Eastern Illinois UniversityThe Keep

Masters Theses Student Theses & Publications

1984

Observations on Relationships Between LimbMeasurements and Mode of Locomotion inMalaysian TurtlesDuayne NyckelEastern Illinois UniversityThis research is a product of the graduate program in Environmental Biology at Eastern Illinois University.Find out more about the program.

This is brought to you for free and open access by the Student Theses & Publications at The Keep. It has been accepted for inclusion in Masters Thesesby an authorized administrator of The Keep. For more information, please contact tabruns@eiu.edu.

Recommended CitationNyckel, Duayne, "Observations on Relationships Between Limb Measurements and Mode of Locomotion in Malaysian Turtles"(1984). Masters Theses. 2846.https://thekeep.eiu.edu/theses/2846

THESIS REPRODUCTION CERTIFICATE

TO: Graduate Degree Candidates who have written formal theses.

SUBJECT: Permission to reproduce theses.

The University Library is receiving a number of requests from other

institutions asking permission to reproduce dissertations for inclusion in their library holdings. Although no copyright laws are involved, we feel that professional courtesy demands that permission· be obtained from the author before we allow theses to be copied.

Please sign one of the following statements:

Booth Library of Eastern Illinois University has my permission to lend my thesis to a reputable college or university for the purpose of copying it for inclusion in that institution's library or research holdings.

Date Author

I respectfully request Booth Library of Eastern Illinois University not allow my thesis be reproduced because ������-��������

Date Author

m

OBSERVATIONS ON RELATIONSHIPS BETWEEN LIMB MEASUREMENTS

AND MODE OF LOCOMOTION IN MALAYSIAN TURTLES

(TITLE)

BY

DUAYNE NYCKEL

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

M,S. IN ENVIRONMENTAL BIOLOGY

IN THE GRADUATE SCHOOL, EASTERN ILLINOIS UNIVERSITY

CHARLESTON, ILLINOIS

January, 1984 YEAR

I HEREBY RECOMMEND THIS THESIS BE ACCEPTED AS FULFILLING

THIS PART OF THE GRADUATE DEGREE CITED ABOVE

/OjTE ADVISER

s 1�/J-y DATE pOMMITTEE MEMBER

DATE COMMITTEE MEMBER Z�lffr D TE DEPARTMEfllT CHAIRPERSON

2

Table of Contents

Acknowledgements • • • • • • • • • • • • • • • • • • • • • • • • • • • p. 3

Abstract • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • p. 4

Introduction • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • p. 5

Methods and Materials • • • • • • • • • • • • • • • • • • • • • p. 9

Results • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • p. 13

Discussion • • • • • . • • • • • • • • • • • • • . • . • • • • • • • • . p. 18

References Cited. • • • • • • • • • • • • • • • • • • • • • • p. 22

Appendix • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • p. 23

ACKN<:m.EDGEMENTS

I wish to thank the members of my graduate committee:

Dr. Charles Arzeni, Dr. Michael Goodrich, Dr. K�pp Kruse, and

Dr. Edward full.

Special thanks goes to Dr. Moll for the idea to do this

study and especially the many hours he spent on critical

analysis of my writing style. Also, special thanks to Dr.

Kruse who encouraged me to take on a project of this size

and who spent a good deal of time helping me in the statistical

analysis of my data. Because of job and do.mes tic responsibilities,

most of my thesis study was done while away from campus, this

fact made it difficult to meet with my major professors. I

want them to know I appreciate their patience, and consideration

for the several short notice meetings and calls to their homes

on weekends.

3

ABSTRACT

Twelve Genera (spanning three Families) of Malaysian

Turtles were studied to determine whether bone lengths could be

correlated to mode of locomotion. Ulna, humerus, tibia, femur,

coracoid (among others) were measured and results were presented

graphically. Five of the genera were observed singly in a large

aquarium. Data were recorded on the amount of time each spent

in different locomotor modes. Both limb length data and aquarium

observations suggest that turtles can be grouped into three main

locomotor habits (i.e. terrestrial walking, bottomwalking,

and swimming). The terrestrial group includes the genera,

Geochelone, Kinixys, and Gopherus. Bottomwalkers include

Siebenrockiella, Cyclemys, Notochelys, CuGra, Heosemys, and

Orlitia. Highly aquatic turtles include Trionyx, Kachuga,

Callagur, Batagur, and Pelochelys. The combined mean lengths of

the three groups were then compared. Forelimb measurements

(ulna, humerus) were found to be significantly different among

each group, Ulna and humeri were found to be shorter in swimmers

and longer in terrestrials. Hindlimb measurements (tibia, femur)

were found to be indistinguishable for the three groups. These

data point towards divergent evolution of forelimb bones while

hindlimbs seem to have remained evolutionarily conservative,

These results could be easily tested on a similar group of new

world turtles.

4

INTRODUCTION

Turtles are believed by most to have evolved from an archaic

land-dwelling ancestor sometime during the Permian Era. Their

shell and limb structure has remained basically unchanged for:

200 million years. Many important radiations of plant and animal

life occurred during the Mesozoic. However, as far as reptiles

are concerned, only the turtles, alligators and a few snakes

survived the leap into the Cenozoic alive and unchanged. The

facts of what happened during a period lasting 150 million years

(early Triassic to Cretaceous) are not entirely understood. The

shell, already present by Triassic time.s, was of obvious survival

value to turtles. Its weight probably caused these ancestral land

turtles to evolve a largely herbivorous diet (Stahl, 1974).

Consequently, where ver they could find warmth and abundant

vegetation, they could survive. Later, the ability of some to

move into aquatic and semi-aquatic environments, in and around both

fresh and salt water, must have been important for their continuing

survival as selective pressures increased on land. The habitat

5

shift into aquatic environments allowed turtles to better cope with

the weight of the shell. More importantly, the investment in a bony

shell limited the directions and manner of chelonian diversilication

(Zug, 1971). Limb bones and digits must have evolved from those

used solely for walking to types allowing turtles to maneuver

through moist substrate, bottomwalking and eventually swimming.

M:>dern turtles utilize three main locomotor methods: (1) terrestrial

walking, (2) bottomwalking and (3) swimming.

These modes of turtle locomotion are similar to forms of

quadraped movement exhibited by other reptiles. Terrestrial walking

was studied extensively by Zug (1971). He believed the gait of

Gopherus to be a true adaptation to terrestrial locomotion and not

a compromise pattern of an aquatic or semi-aquatic turtle.

Bottomwalking received little more than cursory attention before

Zug, who defined it as propulsion resulting from feet pushing

against the bottom. Locomotionwise, it is nearly identical to

terrestrial walking. True swimming, however, is achieved through

a completely different series of limb movements. Furthermore, the

limbs do not touch the bottom during swimming. Zug made no attempt

to describe the swimming ability of P.rimarily terrestrial forms.

Turtles today may be thoughtr of:.::as:: iilhabit·:ingr'magy:-hab.itacs

along a continuum spanning from completely terrestrial to almost

entirely aquatic. Zug (1971) and Walker (1973) believe the

hindlimbs are the primary pair to consider when studying terrestrial

or bottomwalking. They state that terrestrial turtles use their

hindlimbs primarily for drive and rely on their forelimbs for

steering and support. Conversely, aquatic turtles (especially sea

turtles) use their forelimbs for drive during the swimming stroke

and steer with their hindlimbs. Large river turtles do, however,

begin their swi:mming movements from ·a floating position by utilizing

powerful bindlimb thrusts (Moll, pers. comm.).

This brings about the central question and main purpose of

this study. How have the old world chelonians adapted to their

present mode of living especially in regard to their limb morphology?

6

The specific objectives of the study were: (1) Identify through

review of the literature and by experimentation which habitats

and mode(s) of locomotion are utilized by various Old World species.

(2) Determine by comparison of limb bones what differences exist

between species inhabiting the various habitats and utilizing

dissimilar modes of locomotion.

The turtles sttudied include 14 genera from three families.

From the Family Testudinidae the rep res en tatives are: Geochelone

emys, Gopherus �· and Kinixys erosa (data on t�e latter two taken

from Zug, 1971). From the Emydid�e (Subfamily: Bataguri-nae) the

following gen�ra: eyclemys dentata, Heosemys grandis, Siebenrockiella

crassicollis, Notochelys platynota, Cuo�a amboinensis, Kachuga �.,

Batagur baska, Callagur borneoensis and Orlitia borneensis. From

the Trionychidae: Trionyx �· and Pelochelys bibroni.

The testudinids, which are entirely terrestrial, have some

primitive features, especially Geochelone (Pritchard, 1979). This

family seems to make little use of watery environments with the

exception of some Galapagoes tortoises which carry out some

thermoregulatory activities in shallow pools (Pritchard, 1979).

The family Emydidae comprises the greatest number of genera and

total specimens measured. The most aquatic of the group appear to

be Callagur, Batagur and Kachuga. The majority, however, are 'Dest

categorized as semiaquatic and include Cuora, Cyclemys, Heosemys and

Siebenrockiella.

The last family is the Trionychidae, and includes Trionyx and

Pelochelys . . Trionyx is the only member of this study to have members

7

in both the Old and New World. Trionyx is a good swimmer, but

spends a· great deal of tilne on the bottom buried in sand or mud.

a

METHODS AND MATERIALS

Eighty (80 ) skeletons (16 species) were prepared and measured

from the collection of Dr. E.O. Moll of Eastern Illinois University.

These specimens were obtained from field work in Malaysia during

1975-76 and later while in India 1982-83.

The majority of these skeletons had to be cleaned of dried

muscle and other extra tissue before they could be measured.

Bones were soaked in a jar of tap water for 24-48 hours. If

left longer than two days, the water was changed daily to prevent

accumulation of odor causing bacteria. A small axoount of detergent

or dishwashing liquid, added to the water when soaking, helped

to remove oil from the bones. Dental instruments and. single-edged

razor blades were used for scraping and cutting apart the joints,

respectively. An old pair of scissors were used on larger bones

to scrape off the thin, but tough periosteum. Forceps were also

O·ccasionally helpful.

After removing most of the extraneous material, the bones

were reintroduced into the soaking jar with fresh water and s.oap

for another 24 hours. They were theri reexamined, dried, boxed

and labeled. Care was taken not to soak skeletons· for extensive

periods as soft bone ends may begin to disintegrate. Only one

skeleton was prepared at a time to prevent mixing of bones from

different specimens. As a general rule, 2 to 2� hours were

allowed for the first cleaning and 45 minutes to l hour fo·r the

second.

9

By becoming acquainted as to where the membranes are attached

and where they are the toughest, it was possible to use an alternative

method in combination with the soaking and scraping technique. New

single-edged razor blades were used to carefully cut and scrape

<Nay large amounts of dried meat and membranes- prior to soaking.

Bones were then soaked and scraped a second time as previously

described. This technique saves considerable time, as soaking

time is eliminated.

All measurements were taken with Vernier calipers to the

nearest tenth of a millimeter. Several skeletons were remeasured

to check for precision and it was found the checks varied no more

than + . 1 mm from the original. Only the bones on the right side

were measured, unless the right side bones were not available.

A complete listing of the measurements taken and their abbreviations

follows.

CCL - Length of eighth: cerv.i:cal: vertebra

HL Length of humerus: the straight line length from proximal to distal surf ace

UL Length of ulna: the straight line length from distal to proximal extremity

FL Length of femur: the distance from the notch fonned between the head and trocanter major and the distal end of the condyles

TI. Length of tibia: straight line length of tibia between proximal and distal surf aces

PW Proximal width of the femur: the maximum distance across the flare of the trocanters measured perpendicular to the long axis of the shaft

FHW -- Width of femoral head: the maximum distance across the center of the femoral head

10

•

FHL - Length of femoral head: the maximum distance across the center of the femoral head

WPC Greatest width of the coracoid

NPC Narrowest part of the coracoid, taken in the same plane as the WPC

LC Length of coracoid: the maximum longitudinal length from the proximal to distal surfaces, not including the distal cartilage

IDI -- The maximum inner distance between the ilia

TL + FL - Length of tibia plus femur

HL + UL - Length of humerus plus ulna

Because species and individuals dilfer radically in size,

measurements have oeen converted into ratios in order to provide

a standard for comparison. Zug (1971) has discussed pToblems

with using ratios. He selected the eighth cervical vertebrae

as the best standard divisor of his raw measurements because

of its presumed low variability.

A microcomputer with print out was utilized to obtain

standard deviations and means for all measurements ootained.

One-way analysis of variance statistical procedures were used to

establish whether means. were significantly different from one

another. If any significant dilferences were found in the ANOVA,

a Student-Neumann Kuels mean comparison test (o<.. = .05} was

performed. (Schefler, 1980)

Below is a statement of the null hypothesis:

H = no significant differences exist anxmg 0 relative limb measurements of terrestrial turtles, bottomwalkers and swimming forms.

In order •

to conclusively establish that certain species were

11

indeed bottomwalkers or swimmers. One member from each of five

genera (Callagur, Cuora, Notochelys, Orlitia and Siebenrockiella)·

were observed. Each specimen was placed separately in a 225 liter

aquarium. After an acclimation period lasting 30-60 minutes, data

were recorded with the aid of a stop watch on the number of

seconds each spent performing the activities recorded in Table 1.

At least 145 minutes (including a.m. and p. m. observational

periods) of observation was recorded for each of these five

subjects. A Chi-Square test was per£ormed upon these observational

data. The results will be explained more thoroughly in the next

section of this paper.

12

RESULTS

The observational data summarized in Table 1 supports

the alternate hypothesis which states that among the five genera

(Cuora, Callagur, Siebenrockiella, Notochelys and Orlitia)

statistically significant modes of locomotion do exist. Callagur

showed a significant preference for swimming, spending 67.3% of

its time in this mode. Whereas, the four other species spent a

significantly greater amount of time bottomwalking (Cuora, 42%;

Notochelys, 49%; Siebenrockiella, 50%; and Orlitia, 47.3%).

Some spent a little time bumping into the sides of the aquarium,

but most assumed a standard style of locomotion within minutes of

being introduced into the tank. The latter four species of ten

spent their bottomwalking time in a sort of tip-toe posture, moving

their limbs with a slow sweeping motion, with only the tips of their

claws encountering the bottom. 2 A Chi-Square Test (X 1 = 155.15 ca

d.f. = 12, p<.001) suggests that Callaaur spent a statistically

greater amount of time engaged in swimming activities than did the

other species. A similar test showed the other four genera were

spending a statistically greater amount of time in the bottomwalking

mode. Furthermore, another Chi-Square Test suggests that the proportion

of time spent engaged in the various types of locomotion is irrespective

of the time of day (a.m./p,m,), except in the case of Siebenrockiella

2 (Xcal = 9.981, d.f. = 12, p:;>.45), A comparison of a.m./p.m. totals for

13

this genus for each mode revealed an increase in swimming and bottomwalking

during a.m, periods and an increase in sedentary '(not moving) modes

during the p,m, observation periods, (See Table 2)

Based on the findings above, review of the literature,

and information provided by E. Moll, the following groups have

been established for comparison. Group A (strong swimmers)

includes: Trionyx, Pelochelys, Callagur, Batagur, Kachuga.

Group B (bottomwalkers) includes: Siebenrockiella, Notochelys,

OJora, eyclemys, OrlitPia and Heosemys. Lastly, Group C

(terrestrial) includes: Geochelone, Kinixys and Goph·erus. (See

Fig. 1).

The remaining results deal with osteological observations.

The mean (humeral) length ratio (HL/CCL) (Figure 2 ) divides this

group of turtles into two major groups, possibly with one subgroup.

Siebenrockiella, Orlitia, Kachuga, Callagur, Trionyx and Cuora

form a group which has measurements significantly smaller than

Geochelone and Notochelys. eyclemys and Heosemys tend to bridge

the gap between the two groups but, by themselves do not form a

statistically significant group. The Student-Neumann-Kuels or

SNK Test was utilized to establish when ranked mean values became

significant. A one-way analysis o.f v;a riance test performed on

turtle Groups A, B and C, indicated significant differences do

exist among the three regarding this measure. The terrestrial

group had longer humeri than bottomwalkers which had longer

humeri than the swimmers (F table for three groups see Table 3· ) • Relative ulnar length (UL/CCL) also yielded significant

"F" values. (Fig. 3). After ranking each mean and applying

the SNK Test, the same two major groups appeared as in the HL/CCL

measurement. (The species by species rank in both these cases

14

are almost identical); Notochelys and Geochelone appear to be

linked with statistically similar measurements, as are Heosemys

and Cyclemys. These groups though statistically different, do

occur on the same end of the continuum of mean lengths.. The

remaining genera form a third statistical unit that includes

means which are less than the other two groups. (See Fig. 4

for SNK Summary)� Again, in regard to the three earlier established

Groups A, B and C, terrestrial and bottomwalkers have significantly

greater ulnar lengths than swimmers. Also, the terrestrials'

measurements are significantly greater than bottomwalkers.

Furthermore, figures 2 and 3 are strikingly similar.

15

TL/CCL and FL/CCL measurements do not show statistically significant

differences among the genera. Thougrr the two graphs(figures 5&6) are similar,

neither measurement is significant when the three groups of

turtles are compared.

The proximal width of the femoral trochanters (PW/CCL) does

show statistical significance among the three locomotor groups. Swimming

turtles have a significantly larger flare between the trochanters

than do the terrestrials. These results are congruent with those

of Zug (1971) in his observations on new world trionychids.

Terrestrials and bottomwalkers are statistically indistinguishable

with regard to this character. · Trionyx, Callagur and Batagur are

closely linked and significantly different from all other genera.

Figure 7 suggests a pattern for the above mentioned genera. An

overall look at the ranked means (SNK Test) has aquatic turtles on

the upper end, bottomwalkers in the middle range and terrestrials

near the lower end of the rank. Geochelone, Cuora and possibly

Cyclemys seem to be the only ones which could be deemed out of

line and, as previously mentioned, cause the pattern to be somewhat

distorted. (See Fig. 8 for SNK summary).

Relative coracoid length ( LC/CCL) is greatest in Trionyx

of all the genera examined. Zug (1971) noted swimmers have the

longest co raco ids. On the other end of the ranked means , Cuora

and Geochelone have relatively short coracoids. The remaining

taxa fonn a central group in which no statistical differences can

be found. Within the life modes, aquatics cannot be distinguished

from bottomwalkers. However, both aquatics and bottomwalkers as

a group have measurements significantly greater than terrestrials

(See Figure 9).

The values for NPC1WPC (Fig.10) group Callagur and Batagur

at the uppermost extreme. A second group is apparent, including

Trionyx, Kachuga and Siebenrockiella. These two groups include

most of the swimming turtles. Toward the lower end of the rank,

five bottomwalkers and Geochelone form a statistically significant

set. Groups A, B and C show statistical significance. Aquatics have

16

signi£icantly .�eater ratios than both bottomwalkers. and terrestrials (fig.8).

However, bottomwalkers measurements are not statistically

distinguishable from terrestrials.

HL+UL The combination value CCL (Fig.11) shows another interesting

spectrum of rank. Geochelone and Notochelys have the longest limbs

differing significantly from all the rest. Two aquatic turtles,

Callagur and TrionY?C have the shortest limbs. The SNK Test

indicates that the forelimbs of terrestrials are significantly

longer than bottomwalkers and bottomwalkers are significantly

longer than aquatics. The same analysis is also true of the

UL+HL combination value TL+FL in regards to the three life mode groups.

(See Fig.12 SNK Summary). Again the terrestrials have the highest

17

values .while aquatics have the lowest(Fig. 12). Cartain bottomwalkers(Cuora,

Siebenrockiella, Orlitia) are very close to the aquatic turtles

in Fig.11 while other bottomwalkers (Cyclemys and Notochelys)

range closer to terrestrials in Fig. 13.

Another significant measure is UL/HL because it shows

sharp statistical divisions between the group of turtles

comprising Notochelys, Cuora, Cyclemys, Geochelone and another

group made up of Heosemys, Kachug�, Siebenrockiella, Callagur

and Orlitia. (See Fig.14 ). TrionY?C is again isolated statistically

from all other taxa (See Fig, 12 ). Bottomwalkers are statistically

indistinguishable from terrestrials. However, bottomwalkers and

terrestrials both have values significantly greater than aquatics.

TL/FL Both TL/FL and CCL measurements show no di�f erences with

regard to the three bife style groups. Ranking of these values

yields significance only at the extreme ends. Most turtle

genera are grouped in a large, statistically indistinguishable

area in the middle of the ranking (See Figures 15 and 16) •

DISCUSSION

In order to aid the reader :in keeping track of the large

number of turtle genera under study, Fig. 1 provides an

illustration of them placed along a continuum frolll terrestrial

walkers to swimmers. Bottomwalkers with swimming t.endencies

have been positioned closer to the known swimmers. Conversely,

bottomwalkers with terrestrial tendencies were placed farther

toward the right and therefore closer to the terrestit.Al group.

Laboratory aquaria observations of the four turtle species

in Group B reinforce the assumption that they are1in fact,

primarily bottomwalkers. Callagur (Group A) observations add

evidence :in support of its classification as a swimmer.

Siebenrockiella observations, . · revealed it was active a

18

greater amount of a.m. time swimming and bottomwalking. Furthermore, the· same

genus, was observed. .p,a ssive a greater amotmt of time( not movin,g)

during p.m. periods. Little is known of this species' life haoits,

so it is difficult to confirm or deny this statement. However,

based on these findings, Siebenrockiella may be an early 1D0rning

forager. This statement has yet to be substantiated by field

studies. (M::>ll pers. comm.)

If we can make the assumption that the genera of turtles

studied are correctly placed along the aquatic continuum, (i.e.

terrestrial - aquatic) then the osteological morphometric data

are interesting. Significant statistical differences do exist in

more than half the measurements taken. This iS not surprising,

but the importance here lies in the pattern in which measurements

were significant.

An easily recognized pattern emerged consistently as these

data were statistically analyzed. Measurements involving

hindlimbs (TL, FL) excluding the pelvic girdle (IDI) were

statistically indistinguishable from one another for almost the

entire range of genera and also for the three turtle life mode groups.

Similarily a.djusted femur length (FL/ CCL) , tibia length (TL/ CCL) ,

tibia length divided by femur length (TL/FL) and tibia plus

femur length ��i: were not significantly different.

Zug's (1971) work on chelonian limb structure made few

correlations between skeletal morphology and life mode

( t.errestrial, aquatic or bottomwalker). My findings suggest

that Dr. Zug may have slighted the most significant bones by

concentrating on the hindlimbs and pelvis. Zug did not report

any data dealing with the bones in the pectoral girdle or

forelimb. My study suggests that these are important from a

mode of existence standpoint.

The question of why hindlimbs don't show significant

variation among modes of existence remains. The answer could

be related to the simple fact that all turtles must locomote

to some degree on land in order to lay their eggs. Therefore,

it would not be advantageous for them to evolve a hindlimb

completely adapted to aquatic locomotion. They could become

over-specialized and cause themselves locomotor difficulties

(as sea turtles have when on. land). In addition, almost all

19

turtles use their hindlimbs for digging the nest and need

to retain limbs adapted to perform this operation. For these

reasons, the hindlimb has remained relatively evolutioncµ-ily

conservative.

It may be interesting to note that al though hindli:mb values

were not significantly different, that in all four measurements,

the average mean values (using Groups A, B and C) usually showed

that the terrestrial and bottomwalker values were closer to

each other than either was to the swimming value. This firid:ing

suggests either that bottomwalkers evolved from terrestrial

turtles that took to water or that terrestrial turtles '.evolved

from aquatic bottomwalkers. Similarly, in the measurements

found to be statistically significant, bottomwalke.rs· axe less

like swimmers and more like terrestrial forms.

The most c onsistent significant measurements were found

to be values dealing with the forelimbs. The humerus and ulna

are relatively longer in terrestrials and botto11JWalkers •. At

firs.t glance, this may seem unus�al because of: the extxeme

lengths achieved in forelimbs of sea turtles. There axe two

possible explanations for this. First, my measurements were

confined to the stoutest limb bones and therefore excluded

the radius, a bone which often appears to be longer than the

ulna. Secondly, the total length of the limb itself w ould have

to include th.e addition of the many bones :in the hand which

would in many cases, undoubtedly extend the length consideraoly.

These additional bones would have to be considered· oefore linal

20

conclusions are drawn . It is possible that these stouter bones

(ulna and humerus). have remained shorter in swimmers to retain

strength in propulsion. Overall, the significance of the forelimb

measurements point toward divergent evolution occurring in the

forelimbs and pectoral girdle. Although conjectural at this

point, it appears that selective pressures associated with an

aquatic existence and concurrent locomotor habits caused the

lengthening of the forelimb.

The osteological morphometric data of Cuora deserves

special ment ion because of its unusual yalues.� Although

Cuora spent most of its time bottomwalking (during the

observational data taking periods), its measurements are

consist ently lowe:zr than other bottomwalkers. This could again

be explained by saying this genus is generally a bottolllWalker

with swimming tendencies. Furthermore, Cuora could be a special

case because it is the only member of the study with a highly

kinetic hinged plastron .

In conclusion, it appears that there is a consistent

relationship between the pectoral girdle assemblage and the

mode of existence in old world turtles. This is easily

testable with turtles whose life history is well kno wn .

21

LITERATURE CITED

Ashley, L.M., 1955.Laboratory Anatomy of the Turtle. Wm. C.

Brown Co., Dubuque, Iowa.

Bojanus, L.M. (1819-1821)."Anatome Testudinis Europaeae."

Vilnae (also 1970 reprint Soc. Stud. Amph. Rept.).

Pritchard, Peter, 1979.Encyclopedia of Turtles. TFH Publications,

Inc., Neptune, N.J.

Romer, A.S., 195 6, Osteology of Reptiles. The University of

Chicago Press, Chicago IL.

Schefler, W.C., 1980. Statistics for the Biological Sciences.

Addison-Wesley Pub. Co., Mass.

Sokal, R.R., 1981. Biometry. W.H. Freeman & Co., San Francisco

Stahl, C.C., 1978. Turtle Evolution. p. 88-158 in. Taxapamy,

Evolution and Zoogeosraphy, TFR Publications, Inc. Neptune, N.J.

Walker, W.F., 1973. Locomotor Apparatus of Testudies. p. 1-100

In Biolosy of the Reptiles, Academic Press, New York.

Walker, W.F., 1979, Turtles Perspectives and Research. Harless and

Morlock, eds., John Wiley & Sons, New York.

Williams, T.L., 1979. Experimental Analysis of the Gait and Frequency

of Locomotion in the Tortoise Testudo graeca. J. Physiol. (Land. )

310: 307-320, Vol. 71 #78159.

Zug, George, 1971. Buoyancy, Locomotion, Morphology of the Pelvic

Girdle ·and Hindlimb, and Systematics of Cryptodiran Turtles.

Miscellaneous Publications, Museum of Zoo, University of Michigan

No. 142, pp. 98.

22

Appendix

Tables, Graphs, and Statistical Data.

Table 1 • • • • • •. • • • . . • . . • • • . • • • . • • • • • • . • • • • • pp. 24-25

Table 2 • • • • • • • • • • • • • • • . • • • • • • . • • • • • • • • • • • pp. 26-27

Fig. 1 . • • • • • • • • • • • • • • • . . • • . . • . • • • • • • • • • • • pp. 28-29

Fig. 2 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • pp. 30-31

Tables 3 & 4 . • • • • • • • • • • • • • • • • • • • • • • • • • • • • pp. 32-35

Fig. 3 • • • • • • • • • . • • • • • • • . • . • • • • • • • • • • • • • • • pp. 36-37

Fig. 4 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • pp. 38-39

Fig. 5 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • pp. 40-41

Fig. 6 . . . . . . . .. . . . . .. . . . . . . . . . . . . . . . . . . . . . pp. 42-4-3

Fig. 7 • • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • • • • • pp. 44- -45

Fig. 8 • • • • • • • • • • • • • • • • • • • • • • • • • • • • . • • • • • • pp. 46---47 Fig. 9 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • pp. 48,-49

Fig. 10 • • • • • • • • • • • • • • • • • • . • • • • • • • • • • • • • • • pp. 5.0-51.

Fig. 11 • • • • • • • • . . • • • • • • • • • • • • • • • • • • • • • • • • pp. ,52,-.53

Fig. 12 • • • . • • • • • • • • • • . • • • • . • • . • • • • • • • • pp. 54-55 Fig. 13 • • • • . . . . • • • . • . • • • • • • • • • • • • • • • • • pp • .35 -51

Fig. 14 ................................... pp. 58-59

Fig. 15 • • • • • • • • • • • • • • • • • • • • • • . • • . . • • • • • • • pp. 6.Q-61

Fig. 16 • • • • • • . • • • • • • • • • • • • • • . . • • • • • • • • • • • pp. 62-63.

23

Table 1: Number of minutes spent in each

of four locomotor 100des by r epresentatives

of 5 genera of Old World turtles. Percent

of time is in parentheses.

24

GENERA

Callagur

.Siebenrockiella

Notochelys

"

Cuora

Orlitia

, I . !

TABLE 1

NOT MOVING

BOTTOM TOP·

11(7.3%) 20(1J.J%)

20(1J.3%) 10(6.7%)

J4(23.4%) 16 ( 11%)

18(12%) 23(15.3%)

42 ( 28%) 16(10.7%)

75( 50%)

71(49%)

·63(42%)

..

71(47.J%)

25

45( JO%)

�--] 2�(16.5%) f ·-·

46(30.7%) .

21(14%)

-·

j ' ' '

I.�

" •

Table 2. Shows the relationship between

the number of minutes spent in each

locomotor mode and time of day.

26

27

· - -

NOT MOVING MOVING BOTT07v1 TOP WALKING SW If\';'f�1 ING TOT. x� cal.

A.M. 9 14 28 24 75

Cuora 1.951

P.M. 9 9 35 22 75 n.s.

A.M. 25 9 31 10 75

Orlitia 3.005

P.M. 17 7 40 11 75 n.s.

.

' A.M. 7 10 12 46 75

Call-agur 3.620

P.M. 4 10 6 55 75 n.s.

A.M. 5 3 41 26 75

.-B�fH���Ia •

9.981

P.M. 15 7 34 19 75 sign.

A.M. 18 6 37 9 70

Noto-ch ells 2.520

P.M. 16 10 34 15 75 n.s.

n.s.= not statistically significant p>0.05

Fig. 1 Fourteen genera of chelonians on

a habitat centinuum from aquatic to terrestrial

ha.bits.

28

Cuora

Batagur

Callagur

Sieben rock ie Ila

T r iony Kachuga sp' Drlitia

_....;._:...:.:._ ]J , I Sp. •

Heosemys

Notoche I y s

Cyc.lemys

• QUA�ics \ SEMi-AQUATiC [BOTTOMWALKERS I swimmers

Gopherus

6eochelone

Kinixys

N '°

Key:

Fig. 2. Humeral. length (x ± s) for each

genus under study. The number above each

range indicates the number of specimens

measured. in each genus.

30

GE = Geochelone emys SI = S iebenrockiella crassicollis

KI = Kinixys erosa OR = Orlitia borneensis

GO = Gopherus � KA= Kachuga �

HE = Heosemys grandis CA = Callagur borneoensis

NO = Notochelys platynota BA = Batagur baska

CY = Cyclemys den ta ta TR= Trionyx �

cu = Cuora amboinensis PE = Pelochelys bibroni

Subscripts:

a = aquatic . b= bot tomwalker c= terrestrial

5.5r I 5 I

. ii 4.sL · -

= .... c 2 4 ...., _..

....

u .... " cc 3.5. a: ...., :s =

· = 3

2.5

3 T

f 4 l. I

8

I

· G E.c H Eb tJ 0 b

9

1 13 5 T 3

I I 13

I

I C Yb C Ub Sib

GENERA ORb KAa

6

10

'""'�.�

1

I · -

fi �

·-'

CA1 TRa PE3 w .....

J

) Table 3. Statistical data for combine mean

lengths of the three groups of turtles for

all measurements taken. (p< 0.05) The number

in parentheses indicates the number of

turtles measured in each group.

32

f source o variatio

total

treat.

error

source

total

treat

error

source

total

treat.

error

source

total

treat.

error

source

total

treat.

error

- source

total

treat.

error

source

total

treat.

error

n measure

FL/CCL

F=.68

measure �

rL/CCL F= 1.60

measure

LC/CCL ..

F==4.96

ffiP::I �ll'Y'f:

NPC/WPC *

F==49.86

measure

UL+HL TL-+FL F=57.7* ........_., _ _ .-.. - - . ..,

rneas11rP

TL/FL

F= 2.95

measure

TL+FL CCL

F= 2.05

d.f.

86

2

84 . -

d.f.

74

2

--7? d. f_. 76

2

74

n f'

74

2

72

d.f.

69

2

6 7 � ..... . .,. - eo --rL f' 8 5

2

83

d. f �--

72

2

?()

SS

29, 952

. 4784

29 . 474 - ·--

SS

8 . 791

, 3 74

8 . 417

SS 2 5 . 497

3 . 01 2

?? lJ.R �

C! C!

,596

. 346

. 25

SS

. 354

. 224

.130 � - - · ... ... __ ,,

��

. 36 2

. 024

, 338

--�s - - --

53 . 108

2 . 931

�() 1 ??

ms

, 3483

. 2392

, 3509 - -

ms -

. 1 188

. 18 7

. 1 169

m� , 3355

1 . 506

'1 () '1A

_,...

. 00805

. 1 7'.3

. 00347

ms

. 00513

. 1 12

. 00194

_ _ ms___,

. 0043

. 01 2

. 00407

_ms . . _ , 7376

1 . 47

_?1h8

33

Group x

A J.57 (23)

B J.8J (50)

c J.94 ( 14)

Group x

A 2.75 (2J)

B 2.89 (49)

c J.05 (J)

Group x

A 3.55 (23)

B J. 18 (51 )

c 2.61 (J)

Group x

A .404 (2J)

B .266 (49)

c .233 (J)

Group x

A .86 (21)

B .96 (46)

c 1.10 (3) Group x

A .772 (2J)

B .762 (46)

c • 798 ( 17)

Group x

A 6.JJ (2J)

B 6. 71 (47)

c 7.26 (J)

Table 4.(continuation of Table J) Statistical \.

data for combine meaz.i lengths of the three

groups of �urtles for all measurements taken.

(p< 0.05) The number in parentheses indicates

the number of turtles in each group.

source of variation

total treat.

error

source

total

treat.

error

source

total

treat.

error

source

total

treat.

error

source

total

treat.

error

source

total

treat.

error

source

total

treat.

error

measure d.f.

74

HL/CCL 2 F=15.47• 72

measure d.f.

74

UL/CCL 2

F=20.91* 72

measure d.f.

8 5

PW/CCL 2

F=17.87* 8'3�-measure d-f.

87

FHW/FHL 2 F=12.03* 8 5

measure d. f.

70

ID I/CCL 2

F=9.16• 68 >· ·�""'-�--- - ,,.. ... ----

measure d.f.

70

UL/HL 2

F=14.21 * 68

measure d.f.

70

UL+HL 2 CCL

F=20.1 • ?.A

J5

SS ms Grouo . x

20 . 93 . 2828 A J.42 (21)

6 . 289 J . 145 B J.91 (51)

14. 64 . 20JJ c 4.84 (3)

SS ms Group x

1 2 . 63 . 1 707 A 2.04 (2J)

4 . 641 2 . 321 B 2.54 (49)

7,991 . 1110 c 3.18 (J)

SS ms Group x

4 . 1 1 6 . 0484 A 1.36 (2J)

1 . 240 . 6 20 B 1 .11 (50)

2 . 876 . 0147 c 1.13 (13)

s� m� Group x .408 . 0047 A .612 (23)

. 09 . 045 B .646 (48)

. 318 . 00374 c .686 ( 17)

SS ms Group x 30. 71 . 439 A 2.62 (21)

6 . 52 3 . 26 B 3.39 (47)

24. H . 356 �--·� · · ·- -� ---- � ·

c 2.93 (J)

SS m� Group x

. 184 . 0026 A .59 (21)

. 0 54 . 02 7 B .65 (47)

. 1 30 . 0019 c .66 (J)

SS ms Group x

60 . 37 . 8624 A 5.42 (21 )

2 2 . 44 1 1 . 22 B 6.4J (47)

�'7 o� <<'7R c 8.0J (J)

Fig. 3. Ulnar length divided by CCL

(f'± s ) for each genus under study,

The number above each range indicates

the number of specimens measured for

the genus,

36

3.5 - 3 - 3

_,.I 3.0 _I I E E �

= 8 ,_ c:s z 2.5 I .... -

-u :1u

:2.0 -=

1.5

1.0 &EC HEb NOb

9

- 13 12

I 3 I

CYb cub SI� DRh

5 T I

Kl.

2 10

I T 6 .

ca. BA. TR. w -..:>

Fig. 4. Mean values are ranked from highest

to lowest . (Key to genera same as p. JO )

Dark bars indicate similar measures and

breaks indicate where statistical differences

are present.

J8

39

HL/CCL

· CY 8 KAA CAA

4.84 4.57 4.25 3.97 3,73 3.65 3.62 3.35 3.29 3.26

FL/CCL

Kic N08 CY8 GEc KA4 HE8 SI8 OR8 CAA BAA TRA GOc CU,

4.5 4.4 4.29 4.22 4.12 3.96 3.66 3.6 3,5 3,9 3.3 3.1 3.09

UL/CCL

KAA BAA CAA TRA

3.24 3.18 2.83 2.47 2.27 2.23 2.22 2.21 2.12 2.05 1.8

TL/CCL

KAA BAA CAA TRA

3,37 3.1 3.09 3.05 2.84 2 . 81 2.76- 2.75 2.74 2.56 2.33

•

Fig. S. Tibial length divided by CCL

( x ± s ) for each genus under study.

The number above each range indicates

-the number of specimens measured in

the genus.

40

4 3.5 9 5

-

ii 3.25 .--- 3 -

� 2 t 10 I ! �I 12 3.0 I I 8

t 3 1 I ...

. l �lu 2.75 1- 12

I I 6

c u

T

l -m -

l

....

I I 1 I ..

2.50

2.25

2.0 &EC HEii Nob CYll cub Sib ORb KA, CA, BA, TR, + ....

_G_E.N ER A

42

Fig. 6. Femoral length divided by CCL .

( x ± s ) for each genus under study.

The number above each range indicates -

the number of specimens measured in

that genus.

5 I

4.51--1 = • ---

4 • -c -� =-' -I.a ...

_, _.lu35 �u· c � -!: Loi ... . 3

2.5

3 9

I 3 T 5

l 4

I 8

I I I 12 -J.. 8 l I 10 6 3 T 2

I 13

Kie &Ee &De HEb NOb CYb CUb Sib ORb · KA1 CA1 BA1 TR1 GENERA

.{::" 'vJ

Fig. 7 . Proximal width of the femoral

trochanters divided by CCL (X ± s) for

each genus under study. The number above

each range indicates the number of speci­

mens measured in that genus.

44

.asL ..... -4...,...,...,_,��r.��r.ir Kie &Ee &De HE11 N011 CYl. CU11 Sl11 0811 KA, CA, BA, fR1 GE N t R A

-.....-. .-. ..... � .a...u.a. • ........,. ••-•

�­Ill

Fig. 8. Mean values are ranked from highest

to lowest for the turtle genera under study.

(Key to genera same as p . 30 ) Dark bars in­

dicate similar measurements whereas breaks

indicate where statistical differences occur.

NPC/WPC

,47 . 43

LC/CCL

'I'R A

4 . 34 3 , 6 •

1.43 1.37

. 38

KA SI · CY A B B

, 34 , 32 . 28 , 27 . 26 . 26

47

. 23 . 20

J . 4 3 . 4 3 . 37 3 . 3 5 3 . 2 J , 06 3 . 0 2 . 6 2 . 5

ORB HE B NOB KI' GO' CY B CUB 1.36 1.30 1.27 1.23 1.20 1.14 1.11 1.06 1.04 1.03 0.92

Fig. 9. Length of coracoid divided by CCL

( x ± s ) for each genus under study. The

number above each range indicates the

number of specimens measured in that genus.

48

6

5.0

-

E s l 4.5 ._...

c::a -= . 13 4.0 I· c.:» I

I 5 : 1 c 9 c.:» .

3.s l ...... 4

t 2 c.:»

I � I · 10 I � l e.:» 3 T 8 T =

3.0 l I t-c.= z .., �

13 3

2.5 I- -I 2 .o E '"i'if.I!'/ ·"' .... "'•'i'·"{'1·:� w-y� ... 4**T®"1\!0@�ma>··ri'***r*"*F''*ll -!>

GEC HEb Nob cvb .c�b S i b ORb KAa CA1 BA1 G E N E R A

TR1

Fig .IO • . Narrowest part of coracoid divided

by the widest part of the coracoid for each

turtle genus measured. The number above

each range indicates the total number of

specimens measured in the genus . (x ± s )

50

..... E ..... E E - E "Cl

-- "Cl 0 ·-u 0 • u a. • 0 a. u 0

u -0 -c: 0 0 - c: - 0 a. 0 ·--A. a. 0 - A. .. • -• M 0 • a. "Cl a. ·-• • z

.55

.50

.45 .-

• .4o r .35

I . . 30

. . 25

.20 I-

I .1 5

... ---...... __.., __

3

I I &Ef

-·-- -

8

I I

llE ..

5 12

9 4

13

l l 3

I

I I . I I I - · - I "�" c ,., cu., s 1 .. °''- . • •• G �N�RA

10

I

I

I c�,

2

I 6

I . . - I - -

··� , _! �

I /9\ . • ,. •

Fig. I l Ulnar plus humeral length divided

by the CCL for each turtle genus measured.

The number above each range indicates the

total number of specimens measured in each

genus.

-E

� I CD

; , :: 1 a: ....

:& I .... = u :::c u

� I a: c z .... =

9 . 0 I

8.0

7.0 -

6. 0

5. 0

4.0

3

3

I 9

5 8 ! 12 I t 3 I I 10

12

I

GEc H Eb N Ob CYb C Ub SI._ DR._ . KA. C A. GENERA

6

• TR a \,/\ \..J

Fig. 1 2 . Means values are ranked from

highest to lowest for each turtle genus

under study. (Key to genera same as p . � )

Dark bars indicate similar measurements

whereas breaks indicate where statistical

differences occur .

UL/HL

GE c

, 706 .685 . 666 660 .647 .618 . 6 1 0 .604 603 . 540

UL+HL CCL

GE, NOe

8 . 03 ?.83

UL+HL TL+FL

1.103 1D06 -

-

CY8 HE8

7.08 6.J8

HEB

SI8

601 • OR8

5.84 5�4

. OR B

•

5)7 5D6

. 970 . 965 .958 .933 , 920 905 8 59 . 8 1 2 . .

55

Fig. 13 . Ulnar plus humeral lengths divided

by tibial plus femoral lengths for each

turtle genus� measured . The number above

each range indicates the total number of

specimens measured in each genus . ( x ± s )

56

57

,_. , ,r • I

"' tO a:

t-

"' S21 • c �

"'

tn I e cc :::.::

J::I M t-+-1 ·' a:

=

.Q T'" -T'" en

<: .c Q: N t II = 1L

T'" Co:» 2 · LL

Ol I • >-D. � �

D. 1il (") = !1

z

D. co .... =

u ..... c.::J

0 LO 0 LO 0 LO 0 LO ""': 0 0 0) 0) co co I'-,... • . . • . . . . • • •

,... ,...

( W W) H O W 3 � S R l d V I 9 11 �o H19N31 l W W) s ruuwnu S O l d V N l R � o H 1 9 N 3 1

Fig. 14. . Ulnar length divided by humeral

length for each turtl!e genus .

The number above each range indicates

the total number of specimens measured

in each genus. ( x ± s )

58

s l e · E e =

:c ..... ..... = = :z :z """'

..... """' ..... _...

< a: a: < """' :z :E _... = = =

�•I .75 ... 3

I � I 1 2 ff;

. 10 H- l 9 I 3

I 1 8 l ·.

. 65 1- 1 2 5 ft' , .

I I I 10

.60 l 1 . I t . 3 l l • I . 6

.5 5

.50

. 45 �;••"'\ "'f\. • ' �I '.�,}�II'>;" �t ;.,.., ;;,.;Y,_�.-./_.••-· ' J - :•• I : ' ;�""'.;"\�·' 'i'"'._,�� 1.,:�''.• ---; -·· ')_>1-!J#'�� .. - ,.-f'/�""7' ";"1-;f .. '.�.(" 4 •t�'f:•··• � " 1;/-.t..!J�'.t)\•...:...�.,,1J41-..

a Ee · HEb N ob cvb cub G E N E RA

S ib ORb KAa CAa TRa \J\ '°

Fig . 1' . Tibial length divided by femoral

length for each turtle genus measured . (plus

or minus one standard deviation) . The number

above each range indicates the total number

of specimens measured in each genus . � x t s )

6o

. . 65

,�� . 6 0 I ?t¥1·#1M·¥fNftmtweptMf'i'E¥4&Wlt#fhJlltl#ttW8-#i"!l&?ftR�'tse;*®Pri!fUiWAt1W§�

Kie GEC GOC HEb NOb CYb c ub Sib ORb KAa CAa BAa TRa °' ....

G E N E RA

Fig . 1 .6. Tibial plus femoral length divided

by the CCL for each genus measured . (plus 1

or minus one standard deviation) . The total

number of specimens measured in each range

is indicated by the n'umber above each line.

( 5C ± s }

62

4 9 5

8.0 3

7.5

4.5 �� �;·t .�.-... " ••• -, .. ...... , • ';· •• • 4 .. � ::.· .. ·��· ;a.;:-, .... I�� ... , ,,,�-- .... '"'' ... •ipi:·""":'t ".'"··' t'.#·7:1,,."';0::1"1: . . �;.::.._._:..,..-: 1 .·'¥*MiW4MfMkiiCl@l.fi4�

. GEC HEb NDe; CYb C Ub . Sib G E N E R A

O Rb KAa C A a B A a THa °'

\..J