Studies of the systematics and reproductive cycles of the genus Lepidochelys

Material Information

Studies of the systematics and reproductive cycles of the genus Lepidochelys
Pritchard, Peter Charles Howard, 1943-
Copyright Date:


Subjects / Keywords:
Beaches ( jstor )
Coasts ( jstor )
Eggs ( jstor )
Eggshells ( jstor )
Female animals ( jstor )
Sea turtles ( jstor )
Seas ( jstor )
Skull ( jstor )
Species ( jstor )
Turtles ( jstor )
Gulf of Mexico ( local )
bibliography ( marcgt )
non-fiction ( marcgt )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000985820 ( AlephBibNum )
17625093 ( OCLC )
AEW2233 ( NOTIS )


This item has the following downloads:

Full Text







It is difficult to convey sufficient thanks to Archi.e

Carr, the chairman of my supervisory committee, for allow--

ing me to work with hir. despite my purely physical science

background and our lack of previous personal acquaintance.

While giving' me a hi;h degree of independence throiu-hout

the project, he was a never-failing source of knowledge

and inspiration whenever needed. I ha e also benef ited

greatly both from discussions and from formal courses with

the other members of ry supervisory committee Walter..

Auffenber, W'iliiram Carter and Frank Nordlie; 1 have

learned much from all three.

I am very grateful for financial support fr-, th

World Wildlife Fund, from the Brcwne Fund of the Royal

Society, and from the Department of Tropical Agricul tur

and the Center for Latin LAerican Studies of the University

of Florida. In addition, Archie Carr'-v sources of research

support (JSF Grant 0G 3930; ONR Contract 10O (12) with

the University of Florida; Caribbean Conservation Corpor-

ation) furnished turtle t.g-ging eq cipment and funds for

tag-return rewards. I an_ grateful to the Graduatc Schoco

of tho University of Florida for the award of a fell wslij "

for the 1967--65 academic year, and to he Dcp'rt .nts of

Zoolcrg and of Biol ical Scieinc_ fcr teaching as:istn. C"-

shiis in other yco-s.

My field work in Surinam would have been impossible

without the assistance of J.P.Schulz and F. v3bberian of

the Surinam Forest Service. Many other members of the

Forest Service too were of great assistance in the field;

in particular I would like to mention G.Plak, A.V!olf and

E.Donner. The ex-Governor-General of Guyana, H.E. Sir

Richard Luyt, very kindly offered me accJ- tions N

much assistance in Georgetown. A number of Guyana

government officials, in particular Prs. UWiifrir Gaskin,

C.McA.Ashley, '.L.Persaud and D.Shaw, were of invaluable

help in different ways. Kahamad Ianif and NO.Pocnai

of tho Guyana I cseum greatly facilitated rmy third trip

to Shell Beach, while Dennis Joaquin of the Waini River

was kind enough to rescue me when my companion on Shell

Beach departed takin; the only boat. P.W.Kent, Miy

biochemistry supervisor at Cxford, was of great assistance

with the Royal Society granb application, as was I~~cfh

Popenoc with the University of Florida grants.

In M..exi. co, Antonij.o Iontoya, Hu.mbeorto Chavez and Rene

Iarquez were courteous and helpful hosts. Susan Paredes

was a sti:nulabinZ field companion in Honduras, while Steve

Bass w;;s of great help on the Tam ul i1as trip, as w:ere To

Stubbs and T'c Lesurme in Surin n in 1 '. In !''7 and

19l I was very fort :nate in havinC 'ilin Grel' ioc as

a co-worker in w in.r : I have bcinfi td ;':oc'tly bth fr

his joining" in a Od' iu s ;'c]l o u lrt. bcchs n"! also

frcn his r. t in ". '-hs biolc in t co1al and the w'ny.

ii i.

of sea turt]ae in particular.

Ross Kiester read and commented on several sections

of the manuscript ani I was fortunate in having the

benefit of his broad knowledge of turtle matters and of

evolutionary biology. Paul Laessle was generous in

lending me various items of drawing equipment and taking

much time to instruct me in their use.


ACKNO'I! LEDG ':ET'S . . . . .. .. . . ii

LIST OF TA-- :':, . . . . . viii

LIST OF FIGURES . . . . . . . . ix

INTRODUCTION . . . . . . . . I


LEPDOCELY 1S (1';0 . . . . 17

FROM SU-RIA!! A, L.liT A' :'..(:. PACIFIC
HCT,"i A ND ]E I 0 .... . .. . . 28

Shell Characteristics . . . . . 2

Absolute Carapace Longth . ., . . 2

Relative Carace Width . . . . 3

Relative Pla;bral Length . . . . 35

Relative !idth of Bridge . . . . 4

Absolute Carapnce Len ;th of Hatchlin . 4

Relative Car apce 'Jidth of Match '- . . 147

Lamina Counts Genral Discussion . . 17

Lamiai Coaint of Sea Turtles . . 51

Possible Sirnific nc of i ultil: inti on . 53

Centr ]. Lamrin Co : . . . . .. .

Costa C .. . . .. .

\ri;ti i Cai. race . .

VaKI in P3 j am V irac . .. . . 71

Infra!iarginal Counts .n d Significance of
Infra :ircinal Fores . . . . . .

Possible Significance of Differences in
Shell Shape . . . . . .

Carapace Ostoolony . . .

Skull Characters . . . .

Comparison of Si;e of Skul3s of : ature Females

Comparison of Relative 7'idth' of Skulls of
Iaturc Fen les . . . . .

Relative Helight of Or

Relative Width of ?te

Structure of the Alve

Characters of the Lowe

Significance of S'-:ll

1- 0 -






Scalatbon of the Po. .

Color . . . .

Habitat and Diet of /dul .

Sexual Di 'Cruhis ', ..

Copulation . .

Choice of osti ngi- :achs

Nesting Process ..

Le cV: "1.v L'

Emr' ce fr "' AC

Prepa ratin of .. r.

Layin: of D-q . ..

Fin; I jnr Of

d t 0

y coid Bridge .

Ur Surface

Sal . .

.ifere nc..

T ]-es c

C 0 f C 0 C

* 6 C 0 V

c . t e I C

. e



. 105





. 12-


. :!33

. 135




. 1 '147

Duration of :estin; Stages .. . . .. 1

Geographical Differences in '!estinn P'rocess . 1

Clutch Size and E Dimensions . . . 152

Egg Dimensions. . . . . . . .

Intervals Between Iesting E-.r-'ces . . . 160

Reproductive Cycles . . . . . . . . 169

ARRIBADA F ATIi . . . . . . . 170



i) Relationships Within the Fa mily Cheloniidae . 196

ii) Relationhijps w',ith Other Sen Turtle Families 199

DISCUSSo.ii ON .. . . . . 205

L TE CITED . . . . . . . 216

BIOGRA ICAL ,TCH . . . . . . . .. 224


Ta ble Pa se

1 Recorded Central Lanina Counts of Hatch- 57
lin; LenidochPlvj olivacrf from Several

2 Frequency of Different Costal Lnmina Counts 61
for Three Populations of i:c

3 Costal Lamina Counts for Three Froods of 68
Lenidochetl.y olive aca from Isla de
Ratones, on idur s

-4 I-arginal Lanina Freqje icies for Three 70
Populations of Ievidochelvr

5 Frequency of Different Into, ', -r Counts 71
for Three Populations of LT Ji cocje]. r

6 Infranarg;ial Countqs for Lo/- rh TuC c rtles 74
(Carjtta c .? ) from Three Areac

7 Infra:arginal Counts for Three Pcpulations 75
of Le nidnchOelvs
8 Diameters of Randon E gs laid by Le2l:i(~pcr:i" 157
olivacea at Eilant;:i Surinaj.m

9 Observed Nesting Agcroegations of Lmidcc0.3 el 172
keni in 1966 (fide lontoya, pens, ccrn:)

10 Estimated 'Nubers of Turtles in in, :'ri- 184
badns of ]ye idoc'nlys ol ivaccr RecG' ed
in 1905 ( e o co. Tcrs,, cc'-.i

11 Recovery Data for 2: al.e Locj.i cole.]- 190
olivace-l taLEcd at EilAnti, i



FiMgure Pa ge

1 Distribution of strai h-,-line cara- 30
pace lengths for 203 mature female
LeidoPche7!-7 ke:ni from Rancho
Nuevo, : ulipas, I-exico (fide
Chavez et al., 1967)

2 Distribution of straight-line cara- 30
pace length for 241 mature fcmAle
,Lei6dochelys olivaccei from '; _11
Beach, Guyana and Bigi Santi and
Eilanti, Surinaw

3 Distribution of straig~ht-line cara- 30
pace I.on, '.,s for 99 nature female
Lep idCoc oLs I j1Lva !c from the
Pacific coast o.f onduras

4 Distribution of carapace lc:--ths for 30
201 Lrei 'c i cVc .oJliva canFht at
soa off norrh --:c. stern i. eico a s, adod
a:ras: males unshaded areas: fe-
nales (fi.e Anon, .1966a and b)

5 Carapace di:r:ension (width plotted 34
against lenr 1e'h) for Len~ ioc eil s
k i 1. (sol id circles) 1
ch0vis ol piv ceC (open ci.rcc:)

6 Caralacoe v:r'th pl ott&d a Aint' c'ra- 37
pace lenth for 107 .tre c-'ale
]3.Lido c (. i, iv-cco fro1 ''Ce Gulf
o:' onscca, io dua:

7 Pl.astiraL li: Oltiv. to c ar p, loAi:;fch rc," f 2 crtnro fo al id'l1"
cho- i . O:- 3C L. ... r' s

SuC narl (C) ci ;i Q

Awl Van al.*'/ cunl~lr::ic~lc -i"teral for';' ^3
T)f>'oe '. ; C'*cpu L. o'is (1 lt)
,i^ C" ,,c' .I? )" f"r.;( 'i "0
r'. c '1 1."

9 Distribution of sitrri:ht-line carapace 46
lengths for 124 hatchlings Le~ido-
ch I-n'miri j. 199 hatchlinr: L.oli-
vwc'. from. Surinam; and 94 hatchlin:,
L, oli vacen from Pacific !onduras

10 Carapace width plotted against carapace 49
length for hatchlings L, o]ivaocca from
Honduras and from Surinam

11 Relative frequency of different central 59
lamina counts for four populations of
Sr,i cido c hel3 -vs

12 Frequency of left and right coastal laminae 64
for four populations of Le'iidochllys

13 Graph showing average number of left and 66
right coastal laninae for four pop-
ulations of Le-.idochnolv

14 Graph showing avera number of central 66
laminae for four populations of

15 Transverse and frontal sections th.rou 78
plantron of mature fcnmale. Lc"ridcc hc v
knri in region ilmmediately under-
lyinL an inframarLinal pore

16 Iid--rnarginal laminae of Lkempij from 81
Roncho Nuevo; mid..-marlinal laminae
of ,o]ivarCSe from Shell Beach,

17 Posterior marginal and supracaudal laminae 84
for from Rancho Nuevo; same
for oloiva cep from Guyana and Surinam

18 Anterior carspace profiles of L.komni 86
from PRncho .uovo; same for Leoli-
vacon from Guynna and Suri.nam; iane
for L],oo.ivac?. front: Pacific :cxiLco

19 Neural bones of L.kc-pi from Viracru,, 89
Nexico, and Cecar F, Flori.da; sme
for JolJ vnc_ .n. fr.oi Pacific I.'-exico;
sane for ,..olirac.: .from. Ceylon (after
Dera"ni: a -, 1-93.94)

20 Dorsal view of skulls of nature female L. 94
pliva_,_ce (left) and L,,kmrj- (ri- '..)
21 Ventral view of skulls of mature female 94
L.olivacoa (left) and L.foemri (ri 't)

22 Width/basicranial len '..h of skulls of 96
mature female L.kieli and L.olivacea
from Guyana and Guerrero, IMexico

23 Lateral views of skulls of mature female 99
L.4 kej_ (above) and L.olivacea (below)
21:. Vertical height of orbit (relative to 102
minimum vertical height of maxilla
below orbit) for skulls of three pop-
ulations of Lenidochelvs

25 Width of ptery (cid bridge (as fraction of 104
basicranial length) for three porpua-.
tions of LeC:idocilrys

26 Upper rhamphothecae of mature female 106
Seio e h.1 keP r i ( left) and ._erjcp-
chol0. olJivyLcO c righti)
27 Posterior viewer of skulls of L,.' erii 1.07
(above) and .I.olivacea (below), o-
inI presence of strong bony alveolar
ri(c iln 1i only

2S Lower jaws of L. :e o ( left); of Lo.l 'r_- 108
cea from Guerrc o (aboe center and
right)i and cf Loli-mvac fror Gu-yan
(below center and rij t)

29 Plot of sy:Tph-ysc.a len aJ inist to al 110
width for 1 lower aw o nature f
L.o ivace. freon Gue rero

30 Lower rh r! of i (ri '"',) 111
and LogO i :' (1 ",)

31. Lateral vie.e: lov e ja,'w o 1 ~::i. 112
(above) and J j.. ijLcj" ("o .....

33 Dorsal hend s,, ri.e. fo r: A: C:, i g !11
13: CQ] e].n n do :." .. C f.el' .c.'-' ':. .''
.j32I 1:. ..13.
331 h c,

34 Dorsal head scale ara'r- -ents for: .1: 122
IeCidoc>Kls h i 'iL : L. oljiJvce
from Hond 'Es 0: Lorlivac fi ro
Surina: thronee hatc'-,linJ-s and three
adults in each case)

35 Color of ridley turtles; A: adult female 126
L.....0 e i 3 y vo11 n k r!-
C: adult f:~- l ; c 1 y .o '....o
D: adult ferae1 L,.o2.i nacca fro ondi.ouras

36 Plot of clutch sioz a ~a~ist carapace 156
length for !.5 mature ft de r-l
chJo]ls ol :ivJcac ilromn I Hondilr

37 Histogram showing frequency of intervals 163
(in days) Cbetw.o: nest ing e or enoe O
of IpJ^jie^ er.lj iv- J^.' i- i= jti,
Surinam in 19r./ (s -c/o ) and 19]S (un-

38 numbers of turtles nlstin C; ch niht at 176
Eilanti, Surlna in the 1967 season

39 IurTmers of turtles na 'in ach. i ah t 178
Eilanti, Surinan', in 195 s a.on

40 Recovery locations for frerole ,L ej io]y vis 193
o]li.v cen to, t .q ila.t" i, nar:l
arro',ws do riot shoi routes a at .i L? l 'v-


Jepidoch-lvys is one of the five genera of living sea

turtles, and is currientlyv thought to include two forms

(kemp and olivtcea), .which are given full specific status

by most modern authorities, though only subspecific status

(within the species o,! Ivce) by others. Herpetologists

cor ,only call both species ridleyss a vernacular name of

unknown origin used by Florida turtle fishzrren.

Le;ico 0o l.._ ..a -.kerm:i. breeds only on the Gulf coast of

iexico, though non,-breding individuals are found through-

out the Gulf o. Mexico and on both sides of the North

Atlantic Ocean. n L.pochlv, oliva.c.a is virtually circum-

tropical in distribution, but there are odd areas of dis-

continuity in the range which are discussed in detail later.

Much of thi.E thesis is taken up with an evaluation of

the degree of difference both betwee.-n o.iva-ca and L.kempi

and also beto:eon the ol ivac popnulations at the extreme

ends of the narly c.rcua lobal range the populations in

the East Pacific and in the est South Atlantic. Manv

criteria external, osteological, behavioral and physio-

logical were u1, in an attempt to decide whether the

separation of cj.:'.ya1"_ and cio)J. as full species was justi-

fiable and -'hethr. ,AL ext reme populations of ol jvycoa

shov;cw: suf.ic.ien' divrence to :justify subspecific or other
-,:r I d L -..L

separation. If differences were found, an attempt would be

made to rationalize them in terms of function; on the other

hand, absence of even slight differences would constitute a

remarkable testimonial either to the slowness with which sea

turtles evolve, or to the uniformity of the marine environ-

ment in widely separated parts of the world.

When the writer started this study in 1965, there was

no published record of any Lenido~celys having been tagged

and later recovered, either at sea or on a nesting beach,

Consequently the nesting cycles, both within a season (all

other sea turtle genera are known to nest several times in

a season) and from year to year were completely unknown.

Nor were any data on migrations available, although it was

presumed that these took place, since ridleys were known to

form huge simultaneous nesting aggregations ('arribadasf) in

certain places, with many thousands of turtles being assumed

to be drawn from considerable distances. The discovery of a

small beach in Surinam where large numbers of ridleys nested

enabled me to tag over 1000 female turtles and learn much

about all these previously unknown aspects of ridley biology.

These_, drta are unfortunately not comparative, since for lo-

"stic reasons it was impossible to tag more than an insignifi-

:.u fraction of the East Pacific: turtles in the Honduras-

nesting population that was studied; consequently recovery

data fo:r the latter population are excceding]y scant.

Since a formal taxonomic sbudy of any animal group must

be based on a sure knowledge of what names are available'

the comparative section of this thesis is preceded by a

nomenclatorial history of the genus Lepidochelvs, which

takes careful note of all namesin the literature which

may appertain to members of this genus and attempts to

evaluate their merits.

The possible significance of certain of the phenomena

described (e.g. lamina variation, skull differences, arribada

formation, interesting intervals) is discussed at the ap-

propriate point in the text. Others (such as color) which I

am unable to rationalize or discuss in a convincing fashion,

I have still included for completeness' sake.

The discussion, which follows a literature survey of all

fossil evidence which may bear on the problems in hand, is

taken up with a 'value judgements of the similarities and

differences between different Lonidochejvs populations and

their bearing on the nomenclatorial situation, and also with

a marshalling of all available evidence that applies to the

problem of whore and when the genus Lepidochelvs originated

and how the living forms reached their present distribution.


Despite the numerous well-defined points of difference

which are now known to separate the genera Lenidochelys and

Caretta, the ridley turtle of the Indian and Pacific Oceans,

now known as Le idochelys oivacea, was until comparatively

recent years confused with the Pacific loggerhead turtle,

Caretta caretta Uigas, while both the South Atlantic enclave

of the same species and the North Atlantic Le),idochelys kemoni

have boon confused with the Atlantic loggerhead, Caret:ta

caretta caretta. Consequently, in any discussion of the

nomenclatorial history of leo)idochelys it is important to

take note of all species described under the generic name

Caretta,as well as others which are customarily synonymiied

with Caretta caretta,

Linnaeus (1758: 19,) included two sea turtles in his

System Naturae, calling them Testudco mydIs and Testudo

caretta. To these T.coriace and T.i-bricata were added

later (Linnaeus 1766: 351)l There is little doubt that the

first of these names refers to the green turtle, the third

to the leatherback and the fourth to the hawkshill, although

in no case do types exist. Tc:jide caret;a, is usually

thought to have been a compost of the loggerhead and the

hawksbill, which Linntaus realized in proposing the nr'.w name

imbricata fr the latter in 1.766. It is unfortunate that

the name c-retta, based on the widespread vernacular name


carey for the ha:ksbill., was not reserved for that species,
but the rules do not. allow any further change to be made.

It vas argued by Deraniyagala (1939b) that Linnaeust

description of Testudo c:retta referred with equal accuracy

to Caretta caretta and to Lei)idochelvs kempi (then known as

Thalassochelys (Coloochelvs) kenroi), and that the proper use

of the name caretta must be based on the turtle illustrated

by Schoepff (1792, pl. 16), which was labeled Testudo caretta

and was accompanied by the first specific account of the

species. Deraniyagala identified Schoepff's figure as ker.oi

on the following grounds:

1) The presence of four inframiarginals (on one side)

2) The presence of a single enlarged mandibular scale

on each side

3) The subcircular shape of the carapace.
Parker (.939) challenged these conclusions, remarking

that the 4-3 inframarginal count had been found in several

specimens of the loggerhead but never in the ridley. He also

pointed out that two accounts preceding Schoepff (those of

Lac6pede, 1788 and Bonnaterre, 1789) referred to the logger-

head equat inr the t na. e T~c tudo carett ,a with the m ore tsuit-

able' name Thstuiqdo caoun:jf, a turtle described as yellow, in

color and therefore definitely not a ridley, Brongersma

(1961) took Park.e's r:ide, citinl- fu:'her dat-. to sh t sl hat

the inframnarginal count of Schoopff's turtle e '. .:.'stcd Cr etta,

that there vwer three mandibill:ar scales in the drawing, not

one, on each s:idr ( inothe.r fairly good Car".tno chlaract r),

and finally that the carapace, although rather wide, was

nevertheless within the known range of variation of Caretta.

We need, therefore, concern ourselves no longer with the

name caretta in our nomenclatorial history of the genus

Lek.idochel s.
The name Testudo viridi*-squa mosa Lacepede (1788) re-

quires some comment. Some of the names used by Lacepede in

this work are not binomial, and so there is some question

as to whether any of the names are available nomenclatorially.

Wermuth (1956: 405) identified Testudo viridi-squamosa with

Le.idochelys ke1rr.i, and restricted the type locality to Bocas
del Toro, Panama an unfortunate choice since this is well

outside the range of kempi. Wermuth based his conclusions

on Lacepede's description of the small, rounded head and

greenish color of Testudo viridi-~suamosf. It was later-

pointed out, correctly, by Brongersma (1961: 25) that the

ridley has a large triangular head and is gray in color, and

that, moreover, almost all of Lacepede's localities are weell

outside the range of k;eRpi. It was concluded that Lacepede

was probably referring to the green turtle, but that certainty

was impossible and the name would best be suppressed. Tejstud

chloronotus Bechstein (1800) is a junior synonym of Testudo

viridi-squemossa Bechstein was apparently unaware that

Lacepede's turtle had already received a specific name.

Testudo m ydas minor Suckow 179, was considered by Wermuth

(1956: 413) to be synonymous with LenRidochelvy knemji. The

type locality was given as Cape Blanco, K.exico. Brongersma

(1961: 26) equates this with Cape Blanco on the Pacific coast

of Costa Rica, an impossible locality for kemji, though pos-

sible for olivacea. In view of the impossibility of making

a certain designation, it seems best to suppress the name.

Chelonia multiscutata Kuhl (1820: 78) is the first name

that can be associated with a species of L pidochelvs with

reasonable assurance. The type, originally said to be in

the collection of a Mr. Kuypers at Groningen, must now be

considered lost; but according to the description, the type

had eight costals on each side, and nine vertebrals. This

scalatien is moderately common for Lenidochelvs olivaJcea, in

which both costals and vertebrals may vary from five to nine

or more. To the best of my knowledge, however, it is un-

known for any other species certainly it would be extremely

rare. The carapace laminae were stated to be slightly im-

bricate, a condition present in hatchling LepDidcochelys.

Moreover the color was described as blackish-brown, a co3or

frequently found in preserved hatchlings of LqeLidoclhelys but

not in those of other species of sea turtle. The action of

Mertens and Wermuth (3.961: 233) who followed Boulenger (1889:

185) in placing Chelonia mul.tiscutata in the synony:r.y of

CaretGta caretta, must surely have been based on ignorance of

the description of the animal on which the name is basic d

Nevertheless it :would be wise to suppress this v-ry appropri-

ate name, since the name olivicpa has becor.'o well established

for the species, and in any case it would be unec.s. ory to

select a ncoi;ype for multi sclt..'

The name olivacea itself, then included in the genus

Chelonia, was proposed by Eschscholtz (1829: 3) for two

specimens a 7-inch juvenile and a 281-inch adult male -

from Manila Bay, Philippines. The smaller individual was

figured, the description was detailed, and there is no doubt

that the description indeed refers to the Pacific ridley.

In a footnote, Eschscholtz% mentioned that olivacea differs

from multiscutatan in having fewer carapace laminae, 19 to 21

as opposed to 25 (not counting marginals; the variation

between the two specimens failed to suggest to him that the

number of laminate in the carapace was a highly polymorphic

character in this species, and that the full range of variation

in this number actually included that found in the type of

multiscutata0o Eschscholtz was, however, correct in differ-

entiating olivcea from Chelonia cephalo (i.e. Caretta caretta)

on the basis of the greater number of carapace laminae. The

name Cheloni, dussumierii was substituted for Chelonia oliv-

acca by Dumeril and Bibron (135: 557), and must be consid-

ered a junior synonym of it.
The generic name Lpdochels was first proposed by

Fit-zinger (184l3: 30) to receive Eschscholtz' Chelonin oliv-:

aceao The combination Leoidochelvs olivacea thus first

appeared in 1.43, and was used again by Girard (1858: :.'75),

but did not gain general acceptance for about a century.

Gray (184/:: 53) mentions Caou na ruppellii as a relative

of the loggerhead, Caouangaa caretta, and the olive caouanne,

C:ou.ana olJvyac~ ao This name was listed by Wermuth and Mcertens

(1961: 240) as a nomen nudum and a synonym of Lepidochelvs

olivacea. It is however not quite a noren, nudumr the de-

scription refers to the very large head (i.e. not a green

turtle or a hawksbill), to the ridges on the upper jaw (i.e.

Lepidochelys rather than Carctt,^), while the locality is
given, with a query, as India, which would suggest L.olivacea

and not L.kempi. On the same page the name Chelonia subcari-

nata, first used by RUppell in a manuscript now in the Frank-

furt Museum, is listed as a synonym of Caouana olivacea. This

is a true nomen nudum, having never been published with a

description, The name Chelonia subcarinata again appears in

Gray (1873: 1:05) as a fossil from the London Clay, apparently

allied to Caretta.

The names Chelonia Lolvasnis and Cheloni. dubia are also

nomina nuda. However the types of both are still in the

British Museum (B.M1. reg. nos. 63,32,4.119 and 63.12,L.122),

and both are clearly recognizable as hatchlings of IeCidochelvs

olivacea. Both were included in a collection of reptiles from

Java sent in by Bleeker the former of the two narn:e was in-

cluded, though without description, in his list, of species

collected (Bleeker, 1857: 239).

Cenhalocheovs oceanic was a name proposed by Gray (l 73c :

91; 1873b: 408) for a preserved specimen consisting of hord,

neck and forelimbs only, purchased from a de.lcr who said that

it; probably came from the pacificc coast of 7c..o ico Accordin

to Gray, "Ithe size of the brek ;nd tio shied; of th,- head

lcave no doubt that it is distinct.:f The sr.ncirren is -now in

the British Miuseum, and is clearly a perfectly normal adult
Lepogidocl]ys olijvacea, for which the origin 'Pacific coast

of Mexico7 seems very reasonable, Cenhhlochelv, poeanica

is thus a junior synonym of Lenidoche]vs olivacea: however

if one were to designate the East Pacific population of

olivacea as subspecifically distinct (which I do not think

is justified for reasons given below), the subspecific name

oceanifca will become available.

The Atlantic ridley, Lepidochelvs kemoi, was first de-

scribed by Garman (180: 1.23), who included the species in

the loggerhead genus, Thalassochelvs, He based the species

on two adult specimens received from Richard M.Kemp of

Florida, and noted that the species was commonly considered

a cross between the green and the loggerhead, being known

as ,astard' in the Gulf of Mexico, GCaran wrote that keno.

was "distinguished from T.caouana [iLeo the loggerhead] by

the short, round body, low humps, itarginal plates, nar owness

of head across occiput, and swollen jaws; from T .oiv, ace.e

[i.e. the Pacific ridley] by shape of hesd, swollen jaws,

and plates of the carapace. The compression of the anterior

portion of the head of T.olivaces at once separates the spe-

cies." In fact, the skulls of kemei and olivJcef, a re not as

different as Garruan suggests, except perhaps in very old in-

dividuolsE,. Gari-an tent.atbively suggested that k I-eri was of

more than specific distinctness, and placed it in the new.

subgenus Col.nochelj of the genus Tbh- ...sochlvys. However, he

also wrote '"this wiill give to the species tne name Ci o D, 9 .c.0cs

kem.rpi, KempPs Gulf Turtle," i.e. as if Colpochelvs were of

full generic status, and this was repeated by some later

writers (e.g. Schmidt and Dunn, 1917: 50). Nevertheless

Garman showed uncanny insight in this paper, written almost

90 years ago, in recognizing (though not always by their

present names) all the species of sea turtle recognized today,

including Chelonia depressa, which has only recently been

re-instated (Williams, Grandison and Carr, 1967: 1). However,

the wisdom of this paper was largely ignored for many years

afterwards; Chelon.ia dejrssa was relegated to the synonymy

of' Chelonia mvjdas. and the Pacific ridley and the Pacific

loggerhead were amalgamated as one species. The latter cus-

tom destroyed the validity of much thoughtful work by Gadow

(1899: 207), who studied orthogenetic variation in the shells

of Chelonia, and attempted to rationalize the apparent reduc-

tion in lamina count as concurrent with the maturing of

loggerheads (actually, his hatchlings were the multilaminate

olivacea and his adults the pentecostal carietta).

Boulenger's 'Catalogue of the Chelonianst (1889: 186)

acknowledged the existence of an Atlantic ridley, Thalasso-

chelvs kemI., but otherwise included nli loggerheads and

ridleys in the world--wide species .Tiia,.ssochevs caretta. on

the following grounds: The cnornous amount of variation in

the lar ge, series of Logg erhead Turtles in the .Museum leave no

alternative but to further multiply the- number of species, or

to acdimni only one, I 'have adopted the I-tter course." This

ignoring of the perfectly good species Lenidoch'cyvs olivaceo

was to set a precedent for a widespread habit which lasted

over forty years. Part of the confusion over the loggerheadsv

of the Indo-Pacific region may have been due to the fact that

the true loggerhead, Caretta caretta (subspecies &igas) is a

distinctly rare animal over most of the Indian and Pacific

Oceans, while the ridley is widespread and, in many places,

common. In addition, Gadow's misconception carried a lot of

influence; he was a famous man, and he had seen a large

number of specimens.

Boulenger's plan of only two genera (both the green turtle

and the hawksbill were included in Chelone, and the loggerhead

and Atlantic ridley in Thalassochelys) was quickly challenged

by Baur (1890: 486), who proposed separate genera for the

green turtle (Chelonia) the hawksbill (CareLa), the loL ),r-

head (TCalassochelvs) and the ridleys (jLeijdocheijs). Baur

was the first to realize the close relationship between oliva-

cea and keIr.pi, and the first to include the latter in the

genus Lenidochelvs. Eaur also recognized tie Pacific green

turtle as distinct from the Atlantic form (although not on

grounds that now seem valid), correctly allocated Che.onia

nulti-scut ata to L ciochel!s olivacea (though he considered

it an abnormal specimen, being unaware of the wide range in

the nu.nber of costal laminae found normnll- in L.i.jcca) and

realized that Chelcnija deropessa was a frequently overlooked

form meriting further study.

Philippi (1887, 1.899) published two new names,. Thal.- so-

chelvs contra ver-a and Thalassochelyv ta'-:Cfina, for t':o

Chilean sea turtles. Yaiez (1951) has since shown that both

were based on more or less normal specimens of Lenidochelys

olivacea, but Wermuth and Mertens (1961: 233) erroneously

include both forms in the synonymy of Caretta caretta ig.s.

Hay (1908a) described a sea turtle from La Ventosa,

Tehuantopec, Mexico, as representing a new species, Caretta

remivaga. However, he limited his comparative material to

Caretta caretta and to ?Carettal kempi, and did not realize

that his specimen was indistinguishable from Lenidochelvs

olivacea. (Even if the East Pacific ridley were to be con-

sidered subspccifically distinct from typical L.olivacea
from the Philippines, the name olivacea (Gray, 1873) would

have to be used, not remivnao.)

Schmidt (1953: 107) perpetuated both an unfortunate error

and an unfortunate opinion in classifying the Atlantic and

East Pacific ridleys as Leidochelvs ol.ivacea kemrni and L _j-

idochelys olivacea remivap:a respectively.

Siebenrock (1909) considered Caretta. remivapYa to be

to be merely a half-grown specimen of Carj(tt, caret.ta. He

admitted the validity of the Atlantic ridley (though only

as a species of Caretta), but his confusion of Indo-Pacific

Lepidochelys olivacea_ and Caretta a serious error which

persisted until Deraniyag;.la (1933) show.'ed the two to be

distinct. Stejn .r (1907), showing somoFTh;t. more insight,

called the Atlantic logrcnrhead C .'e.. c';,, overlooked

the presence of this rare species in j. Asitic anatcr-s, and

called a-ll. Pacific t Crettid;te Carctt., ojlivacea.

Much name-.juggling between Caretta, L enidochelvss and

Colpochelvs occurred in the first half of the twentieth

century. Ditmars (1910), although acknowledging the distinct--

ness of caretta, olivacea and kempi, placed all in the same

genus (Caretta), Pope (1939) and Deraniyagala (1939) used

all three generic names. Deraniyagala. although having un-

rivalled knowledge of LTeidochelvs o~L ivacea, had little first-

hand contact with ken oi. Thus he distinguished Colpocheljvs

and Le~idochelvs on the following grounds:

1. Dorsally olive-green. Each inframarginal with a
pore; costal scutes more than five pairs; limbs two or one

clawed . * . . Leictochol clivacea

2. Dorsally dark grey. Inframarginals poreless;

costal scutes in five pairs; lirnbs three clawed . .

ColcocI e lys kemper
The first of these characters does not hold; Derani-

yagala himself admits (1939a: 144) that h.lf-grown olivacea

are dark gray in color, while we now know that adult ko .li

are olive-green. The second character is also invalid: kempi

has a pore in eachf infrainarginal, and it is hard to see how

this could have been overlooked. The third character is

valid, though opivancen does occasionally have five pairs of

costals, as Dera niyagrla Isel' realized (1939a; 1.37)o The

fourth i s invalid; both species h-ave- o:e large claw: and one

tiny, almost indistinguishable from an overlappin.-g scale

margin, on each flipper,

Later Doraniyagala (1943) swung to -the other e:treC; and

lumped the two forms together as subspecies of Lepidochelvs

olivacea; the only difference between the two forms quoted

was coastal scutes usually in five pairs: kemni; costal

scutes usually in more than five pairs: olivacea. This class-

ification is followed even today by many writers, though for

the most part by ones whose knowledge of the genus is second-

hand or incomplete.

Carr (1942) was the first writer since Baur to recog-

nize the close relationship between kempi and olivacea, and

their distinctness from the loggerhead. In this and a later

work (Carr, 1952) he lists the following characters separating

the genera Caretta and ,Le.idocheolys:


Frontal bones entering orbit

M.axillary bones in contact

Descending process of prefrontals
in contact with palatines

Choanal fontanelles far forward,

External orbital openings visible
in ventral aspect

Posterior edge of nasal septum

Expanded ectopterygoid processes

InfranCarginal count

Inframr. ginal pores

Enc d mandibular scales

, evidcche v-,





typically 3

typically 4


In addition Caretta can be distinguished from L,eji.o-

chelvs in terms of shape, color, adult size, the pattern of

neural bones, frequency of nesting both within a season and

from year to year, and numerous features of nesting behavior.

The differences between the two species of LePidochelvs

are discussed at length in later sections of this paper.



Lepidochelys kemp is known to form large breeding

aggregations on the coast of Tamaulipas, Mexico, usually

within a few miles of Rancho Nuevo, in the Municipio de

Aldama (Carr, 1963: 298; Hildebrand, 1963: Chavez et al.

1967). Carr and Caldwell (1958: 246) report that fisher-

men throughout the central part of the coast of Veracruz,

from Tuxpan to Alvarado, are familiar with kempi as a nest-

ing animal. Carr (1961a:10) extended the known nesting

range a little further south, to M-ontepio, Veracruz. Occas-

ional nesting has also been recorded on the coast of Padre

Island, Texas (Carr, 1961a: 11; Wer].er, 1951: 41).

The statement by Deraniyagala (1957: 110) that ridleys

were known to noYt in tho vicinity cf -"ia"ni. Florida, is

apparently entirely without substance. In addition, his

suspicion (Deraniygala, 19.'3: 89) that the snocies breeds

in the Azores is without real foundation, being based solely

on the discovery of a sing].o 10 cm juvenile in that area.

De Sola and Abrams (1933, 12) mentioned finding egms in two-

foot-long Georgin specimen., but their statement that these

individuals wei ';cd only "about eight pounds' is co:fusin7.

to say the least, as Carr (1952: 102) points out. Caldwell

et al. (1959) :',ke no fn 'iJon, cf ', ny ridleys nesting with

the nuoerou, lo, -;hernids in the Gc ia coas-tal s ..nds

Kem.p (quoted by Garman, 1880: 121) reports that "wle know they

come out on the beach to lay (in the Florida Keys) in the

months of December, January and February, but cannot tell

how often or how many eggs.' Carr (1956: 18) queried this,
as he could find no turtle fisherman in the Keys who had

ever seen one nesting at any time. It becomes doubly un-

likely when we consider that nesting in Mexico is now known

to take place in May and June only.

Atlantic ridleys have been caught at sea throughout the

Gulf of M4exico, Chavez (1968) found that females tagged at

Fancho Iuevo while nesting were recovered both to the north

and to the south; no less than eight were found near Ciudad

del Carmen, Campcche, while others ,were found at San Pedro,

Campeche; Dos Bocas, Tabasco; and Alvarado, Veracruz, To the

north, tagged specimens were recovered from: 36 kn SE of

Vermilion Bay, Louisiana; 37 km E of Cameron, Louisiana;

11 km from Empire, Louisiana; 10 km and 15 km E of Les Isles

Dernieres, Louisiana; near Freeport, Texas; near Brownsville,

Texas; and 41 km SE of Freeport, Texas. One specimen was

found almost due east of the point of tagging, being caught

between the Dry Tor -ugas and the Marquesas Keys.

Smith and Taylor (1950: 15) record the species from Isla

de Iiujeres, Quintana Roo, Mexico.

The ridley is found in sufficient numbers off the Gulf

coast of Florida in the summer months (together with the

groen turtle) to support a small turtle fishing industry in

the region between the youths of the Withlacoochee and Crystal

Rivers (Carr and Caldwell, 1956), The species is found along
the Gulf Coast right down to the Keys (records from Key West
and Sand Key, Carr, 1942: 10). Ridlyvs from the Florida Gulf

Coast are for the most part a few inches short of mature di-
mensions, but according to Carr and Caldwell (1958: 246),

occasional larger individuals are found, usually further out

to sea than the area patrolled by the turtle fleets.

Ridleys are unknown from the Bahamas. On the Atlantic
coast of the United States they are very scarce south of IM.el-

bourne, Florida, though there is a sight record from Salerno,

Martin County (Carr, 1912). The species has been recorded

from Georgia, Southern Coast (De Sola and Abrams, 1933);

North (U.S.NatMus. 52015. 029244; Beaufort: USNMI

55735 and Coker, 1906); Virginia, Northumberland County,
near Reedville (USNM 86811h); New Jersey, Atlantic City

(Hay, 1908a); New York, Lower New York Bay (De Sola, 1931).
There is no diminution in records as we proceed north. Dodge

(1944) lists seven records of ridleys from the coast of Massa-
chusetts, while Carr (1957) mentions an astonishing observation,

passed on to him by William Schevill of the Woods Hole Oceano-

graphic Institute, of several dozen yearling ridleys out of
a '"whole fleet of such turtles" being stranded on .oods Hole

beaches vhile traveling out of Buzzards Day into Vineyard

Sound. There is also a record from the coast of Iaine, four

from Nova Scotia, one from Cape Breton Island, and one from

Cape Race, N Tewfoundland (i.eakncy, 1955, 1965 ) There is also

one record firoum ermnudn (Kowbray and Caldwell, 1953), one froms

the Azores (Deraniyagala, 1939b), two from Ireland (Galway
Bay and Miltown 1Malbay, County Clare, Deraniyagal_, 193oa
and b); one from the Scilly Islands (Brit. Mus. Nat. Hist.

1925.12.23.1), one from the Channel Islands (Beaumont,
Jersey, RM 1950.1.2,70), and several from Great Britain

(Cape Wrath; Scotland; Ayr, Scotland; Tremadoc Bay, Wales;

Milford Haven, Wales; Bristol Channel; Portreath, Cornwall;

Pagham Beach, Sussex; Taylor,1963). There is one record from

the Netherlands, from Scharendijke, Schouwen Island (Brong-.

ersma, 1961: 11, 33), and one from St. Joan de Luz on the

Basque coast of France (Brongersma, 196: 439). There is also

a record from Malta (Despot.t, 1930; identification corrected

by Carr, 1957: 4h8)

There is no good record for the Caribbean; Dunn's (1918)
record from Jamaica, although repeated by Pope (1939: 288),
is refuted by Lewis (1940: 56), Grant, in Carr (1952: 397),
and Caldwell (1961: 277). The Venezuelan records of Donoso-

Barros (1964) and Flores (1966) clearly refer to Lepjidochelvs

olivace-, and will be considered later. The record of a ridley

from Gibara, Cuba (Aguayo, 1953) is" thought by Carr (1957:

46), who examined the specimen, to refer to an olivaceca. not
a kempi. Lover:ige and Williams (1957: 496) refer to "inter-

grades" between kicm-i and olivaco.e. from the Cameroons. These
were merely hatcnlings of L.oliv-cery ,sone of which had five

costals on each side of the carapa-ce (the coastal count char-
acter,, the only way. of di:stinuishin the juveniles
of, here cle'.rly breaks cdoun0 ilovevr adults

may be separated by a number of characters, quite apart fron

the costal count).

It is very interesting that only adult ridleys have been

found in the Gulf of Mexico, with the exception of near-

adults on the Gulf coast of Florida, while all extra-Gulf

(i.e. true Atlantic) records are based on juveniles, typically

20-30 cm in carapace length. The conclusion is inescapable

that hatchling ridleys, after reaching the sea from their natal

beach in Taraulipas, Mexico, swim, or more likely are passively

conveyed by the current, through the Straits of Florida and

up the Atlantic coast of- the United States, where they remain

until they approach or reach maturity. Those individuals

which arc swept right out into the Atlantic and across to Eur-

ope are usually picked up in a dead or feeble condition, and

probably all are lost forever from the breeding population,

but those which remain in A'.&lantic waters of the United

States, even as far north as Massachusetts, are usually quite

healthy when caught and may well grow normally and migrate

back southwa.rd into the Gulf of Mexico when they approach

maturity. Once reproductive cycles are established, the

mature turtles fan out throughout the Gulf between breed.:-

seasons to avoid overcrowding of feeding grounds, but by this

time they are sufficiently strong swimrners to avoid passive

driftjing out of the Gulf with the Gulf Stream.

Lj, i, i.lV l. oli_,race:a occurs in the Atl-nt:i.c Pacific

and Indian Occtn.. Eart I/tlantic recordoU are as fol. lowKs:

Murt.'.nia:~ Port Et:Liennle; Scen' .,l: Hann, ; M pr.. Jo.l and

Goree Island; Liberia: north of Point Marshall; Ghana:

Tensa; Ivory Coast: Tabou; Cameroons: Victoria; Gabon;

Congo: Banana and IMoanda (Carr, 1957: 49). Breeding is known

to occur at least from Senegal to the Congo.

West Atlantic records: nesting is known to occur in

Guyana (Shell Beach and Dauntless Point, Leguan Island),

and in Surinam at Bigi Santi, Eilanti and Babboon Santi

(Pritchard, 1969). Nesting is very probable on the western

shores of French Guiana, but has not yet been demonstrated;

there is an adult shell from French Guiana in the Paris M'us-

eum (-i''IMHN 03.226), but there is no information as to whether

the turtle was caught on a nesting beach or at sea (Brongersma.

1968: 441),

Non-breeding records to the east and west of the breeding

range are as follows: Cuba: Gibara (Aguayo, 1953, identifi-

cation corrected by Carr, 1957: 46); Puerto Rico (Erdman and
Caldwell, in press); Venezuela: Cumana (Donoso-Barros, 1964);

Trinidad (fishermen's reports mentioned by Carr, 1956, later

confirmed by a head and shell sent to Carr by F.Assam).

Returns of tags from females found nesting at Eilanti, Sui.rinam

in the course of the present study give the following addi-

tional localities: Brazil: Natal; open sea near 50'i 4.5;

bet ;een Amazo-n and Oyapcquje Rivers (two records); Venezuela:

15 miles N of Punta Bari; Carupano (where the tagged individ-

ual was one of a large group of turtles seen) ,nh ther the records infringe the generalization that ricdlys

are abricnt from the Caribbean depends on whether we consider

Trinidad or Isla de Margarita to represent the eastern limit

of the Caribbean on the South American coast.

East Pacific records: the northernmost record on the

American Pacific coast is that of an adult specimen from

Oregon (Carr, pers. commr), There is also a record from

northern California (Houck and Joseph, 1958: 219), and from

central California (Mionterey; G.Victor Morejohn, in letter

to Carr, 1/31/68). Carr (196la and b) gives definite records

from the middle and southern waters of the Gulf of California,

while a single skeleton from San Felipe, in the northern Gulf,

is in the Los Angeles County Museum (Caldwell, 1962: 23).

There are numerous records for the southern tip of Baja

California, both near Cabo San Lucas itself and also from the

southern part of the Pacific coast, at least up to Isla

Margarita, and the extreme southern waters of the Gulf (Anon:

1967a and b). The northernmost nesting records for this

species are from the vicinity of Topolobampo ("-nrquez, in press),

and nesting probably occurs more or less abundantly along the

entire Pacific coast of Me]xico south of iMazatlan. Specific

records arc a follows Dahhii de Bandcras; Nayarit/Jalisco

border (pess, ob:,)); according to I.:ontoya (pers. comm.) nesb-

ing in Jalisco just south of I hia de Banderas so'ctimres takc3

the form of lre .r>: s Colima: iianzanillo and Boca de

Apiza; Iichoacan: Sa-n Juan de Lima and Playa El Ticuis (Anon.,

1966); Guerrroo: San Luis do la Loma; Picdra do Tiacoyunque;

Playa Ec.n: ada and Pie de la Cuesta (pers. obs. and Carr,

1961a: 9); Oaxaca: between Puerto Escondido and Puerto Angel

(Montoya, pers. comm.); the latter is also a known arribada

site. Chiapas: the species is said to be very abundant in

the Mar Muerto (Alvarez del Toro, 1960: 38), but no definite

nesting record is available.

Little definite information is available for nesting in

Guatemala, but people at the port of San Jose told Carr (1961a:

8) that ridley nesting was frequent there. Definite published

records for El Salvador are also scarce, but I am told by Mr.

Al Chable of Guatemala City that nesting occurs extensively

along most of the coast. Extensive nesting also occurs in

Honduras, in the Gulf of Fonseca (Carr, 1948 and pers. obs.),

centered probably around Isla de Ratones. Some nesting also

occurs in Nicaragua. Nesting has been recorded from near

Punta Arenas, Costa Rica (Carr, 1961a: 8), and from the Iicoya

Peninsula, Costa Rica (Caldwell and Casebeer, 1964). Definite

records south of Costa Rica are hard to locate, but nesting

almost certainly occurs in Panama and probably at least as

far south as Guayaquil, Ecuador, At least three ridleys have

been found in Chile, from near Arica, near Copiapo and near

Coquimbo (Carr, pers. cormm.).

Central and West Pacific records: ridleys are apparently
unknown from the mid-Pacific islands, However in Oceania there

are a few records of this species. McCann (1966a and b) makes

the old mistake of confusing Le:T,.idochelv with Crt-: thus

he mentions the gract variaSility in number of costals in New

Zealand "loggerheads," and also says that some of his specimens

were rod-brown (i.e. true Careta) and others were drab olive

(i.e. Lenidochelvs). An illustrated specimen from Flat Point,

Wellington District, Tew Zealand is clearly L.o!ivacea1 (see

MicCann, 1966b: pi. iv fig. 3; pl. v figs. 2 and 5).

Deraniyagala (1939a: 1.62) mentions young ridleys, poss--

ibly hatchlings, from Bundaberg, Queensland, Australia; there

is also a record from Cape York (Boulenger, 18 9: 1i6))

Cochran (1952: 682) figures a sea turtle fror Arnhem Land,

Australia, which, though labeled "green turtle, is clearly

a ridley.

For the Philippines we have the type locality (Manila

Bay): breeding is also known to occur in Ne'.- Britain (Gadow,

1899): Celebes (threc. hatchlings in British Muscum, nos.

71-9-1-48 to 50); Kuching, Sarawak (B.1M. 1966.244!-249);
Talang Islands, Sarawak (Hondrickson, 1958: 495); Kelantan,

Trangganu, Pahang, Tioman Island, Johore, Perk, Langkawi,

and Penang, Malaya (Hendrickson, 1961: 215; Hendrickson and

Alfred, 1.961: 195); Bay of Bengal (Hatchlings in PI, nos.

68-4-3--144 to 14,9); Halmahera; Flores; Bonin IslandsS

southern Japan (Awa, Hondo, Toza and almost every suitable

sandy bench in the area); Stejneger, 1907. Pope (1935: 24)

writes that the ridley "abounds" off Chirna and South Japan.

Nesting has been reported on the coast of South Viet;-n

(Carr, pers. com;T.).

Indian Ocean records: the species ir known ror 11 Krachi;

Seychelles; Ceylon (Dohivala, Karaduva, Mode :m, Airipn,,

gama, C.,lkissa, Bentota, M.orauva, T: -,na, Clc. i l, Tlavilc,

Egoda, Talaimannnar, Uvana, II.'ndc1la, ;Machichchuk.te, lipand.iiC...

Kelpitiya, Udappuva, Chilaw, INegeombo, Natara, Hambantota)

according to Deraniyagala, 1939a. The species is very rare

along the coasts of East Africa and the Arabian Peninsula;

there are no breeding records, and only four recorded cases

of individuals having been caught at sea or found dead on

shore, from the Gulf of Oman; Kassaua, Eritrea; Lindi, Tan-

zania; Socotra (Loveridge and Williams, 1957). The record of

Hughes et al., 1968, for Natal, was based on a misidentified

kyphotic loggerhead. However, the species may possibly nest

on the coast of Somaliland; Travis (1967) describes a turtle

butchery he established on the coast of this country, most

of the turtles used being nesting females. Although this

concern was centered on the green turtle, the smaller shell

in the picture of turtles being butchered opposite page 81

looks very like that of a ridley,

It is curious that, while our knowledge of the Atlantic

ridleys was, until the ltst. few years, based almost entirely

on juvenile specimens, that of the Pacific ridley is based,

apart from one of the types, almost without exception on

adult specimens. Immature Pacific ridlcys see- to be caught

extremely rarely, and the few writers who have figured or

described such specimens (eg. Deraniyagala, 1939a; iar'que%,

in press) have used specimens raised in captivity from the

hat.chling stage The rea-son for this Nay be, in part, that

ridleys reach maturity in a very short time. D.raniyagala

(1939a: 162) raised a spocir.-en in captivity which 2;::ied 42
pounds when 15 mcnt;h old. This sup:ests (though of course


by no means proves) that ridleys might reach maturity, or at

least mature size (about 65 pounds), after only two years. If

their total life span is several decades, only a small fraction

of the total population at any one time would be immature.




Shell Characteristics

The following parameters were analyzed in the three

populations: absolute carapace length of mature females;

relative carapace width; relative length of plastron:

relative width of plastral bridge; absolute carapace length

of hatchlings; relative carapace, width of hatchlings; number

of central laminae; number of costa laminate; number of

marginal laminae; number of intergular lamin're; n-Iumber of

inframarginal laminae; relative vw.idth of i.ddle marginal

laminae of adult females; Ehape of sup acaudal and prosterior

marginal lami.nae; anterior carapace profile.

Absolute CaraIace LeJf h
Figure 1 shows the straight-line carapace length for

203 mature female Lenidochely.'. V from Rancho Nuevo,

Tamaulipas, Mexico (adapted from Chavez et al., 1967: graph

2, page 1 4). Figure 2 shows the same for 2!1d mature female

Loeidochelvs olivacea from Sh;31l 'bo.:ch, Guyana. Bigi Santi,

Surinam and Eilanti Surinam Figure 3 shows the s1me. for 99

mature female LenidocheJlys ol@ivya from .the Pacific coast

of Hondura s Figure 4 shows the carapace length-s of 201

Le.idoch-e1vs ol.vaocea of both sexes (males shaded, females

unshaded) caught at sea off the northern half of the Pacific

Figure 1. Distribution of straight-line
carapace lengths for 203 mature female
Lepidochfil:.3s keri from Rancho huevo,
Tarmaulipas, IJexico (fide Chaves et al,

Figure 2. Distribution of strairht-line
carapace lengths for 241 mature foe .le
Lqpjidochelvs olivacea front Shell Beach,
Guyana, and TBigi Santi and Eilanti, Sur-

Figure 3. Distribution of straight-line
carapace Ionrths for 99 nature female
Lpidochls c i loii from the Pacific
coast of Honduras

Figure 4. Distri bution of cara1 ce 1en-ths
for 201 L..ic~PS o clivac! caught, at
sea off north h--.: ner:. I-;o: ; i.caded areas:
mal es0 unsha ded areas fo ales (fi C
Anon., ]1963i ad b)

59 60 61 62 63 C4 CS


66 67 68 69 71 74


L. olivocea (Cuyon oand S'rinna


62 63 64 65 66 67 68 69 70 71 72 73 74

L. gliac (ondMwuos)


58 59 60 61 62 63 C4 65 66 67 68 69 70 71 74

J. Cplioce (Mexico)

56 57 60 G6 62 63 64 65 C6 67 68 69 70 71 72 73 74 79 78

r l 1 7



coast of Mexico (JIalisco, Colima, Baja California and Sin-

alca); data from Anon., 1966, 1967a and b.

The first conclusion from the histograms in figures
1 to 4 is that the normal length of mature (i.e. nesting)

female Lpridochelvs kerpi is less than that of mature female

L.olivacea from Surinam; the respective modal length classes

are around 63-63.9 and 68-68.9 cm respectively, while e the

minimum lengths are 59 and 62 cm respectively. The maximum

lengths (74 cm) are the same, but the histograms suggest that

74 cm ke,2ni are odd giant individuals, well separated from

the normal range of carapace lengths, while the single 74. cm

olivacea from Surinam is backed up by six 73 cm individuals

and 10 of 72 cm.

The distribution curve for shell lengths of Honduras
olivacea is amazingly similar to that for kenii; again we

have a few giant 74' cm individuals, none of 73 or 72 cm,

while the modal length class is 65-65.9 cm and the minimrnm
54 (excluding one deformed specimen, not plotted, only 52

cm long). The minimum breeding length in Honduras is thus

4 cm less than that in Surinan, while the modal length is

also about 4I cm less.

Because the Pacific '.r::ican turtles on which the fourth

histogram was based were caught at sea, r2l :; ,'ere alo in-

cluded (shaded arcas), and irnmture specimns mry h ave been,

although it -i striking! how large a proportion of the turtles

are at least of mature size. The method of I.:csure'me:t was

not stated, and it may woel have boen bIy S!iTns I o a af3cx:ilC

tape over the curve of the carapace. The discrepancy in
such measurements is around 4 cm, which would bring the 78 cm

individual down to a more reasonable 74 cm. (There are

apparently only six reports of ridleys of carapace length

more than 74 cm in the literature, and at least; five of

these records are suspect. Deranivyagla (1939a: 162) mentions

a 79 cm Ceylon ridley, Hughes et al. (1967: 51) record an

80 cm shell, but, Hughes and i have recently re-examined photos

of the specimen and agree that it is in fact a kyphotic logger-

head. Caldwell et' l. (1969: 23) mention a female Surinam

ridley of carapace length 83 cm and weight 30 kg, and three

males with respective carapace lengths of 100, 94 and 100 cm,

and weights of 35, 34 and 35 kg. Since these weights are

merely average for mature ridleys (Pritchard, 1969: 113), and

since the males (which of course have longer tails) are so

much longer, but little heavier, than the female, I feel sure

that in all four cases overall length~ and not carapace length,

was measured,) Whatever the method of measurement used for

the Mexican specimens, however, the figure does demonstrate

that the males and the females are of closely comparable


Relative Crai)aac, Width

Figure 5 shows a plot of carapace width: against carapace

length for Le j.idochel'ys and LCeideoclsL ..ol vace

Hatchling and adult diT':.:nsions fo(r Lc;.>.Loci -'s kS:1i, all

from Rancho NVuevo, are tiakrn from Chi vez et al,, 1967 12, 13;




cC 'd
bfl *j

'O -
S4- r-4

r 0

oC o
%cH )H
oH rO

C) 0

-- ,C O -
0 040 'd
*t-O >
ra Mi-
o co
r >,a


(nWO) HlaM 33'dVPrt

those of between hatchlings length and 26 cm are taken from

Marquez (in press); those between 26 and 63 cm, all from

the west coast of Florida, are taken from Carr and Caidwell

(1956: 21). All values for adult Lenidochelys olivacea are

based on my own measurements of Surinam specimens; the three

immature olivacea measurements are taken from iMarquez (in


It may be seen that the carapace of kemni is on the

average relatively wider than that of olivacea, and the

overlap is so slight that no statistical test of significance

is necessary. From the graph we may estimate the average

carapace width of mature fenmle kempi as 96.9 percent of the

length, and that of mature female olivacea from Surinam as

88.2 percent of the length.

Figure 6 shows a plot of carapace width against carapace

length for 107 mature female Lepidoc:helvs olivacea from the

Gulf of Fonseca, Honduras. The line on this graph represents

the expected value of width/length for Surinam specimens,

based on the scatter diagram shown in figure 5. It is clear

that this line approximates closely to expected width/len th

values for Honduras ridleys also,

Relative Plastr~a Lenrth

The scatter diagram in figure 7 was constructed by

plotting values of plastral length against carapace length

for 72 mature, female Loeidochelvs olivnceo froi the Gulf of

Fonseca, Honduras. a linear relationship) the best

0) 0)


*H *)-


S0 0

0Ct 0

*O C

*O o

H 0 0
4-) C5

C -i


cO 0

tl tl
*ro 0)

H C-


S *


* .

* |*

;. 0.

e e


( l 'L) 4't \ * * *V t

: \ Q

\ *'


* *S. rs

\ u>-~--------

n~* ^\r

\OW"~O~" .

(i* ~ )10\\^)~~

4-3 E
bO 0

0- ,C

0 Ef

o o
ci ct

0 O'd

0 > r-0
Sr-! 0'
<)0 *r
> 0
fl c-0

S0 C) *-)
0 ,C O
C-I co H

C?.-6 k

- r-1- 0 0
cs cd x-
4C 0 C -
-4 E
4.) ( ',

.O -
ca 0 V)

* CH

-r-i o o *
HO 0 C*


. 0


\ o




00 0




0 s
O *

* 0


o m co -- co '0 '--j- rr c' o G co I__-

0 (n CO I O v+ r N C-) 0) C- I- T,
S10O LO )in tn (n LO I Pn tO o *," 'q
(1'3C) -119 v1 VU.LSV'ld

possible straight line was droa.wn through these points, giving

the formula- expected plastral length = 0.783 X carapace

length. Values of plastral length against carapace length for
36 mature female Lol.ivacea from Surinam were then plotted.

It can be seen from the figure that the smaller Surinam

turtles (carapace length less than 69 cm) distribute themselves

rather closely around the line of expected values for Honduras

specimens; but for those longer than 69 cm, plastral lengths

are in most cases distinctly less than the expected values,

and the regression line for all the Surinam turtles thus has

a shallower slope than that for the Honduras turtles. However,

before interpreting this as a possible average difference

between the populations, we should realize that Surinam turtles

are on the average somewhat longer than those from Honduras,

and that relatively few Honduras specimens over 69 cm in length

are available for comparison. It may well be that the longest

Surinam specimens are not larger in all parameters, but merely

have a disproportionately long carapace not reflected in other

parts of the body (e.g. the plastron); that is, that the

relationship between carapace length and plastral length is

not perfectly linear.

relative Width o` Brid[7e

The relative width of the plastral bridge is expressed in

terms of the Bridge Index, a parameter defined as the shortest

distance across the bridge, between the axillary and inguinal

notches, j.divid.ed b. y half the maximum breadth of the plastron

(Zangcrl, 195": 52). Since most of Zangerl's measurements

were taken on skeletal or fossil material. it is presu-ied

that the width across the bridge refers to the width across

the hyo- and hypoplastral bones at the slight constriction

adjacent to the anterior and posterior plastral lobes, In

living or spirit specimens the anterior limit of measur'ement

is a little difficult to fix, as the numerous axillary lam-

inae tend to obscure the margin of the hyoplastron. Usually,

however, the bones are just visible as a darker area beneath

the laminae, and it was across this darker area that the

measurements below were taken. The width of the plastron

is easy to measure, especially in mature Le pi.docholvs in

which the outer margin of the hyo- and hypoplastra is relatively

smooth and not extensively indented as in many other sea

turtles, particularly in juveniles,

Samples of mature females from three populations of had the following bridge indices:

L.kempi: 0.59; 0.61; 0.62; 0,63: 0.64; 0,65; 0.65; 0.67;

0.69; 0.70; 0.72 (mean = 0.65; n = 11; S.D. = +0.04;

95% confidence interval for mean = 0.65 1 0.0235)
JL.olivacea (Surinam): 0.61; 0.65; 0.65; G.65; 0.67; 0.67;

0.69; 0.69; 0.70 (mean = 0.66; n 9; S.D. -n .IO~o;

951' confidence interval for meon = 0,66 1 0.026()
iL.olivacea (Hlonduras): 0.56; 0.5;; 0.5; 0.64; 0.66; 0, ';

0.72 (mean = 0,63; n = 7; S.D -= 0.23; 95,, confid-

ence intcrval for rncon = 0.633 1 0.1593)

b .0 C-
O 0

*H r- 1
H)c :s

d 0 )

C)1 C
o 0>
C) C)
*H >


O Cd ,c-

S) 0

cD 0 .r-I

)3 'd -
SC 0* *
01'd ,c

Q) C) -1 r-

> 0f 0

C 0- O
> 0

rr CH o'-





r -

0 0


g : tO
C ')

' -Jljl

The mean values, 95 percent confidence intervals for

the mean values, standard deviations and overall ranges for

the three populations are plotted in figure 8. It may be seen

that the confidence intervals for the mean values for each

population overlap, and the differences between the mean

values of our samples are therefore not statistically sig-


Zangerl (1958: 52) gave plastral indices for two speci-

mens of each species of Loidoclhely s the two figures for Jkcm~1

being distinctly lower than the two for olivces,. However

since the specimens of kemni were both immature and the two

olivacea mature, these data are not useful for comparative


Absolute Caranrac e Iength of IHat chlins

Figure 9 shows carapace lengths for 1.214 hatchling LeT--

idochel-vs kemni from Tamaulipas, M.exico; 199 hatchling

L.olivakcea from Surinam; and 94 hatchlino L.olivace q from

Pacific Honduras.

The minimum length of the ker)i hatchlings. is between

38 and 38.9 mm. The maximum is between ,46 and 46.9 um. The

modal length class is 42-!:2.9 mm. Surinam. olivacca nhtch.lihn's

are fractionally larger, the smallest being: 39-39.9 ma and

the largest 47-47.9 mm in length, The modal length class is

43-43.9 mm. The mean values arc 413.15 end 43.3 i3m respectively,

and th iedifernce between these, 0.15 lin, is so siii. 1 relative

to the limits of accuracy in measurement that it cannot be

regarded as signif giant.

Figure 9. Distribution of strai-'t-line
carapace lengths for 124 hatchlings L.-
idochelvs kem i (above); 199 hatchlings
L,.olivracea from Surinan (middle); and
94 L.coivacea fror Pacific Honduras

30 39 40

4! 42 43

L. kcn ri






4 J

37 38 39 4O

4 45 4.6 -47

L. CliurcCo

i( a d.urasC )

__ _, __ J __ 1 ._ ...

41 42



6I .'t


c 20


Hatchling ridleys from Honduras, however, are substan-

tially smaller, on the average, than those from Surinam, the

smallest being 37-37.9 mm in length and the largest 42-42.9 imm.

The modal length class is 40-40.9 mm; the largest Honduras

hatchlings are thus smaller than those in the modal length

class for Surinam hatchlings.

Relative Carran ace Width of Hatchlinos

Figure 10 is a plot of carapace width against carapace

length for series of hatchling Lenidochelvs olivacea from

Honduras and Surinam. The difference in average dimensions

between the two groups is striking and indeed there is rather

little overlap. The average width/length ratio for the

Honduras turtles is 0.813, while that for Surinam turtles is

0.833. This however must not be interpreted as a simple

indication that Surinam ridlevs are not only larger but are

relatively wider on average. They may simply have hatched

at a slightly more mature stage, which means they would not

only be larger but would also have commenced the lateral ex-

pansion typical of half-grown and mature ridleys. Unfortun-

ately no I!onduras ridleys in the 43-47 mm bracket (i.e. a

week or two old) were available for comparison with Surinam

hatchlings in this size range

,' I.' Counts t .n. -" Disciussi.(n

twith th obvicus exception of the ,cuteless Trionychidae

and D[:rrinochol ide, In>.in variation h's been found in all




0 Q0 (1


r- 0 P3
)0 r-

o (DC)

CI. 0

U :I0
o- H F

0 b0 S

*LO51 p
Ct bO
p.ii E(

A b-O


0 0


0 CO 0

0 0

o g o 80

0 oo o
o 00 0


00oo oo 00

0 0 0 00

0 0 00
o o 0o Oco ~O oo

0 *
o o


0 0


a *
c so

* *

o o *0
0 0 *


- a0) C' (a C Oi
T f ( ri rQoI)


N 0
141 141








* *

J m & --


turtle species for which series of reasonable size are

available. Nevertheless, omitting sea turtles from discussion

for the moment, it is possible in all cases to designate a

certain lamina condition as the normm, deviations from which

are not only relatively few in number but also are frequently

asymmetrical and clearly abnormal. The commonest configuration

for the majority of turtle species is as follows: the median

dorsal series is composed of five laminae, the centrals or

vertebrals, which are flanked by two series each comprising

four costals or la.terals, while the rim of the caranace con-

sists of an anterior nuchal or nroneural, two lateral series

of eleven marurinals, and posterior paired elements, the sunra-

caudals. The plastron consists of six pairs of laminate, called

(from front to back) the o ilars; h~umerals, pectorJls, abdom-

ina3s. femorals and annals, each being named after the organ

or structure lying beneath. The plastron also contains a

small lamina in each axiliary notch, the axillarv, and a corres-

ponding lamina in each inguinal notch, the inguinal.

In certain turtle species or groups of species the normal

condition is slightly different from this, Thus in the two

species of Hydromeduso_, the nuchal is expanded laterally and

displaced posteriorly, so that the anterior marginals are in

contact in front of it; while in several groups (c'og mwny

tortoises, genus Geochelone: certain sicidencks, family Pelo-

medusidac) the nuchal is absent. The central series norr.m2lly

numbers six in one gonus LoocevheG), an ex:tr ~a sall lamina

being intercalated betlw.ecn the honiolog of the fourth and

fifth centrals of other species a not unconmmon variation

among many species for which the norm is five centrals. The

marginal series is reduced to ten in mud (Kinostern-

idae), and increased to twelve in one tortoise species
(HJomonus bouleneri). In many tortoise species the supracau-

dais are fused into a single lamina. A single gular only is

present in Kinostern.., n, and both gulars and hum:erals are

either absent or vestigial in Staurotypun ind Claudius. The

gulars are partially or completely separated by a median inter-

gular in side-.neck turtles (Pleurodira). The abdominals are

widely separated in the Chelydridae. The axillary .laminae

are missing in certain forms (e.g. Kinosrnteon abaxillare)

while both a-xillaries and inguinals are missing in Cyclemyrs,

Cuor, Torrpeie., Cgeoevrda snenrleri and many Pleurodires.

In other forms (,Der.aitemrvs, Platvsternum) one or more infra--

miar-inals are intercalated between the axillary and inguinal

on each side, separating the main plastral laminae from the

marginals. Iacrociemm!:'s has a row- of several sunramrarninals

between the lateral marginal and the costals.

T1 r'i.,- Counts o0 -S, T.-rtl ,s

Tho four Chcloniid sea turtle genera (Chelonia, Eretmo-

chols, C:ret',ta and )Len"!ido_1cho'ly s) arIso show several deviation:

froi,.th tvye ic;l chelonian lamina condition. Most conspicuous

is a row of three or four inframnrginlrs, which separate the

main ,,Iil el~ ncnts from; the marginal. Also the ax1illari c

ac fra, 'rt.d inito iJUrous lan.inae of dj.ffcrc:!t, sizes, so

that there is no external demarcation of the point at which

the integument of the axillary region becomes the anterior

part of the plastral bridge.

The green turtles (Chelonia) have a triangular inter-

gular lamina, while in the hawksbills (rotmochelv.s) an

internal lamina is also present and the intergular nay be

longitudinally divided. In Caretta and Lopeidochelvs the

intergular may be single, paired or absent, the internal

present or absent.

The Chelonine sea turtles (Chclonia and Eretmochlilys)

shove; no alterations in their typical carapace lamina conform-

ations. Hol.ever in Caretta and Lenidochelvs an extra, small

costal lamina is intercalated on each side in the area between

the nuchal, the first two marginal, the first central and

the first large costal. These extra costals are .flanked by

at least one extra marginal. Finally, a single species

(LJeoidochelvs olivacea) has a marked tendency to subdivide

the four large costal laminae, so that the actual costal count,

in normal individuals, may range from five to about nine on

each side. In fact a count of 5-5 is so rare that it is no

longer possible to speak of this as the normal condition, and

since no other particular coastal configJur.ation is predom.inant,

Lepidocheoiys olivacea mush be considered unique arong tur',les

in having a truly polymorphic carapace. The centra. lamin. e

of this species also show poly!o. rphisim, and may vary in number

from five to cicght or nine

The costal laminae of LolI'ice~ are clearly divisible

into 'whole laminae' and half laminaet, the whole laminate

being homologous with the five costals of L.kemnj., which,

by contrast, is unusually stable in respect to the numbers

of carapace laminae. Displacement of the homologs of the

seams of L.kemui is usually slight in olivacea, though in

cases of extreme splitting to eight or nine costals, the seams
become displaced to lessen the size df the small first coastal

and the largo last central. Division is typically by an

almost perfectly straight, transverse break, but occasionally

a split, though beginning in this fashion from the marginal,

turns sharply aside before reaching the central series, and

terminates at an intercostal seam. Incomplete division of

laminae also occurs from time to time. In almost every case

division takes place in the rearmost centrals; for example

a 6-6 count is produced by division of the fifth costals on

each side, or an 8-8 count by division of costals 3, 4 and 5.

One exception was a Surinam shell with a 7-6 count in which

only costal four on the right hand side was divided. Division

of central laminae is concentrated in the region of the homo-

logs of the third and fourth centrals of L.keni., The last

central, which is twice or three tirnis the width of any of

the others, is rarely divided, nor is the short first central,

Possible Si gnificance of Itilamination

It seems impossible to rationalize the multilai-iinote

condition in terms of function: the extra seams 1 in no \:'y

alter the hydrody:namic form of the aninil, and the IDRin::e

are so thin that their precise arrangement could have no

effect on the overall strength of the carapace. Nor could

it cause any desirable alteration in the overall appearance

of the animal, as one usually has to look very closely at

adult ridleys to see the lamina boundaries at all. We can

also presumably rule out atavism; as far as is known, even

the earliest (iMesozoic) turtles had the typical pattern of

five centrals and four costals, and apart from a few relat-

ively recent Carettine fossils with five pairs of costals,

and the single genus C].emnvdoosis with only two pairs, no

fossil species is known with other than four costals in the

modal condition.

Carr (1957) suggests that the number of carapace laminee

in Lenidochelvs olivacea may be geographically correlated,

Accordingly series of this species from the South Atlantic

(Guyana and Surinam) and from the East Pacific (Honduras and

Guerrero, Mexico) were studied and their carapace laminae

counted. All available counts for Leoidcc.hel]vs kemnii were

also obtained for comparative purposes. The following results

were obtained.

Central LamrinLa Cou.nts

The numbers of central larina. e of 154 pmture female-

L.enidpchelv.e! Jkem)i found on the bach at ancho Uevo,

Mexico, by Chaves et al. 11967:. 16) wre distributed as follows:

in 135 cases there were five centrals, in 15 therc wer six,

and in four they were uncountable rd'ic to damage to the central

part of the carapace. To these we may add the counts for 59

hatchlings from several clutches from the same beach, sent

to the writer by Ren. Marques. Of these, 47 had five centrals,

11 had six and one had eight. The somewhat greater frequency

of six centrals in the hatchling series, as well as the presence

of one specimen with eight, is probably attributable to slightly

abnormal conditions in the artificial nests from which the

turtles were hatched. Combining the two series (neither of

which constitutes a true random sample of the species) we get:

182 turtles with five centrals; 26 with six; and one with

eight. In all but two of the eleven cases of six centrals I

examined personally, the extra central was a small square

element intercalated between the fourth and fifth c.-nt~rals,

similar to that found in normal specimens of Notochiol.s pat-

nota. In one of the other cases the extra lamina was an

asymmetrical triangle between the somewhat distorted third

and fourth centrals, and in the other the large last central

was fragmented into four two median clements classified as

centrals and two lateral elements which were considered extra


A series of 474 J,eidoche.lys oljvac'a from Surinam, comi-

posed in approximately equal part of adult fei and randc

hatchlings from many clutches, gave the following central

lamina counts: 72 individuals had five ccntralr, 200 had six,

185 had seven, 16 had eight and one had nine.

A series of 102 adult female olivacea( frcp HonduraT's a'c

the following counts: 63 had five contr!~ls, 37 had six and

two had seven. 126 hatchlings from Honduras (from several
clutches) gave the following counts: 69 had five centrals,

39 had six, 17 had seven and one had eight.

For Surinam olivacea there is a clear Poisson-type dis-

tribution. A fair number of individuals have the minimum

number of five centrals, but the typical counts are six and

seven. Eight is rare, and more than eight almost unknown.

The mean number of centrals is 6.31.

Combining the two Honduras series, we get the following

totals: 132 had five centrals, 76 had six, 19 had seven and

one had eight, This too gives a clear Poisson distribution,

The minimum value is still five, but only half as many, not

three times as many, have six, while seven is definitely rare,

and m6re than seven almost unknown. The mean number of centrals

is 5.51, O.8 fewer than in Surinam. Since our samples number

in the hundreds, while standard deviations are less than 1.5,

standard errors of the means are sc small that there is no

need for further confirmation of the significance of this


Deraniyagala (1939a: 137) supplies lamina counts for a

series of 377 L.o.ivace from Ceylon0. Whether these were
hatchlings, adults or both was not stated. Of these, one had

only four central, 67 had five, 3.71 had six, 127 had seven,

ten had eight and two had nine. This distribution is very
similar to that for Surinam olivacea. The m-an number of

centrals is 6.?, only- 0,07 different fromi the Surinam mean.

No large cisries of 1,e pidoch1l0vs from other areas a.e avail--

able in museums; nor have extensive tables of carapace lamina

counts been published. However, a few hatchlings specimens

from other areas, now in the British I-useum, were examined.

Central counts are shown in Table 1.



Banana, Congo

Fort M'arshall, Liberia

Tensa, Ghana

Manado, Celebes


Bay of Bengal

Kuching, Sarawak

Central lamina counts

5; 5

7; 5


5; 5; 6

5; 5; 6; 7; 7

5; 5; 5; 6; 7; 7

5; 6; 6; 6; 6; 7

These samples are, of course, much too small for compar-

ative purposes. All we can say is that the only counts re--

presented, 5, 6 and 7, are kno-.:n to be common in specimens

from all three oceans.

The counts for L- k-Cni and for L,.ol'.Ziace- frol-,: Honduras,

Ceylon and Surinam are plotted in histo;ra- form. in f:. .e 11.

A progressive ii cease .n Pvcrage cntra ccoulnt; is app rentc

as we proceed w t lard fr,, the Gulf of iex:ico throu ', the Es

Pacific, Indian a ind Sou~th Atlnti:c Oceans.

~-~~'I~-----~I'-~c~'~- 1--1111--------

Figure 11. Relative frequency of different
central lamina counts for four populations
of Lenidoch el-ys



L. olivncco

4 5 6 7 8 9

L. clivoceo

8 9

L. clivaceo


L. kcmpi

6 7 8 9




















4 5

4 5 6 7 8 9

-- ------------ ~~ --- -- -- I - -- --- i l ________ _ _ ______________ !.I



Costal lamina Counts

The relative stability of the typical 5-5 costal count of
Lepidochelrs kemoi has already been mentioned. Carr and Cald-

well (1956: 23) examined 96 specimens of immature kemoi from
the west coast of Florida. Of these only one deviant from a

5-5 costal count was encountered, an individual with a 6-5
count (in all cases the number of left costals is given first).

Chavez et al. (1967: 16) counted costal laminae in 154 adult

female ker !ip from the nesting beach at Rancho Nuevo; 5-5: 147;

5-6: 4; 6-5: 1; 6-6: 1; and one was uncountable because of
indistinguishable seams. I counted the ccstals of 59 hatchlings

from several clutches from Rancho Nuovo; 5-5: 54; 5-6: 1; 6-6: 4.

The 6 count in the 5-6 individual and in two of the 6-6 indi-

vilials were caused by breaks across the lateral extensions of

the large posterior central. In one of the other 6- 6 counts

the central series was fragmented into eight asymmetrical

elements, and in the other the fifth costal on each side was


Combining these three series, we get: 296 5-5 counts; five

5-6's; two 6-5's and five 6-6ts (n 308).
A series of 743 Lenidochels .olivace, from Surinsri, was

exawinned and their costal laminae counted (approximately 400

of these wrce !mature females and the remainder hatchilinr'. from

several clutches,) Counts obtained are shown in Table 2, as

are counts for 378 L.olivsacea of unspecified age and sex from,
Ceylon (fide Deraniyagala, 1939a: 137) and for 102 mature ft. '7i.s

from the Gulf of Fonseca, Honduras.



Costal lamina





n = 743

Mean left costals: 6.881

Mean righL t costals: 6.781




n = 378

6..75 4






n = 102



Large series of costal counts for Tcuidoc. o_-_q from other

arc not availablee at present. Hov.evor Carr (1957: 49) tabulated

all available. counts for ridlcys front the ":cct coart of iA.'j.ica

(Maurctania, Seneal, Liberia, Ghana, Ivory CoaC'.,, CocIcroons;

_ ___~ _I_ __~_ --------------------- --

Gabon and Congo), as follows: two 5-5; two 6-5; ten 6-6; one

6-8; seven 7-6; seven 7-7; one 3-7. A series of six from

Kuching, Sarawak included one 6-6; three 6-7; one 7-7; one 8-6.

Six from the Bay of Bengal included four 6-6; one 7-6; one 7-7.

Five from Karachi included one 5-6; two 6-6; one 6-7; one 8-8.
Three from HIanado, Celebes included two 6-6; one 7-7.
The histograms in figure 12 show the frequency of part-

icular costal counts for the four Lenidochel.ys npopnulations of

which large series have been analyzed those of the Gulf of

M!exico, of Pacific Honduras, of Ceylon and of Surinam. A gradual

increase in average costal count is apparent as we move west

starting from the Gulf of Mexico, Thus, the ke oi population

of the Gulf of Mexico shows a massive preponderance of speci-

mens with five costals, with a very small number with six.

In Honduras we fi.n.d he count of six is a little more than

twice as common as five, while few specimens exceed six. In

Ceylon the 5-count is rare, while 6 and 7 are of comparable

frequency, and eight less common but still quite frequent. In

Surinam the 7-count is the cormmonest by a large margin, while

6 and 8 are much fewer and 5 and 9 almost unknown.

Average values of costal and central counts for the four

populations are plotted in figures 13 and 14.

It is evident from the tabulations of costal counts that

individuals of Lolivacea bilaterally asymmetrical in this

respect are nearly as common as symmetrical ones. The 743

Surinnam secirmens include 418 symmetrical individuals and 325

asymmetrical ones: of these 198 have the greater number of

Figure 12. Frequency of left and right
costal laminae for four populations of
L epdocheIv s


3 4 5 6 7 8 9

3 4 5 6 7 8 9

LH. pivoceo

V 7

3 4 5 6 7 8 9


3 4 5 6 7 8 9

L. olicco

~I---~---- j .j

3 4 5 6 7 8 9

3 4 5 6 7


3 4 5 6 7 8 9

8 9

3 4 5 6 7 8 9

L -kI .pi

z 160.
o 120.
u. 840.






a P







1 1

I I I = e






Figure 13. Graph showing average number of
left and right costal laminae for four
populations of Lepidochelys

Figure 314 Graph showing average number of
central laminae for four populations of







Guif of







Gulf of
MNe Ico

Pcn u'




I :

laminae on the left, and 127 on the right. A similar break--

down of the 378 Ceylon specimens gives 217 syr.metrical, 85

with more laminae on the left and 76 with more on the right.

The 102 Honduras specimens include 63 syrnmtrical, 16 with

more larinae on the left and 23 with more on the right. The

respective percentages of symmetrical individuals in each sam-

ple are: 56.1; 57.4; 61.8. Asymmetry nearly always involves

a difference of unity between the left and right coastal counts;

however two of the 36 asymmetrical Honduras individuals showed

a difference of two, and one was asymmetrical by three. The

140 asymmetrical Ceylon individuals included 18 asymmetrical

by two and three asymmetrical by three. The corresponding

fi- -s for the 291L asymmetrical Surinam turtles were: 29

differing by two; none differing by three.

Deraniyagala (1939a: 143) writes that a low average

lamina count is a tendency of certain broods of hatchlings.

That such rmy a. o be a result of different incubation condi-

tions seems borne out by an analysis of three clutches of hatch-

linCg" from Isla de Ratones, Honduras, shown in Table 3, two

of which were rrn't hatch'-d under natural conditions. Breed 1

was hatched in Fl'.oridi aftr the cggs had been bron 't back

by car, while t.he eS'?; oro which Brood 3 was hatched were

taken 20 :nilrs alon- a. bury road before being re-.burird.

The brood 2 eg'", h -'. v-/ ', were transferred i .m:-.n ; ,i

after lay I to a n n ::y hatchery.

The ox(:l 'cr con'ra cho n .ro. '. 1 and 3 coulJ urol1

have be. i prodi uc, ied b,' tthe e, s. Lynn ,~ind U1:'ich



Costal lamina

Brood I

5L- 4

n = 26

Mean loft costa3.s: 4,808

Mean right costals: i4..8,6

Brood 2



n = 116



Brood 3






n = 101



(1950) produced hatchlings C-hryv,-ys and

Celydr' with sil.iilar abnormal iieos by subjectil- tile e g
..... <., qtl~t; e r ,

_~I~YU_____ ~--------~---- --.-----C--ll~-------- -- ----
---- C-l-~~_-_l~--~--------1-1--1--- I-Y----^-~-L---------------

to suboptimal moisture conditions.

In view of the plasticity of the costal lamina character

it might be argued that little importance can be given to

differences in average costal counts for turtles from differ-

ent areas. However the average count for hatchlings from un-

disturbed nests is usually fairly close to the average for

adult females from the same area. Thus, the respective left

and right average costal counts for Brood 2 from Honduras

(which was subjected to no embryonic trauma) were 5.921 and

6.039, which are very close to the average figures for mature

females from that population (5.912 and 5.968). The averages

for a brood of 88 hatchlings from Surinam were 6.659 and 6.6V4,

and for a brood of 118 hatchlings, 7.070 and 6.965 (cf 6.881

and 6.781 for mature females from the same area),

Marginal Counts

The extreme variability of' the central and coastal laminaie

of Lejldochelvs olivacea does not extend to the marginals. A

marginal count of 12-12 is by far the commonest one for all

known populations of both kemni and olivocea; deviations

from 12-12 are probably no more frequent than they are in

Ca.retta. Imarginal count frequencies are given in Table I;.

The supracaudal l r.i:-nne, although serially homolorous

to the t-rue marinals, are not included in rrr :inJ : counts.

Addition of extra iiar-:inan]r;, to give counts of 3.3 and 1.,

takes pcc at the t thanMt ior part of the mar i na.l row, in

contrast to additional centrals and costals, which are alhlost



Marginal lamina


L. kemni



n = 58



n = 395




n = 258

invariably inserted towards the rear. However the rare

specimens with fewer than twelve marginal on one or both sides

usually showed missing seams in the rear marginal region, or

between a posterior marginal and the adjacent supracaudal.

Marginal counts of more than twelve were frequently associated

with high costal counts.

Variation in Other Caraoace Laminae

The nuchal lamina may be split in occasional individuals
from all popul..ations.s Eradication of tLh semn be;;we, the

nuchal and the first central moy also occur, giving the

_ __~_ _I_ __ __
-CIICIII-----------I __ ~ I-

appearance of an absent nuchal. Both typos of variation
are rare and were not studied quantitatively.

No variation in the supracaudal laminate was observed,

apart from the occasional missing seams mentioned above.

variation in Plastral Laminae
The laminae of the plastron are strikingly stable in

nearly all turtle species, including both species of Lepido-

chelvs. No meristic variation wias found in the main plastral

laminae (gulars, humerals, pectorals, abdominal, femorals

and anals) of either species. However the small intergular

lamina may be present, absent or double. Table 5 shows the

frequency of each condition for series of hatchlings of kemin

and of olivacea from Honduras and Surinam.



Population Number of Mean number of inter-
0 1 2 gulars for sample

L.k~iemj 22 23 15 o08

L.,oi vacea 74 31 21 0.58
( HToindur s)

L. olivace 120 66 73 0 2

No clear conclusion can be drawn from these data, since

most of the Honduras specimens were drawn from only two or

three clutches. However it is Dossible that the relatively

frequent absence of the intergular relative to the single

intergular condition in L.olivacea would be substantiated by
larger samples.

Inframarginal Counts and Significance of Inframar"inal Pores

The presence of three inframarginals in the genus

Caretta and of four in Lepidochelys has been considered one

of the better 'key' characters for separating the genera by

many authors. Deraniyagala (1934: 20o-209), in listing

characters to separate Caretta, Lepidcochels, and Col.ochelys,

writes that LTpidochelvs is characterized by four inframar-

ginals on each bridge, Colpochelvs by three or four, and

Caretta by three, without any qualifying phrase on the possi--

bility of exceptions. In later works (Deraniyagala, 1939a:

122; 1943: 87: 1945: 95), in which he includes ColnochcJly

under Lepidochelys, he states categorically in the keys that

Leidocjchelvs has four inframargina3 s on each bridge, and
Caretta three. Pope (1939: 278, 287, 288), Loveridge and

Williams (1957: 492, 495) and Carr (1952: 343) likewise

quote these meristics without qualification. Fraser and

Parker (1953: 36) use the same character, but in each case

add the word 'normally'. Conant (195: 65--69), using the

inframarginal count to difforentiite bet-'en the A.'tJiantic

loggerhead and the Atlantic ridley, mak c no mention that

the count is variable in the loggerhead, but states that the

ridley 'usually' has four enlarged laminate on the bridge.

Brongersma (1967: 9) again uses inframarginal counts in his

key to sea turtle species, but adds in parentheses that

Lepidochelvs kempi rarely has three inframarginals, and in

a footnote that some Caretta caretta have four or five infra-

marginals, but that, when this is so, the number is usually

different on each side.

Authors who have discussed inframarginal counts have us-

ually characterized them either as "enlarged scutes on the

bridge" or have illustrated the plastron of a typical Caretta

or Lejpidochelvs, in which the distinction between the infra-

marginals and the axillaries is admittedly obvious. Never-

theless specimens exist in which one (or more) of the enlarged

axillary laminae is almost the same size as the anteriormost

inframarginal and we need a definition which will make clear

the difference. Most authors seem to have assumed (and Brong-

ersma, 1968: 441, has now quoted) the definition that an infra-

marginal is a lamina on the bridge which is in contact lateral-

ly with one or more marginal laminac and also mesially with

one or more of the main series of pla-tral laminae (nectorals,

abdominals and femorals). This will be the criterion which I

shall use.

The variUability of the infrainargina'l count of At.latic

loggerheads is clear from the counts of 154 hatchlings froS

Cape Ror in, North Carolina, quoted by Pialdwinr and Loftcin, in

Caldwell ct ,-., 1959: 343, shown in Table 6, Counts for 117

Indian Ocean loggerheads from Natal (from Hughes et al., 1967:

36) and for ten Pacific loggerheads from the Solomon Islands
(from Carr, 1952: 394) are also given in the table.




North Carolina Natal Solomon Islands

57 145 7
21 20
14 10 -
56 10 2
1 1
2 -
2 -
1 .


n = 154 n = 187 n = 10

These figures suggest that the 3-3 inframarginal count

is useless for characterizing Atlantic loggerheads, but that

it may be somewhat more reliable in the Indian and Pacific

Oceans. However, full statistical treatment of the data in

Table 6 is not warranted since wo do not know how many clutches

these series were drawn from.

The inframarginal count of Lieiidoch.elv is more sb-able

than that of Caretta. Chaives t al. (196'7) exaind a scrios

of 151: adult female L.kJcmni found on the nostine beach at

Rancho Nuevo; 149 had four inframarginals on each side; one
had 5-4; three had 5-5 and in the remaining specimen the

seams could not be discerned so the count could not be taken.

In a series of 59 hatchlings from the same locality, all had

the 4-4 inframarginal count except one, in which the posterior

inframarginal on the right, being small and triangular and

thus failing to make contact with any of the main plastral

laminae, was not counted. These data, together with the

inframarginal counts for a series of hatchling Lenidochelvs

olivacea from the Pacific coast of Honduras and a series

from Eilanti, Surinam, are shown in Table 7.



Inframarginal LI.kempi L.olivacea L.olivacea
count Honduras) (Surinam)

3-3 2
3-4 5 3
4-3 1 1
4;-4 212 87 91
4"- 5 -4
5-4 1 2
5-5 3 1 -
5-3 1
n 216 n 9" n = 100

Approx:i.m"tely 91 percent. of thl kr-c i- sample, and 09
and 91- percent respectively of the t.:o oliivacea samples had

four inframarginals on each side; however the 4-4 count is

also sufficiently common in Caretta to render this character

of little application for distinguishing the forms. But there

is another inframarginal character which appears to be of 100

percent validity, and that is the presence in both species of

Lepidochelvs of a small pore located near the posterior margin

of each inframarginal lamina, and another on the seam between

the anterior inframarginal and the adjacent axillary lamina.

Almost all authors who have discussed Legjidochelvs have ment-

ioned these pores; however only Carr (1963: 302) has discussed

their possible significance. He writes:

If the exceptionally active nosing of the sand
mentioned earlier is, an it seems, a smelling manoeuvre,
it suggests that the ridley is more dependent on this
sense than any of the other sea turtles. Moreover, the
ridleys are the only sea turtles which have conspicuous
secretary pores at each seam between the inframarginal.
scales. The function of the secretion from these pores
is not known, but the most reasonable assumption is that
it is an olfactory aid to sex or species recognition.
If so, then it may be a signal for the final massing off
the nesting beach, and possibly even a scent beacon that
marks the sand of the shore for arribadas of later years.

It has not been demonstrated that the pores are secrctory

in function, though this does seem their most likely purpose.

There is also no published account of the cytological morph--

ology of the pore structure.

Fig. 15 shows transverse and frontal sections through

the bone underlying an inframarginal pore of a mature female

Le21 ...c.lys ipij. The sun-dried plastron was found separate

from the animal, and no further preservation was carried out.

Figure 15. Transverse and frontal sections
through plastron of mature female L.kemni
in region immediately underlying an infra-
marginal pore




4- --'


I. .<*'

"f i

'- a ---
M* .-.
' -'a

Il-- ^

In may be seen that the plastron in this region is about

12 mm in thickness, the outermost 2-3 mm being hard, compact

bone and the rest spongy and presumably vascular in life.

The sections reveal cavities of capacity about 1 cc beneath

each pore, reaching almost through to the visceral surface

of the plastron, and with a lining about 1 mm thick of compact

bone similar to that on the outer surface. This layer of bone

is perforated in several places, by channels which may have

carried blood vessels in life. The cavities have a thick

inner lining of soft tissue, which was not studied micro-


Pending a proper biochemical and cytological examination

of the cavities underlying the inframarginal pores of

chelys, Carr's suggestion that they are secretary glands

which aid in sex or species recognition seems most likely,

and they may well aid in arribada formation a behavioral

characteristic peculiar to the genus. However, since it is

now known that ridley arribadas at Rancho Huevo, iKeico, do

not always come ashore on the same exact stretch of beach, it

seems unlikely that the sand is permanently marked ,b a sec-

retion frco these pores. However, the secretion may be dis-

charged into the sea by ovi, 'rous females and attendICnt riles

migrating from all parts of the Gulf of ecxico towards Rancho

Suevo, and may aid in the gradual forniatio and also t" stab-

ility of ]..'.:,e flotillas of turtles.

Possible Sjinificmnce of Differences in Shell S.hRe

As we have already demonstrated, the carapace of adult

Lenidochelvs kemi2i is, on the average, wider relative to its

length than that of L.olivacea, as well as being distinctly

lower. I have rationalized elsewhere that the expanded sides

of the carapace of adult kiemji may be a device to compress

the sand over the nesting site more compactly after nesting,

and thus make it less likely that the odor of the new-laid

eggs will attract terrestrial predators such as coyotes. When

watching an Atlantic ridley stamp down the sand over the nest

with vigorous rocking movements of the shell, one is indeed

struck by how well the sides of the shell seem adapted to

this function, and it is certainly important that the eggs

be hidden as thoroughly as possible, considering the number

of coyotes in the area.

The extra width of the shell of kemni reflects itself

in an increased width of the marginal laminae in the mid

part of the series. Figure 16 shows the 7th, 8th and 9th

marginal (or their homologs in cases of abnormalities in

the anterior parts of the marginal series) for mature female

L.kemni from Rancho Huevo, and also for mature female L.oliva-

cea Shell Beach, Guyana. It may be seen that, while the

.th m arginal lamina of koempi is often about as wide as long,

in polivacpa this lamina is usually at least 50 percent longer

than wide,

Other, slight, differences between L. kemni and L, ol.ivac-ea

are present in the region of the last marginals and

Figure 16. (above) mid-marginal laminae of L.kemni from
Rancho iluevo

(below) mid-marginal laminac of L.olivacea
from Shell Beach, Guyana

the supracaudals. Figure 17 shows these lairinne for specimens

of L.kemni (mature females from Rancho Nuovo), and also for

specimens of L.olivacea from Pacific Miexico, Shell Beach,

Guyana, and Eilanti, Surinam. It may be seen that the seams

separating the supracaudals from the posterior central tend

to meet in a straight line in kemni, and at an obtuse angle

in olivacea. Also the posterior corners of the marginal

and supracaudals of keMpi tend to be relatively smooth, so

that the carapace margin is merely slightly wavy, while in

olivacea these laminae have projecting corners, so that the

hind margin of the carapace is frequently somewhat serrated.

The greater elevation of the carapace of olivacea rela-

tive to that of kempi was noticed by Carr (1961: 9), The

actual height of the shell is hard to measure in the field,

and the character wvas therefore studied by means of anterior-

view, ground-level photographs. The carapace profiles shown

in figure 18 corroborate Carrts observation; the carapace of

L.kempi is low and gently rounded, while in most cases that

of clivacea has a strikingly elevated, flat-topped central


The possible significance of the differences in carapace

profile between the two species can, at present, only be guessed

at. Possibly the lateral expansion of the shell of kemni is

caused by a re-direction of grovwh vectors which in olivi-cea

contribute to the progressive deepening of the carapace.

Figure 17. (left) posterior marginal
and supracaudal laminae for I..kempi
from Rancho Nuevo

(right) same for Lolivacea
from Guyana and Surinam

9~1~3 -(S~j~

~ct-r~- ~ti
~tGu~~ ~:_~3~
`(\T~C) j`~2~J;P



.-l 0









S co

$l *H





od O

0 o

, H



Caraauce Osteolorv

No detailed comparative studies of the osteology of the

carapace of Leq.idochelys were undertaken.

The bony carapace of Lecidochelvs consists of a large

proneural, twelve (rarely thirteen) pairs of peripherals,

eight pairs of pleurals, two (or occasionally three) supra-

pygals, the anterior of which has a characteristic crescentic

shape, the posterio-lateral rami enclosing the second supra-

pygal (in adult specimens) and frequently coming in contact

with the penultimate peripherals, and a single pygal.

Great variability is shown by the neural series. The

basic chelonian number of nine neurals has been increased by

transverse (and occasionally longitudinal) division of most

of these elements, so that they now number between eleven

and fifteen. The most frequent type of division occurs when

a curved suture, with the convex side directed anteriorly,

splits an elongated, hexagonal neural bone into a nearly

regular hexagon and a smaller bone, subovate or almost square

in shape.

Figure 19 shows the neural bones of three mature Lepido-

chclys o2livacea from Bahia Banderas, Nayarit, Mexico; of

two mature and one nearly mature L.olivacea front Ceylon (re-

drawn from Deraniyagala, 1939a); and of three L.kcml i (two

mature specimens from Veracruz, Mexico, and one nearly mature

from Cedar Key, Florida).

These series are of course far too small for any statis-

tical conclusions to be drawn' from them. Ieverthelcss it is

interesting to note that frna :-tation of the neural bones