age determination of sea turtle.pdf

Age Determination of LoggerheadSea Turtles, Caretta caretta,

by Incremental GrowthMarks in the Skeleton

GEORGE R. ZUG, ADDISON H. WYNN,and CAROL RUCKDESCHEL

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 427

SERIES PUBLICATIONS OF THE SMITHSONIAN INSTITUTION

Emphasis upon publication as a means of "diffusing knowledge" was expressed by the firstSecretary of the Smithsonian. In his formal plan for the Institution, Joseph Henry outlined aprogram that included the following statement: "It is proposed to publish a series of reports,giving an account of the new discoveries in science, and of the changes made from year to yearin all branches of knowledge." This theme of basic research has been adhered to through theyears by thousands of titles issued in series publications under the Smithsonian imprint,commencing with Smithsonian Contributions to Knowledge in 1848 and continuing with thefollowing active series:

Smithsonian Contributions to Anthropology

Smithsonian Contributions to Astrophysics

Smithsonian Contributions to Botany

Smithsonian Contributions to the Earth Sciences

Smithsonian Contributions to the Marine Sciences

Smithsonian Contributions to Paleobiology

Smithsonian Contributions to Zoology

Smithsonian Folklife Studies

Smithsonian Studies in Air and Space

Smithsonian Studies in History and Technology

In these series, the Institution publishes small papers and full-scale monographs that reportthe research and collections of its various museums and bureaux or of professional colleaguesin the world of science and scholarship. The publications are distributed by mailing lists tolibraries, universities, and similar institutions throughout the world.

Papers or monographs submitted for series publication are received by the SmithsonianInstitution Press, subject to its own review for format and style, only through departments of thevarious Smithsonian museums or bureaux, where the manuscripts are given substantive review.Press requirements for manuscript and art preparation are outlined on the inside back cover.

Robert McC. AdamsSecretarySmithsonian Institution

S M I T H S O N I A N C O N T R I B U T I O N S T O Z O O L O G Y • N U M B E R 4 2 7



George R. Zug, Addison H. Wynn,and Carol Ruckdeschel

SMITHSONIAN INSTITUTION PRESS

City of Washington

1986

ABSTRACT

Zug, George R., Addison H. Wynn, and Carol Ruckdeschel. Age Determi-nation of Loggerhead Sea Turtles, Caretta caretta, by Incremental GrowthMarks in the Skeleton. Smithsonian Contributions to Zoology, number 427, 34pages, 9 figures, 6 tables, 1 appendix, 1986.—Periosteal growth in theskeleton of Caretta caretta is cyclic and produces a record of distinct bonylayers (growth marks). These layers are most apparent in the long bones,particularly the humerus and femur. Correlative evidence supports the pro-duction of one layer (i.e., growth increment) each year in some reptiles, andthe annual nature of each bony layer is an explicit assumption in ourestimation of ages of Caretta from Cumberland Island, Georgia. Sectionsfrom the shafts of humeri were prepared with standard histological tech-niques, and the marks of skeletal growth were counted and measured. Owingto resorption of the early periosteal layers, determination of number ofgrowth marks and, hence, age requires an extrapolation from the remainingmarks to total growth marks produced during the life of the turtle. Variousage estimates are possible; the most reliable are derived from the narrowestaxis of the humerus sections and suggest that, on the average, Georgia Carettaattain sexual maturity in their thirteenth to fifteenth years, with someindividuals possibly maturing as early as six years and others as late as theirtwenty-fifth year. Since the average age of maturity agrees with the resultsof a mark-recapture study in a neighboring Florida population, the value ofthe skeletochronological technique and its assumptions are confirmed. Thetechnique is not advocated as a method for the age determination of individ-ual sea turtles, but does provide statistical age estimates for different sizeclasses of turtles.

OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and isrecorded in the Institution's annual report, Smithsonian Year. SERIES COVER DESIGN: The coralMontastrea cavernosa Linnaeus.

Library of Congress Cataloging in Publication DataZug, George R., Addison H. Wynn, and Carol RuckdeschelAge determination of loggerhead sea turtles, Caretta caretta, by incremental growth marks in

the skeleton(Smithsonian contributions to zoology ; no. 427)Bibliography: p.Supt. of Docs. no. : SI 1.27:4271. Loggerhead turtle—Age determination. 2. Skeleton—Growth. 3. Reptiles—Age deter-

mination. I. Wynn, Addison H., 1955- . II. Ruckdeschel, Carol. III. Title. IV. Series.QL1.S54 no. 427 591s [597.92] 85-600229 [QL666.C536]

Contents

Page

Introduction 1Acknowledgments 2

Materials, Methods, and Rationale 3Turtles 3Selection of Bone 3Preparation of Bone 3Measurements and Counts 4

Anatomy of the Humerus 5Gross Morphology 5Morphometrics 8Tissue Organization 9

Estimation of Age 10Assumptions and Difficulties 10Data Analysis 13

Life History Implications 18Summary 22Appendix 24Literature Cited 31

111

Dedicated to the memory of Chuck Ossola, a friend and outstandingscientific administrator. Chuck served as executive officer for the National

Museum of Natural History from 1977 to 1982.



George R. Zug, Addison H. Wynn,and Carol Ruckdeschel

Introduction

The "lost year" of sea turtles is much discussed.Upon entering the sea, what happens to the tinyhatchling sea turtles until they are seen again ashalf-grown juveniles? Where do they go, what dothey do, and, most importantly, for how long arethey absent? The intense and voluminous re-search efforts of the last decade and a half haveslowly begun to answer these questions and haverevealed many unanticipated and fascinating fea-tures of sea turtle biology, behavior, and ecology.

Of particular interest to us are growth rates,age at sexual maturity, and other aspects of ag-ing. These aspects are critical in the developmentof demographic profiles for the different seaturtle species and populations, hence their con-servation and management. Mark and recapturestudies, such as those of Balazs (1980), Limpusand Walter (1980), and Mendonca (1981), arethe ideal means for obtaining the most accurateage determinations. Such studies are, however,

George R. Zug and Addison H. Wynn, Department of VertebrateZoology, National Museum of Natural History, SmithsonianInstitution, Washington, D.C. 20560. Carol Ruckdeschel, TheSettlement, Cumberland Island, P.O. Box 796, St. Marys, GA31558.

labor intensive, time demanding in the short andlong term, and frequently logistically difficultand expensive. Captive-rearing, as done byWitham and Futch (1977) and Wood and Wood(1980), is similarly costly in time, labor, andequipment. Furthermore, captive rearing givesunnatural results, because these turtles growfaster and reach maturity sooner than wild ones(Balazs, 1979).

The research activities of the Universite deParis "Equipe de Recherche 'Formation squelet-tiques'" and particularly the investigations of Cas-tanet into skeletochronology suggested an alter-native approach that might convert the numer-ous strandings of dead sea turtles into potentiallyvaluable research specimens. They (Castanet etal., 1970; Castanet, Meunier, and Ricqles, 1977)provided a convincing argument that ecto-thermic vertebrates possess cyclic bone growthand that the cycles are annual. Other authorshave offered varied opinions upon annual cyclicgrowth of bone in turtles. Mattox (1936) was thefirst to recognize the potential of bone growthrings or marks as a means of aging turtles; henoted that, in Chrysemys picta, smaller/youngerturtles had fewer rings than did larger/olderturtles. Peabody (1961) was a strong advocate of

1

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

age determination through the use of these bonygrowth rings and he assumed, although he nevertested, the annual nature of the rings. He ob-served these growth marks in Chelydra serpentina,but provided no estimate of age. Suzuki (1963)studied osteogenesis in Trachemys scripta elegans.He accepted the presence of annual growth lay-ers in the long bones; however, he adamantlyrejected the use of these growth marks for agedetermination owing to the extensive resorptionand remodeling that occurs throughout growth.In contrast, Hammer (1969) suggested the useof the bony rings for aging, since he found a highcorrelation (r = 0.94) of number of rings tocarapace length. Dobie (1971) found no suchcorrelation in Macroclemys temminckii. He exam-ined sections of vertebrae and mandibles,whereas Hammer used a variety of skeletal ma-terial, but found the growth marks most evidentin long bones of Chelydra. Castanet and Cheylan(1979) found concordance between the numberof rings and the known age of a Testudo hermanniand high correlations (r = 0.89) for ring numberand carapace length (also scute annuli) in T.graeca and T. hermanni. Hohn and Frazier (1979)also noted a positive relationship between num-ber of bone layers and carapace length in severalspecies of sea turtles. Castanet (1982) proposedthat skeletochronology was promising for mostturtles with the exception of sea turtles owing tothe extensive remodeling in their long bones.

A number of amphibian and reptile speciesmatch age in years with number of bony growthlayers. Hemelaar and van Gelder (1980) foundthat more than 90% of toads (Bufo bufo) recap-tured showed an increase in number of marks ofskeletal growth (MSGs) equal to the differencein years between capture and recapture. Thisconcordance in a wild population is one of thestrongest confirmations for the annual produc-tion of one growth layer in ectothermic tetra-pods. Several studies by the Universite de Parisresearch group (Castanet, Cheylan, de Buffrenil,Gasc, Meunier, Naulleau, and Pilorge) confirmthe one-layer/one-year hypothesis. De Buffrenil(1980) showed that a 4-yr-old captive Crocodylus

siamensis possessed three complete growth layersand was forming a fourth one at the time of itsdeath. Castanet and Naulleau (1974) found thatonly 10% of their Vipera aspis born in captivityand raised under constant conditions match ageand number of growth layers; in contrast, captivevipers exposed to seasonal conditions producedone growth layer per year. Pilorge and Castanet(1981, table 2) demonstrated complete agree-ment between actual age and number of growthlayers. Minakami (1979) demonstrated that thegrowth period (deposition of a single zone) ofTrimeresurus okinavensis occurred from June toSeptember. His results derive from in vivo stain-ing of snakes maintained in the laboratory. Sim-ilarly, Castanet (1982) showed with fluorescentlabelling that one layer was deposited each yearin Lacerta viridis. Studies of Iguana iguana (Zug,Rand, Wynn, and Bock, 1983) revealed that thenumber of growth layers closely match theknown age of marked and recaptured animals.

Although numerous studies in turtles haveconfirmed the concordance of epidermal scuteannuli with age in years, there is only one studywith turtles of known ages that confirms theconcordance of bone layer number with age inyears. Castanet (1982) raised Emys orbicularis inseminatural conditions and with fluorescent la-belling showed a one-to-one relationship betweennumber of bone layers and years. Castanet andCheylan (1979) demonstrated a concordance ofepidermal scute annulus number with bonygrowth layers in Testudo graeca and T. hermanni;they also demonstrated a high positive correla-tion of skeletochronological age estimates andcarapace length. Others (e.g., Hammer, 1969)have also observed a high correlation betweencarapace length and age estimates. The dataconfirming the production of one bone layereach year is largely circumstantial, but convinc-ing. Herein, we summarize our successes andfailures for age determination in the loggerheadsea turtle through the use of skeletochronology.

ACKNOWLEDGMENTS.—This research projecthad its beginning in a series of coincidenceslargely linked through Charles W. Potter.

NUMBER 427

Shortly after G. Zug had read Castanet, Meunier,and Ricqles (1977), C. Potter introduced him toAleta Hohn and her spinner dolphin aging re-search. Then after some tentative histologicalpreparations of sea turtle bone proved successful,C. Potter introduced G. Zug to C. Ruckdeschel,thereby permitting the collection (EndangeredSpecies Permit: PRT 2-2381) and eventual ex-amination of a large, yet geographically re-stricted sample of sea turtles. C. Potter has ourthanks for providing the initial linkage and stim-ulation for the research reported herein. Theinitial histological preparations were made byMargaret M. Barber; however, the histology forthe research sample derives entirely from theskilled hands of Helen F. Wimer. Discussions onskeletochronology with Jack Frazier encouragedour research in the early stages. Nat Frazer andTerry Henwood most generously shared with ustheir current ideas and research data on Carettagrowth and age determination.

The manuscript has benefitted by the reviewsof Jacques Castanet, C. Kenneth Dodd, Jr., NatB. Frazer, J. Whitfield Gibbons, Tyrell A. Hen-wood, W. Ronald Heyer, Peter C.H. Pritchard,and Anders G.J. Rhodin. We appreciate theirvaluable advice, but remain totally responsiblefor the interpretations of the data.

This project was supported by the Smithsoni-an's Fluid Research Fund, NMNH Director'sFund, and the Department of Vertebrate Zool-ogy. Without their support, this research wouldhave been impossible.

Materials, Methods, and Rationale

TURTLES.—During 1979 and 1980, C.R.measured and collected a skeletal voucher speci-men from each sea turtle stranded on the oceanicbeach of Cumberland Island, Camden County,Georgia. A total of 187 Caretta caretta werestranded in 1979 (Ruckdeschel and Zug, 1980)and 222 in 1980 (Ruckdeschel and Zug, 1981).The preferred skeletal voucher was the completecranial and right forelimb skeletons for eachspecimen. Many specimens, however, lacked ap-

pendages, head, or some combination thereofowing to butchery by original captors, sharks, orbeachcombing souvenir hunters. Thus, our re-search samples are less than the total number ofstranded turtles. The humeral morphometric setconsists of 137 humeri, and the histological (agedetermination) set represents 119 humeri, in 81of which MSGs could be measured. These setsderive almost entirely from the 1979 strandings.

SELECTION OF BONE.—Growth increments (=layers or marks) are observed most consistentlyin the long bones of turtles (see, e.g., Mattox,1936; Hammer, 1969). Our preliminary histo-logical examination included a third peripheralbone from the carapace, the dentary (a sectionadjacent to the symphysis and one in the middleof the ramus), a centrum of a cervical vertebrae,a penultimate phalanx of forefoot, an ulna, anda humerus. Regular periosteal layers were evi-dent in the ulna and humerus. In the dentary,growth layers were present, but owing to thespongy nature of this bone, the layers were tooirregularly arranged to permit accurate and re-peatable counts and measurements. The peri-pheral bone, centrum, and phalanx showed nodiscernible pattern of incremental growth of per-iosteal bone, and the ulna possessed only a weakpattern. Growth marks were clearly evident inthe shaft of the humerus and, thus, the humeruswas selected for further examination. The femuralso shows a strong pattern of incremental per-iosteal growth (Frazier, 1982) and is likely as wellsuited for age estimation in sea turtles as thehumerus.

PREPARATION OF BONE.—The entire forelimb,usually the right, was removed, macerated inwater, washed, and air-dried. Prior to sectioning,each humerus was weighed and measured (meas-urements are described in the following section).A transverse section of bone approximately 5-7mm thick was cut with a bandsaw from the hu-merus just distal to the deltopectoral ridge. Thesection was stored in a A% formalin solution forat least 24 hrs prior to decalcification.

Histological preparation began with decalcifi-cation. Several decalcification solutions were


tried. RDO (a commercial decalcification solu-tion) was fast and effective; however, its speedand the unequal thickness of our bone sectionsresulted in over-decalcification. A formic-citricacid solution was very slow and resulted in un-even decalcification. A solution of equal parts of8% formic acid and 8% hydrochloric acid gavegood results. Decalcification in the formic-hydro-chloric acid solution required 4-8 days, depend-ing upon the thickness of the bone and its poros-ity. Complete decalcification was determined byX-ray.

The decalcified bone was rinsed in tap waterand processed in a Fisher Histomatic Tissue Pro-cessor as follows: (1) 95% ethanol for 24 hrs,three changes of alcohol, the last 12 hrs under avacuum; (2) 100% ethanol for 30 hrs, fourchanges of alcohol, the last 18 hrs under a vac-uum; (3) xylene for 12 hrs, three changes ofxylene, all under a vacuum; (4) paraffin (para-plast) for 6 hrs, two changes, both under a vac-uum. The paraffin-impregnated bone wasmounted in a block of paraffin and transverselysectioned (6 fim) on a standard microtome. Thebone was periodically soaked with water (cutsurface of block covered with dripping wet gauzepad) to obtain the best results during sectioning.All sections were stained with Harris hematoxy-lin and eosin and normally differentiated.

MEASUREMENTS AND COUNTS.—When a deadsea turtle washed ashore on Cumberland Island,a series of shell and soft part measurements wascollected. Only one measurement is reportedhere: curved carapace length (cCL), the midlinedistance from anterior edge of cervical scute totips of thirteenth marginal scutes measured overthe curve. Over-the-curve measurements are po-tentially more accurate for bloated carcasses thanstraight-line measurements (e.g., sCL), owing toshape changes; the measurement error increasesin both types as maceration and gaseous swellingbegin to separate the articular surfaces withinthe shell of stranded sea turtles. In spite of thepotential error, these shell measurements canserve as size class designators for comparison with

the nesting female loggerhead population of theGeorgia coastal islands and with populations ofmarked-and-recaptured loggerheads from otherareas.

Prior to removal of the bone section, a seriesof 12 straight-line distance measurements wasrecorded from each humerus, as well as its airdried weight. The measurements are: (1) maxi-mum length (ML), distance from proximal-mosttip of ulnar process to distal articular surface; (2)longitudinal length (LL), distance from proximalsurface of head to distal articular surface, parallelto longitudinal axis of humerus; (3) ulnar processlength (UPL), distance from proximal tip of ul-nar process to juncture of head and process; (4)proximal length (PL), distance from proximalsurface of head to distal edge of radial process,parallel to longitudinal axis; (5) proximal width(PW), distance from preaxial surface of head topostaxial surface of ulnar process, perpendicularto longitudinal axis; (6) radial process length(RPL), distance from pre- to postaxial edges ofprocess, diagonal to longitudinal axis; (7) widthat deltopectoral crest (DpCW), transverse dis-tance of shaft from pre- to postaxial surfaces atdeltopectoral crest; (8) medial width (M W), trans-verse distance from pre- to postaxial surfaces atpoint of minimum width; (9) distal width (DW),transverse distance from pre- to postaxial sur-faces at juncture of articular condyles with shaft;(10) maximum head diameter (MaxHD); (11)minimum head diameter (MinHD); (12) thick-ness (T), minimum depth in middle of shaft,approximately in vicinity of MW, perpendicularto longitudinal and transverse axes.

The histological sections were measuredgrossly and microscopically. The length andwidth of bone sections were measured with acaliper to 0.1 mm. The periosteal layers weremeasured to 1.0 fxm with a Wild micrometer andcounted at 60 X magnification on a transmittedlight microscope. The counts and measurementsof the layers were made along two perpendicularaxes, the long (pre-postaxial) and short (dorso-ventral) axes, of the bone section.

NUMBER 427

Anatomy of the Humerus

GROSS MORPHOLOGY.—The humerus of thesea turtles (Cheloniidae and Dermochelyidae)differs from that of all other extant turtles inseveral respects. Presumably, the differences arerelated to a structural reorganization of the fore-limb for aquatic flight. The sea-turtle humerus,

specifically that of Caretta caretta (Figure 1), isflattened in cross-section as well as in the proxi-modistal plane. Cross-sections through most re-gions of the diaphysis (shaft) have elliptical out-lines. The head lies at an approximately 45°angle from the diaphysis in contrast to the nearlyperpendicular orientation in other turtles. Thearticular surface retains the 180° arc, thereby

Ulnar Process

IntertubercularFossa

Radial Process

eltopectoralCrest

Diaphysis

EctepicondylarForamen

Epicondyles

Radial Ulnar

FIGURE 1.—Entire right humerus: A, dorsal view; B, ventral view.


FIGURE 2.—Growth series of humeri showing the closure of the ectepicondylar foramenthrough growth. Curved carapace lengths of specimens (left to right): 995, 965, 890, 815, 720,660, and 590 mm.

giving the articulation a limited ventral aspect aswell as medial and dorsal ones. In general ap-pearance, the articular surface is egg-shaped,with the blunter end ventral and the long axisdorsoventral. The anterior articular shoulder orflange, so characteristic of the humeral head ofother turtles, is absent; there is no hint of itsformer presence. The ulnar (medial) process (ortuberosity) is narrow and elongate, extendingposteromediad at approximately a 25° anglefrom the shaft; the medial tip of the process isstrongly proximal to the humeral head. The

radial (or lateral) process lies distal to the headand ventrally on the preaxial edge of the dia-physis. The radial process is often stated (Romer,1956:355) to have shifted distally on the shaft,yet a comparison of its position and its orienta-tion in other turtles suggests that the change inhead orientation and the elongation of the ulnarprocess result in a change in the radial process'srelative, but not actual position on the shaft.Another process arises on the preaxial bordercontiguously with the radial process. This processis suitably called the deltopectoral crest owing to

NUMBER 427

FIGURE 3.—Frontal section of the humerus: dorsal half on left, ventral half on right; 660 mmcCL.

its location in the center of the insertions of thedeltoid and pectoral muscles; however, this nameis used by others, such as Romer (1956), as asynonym of the radial or medial process. In seaturtles, they are clearly two distinct processes andshall be so treated in the subsequent discussion.

Immediately distal to these two processes, thediaphysis narrows and then gradually widens(along the pre-postaxial axis) distally toward theepicondyles. The epicondylar surface is smoothwith only a slight medial concavity to separatethe radial and ulnar epicondyles. The radial epi-condyle is notched in subadults and juveniles bythe ectepicondylar groove. The groove graduallycloses with age as the epicondylar area grows

beyond and around the radial nerve and bloodvessels (Figure 2).

Externally, the humerus bears a smooth, densebony surface except for the articular surfacesand a few of the major muscle insertions. Thehumeral head, the ulnar, radial and deltopectoralprocesses, and the epicondyles possess smoothporous surfaces. The head region of the inter-tubercular fossa, the region between the radialand deltopectoral process, and the dorsal surfaceof the diaphysis adjacent to the head are com-monly rugose; other regions develop ridges orrugosities in larger (presumably older) turtles.Internally (Figures 3 and 4), the humerus consistsof a narrow shell of dense compact bone and a


FIGURE 4.—Representative cross-sections of the humerus. The photographs show the proximalsurface of each section, except for the distal-most section, which shows the distal surface.

core of spongy cancellous bone. The dense bonecontains the periosteal growth layers; the spongybone is the endosteal, remodeled periosteal, andperhaps endochondral bone. There is no marrowcavity, although the center of the bone is moreporous owing to the concentration of blood ves-sels.

MORPHOMETRICS.—The change of the ecte-picondylar groove to a foramen suggests differ-ential or asymmetrical growth of the humerus.

Comparison of a hatchling's (48 mm cCL) hu-merus with that of a juvenile (565 mm cCL) doesshow proportional changes in humerus shape(Figure 5). The present sample permits an ex-amination of shape changes from half grownjuveniles (560 mm cCL) to reproductively activeadults (>900 mm cCL); however, the sample doesnot permit segregation by sex. We examine onlya subset (N = 80) of the data to address therelative growth of humerus length to body length

NUMBER 427

FIGURE 5.—Comparison of humerus shape in a hatchling (A;cCL = 48 mm), and an old juvenile (B; cCL = 565 mm).

and differential longitudinal growth of the hu-merus and some of its processes (Figure 6).

Not surprisingly, both measures of humerallength—maximum (ML) and longitudinal length(LL)—are strongly correlated to carapace length(cCL); for example, the correlation coefficient ofLL/cCL is 0.98 (LL = -13.25 + 0.2033 cCL).Similarly, other humeral measurements arestrongly correlated with carapace and humerallength. The two length measurements possess anearly isometric growth pattern (ML = —6.97 +1.1077 LL; r = 0.996). The regression slope isgreater than 1.0, however, and suggests that theulnar process grows slightly faster than the majoraxis of the humerus (i.e., the ulnar process lengthis contained within the ML measurement, butnot within the LL measurement). The ulnarprocess (UPL) grows faster than the radial proc-ess (RPL), as demonstrated by a UPL/LL slopeof 0.24 and a RPL/LL slope of 0.19. Althoughthe differences in growth rate are slight, theytend to confirm the production of the peculiarconfiguration of the proximal end of the sea

FIGURE 6.—Growth changes in humerus: A, whole bone; B,cross-section through the middle of diaphysis. The linesrepresent outlines from actual specimens.

turtle humerus through differential growth ofthe processes.

TISSUE ORGANIZATION.—The following de-scription is based on bone that has been macer-ated, air dried, and hydrated in formalin priorto histological preparation; thus, only the min-

10 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

eralized framework of the bone remains. Fur-ther, the description is based entirely on trans-verse sections through the diaphysis immediatelydistal to the deltopectoral crest, and of individ-uals with cCL greater than 55 cm.

The cortex and medulla are composed of cel-lular, vascular bone (terminology from Enlow,1969). Both the cortex and medulla are formedby periosteal deposition. The layers are distinctin the cortex, but strongly modified in the me-dulla by enlargement of the blood vesselsthrough resorption and endosteal layering ofthese vascular canals. The layers appear to consistentirely of non-lamellar (parallel-fibered) bone.The remodeling of the vascular canals in thecortex results in one to several layers of endostealbone on the wall of the canal; however, we arereluctant to label the resulting structure as sec-ondary osteones (Haversian systems), owing tothe irregular spacing of the osteocyte lacunae.The endosteal layers appear to be non-lamellarbone.

The bony layers of the cortex are the growthlayers used in skeletochronolgy. Each growthlayer consists of a light-staining area and dark-staining area. The light area represents an areaof rapid growth, the dark area slow or no growth.Together, an adjacent light and a dark arearepresent one complete cycle of growth; hencetheir pivotal importance in skeletochronology.Castanet (1982; Castanet and Naulleau, 1974;Castanet and Cheylan, 1979) has established aformal terminology for the layers and their sub-units; his terminology is used here and translatedinto English where necessary. Each growth layeris called a mark of skeletal growth (MSG; = MSCof Castanet) and consists of a zone (broad, light-staining region) and a line of arrested growth(LAG; = LAC of Castanet; narrow, dark-stainingregion).

Estimation of Age

ASSUMPTIONS AND DIFFICULTIES.—Our pri-mary assumption is that each MSG representsone year of growth. Tests of this assumption/

hypothesis have produced conflicting conclu-sions, i.e., both for confirmation and falsification.Our results favor the acceptance of this hypoth-esis, although, as we noted earlier in the intro-duction, some authors (e.g., Dobie, 1971), rejectthis hypothesis. Hohn and Frazier (1979) recog-nized the presence of growth layers (MSGs) inEretmochelys and the positive correlation of MSGnumber to shell size. To test this correlationfurther, Frazier (1982) examined sections of hu-meri and femora in five species of sea turtles. Hisconclusions were that (1) bone sections of Der-mochelys and Lepidochelys are difficult to interpretand other sea turtles, particularly Eretmochelys,have readable growth layers; (2) the interpreta-tion of growth layers in sea turtle bones is notsimple, owing to endosteal resorption, unequallamellar accretion rates, and variable visibilityand routes of arrested growth lines (LAGs); (3)width of adjacent MSGs shows no predictablerelationship; and (4) known-aged Chelonia mydas(14 mo, 18 mo, and 42 mo old turtles) fromCayman Turtle Farm, Ltd. often possess moregrowth layers than their age in years, but fewerlayers than their age in months. Thus, Frazier'sresults suggest a falsification of the one-MSG/one-year hypothesis and his remarks (1982) echothose of Suzuki (1963) and Dobie (1971) inlargely dismissing bone layers for age determi-nation.

Our study of Caretta humeri reveals the samedifficulties in MSG recognition and other onesas well; however, recognition of distinct MSGs ispossible in the majority of specimens. WithinCaretta, the major problem areas are as follows:

Resorption and redeposition. Loss of MSGs, usually theearlier ones; appearance of endosteal layers.

Accessory LAGs. The occasional presence of duplicatearrested growth lines (Figure 7B).

Absence of LAGs. Bone with homogeneous appearance,no apparent layering (Figure 7c).

Discontinuous LAGs. Arrested growth lines that do notcompletely encircle the circumference of the bone orlines which converge and cannot be differentiated (Fig-ure 7D).

Irregular MSGs and LAGs. Occurrence of adjacentMSGs of strikingly different thickness and/or wavy or

FIGURE 7.—Histological cross-sections of humeri demonstrating the various states of the marksof skeletal growth (MSGs): A, distinct MSGs bordering area of resorption; B, accessory lines ofarrested growth; c, no lines of arrested growth (LAG); D, discontinuous LAGs; E, irregulargrowth layers; F, compaction of MSGs.

interrupted, non-contiguous arrested growth lines(Figure 7E).

Compaction of MSGs. Adjacent arrest lines so close toone another as to prevent differentiation of the indi-vidual growth layers (Figure 7F).

Resorption and redeposition of the perios-teal layers is extensive in the Caretta humerus,because its cross-sectional diameter (narrowestaxis) increases tenfold from hatching to adult-hood. In addition, there are changes in pro-portions and in orientation of the longitudinal

axis (Figure 6), thus the original humerus andthe early periosteal layers are entirely replacedby remodeling and endosteal growth. The lossof the earlier periosteal layers, i.e., the MSGs,requires an extrapolation from the width ofthe extant MSGs in order to estimate the num-ber of lost MSGs.

Accessory LAGs are lines probably resultingfrom nonseasonal retardation or cessation ofgrowth, caused by such events as injury, illness,or short-term reduction in food supply. Often


the accessory LAGs can be discerned from the"true" LAGs by less intense or more diffusestaining or by incompleteness of the line. Ac-cessory LAGs potentially cause an over-esti-mate of age, since their development violatesthe one-MSG/one-year assumption.

The absence of LAGs prevents the estima-tion of age. The cause of this condition isunknown. Continuous, more or less uniformgrowth would produce this condition, al-though such a growth pattern seems unlikelyfor members of the Georgia Caretta popula-tion, unless some individuals migrate towarmer waters in the winter.

Discontinuous LAGs and compacted MSGsincrease the difficulties of recognition of MSGsand increase the likelihood of sampling errors.They do not, however, prevent the estimationof age. Compaction is most prevalent in largeturtles and may result from a reduction ingrowth rate following the attainment of sexualmaturity (Carr and Goodman, 1970). Discon-tinuous MSGs may reflect a growth patternwhere one region of the periosteum slowed itsrate of deposition while an adjacent region didnot. The process is analogous to the differen-tial rates of deposition whereby a bone altersits shape by producing a thicker MSG (i.e.,periosteal layer) on one surface than on an-other surface (see Figure 6 for shape changesowing to differential growth).

These problem areas cause difficulties incounting and measuring MSGs and in the es-timation of age from the MSG data. Thesedifficulties are not insurmountable, but theydo require care of interpretation and imposenarrow limits upon the amount of extrapola-tion from the age estimates. We are not pro-posing that we can precisely age every individ-ual turtle. Although our age estimates may beaccurate for individual turtles, our goal is toprovide age estimates for various classes ofturtles, e.g., immatures of 500-600 mm cCLor adults > 1000 mm cCL, and most specificallyto provide an accurate estimate of the age ofsexual maturity for the Georgia Caretta popu-

lation. (We assume that the juveniles livingadjacent to the Georgia coastal islands becomethe adults breeding on these islands.) Thus,our analysis and interpretation depend upon alarge sample of turtles, not on individual spec-imens.

We believe that such an approach overcomesthe six problem areas mentioned above. First,the large sample eliminates the dependency ofinterpretation upon individual specimens andpermits the removal of specimens with anom-alous or peculiar growth patterns which wouldlikely cause erroneous age estimates. The largesample further provides an overview of bonegrowth and remodeling, so that the countingand measuring of MSGs is standardized andobjective. This standardization is important,because data collection is always more efficientand consistent when the collector knows thesubject area well, particularly its variability,and collects the data within a limited and un-interrupted time period. The Caretta data setwas collected by A. Wynn during a single timeperiod (3 mo), and upon completion of theentire set, the earlier specimens and randomlyselected later specimens were re-examined toconfirm the consistency of the entire data set.In this latter respect, re-examination at lowermagnification (10-20 X) was useful in the con-firmation of number of MSGs, although nottheir widths.

Such problems as accessory LAGs and irreg-ular growth layers illustrate that periostealgrowth has periods of slow and rapid growth,thereby emphasizing the cyclic nature ofMSGs. Nonetheless, their presence only cloudsthe question of whether or not the cycle is anannual one. Our data on Caretta cannot ad-dress this question. Further, the inability tolocate MSGs in all individuals of a sample orin all skeletal elements within a single individ-ual is inadequate grounds to falsify the one-MSG/one-year hypothesis. We accept the an-nual production of one mark of skeletal growth(MSG) on several lines of circumstantial evi-dence.

NUMBER 427 13

Our second assumption in the application ofskeletochronology is that the number of MSGslost through bone resorption/remodeling canbe estimated from the number and size of theremaining MSGs. Our method depends uponusing the average width of extant MSGs toestimate the total numbers of MSGs formedfrom hatching to death. The estimate derivesfrom the number of times that the averageMSG width occurs in half the diameter of thesample bone. The formula is

_ (0.5 BD) - HPAW '

where TM is the total number of MSGs; BD isthe cross-sectional diameter of the samplebone; HD is the average cross-sectional diam-eter of hatchling Caretta humerus (HD =1.5063 mm); and AW is average width ofextant MSGs. HD is subtracted, because itrepresents bone diameter at age zero; only thesubsequent growth marks represent years ofgrowth. BD is halved, because the averageMSG width derives from only one side of thebone.

The estimate of total number of MSGs isequivalent to age in years if we accept a cyclicgrowth pattern of one MSG per year and thatthe average width of extant MSGs accuratelyreflects the average width of lost and extantMSGs.

DATA ANALYSIS.—In our preliminary analy-sis (Zug, Wynn, and Ruckdeschel, 1983), wedemonstrated a strong positive correlation ofhumerus length to carapace length (over-the-curve, cCL) as well as significant positive cor-relations of humerus length to estimated ageand of carapace length to age. Those analysesare reproduced in Part 1 of the Appendix,because they provide the analytical base forour present discussion and may be unavailableto many readers of this report. Although thebasic data are the same, the analytical proce-dure differs.

Herein, each age estimate is rounded to thenearest whole number, in contrast to the near-

est hundredths in the preliminary analysis. Theregression analyses examine the data in threeoverlapping sets: (1) IMMATURES, all specimensless than 860 mm cCL, except for individualswith excessively high estimates of age; (2) IM-MATURES AND ADULTS, all specimens, exceptfor individuals with excessively high estimatesof age; (3) ALL, all specimens. The data arereproduced in Part 2 of the Appendix.

Our preliminary analysis (see Appendix,Part 1) used age as the dependent variable,because carapace and humerus length weredirectly measurable variables and age was es-timated. That manner of presentation causedsome confusion or was less easily read owingto the reversal of the axes as commonly pre-sented for the analysis of growth. An argumentcan be given for each manner of presentation;hence we provide both. Importantly, the analy-sis of estimated age as a dependent variableprovides regression equations that more accu-rately predict size at hatching.

A comparison of the preliminary analysiswith the current analysis shows a striking dif-ference in regression equations and correla-tions. (For both analyses, all correlation coef-ficients are statistically significant at 0.01 prob-ability level.) The linear regression equationsfor both analyses (Table 1, 2LI.2; Appendix,Part 1, N — 75*) give unrealistic carapacelengths (407.8 and 580.5 mm, current andpreliminary analyses, respectively) for hatch-lings; the power or allometric equations forthe same data sets (2PI.2; N = 75f) predictmore realistic lengths (33.0 and 71.1 mm, re-spectively). Actual carapace lengths (38-50mm; Baldwin and Lofton, 1959:342) of hatchl-ing Caretta lie between these two allometricequation predictions and are somewhat closerto the current data set's prediction. The pointis not that one data set is better than the other;rather that the analysis of age estimates aswhole numbers versus that using decimal num-bers does not alter the general results.

The age estimates derive from two axes ofthe humeral section. The results are usually


TABLE 1.—Regression statistics for age estimates (long axis) and carapace length. A modelcode is assigned to each regression to facilitate discussion within the text. Models 1L1 and 1P1include all specimens with cCL <860 mm except those with excessive age estimates; models1L2 and 1P2, all specimens except those with excessive ages. Data from Part 2 of Appendix.(Abbreviations: A, K-intercept; B, slope of regression curve; N, sample size; r, correlationcoefficient.)

(X,Y)

Age, cCLcCL, AgeAge, cCLcCL, Age


N

38384444

38384444

A

576.966-17.561615.845-77.325

402.5245.7X10"5

365.4851.6X10"7

B rLinear Regression >5.1630.0593.8080.147

0.550.550.750.75

Power Regression0.1761.9660.2102.870

0.590.590.780.78

Range

X

' = A + BX11-40

565-85511-111

565-960

Y= AXB

11-40565-855

11-111565-960

Y

565-85511-40

565-96011-111

565-85511-40

565-96011-111

Model

1L1.11L1.21L2.11L2.2

1P1.11P1.21P2.11P2.2

strikingly different, with the long axis yieldingestimates typically two or more times those ofthe short axis (see Appendix, Part 2). Thesedifferences arise not from a greater number ofMSGs, but from fewer MSGs of the same ornarrower widths than those of the short axis.Hence, the narrower average MSG width andthe greater bone diameter of the long axisresult in the estimation of a greater number ofMSGs, which convert to older age estimates.Because the number of MSGs seems unreason-ably high and because the number of sectionsshowing measurable MSGs on the long axis isapproximately half the sample size of the shortaxis, the long axis data will not be emphasized.

The short axis data possess a greater uni-formity of estimates, although disparate ageestimates do occur between similar sized indi-viduals and seemingly high estimates occur forother individuals. As defined earlier, the dataare analyzed in three overlapping sets. Thefirst set (IMMATURES, cCL <860 mm) containsimmature turtles and is a particularly valuableset for determining the age of sexual maturity.The second set (IMMATURES AND ADULTS) in-cludes all the turtles of the previous set andthe mature turtles (cCL >860 mm). The selec-

tion of a size for classifying loggerheads asimmature or mature is arbitrary, because in-dividuals of the same sex obtain maturity atdifferent sizes and males and females presum-ably mature at different average sizes. None-theless, a selection is useful for analysis and the860 mm cCL should accurately sort most ofthe turtles in our sample. The choice of thisvalue derives from the range of nesting-femalecarapace lengths, not from mean lengths whichare characteristically greater than 900 mmcCL (e.g., Ehrhart and Yoder, 1978; Stone-burner, 1980). For the U.S. Atlantic coast,ranges are 845-1029 mm cCL (Cape Romain,SC; Baldwin and Lofton, 1959), 800-1150mm cCL (Jekyll Island, Ga.; Caldwell, Carr,and Owen, 1959), 892-1186 mm cCL (LittleCumberland Isl., Ga.; Ruckdeschel unpubl.data), 830-1240 mm cCL (Merritt Isl., Fla.;Ehrhart and Yoder 1978), 775-1067 mm cCL(Hutchinson Isl., Fla.; Gallagher et al., 1972),and 749-1092 mm sCL (Melbourne Beach,Fla.; Bjorndal et al., 1983).

A comparison of age estimates from the longand the short axis samples (Tables 1 and 2)shows that the long axis sample possesseshigher correlation coefficients than the short

NUMBER 427 15

TABLE 2.—Regression statistics for age estimates (short axis) and carapace length. Data fromPart 2 of Appendix. (Abbreviations and explanations same as in Table 1.)

(X, Y)

Age, cCLcCL, AgeAge, cCLcCL, AgeAge, cCL


N

6868767681

6868767681

A]

631.009-13.454657.848-38.976699.103

529.6733.1X10

511.8631.0X10

537.579

B rlinear Regression, Y =

7.4600.0335.6560.0702.149

0.490.490.630.630.60

Power Regression, Y =0.130

"7 2.6160.151

~8 3.2390.128

0.580.580.700.700.74

Range

XA + BX

3-25565-855

3-74565-1010

3-175

« AX"3-25

565-8553-74

565-10103-175

Y

565-8553-25

565-10103-74

565-1090

565-8553-25

565-10103-74

565-1090

Model

2L1.12L1.22L2.12L2.22L3.1

2P1.12P1.22P2.12P2.22P3.1

TABLE 3.—Regression statistics for age estimates (short axis) and carapace length. Age estimatesused herein were based on the maximum number of MSGs, either observed or estimated; datafrom Part 2 of Appendix. (Abbreviations and explanations same as in Table 1.)

(X, Y)



N

6868767681

6868767681

A

628.010-12.550655.892-38.087698.549

519.3811.8X10

500.2672.0X10

530.481

B rLinear Regression Y =

7.5530.0325.6930.0702.150

0.490.490.630.630.60

Power Regression Y0.136

-• 2.3490.159

"7 3.0440.132

0.570.570.700.700.73

Range

XA + BX

4-25565-855

4-74565-1010

4-175

= AX»4-25

565-8554-74

565-10104-175

Y

565-8554-25

565 -10104-74

565-1090

565-8554-25

565-10104-74

565-1090

Model

3L1.13L1.23L2.13L2.23L3.1

3P1.13P1.23P2.13P2.23P3.1

axis sample. This difference is especially no-ticeable in the linear regression analysis.

In calculating the number of MSGs for eachspecimen, the resulting MSG estimate was oc-casionally less than the number of MSGs ob-served. Potentially, these differences might re-duce the predictability of the regression equa-tions which were determined strictly from theMSG/age estimate data. The low frequency ofhigher observed MSG counts versus estimatedMSG counts and the slight difference between

the two when they are different have no majoreffect on the regression equations (Tables 2and 3) and their curves (Figure 8C,D). Thedifferences are sufficiently minor that themixed data set (observed and estimated, Table3) can be ignored.

The estimated data (Figure 8A,B) yield fiveestimated MSGs/ages that are clearly outsidethe main cluster of points. These outliers wereremoved from the analysis (Table 4, Figure8E,F). Their removal results in a striking im-


59

55

UJ 2 0

o

20

AGE

67 75

CARAPACE LENGTH

NUMBER 427 17

TABLE 4.—Regression statistics for age estimates (short axis) and carapace length; same dataset as in Table 2 with the removal of outliers (shown as open circles in Figure 8A,B). Data fromPart 2 of Appendix. (Abbreviations and explanations same as in Table 1.)

(X. Y)



N

63637171

63637171

A

572.778-12.741656.076-61.204

484.2902.8X10

491.2634.3X10

B r

Linear Regression Y = A15.1780.0305.6470.102

0.680.680.760.76

Power Regression Y =0.178

"7 2.6170.171

"" 3.972

0.680.680.820.82

Range

X

+ BX3-18

565-8553-81

565-1060

AXB

3-18565-855

3-81565-1060

Y

565-8553-18

565-10603-81

565-8553-18

565-10603-81

Model

4L1.14L1.24L2.14L2.2

4P1.14P1.24P2.14P2.2

provement of the correlation coefficients anda better fit of the regression curves to thecluster of points. The outliers share no featuresthat might explain their isolation from theother data points. They do not derive fromsimilar-sized turtles, they do not have excep-tionally high numbers of observed MSGs, andthey do not have narrow average MSG widths.

The linear regressions and correlation coef-ficients (Tables 1,2, and 4; Figure 8) providereasonably good fits to the data clusters for theimmature turtles (cCL <860 mm). The addi-tion of the adult specimens to the samples(models 2L2 and 4L2) improves the correla-tion of carapace length with age; however, theaddition of specimens with very high age esti-mates (e.g., model 2L3) suppresses the corre-lation coefficients, but not below that of theimmature turtles. The same summary remarks

FIGURE 8.—The relationship of estimated age to carapacelength. Age is an independent variable in the left column ofgraphs and a dependent variable in the right column ofgraphs. The top row of graphs (A,B) contains the ages (ofimmature turtles, <860 mm cCL) derived entirely fromestimated MSGs (Table 2, models 2L1 and 2P1; outliersshown as open circles). The middle row (c,n) contains theages (of immature turtles) derived from the highest numberof MSGs, either observed or estimated (Table 3, models 3L1and 3P1; outliers shown as open circles). The lower row (E,F)shows the same age data as the top row with the removal ofthe five outliers (Table 4, models 4L1 and 4P1).

apply to the power regressions and correlationcoefficients; however, the fit is better for thepower equations in all cases. Owing to thebetter fit and higher coefficients, subsequentextrapolation to determine body size at sexualmaturity and hatching will be based on powerequations. The inability of the linear regres-sion equations to predict accurately hatchlingsize is apparent from the following calculationsderived from the values in Table 2 wherecCL(mm) and age (yr, 1 day = 0.003). Forexample, sizes are 631.0 mm (Y = 631.009 +7.46(0.003) for model 2L1.1), 407.8 mm(0.003 = - 13.454 + 0.033X for model 2L1.2),657.9 (2L2.1), 556.8 mm (2L2.2), and 699.1mm (2L3). Such predictions are far in excessof the 38-50 mm range for Caretta hatchlings(Baldwin and Lofton, 1959).

Prediction of hatchling size is improved byusing the power equations. The comparablesize predictions are: 248.9 mm (2P1.1), 33.4mm (2P1.2), 212.9 mm (2P2.1) 49.1 mm(2P2.2), and 255.6 mm (2P3) from Table 2.The predictions, however, are biologically re-alistic only for power equations where Y equalsage, which is the reverse of the typical pre-sentation for growth. Technically, regressionis appropriate only for independent variablesmeasured without error. The age data do notmeet this criterion and, statistically, are ana-


lyzed correctly as dependent variables. Visu-ally, the graphs (Figure 8) show that the powercurve for age as a dependent variable matchesbetter the longitudinal axis of the data clusterfor immature turtles. The better fit to the datacluster and accurate prediction also recom-mend the use of the power equations with ageas the dependent variable.

In the determination of age at sexual matu-rity, linear regression and power equationsshow the same general pattern as observed inthe hatchling-size predictions. Those equationswith age as the independent variable alwayspredict an older age than the equations withage as a dependent variable (Table 5). Theskeletochronological age estimates for speci-mens between 801 and 900 mm cCL (Appen-dix, Part 2 and Figure 9) range from 6 to 31yrs with a mean of 14.5 yrs (s = 7.4614, N =11). These data would again recommend theuse of equations with age as the dependentvariable. Furthermore, the equations (both lin-ear and power) with the closest similarity toactual data are those derived from the IMMA-TURE data set. Importantly, the equations Y =(3.1 X 10"7)X 2.616 (Table 2, 2P1.2;Fage,XcCL)F=(2.8x 10-7)X2.617(Table4,4P1.2),Y= (1.0 X 10"8) X 3.239 (Table 2, 2P2.2) yieldrealistic, although small estimates of hatchlingsize (33.4 mm, 34.7 mm, and 32.0 mm cCL,respectively). The equations Y = 365.485 X0.21 (Table 1, 1P2.1; Y cCL, X age) and Y =

TABLE 5.—Age (years) at sexual maturity calculated usinglinear and power equations of preceding tables and an 860mm cCL.

Data

Table 11L2 & 1P2

Table 22L1 &2P12L2&2P2

Table 44L1 &4P14L2 & 4P2

X =

Linear

64.1

30.735.7

18.940.2

Age

Power

58.8

41.631.1

25.226.4

K=Age

Linear

49.1

14.921.2

13.126.5

Power

42.3

14.732.0

13.419.5

529.673 X 0.13 (Table 2, 2P2.2) give hatch-ling carapace lengths of 58.8 and 41.6 mm,respectively; these predictions are within thesize range of hatchling Caretta. The corre-sponding age estimates for sexual maturityfrom the preceding five equations are 14.7,13.4, 32, 58.8, and 41.6 yrs, respectively.

The mean age estimate for different sizeclasses (Figure 9) shows a gradual increase inmean age from 6 yrs for the immature turtles(550-600 mm cCL) to 13 yrs for those nearingsexual maturity. The ages for the mature sizeclasses (>850 mm) are from 20 to 50 yrs.Although the latter estimates seem biologicallyreasonable, the small samples for each classsuggest that the data be treated with caution.This caution is also advocated by the widerange of estimates and the large standard de-viations of the mature turtles. The small sam-ples reflect not only the fewer number of adultsstranding on Cumberland Island (Ruckdescheland Zug, 1980), but also the difficulty of ob-taining skeletochronological data from maturespecimens. These difficulties encompass theentire spectrum from major remodeling withno MSGs to compaction of MSGs. It is thelatter condition which results in the excessivelyhigh age estimates, e.g., 81 yrs for 1060 mmcCL, 142 yrs for 1000 mm cCL, 140 yrs for975 mm cCL. These specimens have very nar-row mean MSG widths, which overestimate thetotal number of MSGs.

Life History Implications

The preceding data center the attainment ofsexual maturity on the ages of 13-15 yrs (Table5 and Figure 9). This result matches Mendonca's(1981) estimate of 10-15 yrs for Merritt Islandloggerheads remarkably well. Mendonca derivedher estimates from the growth rates of recap-tured wild Caretta with straight-line carapacelengths (sCL) of 440-925 mm. The average an-nual growth rate for her sample is 59 mm/yr,which yields an estimate of 13 years to reach 800mm sCL, i.e., sexual maturity. The concordance

NUMBER 427 19

2 5

^ 20

HIO 15

1 0

3 1

7 4

4033.6 d-

5 88 *1T

81 •

4 8 . 5 - -

550-600

601-650

651-700

701-750

751-800

801- 85850 9

1- 901-00 950

951-1000

>1000

CARAPACE LENGTH (mm)FIGURE 9.—The distributions of age estimates for sequential size classes of the 1979 Cumber-land Island sample of loggerhead sea turtles. Vertical line = range; horizontal bar = mean;open box = ± 1 standard deviation of mean; each dot represents one individual.

of Mendonca's and our age estimates of sexualmaturity is important in that the estimates wereobtained by two strikingly different techniques;hence the concordance supports the acceptanceof 15 years as an "average" age for sexual matu-rity of southeastern U.S. Caretta. Frazer andEhrhart (1985) suggest an older "average" agefor maturing loggerheads at Merritt Island.Their results derive from the Mendonca data setexpanded by the inclusion of growth rates fromfemales nesting in successive years and predict arange of 12-30 yrs for maturation. The two

predictions differ for several reasons (Frazer andEhrhart, 1985), but primarily because Mendoncaused a smaller body size for determining age ofmaturation and a linear model (in contrast to thevon Bertalanffy model) with only the faster ju-venile growth rates. As noted earlier, the meanCL of nesting females is more than 900 mm.Mendonca's linear model will predict 16 yrs atthis length, and our allometric equations yieldages of 16 yrs (2P1.2) and 15 yrs (4P1.2). Theselatter predictions are like Mendonca's, becausethey are derived from the IMMATURE data. Using


the combined IMMATURE AND ADULT set, age at900 mm cCL is 37 yrs (2P2.2) and 23 yrs (4P2.2);these latter estimates are more in agreement withthose of Frazer and Ehrhart (1985). Henwood(in prep.) also uses a von Bertalanffy model andjuvenile and adult growth data to predict matu-ration in the mid-twenties for western Atlanticloggerheads.

An earlier estimate of maturity for the GeorgiaCaretta population was 6-7 yrs (Caldwell, 1962)and was based on the growth rates of captive-reared hatchlings. The growth rate of well-fedcaptive sea turtles is typically two or more timesfaster than that of wild ones (Appendix, Part 3),and it is becoming increasingly apparent that ageextrapolations from captive growth rates mustbe used cautiously. The average captive growthrate (from data in Appendix) is 136.6 mm/yr (s= 35.28, N = 8) for loggerheads. This rate wouldconvert a 42 mm hatchling into an adult in 6 yrsas Caldwell (1962) and Uchida (1967) predicted.Use of captive growth rates ignores two aspectsof natural growth, however. It assumes a con-stancy of growth rate in all growth stages, andeven captive rearing of different size classes ofsea turtles demonstrates the fallacy of this as-sumption (e.g., Frazer, 1982; Limpus, 1979; Par-ker, 1926, 1929; Uchida, 1967; Witham andFutch, 1977; Witzell, 1980). Furthermore, ob-servations from captive rearing programs typi-cally record growth from hatching to one yearof age, occasionally for a second year, whengrowth is maximal. Second, the diet of captivesis high in mass and protein and low in undiges-tible components (reported captive diets usuallyconsisted of chopped fish and/or shrimp). Wildturtles probably obtain such a high quality mealrarely and probably do not approach the samemass of food on a monthly or yearly basis. Nuitjaand Uchida (1982) have shown that growth ratesin Caretta are directly linked to diet; more foodproduces faster growth. Third, wild turtles ex-pend more energy in food gathering, so the netenergy available for growth is presumably muchless than that for captive turtles.

A striking exception to the preceding exposi-

tion are the growth data of two Caretta raised incaptivity from hatching to 14 yrs. As theyreached 14 yrs, the turtles were beginning todevelop secondary sexual characteristics. If theycontinued their growth pattern, they wouldreach 860 mm cCL at 16-17 yrs and 920 mm at19-20 yrs. Since they fed year round, althoughat a reduced level in the winter, an "average" ageof maturation for wild loggerheads from the U.S.Atlantic coast would likely fall in the mid-twen-ties as postulated by Frazer and Ehrhart (1985)and Henwood (in prep.).

Most captive growth experiments demonstrateonly that sea turtles are capable of rapid growthwhen given abundant food resources, not thatthey do so in the wild. The reproductive data oncaptive reared green sea turtles suggests, how-ever, that the attainment of "mature" size andthe production of gametes may occur prior tothe animals being fully mature (Witham, 1971;Wood and Wood, 1980), because nesting fre-quency and clutch size of captive-reared femalesare well below normal. In addition, hatchingsuccess (? = fertility of eggs) is extremely low thefirst nesting season and increases gradually there-after (Wood and Wood, 1980). A side, but re-lated aspect to the preceding observations is thelack of concordance between age and marks ofskeletal growth (MSGs) in Chelonia mydas at theCayman Turtle Farm (Frazier, 1982). Since thegrowth rates are abnormally high in captive tur-tles and this rapid growth affects the physiologyof reproduction (as outlined above), the normalpattern of bone growth may be similarly dis-rupted; hence, captive-reared known-aged tur-tles are a poor basis on which to confirm or falsifythe one-MSG equals one-year hypothesis.

Maturity does not occur in all Caretta popula-tions at the same age. Loggerheads from theGreat Barrier Reef may require more than 30yrs (Limpus, 1979), because growth rates of theolder juveniles (>500 mm CL) are only 10-20mm/yr. Such slow growth presumably reflectsthe quality and quantity of the food available tothese turtles rather than a genetically determinedgrowth potential. The rapid growth response of

NUMBER 427 21

all sea turtle species to abundant food in captivityillustrates the plasticity of sea turtle growth. Thisplasticity is evident in the differential responseof turtles in adjacent populations to differencesin food quality and quantity. For example, Che-Ionia mydas populations in the Hawaiian Archi-pelago show growth rates ranging from 10 to 53mm/yr (Appendix, Part 3), and these rates aredirectly correlated with food supply (Balazs,1979, 1980, 1982; see also Gibbons, 1967;Hulse, 1976; and Parmenter, 1980 for the influ-ence of food quality and abundance on growthin populations of freshwater turtles). Caretta pop-ulations may respond similarly, but the growthrates of wild populations are known for so fewthat such patterns cannot be discerned at present.

Growth rates and ages of maturity are notidentical for all members of a population eventhough they may have access to the same foodresources and experience the same environmen-tal conditions. Beginning with hatching, siblingsshow a range of body sizes. With growth, therange/variation of body sizes increases (e.g., seeBourke, Balazs, and Shallenberger, 1977, fig. 1,for captive-reared Chelonia mydas). Such differ-ential growth probably occurs throughout thejuvenile stage, and Carr and Goodman (1970)have hypothesized that the variability of bodysize in nesting female Chelonia is more stronglyaffected by the different sizes (i.e., differentialgrowth) attained at maturity than by differencesin postmaturity growth; Hildebrand (1932)found that maturation age may differ as much asfour years in sibling Malaclemys terrapin from thesame clutch of eggs. Caretta should show similarvariation in size and age at maturation.

Although our data cannot directly address thisindividuality of growth and maturity, the highvariation of age estimates within size classes (Fig-ure 9) indicates its presence. The smallest sizeclass (550-660 mm cCL) has a range of 3-17 yrsand, excluding the 17-yr estimate, an averageage of 4.6 yrs (standard deviation, s = 1.51). Thisaverage age closely matches the median andmodal ages for this size class. With the aboveexclusion, this size class and the next largest

(601-650 mm cCL) have the lowest variation.Age variation is noticeably higher in the largerjuvenile size classes and several magnitudeshigher in the adult classes (>851 mm cCL), al-though the sample size of the latter is too smallto attribute any statistical significance to the in-crease in variation at maturity. The increase invariation is principally caused by the adult classescontaining the full spectrum of mature individ-uals from those nesting for the first time to thosein their fifth or later nesting season. Secondarily,the range of ages at maturity is broad, and thisphenomenon is probably common to all popula-tions of sea turtles. Even though our data arefew in the mature age classes, they strongly sug-gest that most of the mature Caretta (>900 mmcCL) are old (>30 yrs). Furthermore, our datashow one individual with 6 estimated MSGs andanother with 25 MSGs, suggesting a potentialage range of 19 yrs in maturing Georgia Caretta.

Our age estimates can provide estimates ofgrowth rates in the same manner as growth ratescan be used to estimate maturity. In this case, weestimate the growth by subtracting averagehatchling size (45 mm) from the carapace lengthof each specimen and then dividing by its esti-mated age (short axis estimate; Appendix, Part2). The range for the total sample is 4.2-198.3mm/yr (Table 6) and this is equivalent to actualgrowth rates observed in captive and wild log-gerheads (Appendix, Part 3). More important,however, is the lack of agreement with Mendon-ca's data (1981: table 1). Our estimates are 20-30 mm/yr faster than her observed rates. Wesuspect that this difference is partially explainedby our inability to eliminate the earlier rapidgrowth from our estimates. Mendonca's datameasured growth during an interval of late ju-venile life; our estimates measure growth fromhatching to the animal's death and include therapid growth of posthatchlings so the average ishigher. Nonetheless, the mean growth rates (Ta-ble 6) show the anticipated pattern of slowinggrowth with increasing size (and age). In addi-tion, the data suggest a relatively homogeneousgrowth in the 650-850 mm size classes and then


TABLE 6.—Growth rate estimates (mm/yr) for sequentialsize classes of loggerhead sea turtles. Sexual maturation isassumed to occur in the 851—900 mm cCL size class.

Size Class(cCL)

550-600601-650651-700701-750751-800801-850851-900901-950951-1000> 1000

N*

8121713892522

Mean

117.4102.882.077.464.271.447.648.018.436.4

StandardDeviation

51.0437.6929.1836.7416.5130.79

_44.92

--

Range

79.3-181.769.4-198.326.2-162.527.4-139.047.7-88.731.0-129.227.6-67.512.0-122.915.8-21.112.5-60.3

* Four specimens with extraordinary age estimates arenot included in this table; their estimates are (cCL, mm/yr):735, 4.2; 975, 6.6; 1000, 6.7; 1090, 5.9.

a sharp drop in growth with maturity (>850 mm).In comparison to North American freshwater

and terrestrial turtles (Appendix, Part 4), Carettacaretta appears to mature sooner than would beexpected by extrapolation from ages and sizes atmaturation of the smaller turtle species. Thelarger emydines (Pseudemys and Trachemys) andChelydra typically require a minimum of 6 yrs toattain a 200 mm CL. Simple linear extrapolationwould suggest a minimum of 24 yrs for maturityat 800 mm CL. The relationship between turtlespecies' size and their age at maturity is notlinear, however, because smaller species (e.g.,Clemmys, Kinosternon, and Sternotherus mature in4-7 yrs at 100 mm CL or less) possess ages ofmaturity equivalent to those of their biggerbrethren. Clearly one aspect which biases ourexpectation of delayed maturity is that the ma-jority of turtles studied derive from cooleraquatic environments with shorter growing sea-sons than the sea turtles studied. Turtles inwarmer waters do grow faster (Cagle, 1954; Gib-bons et al., 1981; Thornhill, 1982), and whereexamined closely (Gibbons et al., 1981; Parmen-ter, 1980), this faster growth results from twofactors: primarily, owing to the availability ofmore and higher quality food and, secondarily,by higher ingestion rates (likely higher assimila-

tion rates as well) and a longer growing season.Thus, relative to the growth and maturation inthese turtle populations, the growth and matu-ration of the Georgia Caretta do not deviate fromthe expected. Furthermore, the range of wildgrowth rates in sea turtles (Appendix, Part 3)and the high captive growth rates suggest thatthere are no real intrinsic differences in growthphysiology between cheloniid sea turtles and thesmaller-bodied freshwater turtles. Likewise, thelarger-bodied freshwater and terrestrial turtlesshow similar growth rates or maturation ages assea turtles (Appendix, Part 4). This similarity ismost apparent in the giant tortoises (Geochelone),which have estimates of 16 to more than 30 yrsto attain sexual maturity.

Summary

The humeri of Caretta caretta show distinctbony layers, and these marks of skeletal growth(MSGs) are assumed to be annual records ofgrowth. The use of the skeletochronologicaltechnique is not without difficulties, such as ab-sence of MSGs; irregular, interrupted, or acces-sory arrested growth lines; loss of MSGs by boneremodeling. Consequently, we use large samplesand treat each age determination as an estimate.

Two age estimates were made from each bonesection removed from the diaphysis just distal tothe deltopectoral ridge. One estimate derivedfrom the long axis of the section and the otherfrom the short axis. Remodeling occurred in allhumeri and required an extrapolation from ex-isting MSGs to estimate the total number ofMSGs formed during the life of the turtle. Theestimate was obtained by dividing half of thehumerus diameter by the average width of theextant MSGs. Since the basic assumption is thatone MSG forms each year, the total number ofMSGs equals the turtle's age in years. Age esti-mates derived from the long axis are usually 10or more yrs higher than those from the the shortaxis. Also linear and curvilinear regression equa-tions derived from short-axis age estimates andcarapace lengths calculations provide the bestestimates of carapace lengths at hatching. Hence,

NUMBER 427 23

short-axis age estimates are considered the most tions for 860 mm cCL turtles, i.e., an arbitraryreliable. minimum size for Caretta nesting on the U.S.

Sexual maturity in Cumberland Island Caretta Atlantic coast. This age estimate for maturityoccurs most commonly between 13 and 15 yrs matches one estimate proposed for a Floridafrom hatching. These values derive both from Caretta population and is 5—10 yrs less for twothe average of the age estimates for the 800-900 other estimates that use 900-920 mm sCL as themm cCL size class and from the regression equa- size at maturity.

Appendix

Part 1

Table and figure from our preliminary analysis of age estimates in Caretta(Zug, Wynn, and Ruckdeschel, 1983) in the Marine Turtle Newsletter. Theyare reproduced to permit ready comparison with the results of the presentanalysis.

Age estimates, carapace lengths, and humerus lengths. Data for linear regression (Y = A+BX)*and power (Y= AXB)+ equations. Other abbreviations are: cCL, carapace length over the curve;HL, humerus length; N, sample size; r, correlation coefficient. Lengths are given in mm, agein yrs.

(X,Y)

HL, cCL

HL, Age

cCL, Age

cCL, Age

N

78*78+

76*

75*

78*78+

A

-26.7590.078

-45.7388.8X10"7

-51.0811.0X10~9

-108.400.6X10""

B

0.2301.1360.4163.2860.0883.4980.1714.279

r

0.970.970.750.800.690.730.690.78

X

105.5-200.8

105.5-200.8

565-1060

565-1060

Y

565-1060

3.0-80.9

3.0-80.9

3.0-175.3

10D

500 800

Carapace L e n g t h ( m r

1100

FIGURE Al.—The relation of curved carapace length to ageestimates derived from periosteal annuli in Caretta humeri.Both the straight (linear regression) and curved (power) linesare based on 75 individuals; the three individuals with ageestimates greater than 50 years were excluded from thecalculations for these lines.

24

Part 2

Data summary for the skeletochronological sample of Caretta caretta fromCumberland Island, Georgia. Age estimates were made from measurementsof MSGs along the anteroposterior (long) axis and the dorsoventral (short)axis of the humerus section. Abbreviations and symbols: AW, mean width ofMSGs along a single axis; cCL, carapace length over the curve; growth rate,estimated average annual cCL growth from hatching to death using estimatedMSGs (short axis) as age at death; ML, maximum humerus length; t, anoutlier value; *, an excessively high age estimate. The 235000 series specimennumbers are from the USNM series, from collections deposited in theNational Museum of Natural History, Smithsonian Institution; the 79.0.0.0series are field numbers assigned by the collector.

SpecimenNumber

235284

235286

235287

235291

235292

235293

235294

235295

235296

79.5.4.9

79.5.4.10

235297

235298

235301

235302

235303

235305

235306

235310

235314

235315

235316

235318

235319

235320

79.5.18.8

235324

235325

235327

235328

235330

cCL

695690960630935670665745+

975695745700+

780905740730f

735*

935820+

1090*

7601060*

600820715820900590670630850

ML

123.5

143.1

200.8

121.8

198.0

129.4

138.0

146.6

193.0

134.8

143.6

135.4

154.7

168.0

152.4

144.9

143.5

189.0

161.7

250.0

145.4

224.0

112.0

162.1

139.1

151.4

176.6

112.0

132.7

125.2

170.3

Long axis

Obs.MSGs

_

-

8---411196-

5-

8-

4277-23-

10-_--

10--4-

Estim.MSGs

_

-Ill---35409435-

25-30-3034359-

638-169----89--23-

Obs.MSGs

4774944171474648693074415185447114445

Short axis

AW

0.76

0.45

0.10

0.39

0.10

0.50

0.46

0.20

0.14

0.48

0.20

0.16

0.42

0.54

0.23

0.20

0.03

0.29

0.43

0.04

0.46

0.10

0.44

0.34

0.81

0.81

0.19

0.71

0.58

0.44

0.43

Estim.MSGs

4958874811244410182597112516416251751081815563136812

Growthrate

(mm/yr)

162.5

71.7

15.8

73.1

12.0

78.0

56.4

29.2

21.1

65.0

38.9

26.2

81.7

122.9

63.2

27.4

4.255.6

31.0

6.071.5

12.5

69.4

51.7

134.0

129.2

27.6

181.7

104.2

73.1

67.1

25


Part 2.—Continued.

SpecimenNumber

23533223533323533423533623533723533823534023534123534323534423534523534679.6.17.523534823534979.6.17.823535079.7.5.1235356235357235358235362235363235365235366235367235371235375235378235380235381235383235388235389235390235391235392235393235394235395235397235398235399235400235401235407235415235417235418235419

cCL

66061568082058585562064078083010106801000*755665580835975*730630675680730565600745660780640750915640810600830590785760730740640580f

660700780715610740690605

ML130.8119.1142.0145.6113.3157.9115.4116.2146.4159.1210.0121.7217.0141.4123.0-

170.2195.0144.4121.8129.4122.8139.9105.5123.7144.2120.0146.0131.4134.4173.5118.4156.0117.2155.5107.9147.3141.1146.0143.3119.5116.3128.1124.7149.9130.0113.5133.6123.7106.5

Long axis

Obs.MSGs

56611-

984-

7-51111541113--5644---86-

10454__

41084_-74-

54-

74

Estim.MSGs

20122324-251812-37-

2242019171835339--27192012---

2815-

43222213-_

36322324__

3622-

1326_

1811

Obs.MSGs

566115984464591054123010456766447778108444517964874557577

Short axis

AW0.560.790.470.490.980.500.541.400.330.350.420.540.060.500.450.420.410.050.030.680.350.550.630.730.500.380.680.430.660.630.160.520.730.770.340.820.350.310.420.491.020.180.420.450.310.770.580.870.600.74

Estim.MSGs

761010312631515166

1428761114010511774711610774077515415151110417991467575

Growthrate

(mm/yr)

87.995.063.577.5180.067.595.8198.349.052.360.3105.86.7

88.788.689.271.86.6

68.5117.057.390.797.8130.079.363.6102.573.585.0100.721.785.0109.3111.052.3136.349.347.762.369.5148.731.568.372.852.5111.780.7139.092.1112.0

Part 3

Summary of selected references on growth rates and age estimates of sexualmaturity in cheloniid sea turtles. The data listed below may not match thevalues listed in the text for these references, because we often summarizedthe original data; the data could not be standardized to a single method ofcarapace measurement. (Abbreviations: C, data from captive raised turtles;W, mark and recapture data from wild turtles; dash, no data; ?, questionabledata.)

SpeciesLocality

Caretta carettaFloridaFloridaGeorgiaNorth CarolinaColombiaColombiaGreat Barrier ReefFloridaFloridaFloridaFlorida

Chelonia mydasHawaiiOahuNeckerFrench FrigateLisianskiMidwayKure

hatchlingsimmaturesadults

U.S. Virgin IslandsFloridaTortugueroSarawakColombiaTorres StraitGreat Barrier ReefGreat Barrier ReefHeron IslandFloridaHeron Island

Samplesize

672529

13-

1318

25

421

19382468

354

141_48

401240122

Exper.cond.

WCCCCC

W

wccc

wwwwwwwcccwcwccwwwwwc

Growthrate

(mm/yr)

57.082.1

131.4110.3176.8111.4

-

59.0161.7182.9136.4

52.824.016.89.6

15.610.89.6

210.360.1

0.2450.4

110.05.8

104.0148.9

17.413.2

1.413.535.4

120.9

Age atmaturity(estim.

yrs)

--

6-7---

>3010-15

---

11243459365359

----

8-13-

4-6--

>30-

>3025-30

-

Reference

Bjorndal et al., 1983Caldwell et al., 1955Caldwell, 1962Hildebrand and Hatsel, 1927Kaufmann,1972Kaufmann, 1972Limpus, 1979Mendonca, 1981Parker, 1926Rebel, 1974Witham and Futch, 1977

Balazs, 1979, 1982Balazs, 1980, 1982Balazs, 1980, 1982Balazs, 1980, 1982Balazs, 1980, 1982Balazs, 1980, 1982Balazs, 1980, 1982Balazs, 1980Balazs, 1980Balazs, 1980Boulon, 1983Caldwell, 1962Carr and Goodman, 1970Hendrickson, 1958Kaufmann, 1972Kowarsky and Capelle, 1979Limpus, 1979Limpus, 1979Limpus and Walter, 1980Mendonca, 1981Moorhouse, 1933

27



SpeciesLocality

St. ThomasMauritaniaFloridaGrand Cayman

Eretmochelys imbricata

U.S. Virgin IslandsJamaicaJamaicaFloridaSri LankaSri LankaSeychellesSarawakSarawakColombiaTorres StraitSt. ThomasSamoa

Lepidochelys kempii

FloridaMexico

juvenilesadult females

Lepiodochelys olivacea

IndiaIndiaSri LankaSri Lanka

Samplesize

9-

1222

15??911

10347236

2

->10

151931

Exper.cond.

WC,W

CC

WCCCCCCCCC

WCC

C

CW

CCCC

Growthrate

(mm/yr)

51.1-

166.9-

33.650.323.386.8

247.293.0

170.5113.4132.1174.290.6

192.8172.3

42.0

105.035.0

125.2143.352.7

188.4

Age atmaturity(estim.

yrs)_

5.5-

8-10

-------

4----

3.5

_

-5.5

----

Reference

Schmidt, 1916Toquinetal., 1980Witham and Futch, 1977Wood and Wood, 1980

Boulon, 1983Brown etal., 1982Brown etal., 1982Caldwell, 1962Deraniyagala, 1939Deraniyagala, 1939Diamond, 1976Harrisson, 1963Harrisson, 1963Kaufmann, 1972Kowarsky and Capelle, 1979Schmidt, 1916Witzell, 1980

Caldwell, 1962

Marquez, 1972Marquez, 1972

Rajagopalan, 1984Rajagopalan, 1984Deraniyagala, 1939Deraniyagala, 1939

Part 4

Summary of selected references on minimum size and estimated ages atsexual maturity in miscellaneous freshwater and terrestrial turtles. The datalisted below may not match the values listed in the references, because wemay have abstracted the original data or extracted the data from graphs orfigures. (Abbreviation: PL, body size given as plastron length rather than ascarapace length.)

SpeciesLocality

Chelydra serpentinaIowa 99Iowa 66Quebec 99Tennessee

Chrysemys pictaAll 99Michigan 66Illinois 66Louisiana 66Pennsylvania 99Pennsylvania 66Michigan 99Michigan 66

Clemmys guttataPennsylvania 99, 66

Emydoidea blandingiiMassachusetts 66

Geochelone elepantopushoodensis 99hoodensis 66porteri 99porteri 66

Geochelone giganteaMalabar 99Grand Terre 99

Kinosternon JlavescensOklahoma

Kinosternon subrubrumFloridaOklahoma

Macroclemys temminckiiLouisiana

Malademys terrapinN. Carolina captive 99N. Carolina captive 66

Minimum sizeat maturity(CL, mm)

210190200145

120-130(PL)90 (PL)70 (PL)50-60 (PL)

110 (PL)80-90 (PL)

120-13090-105

80-90 (PL)

180-190

550700800950

--

80

8080

340-400

120-14080-90

Age ofmaturity (yrs)

6-104-5

--

-3-42-315-64-56-74-5

7-10

9-11

>30>30>30>30

16-1822-24

5-6

6-75-6

11-13

5-92-6

Reference

Christiansen & Burken, 1979Christiansen & Burken, 1979Mosimann & Bider, 1960White & Murphy, 1973

Cagle, 1954Cagle, 1954Cagle, 1954Cagle, 1954Ernst, 1971Ernst, 1971Wilbur, 1975Wilbur, 1975

Ernst, 1975

Graham & Doyle, 1977

MacFarland et al., 1974MacFarland et al., 1974MacFarland et al., 1974MacFarland et al., 1974

Swingland & Coe, 1978Swingland & Coe, 1978

Mahmoud, 1967

Ernst etal., 1973Mahmoud, 1967

Dobie, 1971

Hildebrand, 1932Hildebrand, 1932

29



Species

Locality

Pseudemys floridanaS. Carolina 99S. Carolina 66

Sternotherus carinatusOklahoma

Sternotherus odoratusOklahoma

Terrapene ornataKansas 99Kansas 66

Trachemys scriptaS. Carolina: Ellent. Bay 99S. Carolina: Ellent. Bay 6S. Carolina: Par Pond 99S. Carolina: Par Pond 66Illinois natural 99Illinois: heated 66

Minimum sizeat maturity(CL, mm)

200-240(PL)100-120(PL)

80

65

110-120100-110

160-170 (PL)80-100 (PL)

200-210 (PL)80-100 (PI)

170-195185-190

Age ofmaturity (yrs)

6-73-4

4

5-6

6-114-9

74734-53-4

Reference

Gibbons & Coker, 1977Gibbons & Coker, 1977

Mahmoud, 1967

Mahmoud, 1967

Legler, 1960Legler, 1960

Gibbons et al., 1981Gibbons et al., 1981Gibbons et al., 1981Gibbons et al., 1981Thornhill, 1982Thornhill, 1982

Literature Cited

Balazs, George H.1979. Growth, Food Sources and Migrations of Imma-

ture Hawaiian Chelonia. Marine Turtle Newsletter,(10): 1-3.

1980. Synopsis of Biological Data on the Green Turtlein the Hawaiian Islands. NOAA Technical Memo-randum NMFS, 141 pages.

1982. Growth Rates of Immature Green Turtles in theHawaiian Archipelago. In Karen A. Bjorndal, ed-itor, Biology and Conservation of Sea Turtles, pages117-125. Washington, D.C.: Smithsonian Insti-tution Press.

Baldwin, William P., Jr., and John P. Lofton, Jr.1959. The Loggerhead Turtles of Cape Romain, South

Carolina. Bulletin of the Florida State Museum, Bi-ological Sciences, 4( 10):319-448.

Bjorndal, Karen A., Anne B. Meylan, and Billy J. Turner1983. Sea Turtles Nesting at Melbourne Beach, Florida,

I: Size, Growth and Reproductive Biology. Biolog-ical Conservation, 26:65-77.

Boulon, Ralf H.,Jr.1983. Some Notes on the Population Biology of Green (Che-

lonia mydas) and Hawksbill (Eretmochelys imbri-cata) Turtles in Northern U.S. Virgin Islands: 1981-83. 18 pages. [Report to National Marine FisheriesService, NA 82-GA-A-00044.]

Bourke, R.E., G. Balazs, and E.W. Shallenberger1977. Breeding of the Sea Turtle {Chelonia mydas) at Sea

Life Park, Hawaii. Drum and Croaker, 17(2):4-9.Brown, Robert A., Guy C. McN. Harvey, and Lesley A.

Wilkins1982. Growth of Jamaican Hawksbill Turtles (Eretmoche-

lys imbricata) Reared in Captivity. British fournalof Herpetology, 6(7):233-236.

Cagle, Fred R.1954. Observations on the Life Cycles of Painted Turtles

(Genus Chrysemys). American Midland Naturalist,52(l):225-235.

Caldwell, David K.1962. Growth Measurements of Young Captive Atlantic

Sea Turtles in Temperate Waters. Los AngelesCounty Museum, Contributions in Science, (50): 1—8.

Caldwell, David K., Archie Carr, and Thomas R. Hellier, Jr.1955. Natural History Notes on the Atlantic Logger-

head Turtle, Caretta caretta caretta. Quarterly four-nal of the Florida Academy of Science, 18(4):292-302.

Caldwell, David K., Archie Carr, and Larry H. Ogren1959. Nesting and Migration of the Atlantic Loggerhead

Turtle. Bulletin of the Florida State Museum, Biolog-ical Sciences, 4(10):295-308.

Carr, Archie, and Donald Goodman1970. Ecological Implications of Size and Growth in

Chelonia. Copeia, 1970(4):783-786.Castanet, Jacques

1982. Recherches sur la Croissance du Tissu Osseux desReptiles. Application: La Methode Squeletto-chronologique. Doctoral dissertation, Universityof Paris, VII, 191 pages.

Castanet, Jacques, and Marc Cheylan1979. Les marques de croissance des os et des ecailles

comme indicateur de l'age chez Testudo hermanniet Testudo graeca (Reptilia, Chelonia, Testudini-dae). Canadian fournal of Zoology, 57(8): 1649-1665.

Castanet, Jacques, Jean-Pierre Gasc, Francois Meunier, andArmand de Ricqles

1970. Calcium et nature des zones de croissance cycliquedans I'os des vertebres poikilothermes. CompteRendu Academie de Science Paris, series D,270:2853-2856.

Castanet, Jacques, F.J. Meunier, and A. de Ricqles1977. L'enregistrement de la croissance cyclique par le

tissu osseux chez les vertebres poikilothermes:donnees comparatives et essai de synthese. BulletinBiologique de la France et de la Belgique,111(2): 183-202.

Castanet, Jacques, and Guy Naulleau1974. Donnees experimentales sur la valeur des marques

squelettiques comme indicateur de l'age chez Vi-pera aspis (L.) (Ophidia, Viperidae). ZoologicaScripta, 3(2):201-208.

Christiansen, James L., and Russell R. Burken1979. Growth and Maturity of the Snapping Turtle (Che-

lydra serpentina) in Iowa. Herpetologica, 35(3):261-266.

de Buffrenil, Vivian1980. Mise en evidence de I'incidence des conditions de

milieu sur la croissance de Crocodylus siamensis(Schneider, 1801) et valeur des marques de croiss-ance squelettiques pour revaluation de l'age indi-viduel. Archives de Zoologie experimentale et gener-ate, 121(l):63-76.

31


Deraniyagala, P.E.P.1939. The Tetrapod Reptiles of Ceylon. Volume 1, Testu-

dinatesand Crocodilians. Colombo: Colombo Mu-seum.

Diamond, A.W.1976. Breeding Biology and Conservation of Hawksbill

Turtles, Eretmochelys imbricata L., on Cousin Is-land, Seychelles. Biological Conservation, 9(3): 199-215.

Dobie, James L.1971. Reproduction and Growth in the Alligator Snap-

ping Turtle, Macroclemys temmincki (Troost). Cop-eia, 1971(4):645-658.

Ehrhart, L.M., and R.G. Yoder1978. Marine Turtles of Merritt Island National Wildlife

Refuge, Kennedy Space Center, Florida. In G.E.Henderson, editor, Proceedings of the Florida andInterrregional Conference on Sea Turtles, 24-25July 1976. Florida Marine Research Publication,33:25-30. Jensen Beach, Florida.

En low, Donald H.1969. The Bones of Reptiles. In C. Gans, T.S. Parsons,

and A. d'A. Bellairs, editors. Biology of the Reptilia,volume 1:45-80. London: Academic Press.

Ernst, Carl H.1971. Sexual Cycles and Maturity of the Turtle Chryse-

myspicta. Biological Bulletin, 140(2): 191-200.1975. Growth of the Spotted Turtle, Clemmys guttata.

Journal of Herpetology, 9(3):313-318.Ernst, Carl H., Roger W. Barbour, Evelyn M. Ernst, and

James R. Butler1973. Growth of the Mud Turtle, Kinosternon subru-

brum, in Florida. Herpetologica, 29(3):247-250.Frazer, Nat B.

1982. Growth and Age at Maturity of Loggerhead SeaTurtles: Review and Prospectus. Marine TurtleNewsletter, 22:5-8.

Frazer, Nat B., and Llewellyn M. Ehrhart1985. Preliminary Growth Models for Green, Chelonia

mydas, and Loggerhead, Caretta caretta, Turtles inthe Wild. Copeia, 1985(l):73-79.

Frazer, Nat B., and Frank J. Schwartz1984. Growth Curves for Captive Loggerhead Turtles,

Caretta caretta, in North Carolina, USA. Bulletinof Marine Science, 34(4):485-489.

Frazier, John1982. Aging Studies of Sea Turtles. 133 pages. [Report to

the National Marine Fisheries Service, NA 81-GA-C-00018, and the U.S. Fish and Wildlife Serv-ice, NA 81-GF-A-184.]

Gallagher, Robert M., Malcolm L. Hollinger, Robert M.Ingle, and Charles R. Futch

1972. Marine Turtle Nesting on Hutchinson Island,Florida, in 1971. Florida Department of Natural

Resouces, Marine Research Laboratory Special Re-port, (37): 1-11.

Gibbons, J. Whitfield1967. Variation in Growth Rates in Three Populations

of the Painted Turtle, Chrysemys picta. Herpetolo-gica, 23(4):296-303.

Gibbons, J. Whitfield, and John W. Coker1977. Ecological and Life History Aspects of the Cooter,

Chrysemys floridana (Le Conte). Herpetologica,33(l):29-33.

Gibbons, J. Whitfield, Raymond D. Semlitsch, Judith L.Greene, and Joseph P. Schubauer

1981. Variation in Age and Size at Maturity of the SliderTurtle (Pseudemys scripta). American Naturalist,117:841-845.

Graham, Terry E., and Timothy S. Doyle1977. Growth and Population Characteristics of Bland-

ing's Turtle, Emydoidea blandingii, in Massachu-setts. Herpetologica, 33(4):410-414.

Hammer, Donald A.1969. Parameters of a Marsh Snapping Turtle Popula-

tion, Lacreek Refuge, South Dakota. Journal ofWildlife Management, 33(4):995-1005.

Harrisson, Tom1963. Notes on Marine Turtles: 13—Growth Rate of

the Hawksbill. The Sarawak Museum Journal,l l(21,22):302-303.

Hemelaar, A.S.M., and J.J. van Gelder1980. Annual Growth Rings in Phalanges of Bufo bufo

(Anura, Amphibia) from the Netherlands andTheir Use for Age Determination. NetherlandsJournal of Zoology, 30(l):129-135.

Hendrickson.John R.1958. The Green Sea Turtle, Chelonia mydas (Linn.), in

Malaya and Sarawak. Proceedings of the ZoologicalSociety of London, 130(4):455-535.

Henwood, T.In prep. Age, Growth, and Survival in Loggerhead Sea

Turtles, Caretta caretta.Hildebrand, Samuel F.

1932. Growth of Diamond-back Terrapins, Size At-tained, Sex Ratio and Longevity. Zoologica,9(15):551-563.

Hildebrand, Samuel F., and Charles Hatsel1927. On the Growth, Care and Behavior of Logger-

head Turtles in Captivity. Proceedings of the Na-tional Academy of Sciences, 13(6):374-377.

Hohn, A., and J. Frazier1979. Growth Layers in Bones and Scutes of Sea Turtles:

A Possible Aging Method [abstract]. American Zo-ologist, 19(3):953.

Hulse, Arthur C.1976. Growth and Morphometrics of Kinosternon sono-

riense (Reptilia, Testudines, Kinosternidae). Jour-

NUMBER 427 33

nal of Herpetology, 10(4):341-348.Kaufmann, Reinhard

1972. Wachstumsraten in Gefangenschaft gehaltenerMeeresschildkroten, II. Instituto Colombo-Aleman,lnvestigaciones Cientificas, 6:105—112.

Kowarsky, John, and Mikael Capelle1979. Returns of Pond-reared Juvenile Green Turtles

Tagged and Released in Torres Strait, NorthernAustralia. Biological Conservation, 15(3):207-214.

Legler.John M.1960. Natural History of the Ornate Box Turtle, Ter-

rapene ornata ornata Agassiz. University of KansasPublications, Museum of Natural History,ll(10):527-669.

Limpus, Colin1979. Notes on Growth Rates of Wild Turtles. Marine

Turtle Newsletter, (10):3-5.Limpus, Colin, and David G. Walter

1980. The Growth of Immature Green Turtles (Cheloniamydas) under Natural Conditions. Herpetologica,36(2): 162-165.

MacFarland, Craig G., Jose Villa, and Basilio Toro1974. The Galapagos Giant Tortoises (Geochelone ele-

phantopus). Part I: Status of Surviving Popula-tions. Biological Conservation, 6(2): 118-133.

Mahmoud, I.Y.1967. Courtship Behavior and Sexual Maturity in Four

Species of Kinosternid Turtles. Copeia,1967(2):314-319.

Marquez, Rene1972. Resultados preliminares sobre edad y crecimiento

de la tortuga lora, Lepidochelys kempi (Garman). InMemorias IV Congreso Nacional de Oceanografia,Mexico, D.F., 17-19 Noviembre 1969, pages 419-427.

Mattox, Norman T.1936. Annular Rings in the Long Bones of Turtles and

Their Correlation with Size. Transactions of theIllinois State Academy of Science, [1935] 28(2):255-256.

Mendonca, Mary T.1981. Comparative Growth Rates of Wild Immature

Chelonia mydas and Caretta caretta in Florida. Jour-nal of Herpetology, 15(4):447-451.

Minakami, Korebumi1979. An Estimation of Age and Life Span of the Genus

Trimeresurus (Reptilia, Serpentes, Viperidae) onAmani Oshima Island, Japan. Journal of Herpetol-ogy, 13(2): 147-152.

Moorhouse, F.W.1933. Notes on the Green Turtle (Chelonia mydas). Re-

port of the Great Barrier Reef Committee, 4(1): 1-22.Mosimann, James E., and J. Roger Bider

1960. Variation, Sexual Dimorphism, and Maturity in a

Quebec Population of the Common SnappingTurtle, Chelydra serpentina. Canadian Journal ofZoology, 38(1): 19-38.

Nuitja, I. Njoman S., and I. Uchida1982. Preliminary Studies on the Growth and Food Con-

sumption of the Juvenile Loggerhead Turtle (Car-etta caretta L.) in Captivity. Aquaculture,27(2):157-160.

Parmenter, Robert R.1980. Effects of Food Availability and Water Tempera-

ture on the Feeding Ecology of Pond Sliders (Chry-semyss. scripta). Copeia, 1980(3):503-514.

Parker, G.H.1926. The Growth of Turtles. Proceedings of the National

Academy of Science, 12(7):422-424.1929. The Growth of the Loggerhead Turtle in Captiv-

ity. American Naturalist, 63(687):367-373.Peabody, Frank E.

1961. Annual Growth Zones in Living and Fossil Ver-tebrates. Journal of Morphology, 108(1): 11-62.

Pilorge, Thierry, and Jacques Castanet1981. Determination I'age dans une population naturelle

du lezard vivipare (Lacerta vivipara Jacquin 1787).Ada OEcologica/OEcologia Generalis, 2(1):3—16.

Rajagopalan, M.1984. Studies on the Growth of Olive Ridley Lepidochelys

olivacea in Captivity. Central Marine Fisheries Re-search Institute (India), 35:49-54.

Rebel, Thomas P.1974. Sea Turtles and the Turtle Industry of the West Indies,

Florida, and the Gulf of Mexico. 250 pages, CoralGables, Florida: University of Miami Press.

Rhodin, Anders G.J.1985. Comparative Chrondo-osseous Development and

Growth of Marine Turtles. Copeia, 1985(3):752-771.

Rhodin, Anders G.J., John A. Ogden and Gerald J. Con-logue

1981. Chondro-osseous Morphology of Dermochelys cor-iacea, a Marine Reptile with Mammalian SkeletalFeatures. Nature, 290(5803):244-246.

Romer, Alfred Sherwood1956. Osteology of the Reptiles. 772 pages. Chicago: Uni-

versity of Chicago Press.Ruckdeschel, Carol, and George R. Zug

1980. Marine Turtles [ 1979 Strandings on CumberlandIsland]. SEAN Bulletin, 5(2): 14-16.

1981. 1980 Marine Turtle Stranding Survey Cumber-land Island, Georgia. SEAN Bulletin, 6(2):20-21.

Schmidt, Johs.1916. Marking Experiments with Turtles in the Danish

West Indies. Meddelser fra Kommissionen for Ha-vundersogelser, Serie Fiskeri, 5(1): 1-26.


Stickney, Robert R., David B. White, and Daniel Perlmutter1973. Growth of Green and Loggerhead Sea Turtles in

Georgia on Natural and Artificial Diets. Bulletinof the Georgia Academy of Science, 31:37-44.

Stoneburner, D.L.1980. Body Depth: An Indicator of Morphological Var-

iation among Nesting Groups of Adult Logger-head Sea Turtles (Caretta caretta). Journal of Her-petology, 14(2):205-206.

Suzuki, Howard K.1963. Studies on the Osseous System of the Slider Tur-

tle. Annals of the New York Academy of Sciences,109(l):351-410.

Swingland, Ian R., and Malcolm Coe1978. The Natural Regulation of Giant Tortoises Pop-

ulations on Aldabra Atoll: Reproduction. Journalof Zoology, London, 186:285-309.

Thornhill, Gary M.1982. Comparative Reproduction of the Turtle, Chryse-

mys scripta elegans, in Heated and Natural Lakes.Journal of Herpetology, 16(4):347-353.

Toquin, A. Le., E. Gamel, and J. Trotignon1980. Morphologie, croissance individuelle et dyna-

mique des populations de la tortue verte (Cheloniamydas L.) au Bane d'Arguin (Republique Isla-mique de Mauritanie). Revue d'Ecologie, Terre etVie, 34:271-302.

Uchida, Itaru1967. On the Growth of the Loggerhead Turtle, Caretta

caretta, under Rearing Conditions. Bulletin of theJapanese Society of Scientific Fisheries, 33(6):497-507.

White, James B., and George G. Murphy1973. The Reproductive Cycle and Sexual Dimorphism

of the Common Snapping Turtle, Chelydra serpen-tina serpentina. Herpetologica, 29(3):240-246.

Wilbur, Henry M.1975. The Evolutionary and Mathematical Demography

of the Turtle Chrysemys picta. Ecology, 56(1 ):64-77.

Witham, Ross1971. Breeding of a Pair of Pen-reared Green Turtles.

Quarterly Journal of the Florida Academy of Science[1970], 33(4):288-290.

Witham, R., and C. Futch1977. Early Growth and Oceanic Survival of Pen-reared

Sea Turtles. Herpetologica, 33(4):404-409.

Witzell, W.N.1980. Growth of Captive Hawksbill Turtles, Eretmochelys

imbricata, in Western Samoa. Bulletin of MarineScience, 30(4):909-912.

Wood, James R., and Fern E. Wood1980. Reproductive Biology of Captive Green Sea Tur-

tles Chelonia mydas. American Zoologist, 20(3):499-505.

1981. Growth and Digestibility for the Green Turtle(Chelonia mydas) Fed Diets Containing VaryingProtein Levels. Aquaculture, 25:269-274.

Zug, G.R., A S . Rand, A.H. Wynn, and B. Bock1983. Skeletochronology of Panamanian Iguana iguana:

A Test of Annual Periosteal Growth [abstract].Program of Twenty-sixth Annual Meeting (SSAR)and Thirty-first Annual Meeting, HL, page 100.

Zug, George R., Addison Wynn, and Carol Ruckdeschel1983. Age Estimates of Cumberland Island Loggerhead

Sea Turtles. Marine Turtle Newsletter, (25):9-l 1.

REQUIREMENTS FOR SMITHSONIAN SERIES PUBLICATION

Manuscripts intended for series publication receive substantivereview (conducted by their originating Smithsonian museums oroffices) and are submitted to the Smithsonian Institution Presswith Form SI-36, which must show the approval of the appropriateauthority designated by the sponsoring organizational unit. Re-quests for special treatment—use of color, foldouts, case-boundcovers, etc.—require, on the same form, the added approval ofthe sponsoring authority.

Review of manuscripts and art by the Press for requirementsof series format and style, completeness and clarity of copy, andarrangement of all material, as outlined below, will govern, withinthe judgment of the Press, acceptance or rejection of manuscriptsand art.

Copy must be prepared on typewriter or word processor,double-spaced, on one side of standard white bond paper (noterasable), with 1 VA" margins, submitted as ribbon copy (notcarbon or xerox), in loose sheets (not stapled or bound), andaccompanied by original art. Minimum acceptable length is 30pages.

Front matter (preceding the text) should include: title pagewith only title and author and no other information, abstract pagewith author, title, series, etc., following the established format;table of contents with indents reflecting the hierarchy of heads inthe paper; also, foreword and/or preface, if appropriate.

First page of text should carry the title and author at the topof the page; second page should have only the author's nameand professional mailing address, to be used as an unnumberedfootnote on the first page of printed text.

Center heads of whatever level should be typed with initialcaps of major words, with extra space above and below the head,but no other preparation (such as all caps or underline, except forthe underline necessary for generic and specific epithets). Run-inparagraph heads should use period/dashes or colons as neces-sary.

Tabulations within text (lists of data, often in parallel columns)can be typed on the text page where they occur, but they shouldnot contain rules or numbered table captions.

Formal tables (numbered, with captions, boxheads, stubs,rules) should be submitted as carefully typed, double-spaced copyseparate from the text; they will be typeset unless otherwiserequested. If camera-copy use is anticipated, do not draw ruleson manuscript copy.

Taxonomic keys in natural history papers should use thealigned-couplet form for zoology and may use the multi-levelindent form for botany. If cross referencing is required betweenkey and text, do not include page references within the key, butnumber the keyed-out taxa, using the same numbers with theircorresponding heads in the text.

Synonymy in zoology must use the short form (taxon, author,yearpage), with full reference at the end of the paper under"Literature Cited." For botany, the long form (taxon, author,abbreviated journal or book title, volume, page, year, with noreference in "Literature Cited") is optional.

Text-reference system (author, yearpage used within the text,with full citation in "Literature Cited" at the end of the text) mustbe used in place of bibliographic footnotes in all ContributionsSeries and is strongly recommended in the Studies Series:"(Jones. 1910122)" or Jones (1910:122)." If bibliographic

footnotes are required, use the short form (author, brief title, page)with the full citation in the bibliography.

Footnotes, when few in number, whether annotative or biblio-graphic, should be typed on separate sheets and inserted imme-diately after the text pages on which the references occur. Exten-sive notes must be gathered together and placed at the end ofthe text in a notes section.

Bibliography, depending upon use, is termed "Literature Cited,"References," or "Bibliography." Spell out titles of books, articles,

journals, and monographic series. For book and article titles usesentence-style capitalization according to the rules of the lan-guage employed (exception: capitalize all major words in English).For journal and series titles, capitalize the initial word and allsubsequent words except articles, conjunctions, and prepositions.Transliterate languages that use a non-Roman alphabet accordingto the Library of Congress system. Underline (for italics) titles ofjournals and series and titles of books that are not part of a series.Use the parentheses/colon system for volume (number):pagination: "10(2):5-9." For alignment and arrangement of ele-ments, follow the format of recent publications in the series forwhich the manuscript is intended. Guidelines for preparing bibli-ography may be secured from Series Section, SI Press.

Legends for illustrations must be submitted at the end of themanuscript, with as many legends typed, double-spaced, to apage as convenient.

Illustrations must be submitted as original art (not copies)accompanying, but separate from, the manuscript. Guidelines forpreparing art may be secured from Series Section, SI Press. Alltypes of illustrations (photographs, line drawings, maps, etc.) maybe intermixed throughout the printed text. They should be termedFigures and should be numbered consecutively as they willappear in the monograph. If several illustrations are treated ascomponents of a single composite figure, they should be desig-nated by lowercase italic letters on the illustration; also, in thelegend and in text references the italic letters (underlined in copy)should be used: "Figure 9fc." Illustrations that are intended tofollow the printed text may be termed Plates, and any componentsshould be similarly lettered and referenced: "Plate 9b.." Keys toany symbols within an illustration should appear on the art ratherthan in the legend.

Some points of style: Do not use periods after such abbrevi-ations as "mm, ft, USNM, NNE." Spell out numbers "one" through"nine in expository text, but use digits in all other cases if possible.Use of the metric system of measurement is preferable; whereuse of the English system is unavoidable, supply metric equiva-lents in parentheses. Use the decimal system for precise meas-urements and relationships, common fractions for approximations.Use day/month/year sequence for dates: "9 April 1976." Formonths in tabular listings or data sections, use three-letter abbre-viations with no periods: "Jan, Mar, Jun," etc. Omit space betweeninitials of a personal name: "J.B. Jones."

Arrange and paginate sequentially every sheet of manu-script in the following order: (1) title page, (2) abstract, (3) con-tents, (4) foreword and/or preface, (5) text, (6) appendixes, (7)notes section, (8) glossary, (9) bibliography, (10) legends, (11)tables. Index copy may be submitted at page proof stage, butplans for an index should be indicated when manuscript is sub-mitted.

«•

,111/.,

age determination of sea turtle.pdf

Documents

theskeleton of caretta

series publications

smithsonian imprint

i t h s o n

series of reports

o g y n u

incremental growthmarks

thefollowing active