and death arnongearly hcminids. Leecn 44 (3): JJ9-24.

---o 1975. Long or short hominid phylogenies?Palaeontological and molecular evidences. In The roleofnatural selectíon in human evolutíon, ed. F. M. Sal­zano. Amsterdam: North Holland.

Tobias, P. V., and Hughes, A. R. 1969. The new Wit­watersrand University excavation at Sterkfontein. S.Afr. Archaeol. Bull. 24 (3-4): 158-69.

Tobias, P. V., and von Koenigswald, G. H. R. 1964.Comparison between the Olduvai hominines and thoseof Java and sorne implications for horninid phylogeny.

360 Referenees

Nature (London) 204:515-18.Toerien, M. J. 1952. The fossil hyaenas of the

Makapansgat valley. S. Afr. J. Sci. 48:293-300.---o 1955. A sabre-tooth cat from the Makapansgat

valley. Palaeont. Afr, 3:43-46.Turnbull-Kemp, P. 1967. The leopardo Cape Town: How­

ard Timmins.van Hoepen, E. C. N. 1930. Fossiele perde van Cornelia,

O.V. S. Paíeont, Novors, Nas. Mus. Bloemfontein 2(2): 12-24.

---o 1932a. Die stamlyn van die Sebras. Paleont,Navors. Nas. Mus. Bloemfontein 2 (3): 25-37.

---, 1932b. Voorlopige beskrywing van Vrystaatsescogdíere. Paleont, Navors. Nas. Mus. Bloemfonteín 2(5): 63-65.

Van Hoepen, E. C. N., and van Hoepen, H. E. 1932.Vrystaatse wilde varke. Paleont, Navors. Nas. Mus.Bloemfantein 2:39-<i2.

Van Lawiek-Goodall, H., and van Lawick-Goodall, J.1970. lnnocent killers. London: Collins.

van Riet Lowe, C. 1947. Die ontdekking van die Sterk­fontein grotte, S. Afr. Sci. 1 (4): 85-86.

Vernon, C. J. 1965. The 1964 black eagle survey in theMatopos, Rhodesia. Amoldía (Rhodesia) 2 (6): 1-9.

---o 197\. Owl foods and otber notes from a trip toSouth West Africa. Ostrich 42 (2): 153-54.

---o 1972. An analysis of owl pellets colleeted insouthern Africa. Ostrich 43:109-24.

Visser, J. 1963. The blaek eagles of Zuurhoek, Jansen­ville. Afr. Wild Lije 17:191-94.

Vogel, J. C. 1969. Radioearbon dating of Bushman RockShelter, Ohrigstad District. S. Afr. Archaeol. Bull. 24(94): 56.

Voigt, E. A. 1973. Stone Age molluscan utilisation atKlasies River Mouth caves. S. Afr. J. Sci. 69:3~9.

---o 1975. Studies of marine mollusea from ar­ehaeologieal sites: Dietary preferences, environmentalreeonstruetions and ethnologícal parallels. lo Ar­chaeozoologicaí studies, ed. A. T. Clason, pp. 87-98.Amsterdam: North Holland.

von Koenigswald, G. H. R. 1953. The Australopitheeinaeand Pithecanthropus. Proc. K. Ned. Akad. Wet., ser.B., 56:403-13.

Voorhies, M. R. 1969. Taphonomy and populationdynamics of an earIy Pliocene vertebrate fauna, KnoxCounty, Nebraska. Univ. Wyaming Cantrib. Geol.,special paper, 1:1-69.

Vrba, E. S. 197\. A new fossil a1eelaphine (Artiodactyla:Bovidae) from Swartkrans. Ann. Transv. Mus. 27 (5):59-82.

---o 1973. Two species of Antidorcas SundevalI atSwartkrans (Marnmalia: Bovidae), Ann. Transv. Mus.28 (15): 287-352.

---o 1974. Chronologtcal and ecological implieationsof the fossil Bovidae at the Sterkfontein australopithe­cine site, Nature (London) 250 (5461): 19-23.

---o 1975. Sorne evídence of chronology andpalaeoeeology of Sterkfontein, Swartkrans and Krom­draai from the fossi1 Bovidae. Nature (London) 254(5498): 301-4.

---o 19760. The fossíl Bovidoe ofSterlifontein, Swart­krans and Kromdraaí, Mem. Transvaal Museum, no.2\.

---o 1976b. Tbe sígnificance of bovid rernains as in­dicators of envíronment and predation patterns. InFossils in the making, ed. A. K. Behrensrneyer

and A. P. Hill. Chicago: University ofChicago Press.---o 1976c. Chronology and palaeoecology of the

South African early hominids: Sorne evidence from thefossil Bovidae of Sterkfontein, Swartkrans and Krorn­draai. Paper presented at the Ninth Congress,U.I.S.P.P., Nice, 1976.

---o 1977. New species of Parmularíus Hopwood andDamalíscus Sclater and Thomas (Alcelaphini: Bovidae:MammaJia) frorn Makapansgat. Palaeont. Afr.20:137-51.

---o 1978. Prob1ematieal a1eelaphine fossils from theKromdraai Faunal Site (Mammalia: Bovidae). Ann.Transv, Mus. 31 (3): 21-28.

Wallace, J. G. 1948. The bam ow1 in Miehigan. Tech.Bul/. Mich. (St, Col/.) Agric. Exp, Sta., no. 208.

Washbum, S. L. 1957. Australopitheeines: The huntersor the hunted? Am. Anthrop. 59 (4): 612-14.

Washbum, S. L., and Patterson, B. 1951. Evolutionaryirnportance of the South African ••Man-apes." Nature(London} 167:65(}"'5\.

Waterhouse, G., ed. 1932.Simon van der Stel'sjournal ofhis expedítíon 10 Namaqualand, 1685~. Dublin: Dub­lin Universíty Press.

Weinert, H. 1950. Über die Neuen Vor- und Frühmen­schenfunde aus Afrika, Java, China und Frankreich.Z. Morphol. Anthropol. 42:113-48.

---o 1951. Über die Vie1gestaltigkeit der Summo­primaten vor den Menschwerdung. Z. Morphol. An­thropol. 43:73-103.

Welbourne, R. G. 1973. Prey of the spolted eagle owl.Witwatersrand Bird Club News Sheet 83:17-18.

WeIls, L. H. 1951. A large fossil klipspringer from Pot­gietersrust. S. Afr, J. Sci. 47:167-68.

---o 1959. The nomenclature of South Afriean fossilequids. S. Afr. J. Sci. 55:64-66.

---o 1967. Antelopes in the Pleistocene of southernAfriea. lo Background to evolution in Africa, ed.W. W. Bishop and J. D. Clark. Chieago: University ofChieago Press,

---o 1969a. Faunal subdivision of the Quatemary insouthern Afriea. S. Afr. Archaeol. Bull. 24:93-95.

---o 1969b. Generic positionof"Phenacotragus" vanñoepeni, S. Afr. J. Sci. 65:162.

---o 1970. A Late Pleistocene faunal assemblage fromDriefontein, Cradock District, C.P. S. Afr. J. Sci.66:59-<i1.

WeIls, L. H., and Cooke, H. B. S. 1942. The associatedfauna and culture of Vlakkraal lhermal springs, O.F.S.3. The faunal remains. Trans. Roy. Soco S. Afr.29:214-32.

---o 1956. Fossil Bovidae from the LimeworksQuarry, Makapansgat, Potgíetersrust. Palaeont. Afr.4:1-55.

Wender, L. 1948. Animal encyclopedia: Mammals. Lon­don: George Allen and Unwin.

Wendt, W. E. 1972. Preliminary report on an ar­chaeologícaí researeh programme in South West Af­rica. Cimbebasia, ser. B, 2 (1): 1-61.

Wheat, J. B. 1967. A paleo-Indian bison kill. Sci. Am. 216(1): 44-52.

White, T. D., and Harris, J. M. 1977. Suid evolution andcorrelation of African homínid Iocalities. Science 198(4312): 13-21.

White, T. E. 1953a. A method of ealeulating the dietarypercentage of various food animals utilized by aborigi­nal peoples. Am. Antiq. 18 (4): 396-98.

,,

---o 1953h. Observations on the butchering techniqueof sorne aboriginal peoples. No. 2. Am. Antiq. 19 (2):160--64.

---, 1954a. Observations on the butchering techniqueof sorne aboriginal pecples. No. 4. Am. Antiq,19:257-59.

---o 1954b. Observations on the butchering techniqueof sorne aboriginaJ peoples. No. 5. Am. Antiq.19:259-62.

Wilkinson, M. J. 1973. Sterkfontein cave system: Evolu­tion of a karst formo M. A. thesis, Department of Geog­raphy, Witwatersrand University.

Wilson, V. J. 1969. The large marnmals of the MatoposNalional Park. Amo/dio (Rhodesia) 4 (13): 1-32.

---o 1970. Notes on the breeding and feeding habits ofa pair of barn owls, Tyto alba (Scopoli), in Rhodesia.Amo/dio (Rhodesia) 34 (4): 1-8.

Wilson, V. J., and Child, G. 1966. Notes on developmentand behaviour of two captive leopards. ZoologischeGarten 32:67-70.

Wilson, V. J., and Grobler, J. H. 1972. Food of theleopard, Panthera pardus (Linn.) in the Rhodes

References 361

Matopos National Park, Rhodesia, as determined byfaecal analysis. Amoldia (Rhodesia) 5 (35): 1-10.

winterbottom, J. M. 1966. Notes on the fcod ofTyto albain the Etosha Pan. Ostrich 37 (2): 139.

Wolhuter, H. 1948. Memories 01a game ranger, Johan­nesburg: Wild Life Protection Soclety, South Africa.

woodbum, J. 1%80. An introduction to Hadza ecology.In Man the hunter, ed. R. B. Lee and 1. DeVore, pp.49-50. Chicago: Aldine.

---o 1968b. Stability and flexibility in Hadza residen­tia! groupings. In Man the hunter, ed. R. B. Lee and I.DeVore, pp. 103-10. Chicago: Aldine.

Yellen, J. E. 1977. Cultural patterning in faunal remains:Evidence from the !Kung bushmen. In Experimentalarchaeology, ed. D. W. Ingersoll. New York: Colum­bia University Press.

Zapfe, H. 1939. (Quoted by Sutcliffe 1970.) Palaeo­biologíca 7:111.

Zihlman, A., and Tanner, N. N.d. Gathering and thehominid adaptation. In Female hierarchies, ed. L.Tiger. Chicago: Aldine. In press.

Index

Acacia albída, 13, 16Achatina sp., 237, 224, 254Album graecum, 57Alexander, A. J., 109Allochthonous assemblages, 23AUopod transport of bones, 139Amblysomus spelea, 239Andrieskraal Caves, 302Anttdorcas, 174Antidorcas australis, 174,242Antidorcas bondi, J74, 207, 218, 242,

246, 253, 259, 332Antldorcas marsupiaJis. 174, 242, 246Antidorcas cf. recki, 175.207,215,235,

253,259Apple-corer, 5Aquila verreauxi, 104, 106, 107, 108,

301-2Archaeological sites, classiñcation oí, 45Australopithecus afarensis, 149Australopithecus ofricanus, 141-49, 150,

199-204,314Australopithecus robustus, 149,229--32.

246,256, 268, 325Autochthonous assemblages, 23Autopod transport ofbones, 139Aves, 189,216,219,237,244,254,260

Barlow, G. W., 248Barn owl. See Tyto albaBearder, S., 62Beatragus sp., 171,242Black eagIe. See Aquila verreauxiBloubank River ValIey, 7Bcft's Farm Caves, 130Bone assemblages, macrovertebrate and

microvertebrate, 8Bone density, 43Bone flakes, 10,69. 141Bone scoop, 5Bovid size classes, 275Bmom, R., 118, 148, 192-94,220,248Brown hyena. See Hyaena brunneaBubo africanus, 120-21. 138,307Bubo capensis, 122-23, 138,262,308Bubo Iacteus, 124, 138Buckland, Dean, 4Bushman Rock Shelter, 34, 281-84Butzer, K. W., 221, 262Buxton-Norlim, 262

Canis, 164,259Canis brevirostrís, 164,206Canis mesometas, 165,206,234,241.

246,252Canis cf. terblanchei, 164.214,252Cape eagle owl. See Bubo capensísCaves, stages in fonnation of. 7--8Celtis trees, 266. 268

Cercopithecidae, 152, 15SCercopithecoides, 155Cercopithecoides wíttiamsi, 155,205,

240,258,316Cheetah, 24-29

feeding behavior, 24feeding experiments, 26-29

Chiroptera, 186Chlorotalpa spelea, 185Clark, J. D., 47. 194Clarke, R. J., 152,221Composite tool. 5Compressional effects 00 bones, 134--37Connochaetes, 169,207,215,235,242,

253,259Cooke, C. K., 30Cooke, H. B. S., 195-96Cordylus giganteus, 183,260Crack-and-twist techniques, 5, 54Crocídara, 185,261Crocodylus nüotícus, 183,260Crocuta, 162Crocuta erocuta, 56-57, 163,206,234,

251bone..cracking,69bone ftakes, 310bone-gnawing, 70-72bone transport, 60diet, 64, 294droppings,64food remains, 293-94food storage, 66huntinglscavenging range. 68jaw acrlon, 69. 295regurgitations, 63

Crocuta spelaea, 163Crocuta ultra, 163Crocuta venustula, 163Crossarchus transvaalensis, 167,252Cryptomys, 186,261Cryptomys robertsi, 239Cvnictis penicillata, 167,242

Damaliscus, 170,207,215,235,252Damaliscus ef. dorcas, 170,218,242Dart, R. A., ii, 3, 4, 7, 54, 57, 109, 147,

195,262Dasymys boltl, 239Davis, D. H. S., 124, 239Daytight Cave, 196Deacon, H. J. and J., 36, 50Dendromys, 188,261Dendromys antiquus, 239Desmodillus, 15, 188Dinofetis, 158,234,258,266,271,273Dinofelis abelí, 159Dinofe/is barlowi, 106, 206Dinofelis diastemata, 159Dínofetis piveteaui, 159

Dínopithecus, 154Dinopithecus ingens, 154,233, 26S, 326Diplomesodon, 186Drakensberg rock paintings, SO

Elephantulus, 184,239.261Elephas recki, 181,20SEloff, F. C., 68Eloff, J. F., 36Epiphyseal closure, 21. 43Equus burchelli, ISO, 215. 218, 253Equus capensis, 179,208,236,243,253Equus quagga, 180,243Euryboas, 161Euryboas nitidula. (62, 234Euryboas sitberbergi, 162,206Ewer, R. F., v, 56,157

Fackeltráger Rock Shelter, 41, 290-91Felis crassidens, 156.250Fire, 54, 140Fractures, 5-6. 140

Galton, Francia, 13, 15Gathered foods, 47. 50Gazeíta, 175,207,259Gerbils, 15Giant eagle owls. See Bubo lacteusGorgopithecus, 155Gorgoplthecus major, 155,250Gow, c., 271

Hakos River leopard Iair. 297Herpestes, 259Herpestes mesotes, 166,252Herpestes sanguineus, 166,241Heterohyrax brucei, 33, 281Hipparion, 179Hipparion Iybicum, 179.236,253Hippotragus cf. equínus, 172.207,253Hippotragus cf. níger, 172,242Hluhluwe Game Reserve, 59Horno, 151,213,240Horno habilis, 30Homotherium, 160,25]Hottentots, 12, 16, 17

butchering techniques. 15-17dogs, 16goars, 18,20,21, 27&-TIvillages, 11, 12, 18,275-76

Howell, F. C., 152,221Hughes, A. R., 8, 57, 109, 112, 148,

195-96,264Human food remalns, 51-52, 140,

292-93. See atso HottentctsHyaena. 141,162Hyaena bellas, 162,252Hyaena brurmea, 162, 234, 240, 259

bone accumulations, 74-75

363

remains from ehannel fill, 246, 333-34remains from Member 1,228,322-24,

327remaíns from Member 2,239.328--29,

330-31site details, 221-23stages in formation, 224stone artifacts, 227, 238, 245

Syncerus, 167,207,236,253

Taphonomic biases, 54Taphonomy, 7Tatera, 188,261Tatera robínsoni, 239Taung, 262Taurotragus ef. oryx, 169,215,243,253Telanthropus capensis. 152Terblanehe, G., 248Testudinidae, 184,216,244,247Testudo sp., 260Theropitñecus, 154

Theropithecus (Simopithecus¡ danieli,154,268,326

Timbavati River, 24-25Tobias, P. V., 8,147-49,151,195Tooth marks, carnivore, 143Topnaar Hottentots, 13Tomewton Cave, 56-57Tragelaphus sp. aff. angasi, 168,207,

243Tragelaphus cf. scriptus, 168,243,253Tragelaphus ef. strepsíceros, 168,236,

243,253Transitory camps, 45Tylo alba, 124-33

diet, 125, 126, 138distribution, 124-26prey, 307-10prey remains as habitat indicators,

127-28, 130-31

Uitkomst, 124

11

Index 365

Uitkomst baboon sleeping cave, 271-73Umfolosi game reserve, 24-25

Valencia Ranch, 26Van Lawick-Goodall, J., 62Van der Stel'sjoumal, 48Varanus cf. niloticus, 183,216Yiverra sp., 166,259Voigt, E. A., 49Vrba, E. S., 9, 221, 249, 255, 268, 273Yulpes, 164Vulpes puícher, 234, 252

Warthogs, 59Washbum, S. L., 3, 149Wenner-Gren Foundation for An-

thropological Research, 221Wilson, V. J., 96Wilton Rock Shelter, 36, 284-89Wolhuter, H., 68