7/30/2019 Book Stein to Heathcote

1/3

REVIEWS I 79

nobody should be under any llusions as o he seriousness of thesituation. Well, perhap s t is, in the sense tha t the biolog ical produc-tivityof heworld s hreatened.But,comparatively,what s heextent of the threat? Is it life-threatening, as in the Sahel? Are wholeecosystems disappearing, as they are in the clearingf tropical forests?Are he akes being killed, as in Norwa y and Ontario? No: Sagesconcern is that the wilderness atmosphere of large areas has beendestroyed, and nobody really knows what effects industrial activityin the Arctic will have in the long term on the fauna and flora.The current, unprecedented rate of degradation of the world, wellillustrated in the report of the World Commissionn Environment andDevelopment ( the Bruntland Com mission), is consequential on pres-sureson he esources of thebiospherecausedby high evelsofresource use by humans and high ratesf human population increase.Solutions,even heoreticalones,ar edifficultan dparadoxical: npractice, the imperatives of political and religious leadership put theproblems beyond the capacity of democratic institutions to resolve.However, the Arctic is as well buffered from these pressures as is anygeographic zone: indigenous pop ulations were extremely sparse untilrecently, and industrial growth has been slow.Now hat hepopulationsofarcticpeoplesar eexpanding,an dmaterial expectations escalating, many look to economic developmentfor their future well-being. Development will indeed entrain someossof wilderness among the costs. However, the decisions mustbe madeby northerners, and not for them. Southerners can be confident thatconserva tion is close to the hearts of their cousins in the North. Theproblem will be one of balance.The Arctic&Its Wildlifes a book most people willike, for its manyillustrations and informative text.t is attractively presented and largelyfree of errors except orCanadianplacenameson page 18). It sdeficiencies are due mostly o ts ambitious scope, ts emphasis oncataloguing information and the weak relationship between the factualinformation it displays and the facile message it attempts to deliver.

Andrew H . MacphersonIndian and Northern Affairs CanadaP.O. ox 1500Yellowknife, Northwest Territories, CanadaXIA 2R 3

EX PLO RA TO RY H U MA N CRA N I O METRY O F REC EN T ESK -ALEUTIAN REGIONAL GROUPSRO M TH E WESTERNARC-T ICA N DS U B A R C T I C OF N O R T HA M E R I C A :AN E WTI O N .By GARY M. HEATHCOTE. Oxford:B.A.R., 1986. B ARInternational Series 301. xiv +332 p . E12.50.For at least ten years, judging by his own entries in the 50-pagebibliography here, Gary Heathcote has been pursuing the history ofarctic populations by consideration of skeletal remains. The presentbook, a revisionof his Ph.D. dissertation (and ty ped variously in picaand elite), is essentially a large-scale co nfirmatory study f this kind .From within a sufficiently rich suite of cranial measurements, heauthor has found a subset that, collectively,rehighly concordant withpopulation distance scores as reconstructed from geo graphic aswell as genetic linguistic criteria. He hopes (p.96) that his f indings,after emporaland patialextension,willenableamore obustattempt than heretofore allowed at unraveling human population histor-ical relationships in the Arctic and Subarctic zones of North America,Siberia, and Greenland.As I am a morphom etrician by trade, my interest was particularlypiqued by the major subordinate theme of thisork the enr ichment ofcraniom etrics within the b ound s f its present caliper-ba sed tradition.The [ovemding] reality is hat researchers in human osteology willcontinue, forome time, to have universal access to only the simple toolsused in this tudy.Pioneeringworks . . . willeventuallycompelosteolog ists to bandon their calipers, but for the immediatefuture, old

A PPRO A CH TO PO PU LA TI O N H I STO RI CA L RECO N STRU C-

fashioned osteology will persist. Certainly, a case ca n be made thatthere is room for improvement within the constraints of the currentlywidespread,caliper-wieldingapproach omorphologicalquestions.This study strives for such improvement. [ p. 63-64 .]In hisaspectofhisproject,Dr.Heathcotes imingwasmostunfortunate. After he gathered his35 OOO measures, but before publi-cation of this volume , there began to appear major revisions of the

founda tions of morph ometrics and its relation to m ultivariate sta tisticschanges that could have saved the author a great deal of effort. Theemphasis on nonstandard measures, of which he is justly proud, isbut a way-station toward the exploitation of strictly patterned sets ofcaliper measures as he equivalent of exp licitly recorded Cartesiancoordinates exploited in turn to construct optimal measurement scfor particular group differences. Heathcotes principal finding (p.93)is that his taxonomically optimal trait battery is dominated by breamainly of the neurocranium. This f inding could very likely have beengenerated wholly automatically by a direct construction of distancemeasures most sensitive to the distinction between Aleut and Inupiaqlanguage groups or between the Kagamil sample and the Kittigazuit.The appropriate method s mean ensor analysis, he ninth of nineother more rigorous approaches to morphological description lisbut not adopted, on page63. And the fiid ing s would then appear in acoherent diagram of typical deformations instead of being a list ofmotley d iscrete variables.Asmuchas I would ike odilate on hene wmorphometricdevelopments (see Bookstein et al. , Morphornetrics in EvolutionaryBiology, 1985), it would be inappropriate to dwell overmuch n themhere. But I must caution the reader not to adopt certain of Heathcoteunconventional variables, notably the perpendiculars from chordsto arcs of the vault. His goal, the representation of curving form, issound, but t is not achieved by a suite of m easures all confounded wthe posit ion of Bregma, Lamb da, or both. should point out that anyanalysis, however modem, of these skulls would be well servedy theauthors immense caution and competence in matters of measurementexecution.The approach toatascreening and precision testing recounhere is superb.It is more useful to tum from the slightly obsolete morpho metricdetailsof hisproject oaconsiderationof hecontribution hatmorphological data, according o whatever biometric canon, mightmake to stud ies of population history and prehistory, arctic or other-wise. Let us inquire generally whether morphology has any specialcontribution to make to such studies. In my morphometric view, theanswer is a somewhat qualif ied no, for tw o reasons .1 . Paradoxically,morphometricsoffers oogreata ichness ofmeasu rements for he a-posteriori association of variable ists withpredetermined classes to be meaning ful. From any reasonably well-distr ibuted scheme of landmarks (the authors 80 measures here areroughly equivalent to the digitizing f 28 separate points) , almost anygroup separation having a biological basis can be corroborated by asuitably constructed morphometric descriptor extracted via analysisfdeformation. But theseare no more automatically meaningful than arethe variables of a precisely analogous set, ratios measured at5 to thefirst set, which are variables on which a pair of populations preciselyagree in mean value: the invariants of the comparison, by contrastwith the covariants found by Heathcote. The existence of both suchsets is guaranteed by theorem, regardlessf the nature f the popu lations.2. The human head is highly constrained in its morphology. Thereexist mutually unintelligible languages, but, so to speak, no mutuallynondeformable heads. The variability of normal heads is quite small,and much of that is epigenetic. Then morphological distances meausing skullsare too unreliable a function of variable selection to serveas evidence of interjacency in lineage studies. Indeed, the subject ofHeathcotes book s in effect the unreliability of morphometric dist

as adumbration of population history.I would argue, instead, that morphology serves most usefully as adependent variable inhuman biological studies. I t is morphology that to be predicted, and ul t imately explained, by group mem bership,not the other w ay round. Heathcote studied skulls deposited before

7/30/2019 Book Stein to Heathcote

2/3

80 I REVIEWS

major population catastrophes, such as European intrusions. Can w efind the evidence of hose intrusions in the skulls? Effectsf climaticchanges? Dietary developments? Advances n hunting echnology?Wars and intermarriages? Would one d o better to study post-cranialremains?As in studies of protein evolution , traits having general selectivevalue can be expected oconflictwith a cladogram, being determinedinstead by convergence in response to environmenta l factors. Thustraits hat em erge as bestcorroborating hetrue linguistic)history of Eskaleut populations, Figure 1.2,will tend to the biologi-cally m eaningless. Even if these traits are truly the right indicators ofthe divergence in question, th ey represen t random selection from traitspace, likely never to be replicated. There is thus no meaning to theauthors findings that could reasonably be expe cted to gene ralize toother studies more extensive in pace or in time.Itwould be better to studymorphologicalcorrelatesofal l theenvironmental andanthropological eatures that mightause modifica-tion of form during descentor migration. The author s aware of suchcorrelates but consi ders hem nuisances that must be subjected ostatistical control to arrive at a taxonomically optimal subset(p . 196). On he contrary, these are the main meat of morphologicalanalysis: morphology as covariate. Thepresent work, while its datareextensive and itstatistical workup s heroic, does not persuadee thatosteometrics will serve as a potent source of independent evidence inpopulation historical studies.

Fred L . BooksteinCenterfor Human G r o w t h and Development

Universi ty of MichiganAnn Arbor , Michigan 8109

U . S . A .

Response from the author:I have heen aware of, and admired, Booksteins ground-breaking works onmorphom etric analysis since the late 1970s. While his review is predictablyscholarly and insightful in many respects, there are some issues raised andallegations made that prompt me to respond:1 . Regarding morphom etric methodology and he characterization of myapproach as slightly bsolete:Bookstein notes correctly thatdiscussed boththe limitations of the caliper-based tradition in osteometry and the advantagesconferred by certain newer ap proaches to qu antifying and an alyzing morphol-ogy, including his own. Herewould like to balance theserank admissions byilluminating some current limitationsf Booksteins m ethodology and strengthsof myown. The distinction between morphometric methodology at thenalyticvs. observational (data production) level is central to this critique.First of all, it should be emphasized that Booksteins mean tensor analysismethod is arguably the most powerful approach to the quantitativenalysis ofcomparative morphology currently available. Concedinghis begs the questionas to why I did not adopt his approach in my study. Bookstein provides thehistorical reason in his review. However, ifredesigned my study today, stillwould not follow his example for essentially two reasons that relate to whatperceive as current limitations of his methodo logy at theobservational level:Bookstein advocates the employment ofa) digitized two-dimensional coordi-nate data, produced on (b) images (e.g., photographs and radiographs) of theobjects under study; in his case, fish (see Bookstein et al ., Morphomerrics inEvolutionary Biology , 1985). In order for a next generation approach toetmore igourousmorphometric nquiry oevolve,abetter ntegrated datageneration and analysis) system is needed. Such a system needs to have thecapability of producing and making sense ofhree-dimensionaldata that can beproduced directly on the objects under study.Intuitively, we can all appreciate that complex three-dimensional objectsre.

best quantified hrough considering heir form n hree-dimensional space.Bookstein n fact reports tha t, while he had not yet written he computerprograms (as of 1985), hismean ensoranalysis heorems are. valid or

three-dimensional space Ibid . ,p. 131). I trust that these computer programs arnow written. With Booksteins kind cooperation,would like to be among thefirst of his colleagues to empirically test the three-dimensional version of hisanalytic too l, as long as its application is no t tied to the use of imagesf, in mycase, human crania.Once I amable oproduce - ccurately,precisely,andefficiently -three-dimensional coordinate data directlyn specimens, I hope to experimentwith he 3-D extension of Booksteins method. Hopefully, he appropriatetechnological developments needed for the production of a lightweight (seebelow) system for automated, direct production of 3-D data on objects willmaterializesoon.Currently available -D sonic digitizers ave not been widelyadopted, owing to their practical limitations, and purely manual ways (longavailab le) of producing 3-D data directly n objects are too time-consuming foranyone working with large numbers of specimens. am encou ragedby the factthat mechanical 3-D coordinate measuring machines with co mputer interfaceare on the market, but they have been designed for industrial quality-controlapplications. While they are, in principle, adaptable for osteometric work, tare currently far too cumbersome for those ofs who must frequently heft ourmeasuring tools to our work sites. H opefully, portable models will soon bedeveloped.While my seeming av ersion to image analysis may seem reactionary, it iagenericaversionbut ather sbased on someparticularistic ealities ofproducing data on, and nterpreting, skeletal craniofacial form at he grossmacroscopic level. The plotting and registration of landm ark and surface oucoordinates rom mages,as ecommend ed by Bookstein,wouldproducelower-quality observational data than those produced by a judicious, caliper-wielding studentof craniofacial form, for reasons that include:a) the frequentambiguity of landmarks and outlines of the component parts of the craniofaciskeleton, (b) frequently encountered visual evidencef cortical bone remodel-ing,especially of a esorptivenatureandmost requently nvolving hezygomaxillary omplex,necessitating he nterpolation of original pre-metamorphic) locations f landmarks and surface outlines, andc) the fact thatcertain dimensions - f my rather extensive and unorthodox measurementbattery, at least- over aspects of form that could not be captured fromstandard anatomic v iew (superior, ateral, inferior, etc.) images, as advocatedby Bookstein.Standard view images of, e.g., fish gross anatomy, may very well serve asadequate sources f raw data in Booksteins morphometric studies. H owever,craniofacial skeletal studies, dependence on images would both (a) introduceadditionalmeasurementerrornoiseand b)preclude he akingofcertainfiner-grained measurementsf the component parts of the craniofacial comIn the former case, this is due to the fact that resolutionf landmarks, outlinesand tell-tale cortical bone surface metamorphosis is usually less problematicwhen one visually inspects the actual specimens, rather than pictures of same.Regarding the latter objection: Because many of my unorthodox measuremenhave known or assumed ontogenetic, functional and taxonom ic meaning , amloathe o abandon an obsolete caliper-based approach n favour ofanynewerapproachnotasamenable o hemorphometricdesideratum ofcomprehensive coverageof the form under study.Booksteins approach, especially its-D extension, hasmuch to recommendit in terms of its morphometric data analytic power. However, its widespreadapplica tion in, at least, human osteology will likely not materialize until betterquality data generation complementation is developed.

2 .Regarding thecharacterizationofmy variablesas motley and discreteHere, I assume hatby discreteBooksteinmeans hat my (bounded,continuous) variables are n ot spatially interconnected, in terms of each beingspatially relatable and referenced o all other variables in two-dimensionalspace. This s true and I admit that t would be far better, morphometrically, fthe case were otherwise (as in the case of mean tensor analysis). Howeveof my exploratory batteries were formulated n the basis of stepwise m ultiplediscriminantanalysis,and heirconstituentvariables epresent hosewithmeasuredcovariationvis-a-visallothervariables.Thesebatteriescontainvariables with properties of demonstrated, powerful non-redundant discrimination, i.e ., they each contribute substantial independent information on inter-groupmorphometricdifferentiation.Theseamibutesdonot eemwhollyconsistent with regarding these variables as motley discrete.3 .Regarding the claims that morphology is ill-suited to he investigation ofpopulation history and is best regarded as a dependent variable in humanbiological studies:This issue is likely f greater in terest to most readersf t h i sjoumal than those above. Afterll, if human osteologists re not able to extractinformationon where did these people come fromnd with whom are theymost closely related,o w erviceabilityas collaborators in population histori-cal reconstruction with archaeologists, historical linguists and cultural anthro-

pologists would be altogether nullified.First of all, regarding Booksteins claim that the head is too highly con-strained n ts morphology o be useful in population historical studies: his

7/30/2019 Book Stein to Heathcote

3/3

REVIEWS / 81

linguistic near-metaphor is overstated. Relatively monotypic fish taxaay wellfi t his claim , but he heads of relatively polytypic humans are remarkablyvariable if you look closely enough. This counterclaim of mine is based, notincidently,onempirical findings basedn large battery of0 craniofacial traitsranging from traditional larger-scale covering measurements of two or morebones to small-scale measuresf the developmental end products of individualossification centres.The more exhaustively and comprehensively one describes the morphometryof skeletal form, especially developmentally complex skeletal components likehuman cran ia, the more one appreciates variety at both the intra- and inter-grouplevels. The very fact that, in certain computer runs, 100% of my study craniawere correctly identified through (blind) classificatory multiple discriminantanalysis bespeaks the morphological distinctiveness of (even) closely relatedregional groupsof Inuit, Yuit and Aleuts.As to Bo oksteins cautions abo ut the formf crania being a joint function ofgene action, epigenetic interaction of developing tissue systems, sundry envi-ronmental effects ncluding nutrition and climate, and - wouldadd -plastic remodeling in response to biomechanicalt r ess , I am ully cognizantof these. Indeed, here are some extended discussions in my m onograph on themean ing of craniofacial form,Le ., the multiple factors that literally shape it.Further, I am fully aware of the dangers nh erent n making phylogeneticinferences from dendrogramdcladograms, owing to converging/diverging effectsof natural selection.Contrary to Booksteins claim,my Figure 1.2 is not a representation of thelinguis tic history of Eskaleuts, but rather is a composite representation ofaffinities based on linguistic and spatial considerations, which furthermore isconsistent with the cultural, historical and archaeological data on nter-groupaffinities. In my monograph I state the caseor considering the benchmarkfinter-group affinities, as presented in Figure 1.2, an accurate portrayal of thepopulation historical relationshipsf these populations. Boas and Sapir need notroll over in their graves, for there are independent lines of evidence that thelinguistic, cultural and biological attributes of p opulations have substantiallyco-differentiated in the westernNorth American Arctic and Subarctic.As to the traits which emerge(d) as best in my study: one particular 18-traitbattery (out of many batteries of variable trait number that were empiricallytested) was found to produce inter-group generalized distances most concordantwith the benchmark (Figure 1.2) of approximately known historical relation-ships. These traits were hardly a random selection, as claimedy Bookstein.Rather, this battery was composedf variables that werea) found a posreriorito be the most pow erful multivariate discriminators, less (b) those that werefound or reasoned to be most prone to imprecision, age-related change andsubtle intentional occipital deformation, lessc) those thatrenot morphometri-cally efficacious, less (d) those that cover areasf the craniofacial skeleton thatare biomechanically related, or physiologically responsive, to the function fmastication. Hardly a random selection of rait s!The above results constituted my principal finding (not the predominanceof neurocranialbreadths,asclaimed by Bookstein), v i z . , thatwhenonecarefully pares down a large trait battery, discarding variables susceptible orlabile to v arious sources of noise, one is left with a reduced trait battery thatyields morphometric distance results that have apparent phylogenetic (popula-tion historical) meaning or, at least, an approximately synchronic groupf lateprehistoric/early historic Eskaleutian skeletal samples.By extension, I argue that this taxonomically optimal battery cane used toadvantage in tracing theh r e a d s of population historical continuity farther backinto the prehistoric past,n a fashion analogous to the direct historical approachin archaeology. I emphasize iny monograph that my particular findings,.e.,the specific 18-trait composition of my taxonomically efficacious battery, arenot generalizable. In other words, this particular battery works for herestricted region studied- o more and no ess.However, the above position may be overly conservative. Could the verybattery , out of many tested, that was constructed of the most powerful non-redundant discriminators, less those traits that were demonstrated or reasoned tohave been noise-modified by measurementerror, age regression, deformation,poor morphometric meaning and biomechanical functions,ust co incidently bethe one that performed bes t in a taxonomic application? While my particularis-tic findings re not generalizable to other studies more extensive in space orntime, as Bookstein correctly indicates,my methodology/approach toqualitycontrol screening f variables may efacto yield a reduced trait subset that maybe generally efficacious, taxonomically.Obviously, andI th nk for good reason,I am not so pessimistic asBooksteinabout the prospects for squeezing information about population origins, migra-tions and affinities out of the sizes and shapes of head bones. That we have notdone an especially good job of addressing these problems, for some of thereasons that Bookstein articulates,would agree. Likewise, fully concur withBookstein that morphology is a covariate, a joint function of many interactinggenetic, epigenetic and extragenetic factors. Buthis eality need notead to the

raising of hands in submission. Rather, the challenge is to figure out w ays fteasing apart those traits thatre informational vs. those that lead us down falsepaths in our attempts at population historical reconstruction.Gary M . HeathcoteMount Sinai Hospital Research Institute

600Unive rsity AvenueToronto, Ontario, CanadaM5G 1x5

WI N D O W O N TH E PA ST: A RCH A EO LO G I CA L A SSESSMEN TOF THE PEACE POINT SITE, WOOD BUFFALO NATIONALPA RK , A LBERTA . By MARC G. TEVENSON. Studies in Archaeol-ogy, Architecture and History. Ottawa: National Historic ParksndSites Branch, Parks Canada, 1986. 145 p. , illus. , appe ndices , refs.Sof tbound. Cdn$8.75; outs ide Canada Cdn$10.50.This volume summarizes the resultsf two seasons f archaeologicalsurvey and assessment at and in the vicinity of the Peace Point Site(IgPc-2) in Wood Buffalo National Park. I t comprises eight chapters,beginning with arief introduction in Chapter I and a short synthesif

th ehistory of archaeological research n northeastern Alberta andsouthwestern Northwest Territories n Chap ter 11.Chapter 111 focuses on the significance of Peace Point in historictimes, on a reconstructionof present and past environ ments in the sitearea and on a summaryf survey and tes t excavations done at the PPoint Site and other nearby sites during the 1980 season. Additionalresults of the 1980 season est excavations at the Lake One D une Site(IgPc-9) and a detailed survey f a 32 km sectio n of the b ank s of thePeace River are reviewed in Chapter IV. I t is argued that the areaaround Peace Point has been occupied by bo real forest-related andplains-related peoples at intervals over the last 000-8000 years.Chapter V, the most substantial sectionf the book, summarizes theresults of the 1981 season excavations at the Peace Point Site. Thischapter ncludes a statement about he research goals of the 1981season, a descriptionof f ield methods anda detailed summ ary of thecultural remains associated with each of the 18 stratigra phic ally dis-crete cultural levels identif ied during the excavation. The latter dission s nterspersed with comments on perceived patterning n hecomposition of individual lithic and faunal assemblages and in theirspatial distr ibutions.Broad emporal rends in technology , ubsistencepatternsan dinter-regional affiliations are reviewed in Chapter VI. As well, con-cepts derived fromrecent ethnoarchaeological researchy L.R. Binfordare employed in this chapter to in terpret patterns f regional mobil-ity and the inferred alternating function f the Peace Point Site as abase camp and as a campsite .Chapter VI1 isdevoted odeveloping wocultural ormationmodels deemed to be potentially useful for future explicationof thebehavioral history preserved at the Peace Point Site. The i t model ,explained in greater detail elsewhere (Stevenson, 1985), examines thedynamics of artefactassemblage ormationbasedona unctionalthree-stage sequence f depositional and post-depositionalvents.. Th esecond model focuses on the marked dichotomyf male/female act ivi-ties characteristic f northern hunter-gatherer groups and speculatesnthe effects thathis behavioral duality may have had on the formationfactivity-related patterning in the archaeological record.The final chapter (Chapter VIII) summarizes the major points rain the preceding seven chapters. The three appendices to this bookinclude the results of a detailed soil analysis of samples gathered in1980 and 1981 from the Peace Point Site and its immediate environs(AppendixA ),amacro-an dmicrofloralanalysis of soilsamples(concentrated on the lower levels from the exposed cliff face at thePeace Point Site) gathered in 1980 (Appendix) and a summary tableof me tric and non-m etric attributes of lith ic artefac ts from selecte dlevels of the Peace Point Site (A ppend ixC).According to the author (p. 9) this work was designed to accomthree things: a) to illuminate the importancef the Peace Point Site, b)