![(A) Least-squares regression between initial egg mass and hatchling mass of African spurred tortoise (Geochelone sulcata), fitting the equation log(hatchling mass) =-0.008 Â [log(egg mass)] 0.952. Broken lines are 95% confidence intervals. Note logarithmic axes. (B) Differences in hatchling mass among different clutches. Values were calculated from log-log residuals to correct for differences in initial egg mass. Lowercased letters signify differences among clutches (a = 0.05). Error bars are ±1 SE.](https://www.researchgate.net/profile/Day-Ligon/publication/237154268/figure/fig3/AS:651215595646983@1532273305335/A-Least-squares-regression-between-initial-egg-mass-and-hatchling-mass-of-African_Q320.jpg)
Content uploaded by Day B Ligon
Author content
All content in this area was uploaded by Day B Ligon on Nov 30, 2015
Content may be subject to copyright.
Incubation temperature effects on hatchling
growth and metabolic rate in the African spurred
tortoise, Geochelone sulcata
Day B. Ligon, Joseph R. Bidwell, and Matthew B. Lovern
Abstract: We tested competing hypotheses regarding the persistence of temperature-dependent sex determination (TSD) in
the African spurred tortoise (Geochelone sulcata (Miller, 1779)), by measuring the effects of incubation temperature (Tinc)
on a suite of physiological and behavioral endpoints, including resting metabolic rate, yolk-to-tissue conversion efficiency,
posthatching growth, and temperature preference. Correlations of these variables with Tinc could lend support to the hy-
pothesis that TSD persists owing to sex-specific benefits of development at specific temperatures, whereas absence of Tinc
effects support the null hypothesis that TSD persists simply because selection favoring alternate sex determining mecha-
nisms is weak or absent. The metabolic rate Q10 value exhibited temporal variation and was higher immediately after
hatching compared with 40 or 100 days posthatching, and mass conversion efficiency varied among clutches. Incubation
temperature correlated inversely with duration of embryonic development, but did not influence yolk conversion efficiency,
growth, or resting metabolic rate. Thus, our results provide little evidence indicating contemporary benefits of TSD, sug-
gesting that TSD in G. sulcata is no longer evolutionarily adaptive but persists because selection against it and in favor of
other sex-determining mechanisms is weak, or that TSD is an adaptive trait but for reasons not elucidated by this study.
Re
´sume
´:Nous avons e
´value
´des hypothe
`ses de rechange concernant la persistance de la de
´termination du sexe sous l’in-
fluence de la tempe
´rature (TSD) chez la tortue sillonne
´e africaine (Geochelone sulcata (Miller, 1779)) en mesurant les ef-
fets de la tempe
´rature d’incubation (Tinc) sur une se
´rie de conse
´quences physiologiques et comportementales, dont le taux
de me
´tabolisme de repos, l’efficacite
´de la transformation du vitellus en tissu, la croissance apre
`s l’e
´closion et la pre
´fe
´r-
ence de tempe
´rature. Les corre
´lations entre ces variables et la Tinc pourraient venir appuyer l’hypothe
`se selon laquelle la
TSD persiste a
`cause d’avantages relie
´s au sexe d’un de
´veloppement a
`des tempe
´ratures spe
´cifiques, alors que l’absence
d’effets de la Tinc appuierait l’hypothe
`se nulle qui veut que la TSD persiste simplement parce que la se
´lection qui pourrait
favoriser des me
´canismes de rechange de la de
´termination du sexe est faible ou absente. Le Q10 du taux me
´tabolique pre
´-
sente une variation temporelle et est plus e
´leve
´imme
´diatement apre
`s l’e
´closion que 40 ou 100 jours plus tard et l’efficacite
´
de la conversion de la masse varie d’un groupe d’œufs a
`un autre. Il existe une corre
´lation inverse entre la tempe
´rature
d’incubation et la dure
´edude
´veloppement embryonnaire, mais la tempe
´rature influence ni l’efficacite
´de la conversion du
vitellus, ni la croissance, ni le taux de me
´tabolisme de repos. Ainsi, nos re
´sultats apportent peu d’appui a
`l’existence d’un
be
´ne
´fice actuel de la TSD, ce qui laisse croire que la TSD chez G. sulcata n’est plus adaptative dans l’e
´volution, mais
qu’elle persiste parce que la se
´lection contre elle et en faveur d’autres me
´canismes de de
´termination du sexe est faible. Il
se peut aussi que la TSD soit un trait adaptatif, mais pour des raisons non e
´lucide
´es dans notre e
´tude.
[Traduit par la Re
´daction]
Introduction
Offspring phenotype is determined by genotype, maternal
investment, and environment. Whereas genotype is fixed at
fertilization, the latter two factors often vary throughout em-
bryonic development. Maternal investment can be highly
variable among oviparous ectotherms, ranging from differ-
ential allocation of nutrients during egg production to varia-
tion in the degree of postlaying care. Among environmental
variables, effects of temperature and hydric conditions are
the best studied. Temperature has been shown to affect dura-
tion of embryonic development, prehatching and posthatch-
ing growth rates, and survival (Yntema 1976, 1978; Bull
and Vogt 1979; Wilhoft et al. 1983; Gutzke and Packard
1987; Brooks et al. 1991). The hydric conditions experi-
enced by eggs has been shown to affect development rates
of embryos, oxygen uptake, and size at hatching (Gutzke
and Paukstis 1983; Morris et al. 1983; Gettinger et al. 1984;
Gutzke et al. 1987; Miller et al. 1987; Packard and Packard
1988).
Arguably the most striking (and most studied) environ-
mental effect on phenotype is its influence on gonadal dif-
ferentiation. Temperature-dependent sex determination
(TSD) is the most prevalent and best studied form of envi-
ronmental sex determination (Bull 1980), a phenomenon
known to occur in a variety of invertebrates and vertebrates,
including some crustaceans, fishes, saurians, chelonians, and
crocodilians (Bull 1980; Conover and Kynard 1981; Con-
Received 22 February 2008. Accepted 14 November 2008.
Published on the NRC Research Press Web site at cjz.nrc.ca on
13 January 2009.
D.B. Ligon,1,2 J.R. Bidwell, and M.B. Lovern. Department of
Zoology, Oklahoma State University, Stillwater, OK 74078,
USA.
1Corresponding author (e-mail: dayligon@missouristate.edu).
2Present address: Department of Biology, Missouri State
University, 901 South National, Springfield, MO 65897, USA.
64
Can. J. Zool. 87: 64–72 (2009) doi:10.1139/Z08-138 Published by NRC Research Press
over 1984; Naylor et al. 1988; Bull and Charnov 1989; Ciofi
and Swingland 1997; Warner and Shine 2008).
Among turtles (and in sauropsids generally), TSD is the
ancestral sex-determining mechanism from which genetic
sex determination has evolved at least six times (Ewert and
Nelson 1991; Janzen and Paukstis 1991; Janzen and Krenz
2004). In contrast, no shifts from genetic sex determination
to TSD are evident (Janzen and Krenz 2004). Partly for this
reason, turtles have been popular model organisms for
studying both mechanisms and evolutionary consequences
of TSD. A benefit of much of this TSD research has been
concomitant investigations of incubation temperature (Tinc)
effects on many other phenotypic traits in chelonians, such
as size at hatching, growth rate, locomotor performance,
and thermoregulation (see review by Rhen and Lang 2004).
The purpose of much of this work has been to identify Tinc-
affected traits that are differentially beneficial to males and
females to identify selective mechanisms for the evolution
and persistence of TSD (Bull and Vogt 1979). Plausible
mechanisms have been described involving some, but by no
means all, of these traits (Janzen and Paukstis 1991; Shine
1999). For example, it has been hypothesized that males of
species that exhibit male-biased sexual size dimorphism
could benefit from hatching at temperatures that produce rel-
atively large and (or) fast-growing hatchlings, if such tem-
peratures exist.
Compared with many of the endpoints described above,
the effects of Tinc on hatchling resting metabolic rate (MR)
have been investigated in relatively few turtles (O’Steen
and Janzen 1999; Steyermark and Spotila 2000). MR is an
integrative measure of biochemical activity necessary to
meet basic maintenance requirements (i.e., absent activity,
growth, or digestion). Therefore, MR represents a nonlethal
whole-animal measure that may identify the presence of
physiological differences (e.g., differences in body composi-
tion or enzyme activity), but not the underlying biochemical
or physiological nature of the differences. If MR is low and
constitutes a relatively small fraction of an individual’s total
energy budget, then a greater proportion of available energy
can be allocated to other processes that may increase evolu-
tionary fitness such as those related to growth and activity.
This difference in allocation, if differentially beneficial to
males and females, would lend support to the hypothesis
that TSD is an adaptive trait. Conversely, a lack of MR dif-
ferences between the sexes would suggest that this unusual
sex-determining mechanism persists in a given species sim-
ply because of evolutionary inertia, or persists because of
factors that do not affect MR.
Our study organism was the African spurred tortoise
(Geochelone sulcata (Miller, 1779)), a testudinid once com-
mon in Africa at latitudes coincident with the Sahel region.
Recent population declines have been steep, and likely are
attributable to climatic and vegetational changes throughout
the region, as well as human exploitation, which collectively
have resulted in inclusion of the species on Appendix II
(controlled trade) and petitioned for elevation to Appendix I
(threatened with extinction) of the Convention on Interna-
tional Trade in Endangered Species of Wild Fauna and Flora
(CITES 2007). This species is particularly suitable for study-
ing TSD because, like all testudinids that have been exam-
ined (Ewert et al. 1994, 2004; Spotila et al. 1994;
Eendeback 1995; Burke et al. 1996; Demuth 2001), TSD
fits a dichotomous pattern wherein males are produced at
low temperatures and females at higher temperatures. In ad-
dition, G. sulcata exhibits extreme male-biased sexual size
dimorphism (Ernst and Barbour 1989), with males reaching
a maximum of 92 kg and females 60 kg (Bonin et al. 2006).
Therefore, this species is a suitable model for testing the
often-hypothesized hatchling size and (or) growth Tinc effect.
This hypothesis postulates that natural selection will favor
TSD over genetic sex determination in species that exhibit
sexual size dimorphism when temperatures that produce
larger or faster growing hatchlings also promotes develop-
ment of the sex that ultimately grows larger (Rhen and
Lang 2004). Although TSD patterns have been examined in
a small fraction of tortoise species, the uniformity of avail-
able TSD data suggests either that it is actively maintained
by natural selection among testudinids or that current selec-
tion on sex-determining mechanisms in tortoises is relatively
neutral.
We measured Tinc effects on MR as a potentially integra-
tive measure of TSD-sustaining mechanisms, as well as
hatchling size, posthatching growth, and temperature prefer-
ence (Tpref) of juvenile G. sulcata. We hypothesize that if
TSD maintains a selective advantage over genetic sex deter-
mination, then differences among Tincs in MR, size, growth,
and (or) activity patterns would be evident, and that differ-
ences would be greater between male- and female-producing
temperatures than among temperatures producing each sex.
Absence of differences that correlate with Tinc and sex, on
the other hand, provide evidence that selection favoring ge-
netic sex determination is absent among those taxa in which
TSD persists.
Materials and methods
Eggs
We obtained eggs of G. sulcata from a commercial tor-
toise breeder (Riparian Farms, Hereford, Arizona). A total
of 107 eggs comprising six nests were collected from nests
produced by captive females between 19 January and
31 March 2003 and shipped to arrive at Oklahoma State
University within 48 h following oviposition. Upon arrival,
the eggs were weighed (±0.1 g) and diameter was meas-
ured (±0.1 mm). Eggs were then distributed among
15 plastic shoeboxes filled with damp vermiculite (1:1
vermiculite:water by mass; –150 kPa) in a randomized
block design with clutch serving as the blocking variable.
Three shoeboxes were assigned to each of five incubators
programmed to maintain constant temperatures (28.5, 29.5,
30.5, 31.5, 32.5 8C). Shoeboxes were rotated daily within
each incubator to minimize potential effects of temperature
gradients. Thermocouple wires placed in each incubator
were used to monitor the temperature of each incubator
daily. Each box was weighed weekly and water was added
when necessary to maintain a consistent hydric environ-
ment. Beginning on day 20 of incubation, MRs (oxygen
consumption) of eggs comprising clutches 5 and 6 were
measured at 10 d intervals until pipping (see below). After
pipping, eggs were placed into individual containers and
maintained in the incubators to retain the identification of
individual hatchlings until residual yolks were internalized.
Ligon et al. 65
Published by NRC Research Press
Individual identification numbers were then written on each
tortoise’s carapace with acrylic paint.
Hatchlings
MR of each hatchling was measured at each of the five
incubation temperatures 9–14 d after pipping (1–2 d after re-
sidual yolk was internalized) to test for metabolic compensa-
tion to long-term temperature acclimation during incubation.
Because activity in the chambers can elevate MR above
resting levels, duplicate measurements were made at each
temperature and the lower of each pair was used to estimate
activity-free MR. Tortoises were likely still subsisting on
stored yolk and thus were not postabsorptive. Therefore,
these measurements do not meet all of the criteria for meas-
uring standard metabolic rate. After completing these meas-
urements, hatchlings were removed from the incubators and
housed in groups of five in plastic boxes (57 cm 41 cm
15 cm). Each box was equipped with an overhead 60 W in-
candescent light for basking and a box of moist peat moss
into which the tortoises could retreat. Lights were on a 12 h
light : 12 h dark cycle; at night ambient temperature
dropped to 25 8C.
Tortoises were fed a commercial tortoise diet (Mazuri tor-
toise diet 5M21; 15% crude protein) daily. Food pellets
were saturated with water to soften them. In addition, tor-
toises were soaked for 30 min daily in distilled water, after
which individuals were randomly reassigned to enclosures to
avoid confounding cage and social effects among cage
mates.
Each tortoise was weighed and measured (midline cara-
pace and plastron lengths) at 10 d intervals for 110 d post-
hatching. A final set of mass and length measurements were
recorded 330 d posthatching, after which tortoises were re-
turned to Riparian Farms.
Metabolic rate
Oxygen consumption was measured via closed-system re-
spirometry (Vleck 1987) and used to calculate MR (Peterson
1990). Oxygen consumption rates of each tortoise were
measured three times at all five Tinc (28.5, 29.5, 30.5, 31.5,
and 32.5 8C): 1–3 d after internalizing yolk and on days 40
and 100 posthatching.
Metabolic chambers were constructed from 169, 322, and
959 mL plastic jars with screw-on lids. A stopcock was in-
serted through the lid of each jar, and a thin film of vacuum
grease was applied on the inside of each lid to ensure an air-
tight seal when the stopcock was closed.
Prior to each MR measurement, tortoises were fasted for
4 d to minimize metabolic costs associated with specific dy-
namic action and growth. On the day of measurement, each
hatchling was weighed and placed into a chamber. Cham-
bers were then placed inside an incubator for 1.5–2 h to al-
low body temperatures to stabilize. With the overhead lights
off to minimize disturbance to the tortoises, chambers were
removed from the incubator. After screwing lids on to create
an airtight seal, pretrial air samples were drawn into 30 mL
syringes, also equipped with stopcocks. The stopcocks on
the syringe and chamber were then closed, the syringes re-
moved, and the time of sampling recorded. Chambers were
then placed back into the dark incubator and removed after
56–85 min, when post-trial samples were drawn from the
stopcock after pumping each syringe several times to ensure
mixing of the air inside. After the final measurement at a
given temperature, the tortoises were moved to another tem-
perature and the process was repeated. Measurements at all
five temperatures were conducted over 1–2 d for each tor-
toise.
Oxygen concentrations of all air samples were analyzed in
10 mL aliquots with a Sable Systems FC-1 oxygen analyzer.
Air was drawn from outside the building at a regulated flow
rate of 100 mL/min and through serial columns of Drierite
and Ascarite to remove water and CO2, respectively. Each
aliquot was injected into the air stream, which passed
through another small column of Drierite and Ascarite and
then through the oxygen analyzer. Oxygen consumption by
each turtle was calculated as the difference between initial
and final volumes of oxygen after correcting for chamber
volume (Peterson 1990).
Thermoregulation
Preferred temperature of tortoises from two clutches
(clutch 5: n= 15; clutch 6: n= 17) were measured in a
three-lane linear thermal gradient. Each lane measured
124 cm 25 cm 63 cm (length width height). The
floor of the gradient was constructed from a 0.6 cm thick
aluminum plate. The plate was bent down at a right angle
20 cm from one end and the resulting vertical tab was in-
serted into an insulated box of ice water. Three 250 W in-
candescent lights were placed under the opposite end of the
plate to create a 14–44 8C surface gradient. Fluorescent
lights set on a 12 h light : 12 h dark cycle were placed
above each lane to produce an evenly lit environment.
All measurements were conducted 136–144 d posthatch-
ing so that disturbance to the animals did not coincide with
growth measurements. After fasting for 4 d to eliminate
postprandial effects on temperature preference, tortoises
were outfitted with a thermocouple (see below) and placed
individually in a lane of the gradient between 1200 and
1400. They then remained in the gradient for ~46 h. The
first 6–8 h of data were discarded to provide a period of ac-
climation to the gradient. Data from the first dark cycle
(2000–0800) were analyzed separately from the proceeding
light cycle (0830–2000).
Body temperature (Tb) was measured by inserting a preca-
librated 18-G thermocouple 2–3 cm into the cloaca of each
tortoise. The exposed tip of each thermocouple was coated
with a small bulb of epoxy to produce a smooth surface and
ease insertion. After insertion, thermocouples were secured
with epoxy applied to the posterior carapace and then con-
nected to temperature data loggers (Onset Comp. Corp., Po-
casset, Massachusetts) that were allowed to slide on hooks
along monofilament strung longitudinally above each lane
of the gradient, thereby permitting each tortoise free move-
ment without risk of getting entangled in the thermocouple.
Loggers were programmed to record temperature at 1 min
intervals. After data were downloaded from the loggers,
mean temperatures were calculated for each 30 min interval
and mean Tbwas calculated for each turtle during each
cycle.
Sex determination
Individuals were sexed when 152–235 d old via laparo-
66 Can. J. Zool. Vol. 87, 2009
Published by NRC Research Press
scopic surgery (Rostal et al. 1994). Food was withheld for
4 d prior to the procedure and tortoises were soaked daily
to stimulate elimination of feces. The tortoises were trans-
ferred to the veterinary facilities at the Tulsa Zoo & Living
Museum, Tulsa, Oklahoma, on the mornings that surgeries
were performed. General anesthesia was achieved with a
cocktail of 10 mg/kg Ketamine and 0.1 mg/kg Medetomi-
dine. After cleaning the site with Chlorhexidine, a 3–4 mm
incision was made posterior to the plastral bridge. A 1.9 mm
laparoscope probe was inserted into the incision and the go-
nads were identified visually. Two observers examined each
individual and agreement was 100%. The gonads of G. sul-
cata in this age range were very distinct; ovaries appear as
long ribbons of white follicles, whereas testes are darker,
less textured, and highly vascularized. Incisions were closed
with a suture, and Atipamezole (0.15 mg/kg) was used as a
reversing agent for the Medetomidine in three cases where
individuals did not show signs of activity within 1 h after
surgery.
Statistics
All mass and length data were log-transformed prior to
analyses to improve the distribution of the data and the ho-
mogeneity of variances. Appropriate assumptions of para-
metric statistics were verified for all tests, and
nonparametric alternatives were used when assumptions of
normality and homogeneity of variance were not supported
(see Results).
We analyzed size using two different methods: first, by
comparing log-transformed mass, and second, by comparing
length–mass residuals among treatments. These size meas-
urements, along with maximum growth rate, were analyzed
using analysis of covariance (ANCOVA), with Tinc as a
fixed factor and clutch as a random factor. Egg mass was
used as a covariate for hatchling mass and mass at hatching
was used as a covariate for analyzing maximum growth rate.
Interactions between the covariate and main effects were
tested to confirm homogeneity of slopes among treatments
(P> 0.05). These interaction terms were then removed for
final analyses.
Metabolic compensation to incubation temperature was
tested in a repeated-measures analysis, with Tinc as a fixed
effect, mass as a covariate, and MR repeated on each tor-
toise over the five measurement temperatures. Additionally,
Q10 values (using MR data from 28.5 and 32.5 8C) were cal-
culated for each tortoise at each of the three MR measure-
ment intervals. The Q10 value was then analyzed in a
repeated-measures design with Tinc as a fixed effect, clutch
as a random effect, and the Q10 value repeated for each tor-
toise over the three measurement periods.
Similarly, effects on growth of Tinc and clutch were ana-
lyzed in a repeated-measures ANCOVA model, with mass
repeated over 11 time intervals and egg mass as a covariate.
Tpref was analyzed by first calculating day-time and night-
time mean Tpref for each tortoise. Day- and night-time means
were then used as response variables in separate repeated-
measures ANOVAs with clutch and Tinc as independent var-
iables and Tpref repeated for each tortoise.
Results
Eggs
We randomly distributed 107 eggs comprising six
clutches (clutch size range = 15–21 eggs; Table 1) among
the five incubation treatments. Egg mass was 51.4 ± 0.6 g
(mean ± 1SE) (range = 37.2–62.4 g) and varied among
clutches (Kruskal–Wallis one-way ANOVA on ranks: H=
89.64, P£0.001; Table 1). Of the 107 eggs incubated, 82
(77%) hatched; however, three tortoises that refused to eat
and ultimately died were excluded from subsequent analy-
ses, reducing our hatchling sample size to 79. Hatching suc-
cess varied among clutches (c2
½5= 39.049, P< 0.001) but
was unaffected by Tinc (c2
½4= 0.658, P= 0.956, respec-
tively). Clutches 2 and 4 exhibited low hatching success rel-
ative to the other four clutches (Table 1). These eggs were
not included in subsequent analyses of prehatching end-
points, reducing our egg sample size to four clutches of 71
viable eggs.
Incubation duration was 72–116 d and was inversely cor-
related with Tinc (Fig. 1). The relationship best fit a quad-
ratic curve, indicating that one-unit increases at higher
temperatures influenced duration less than increases of simi-
lar magnitude at lower temperatures.
Hatchlings
Hatchling mass varied among clutches (Table 1).
Log(hatchling mass) correlated positively with log(egg
mass) (Figs. 2, 3A), and mass-conversion efficiency varied
among clutches (Figs. 3A, 3B). The correlation with egg
mass remained significant for up to (and likely beyond)
330 d, but the coefficient of determination decreased with
time (P< 0.001, r2range = 0.10–0.85; Fig. 2). Tortoises
grew 0.471 ± 0.014 g/d. Incubation temperature had no ef-
fect on growth rate (F[4,50] = 1.54, P= 0.205; Figs. 4A,
4B), whereas clutch did, even after controlling for differen-
ces in initial egg mass (F[5,51] = 11.49, P< 0.001). The
Table 1. Clutch size (n), hatching success, and egg and hatchling size for six clutches of African spurred tortoise (Geochelone sul-
cata).
Egg mass (g) Hatchling mass (g) Carapace length (mm) Plastron length (mm)
Clutch nHatching success (%) Mean SE Mean SE Mean SE Mean SE
1 16 94 40.46 0.33 32.87 0.53 46.52 0.25 42.30 0.31
2 15 40 46.36 0.81 37.19 1.21 49.00 0.75 46.37 0.75
3 20 85 57.38 0.30 33.25 0.28 46.45 0.16 42.13 0.19
4 15 34 53.42 0.26 43.38 0.50 51.68 0.21 47.58 0.34
5 21 95 50.14 0.43 38.50 0.46 51.50 0.26 48.20 0.36
6 20 95 55.74 0.77 47.15 0.72 52.31 0.40 48.98 0.53
Ligon et al. 67
Published by NRC Research Press
maximum growth rate calculated over a 10 d interval was
1.082 ± 0.038 g/d and was unaffected by Tinc or clutch.
Sex determination
Sex was verified for all hatched tortoises. As has been ob-
served in other testudines (Spotila et al. 1994; Eendeback
1995; Burke et al. 1996; Demuth 2001; Ewert et al. 2004),
males were produced at low temperatures and females at
high temperatures. A mixed sex ratio was observed at two
intermediate temperatures, and although no single tempera-
ture in the study produced a balanced sex ratio, interpolation
from the data suggests an interclutch pivotal Tinc of 30.8 8C
(Fig. 5).
Metabolic rate
Hatchling MR exhibited a positive correlation with ambi-
ent temperature (F[4,304] = 175.12, P< 0.001). However,
hatchlings did not exhibit metabolic compensation following
acclimation to a constant incubation temperature (F[16,304] =
1.31, P= 0.1861; Fig. 6). The Q10 value was unaffected by
Tinc (F[4,135] = 0.30, P= 0.8749) but was higher 5 d post-
hatching than at 40 or 100 d posthatching (2.75 ± 0.07,
2.38 ± 0.08, 2.32 ± 0.07, respectively; F[2,143] = 9.64, P<
0.0001). There was a maternal effect (F[3,135] = 3.91, P=
0.0102), owing to consistently higher Q10 values among
clutch 3 hatchlings compared with other clutches.
Temperature regulation
Tinc had no effect on either daytime or nighttime Tpref,or
on the precision with which tortoises maintained a consistent
body temperature (Table 2). However, tortoises from clutch
5 consistently selected lower body temperatures than did
those from clutch 6 (daytime: 23.5 vs. 30.2 8C; nighttime:
23.6 vs. 31.0 8C), and tortoises from clutch 6 maintained
more consistent body temperatures then did those from
clutch 5 (Table 2).
Discussion
Tinc strongly affected incubation duration and gonadal dif-
ferentiation. However, within the range of temperatures
tested, Tinc had no discernable effect on egg viability, hatch-
ling mass, posthatching growth, MR, Q10,orTpref. The ef-
fects of Tinc on hatchling size and growth reported in
previous chelonian studies have been variable, even within
individual species, ranging from approximately linear posi-
tive relationships between temperature and size (Packard et
al. 1987; Rhen and Lang 1995) to no observed temperature
Fig. 1. Relationship between constant incubation temperatures and
incubation duration in African spurred tortoise (Geochelone sul-
cata). Error bars are ±1 SE.
Fig. 2. Changes with age in the relationship between initial egg
mass and hatchling mass of African spurred tortoise (Geochelone
sulcata). Note logarithmic axes.
Fig. 3. (A) Least-squares regression between initial egg mass and
hatchling mass of African spurred tortoise (Geochelone sulcata),
fitting the equation log(hatchling mass) = –0.008
[log(egg mass)]0.952. Broken lines are 95% confidence intervals.
Note logarithmic axes. (B) Differences in hatchling mass among
different clutches. Values were calculated from log–log residuals to
correct for differences in initial egg mass. Lowercased letters sig-
nify differences among clutches (a= 0.05). Error bars are ±1 SE.
68 Can. J. Zool. Vol. 87, 2009
Published by NRC Research Press
effect (Janzen 1995; Steyermark and Spotila 2001). A num-
ber of studies have reported little effect of Tinc on hatchling
size or growth rate except at extreme temperatures that also
negatively affected egg viability, indicating that those tem-
peratures pushed the limits that embryos can physiologically
tolerate (Brooks et al. 1991; McKnight and Gutzke 1993;
Bobyn and Brooks 1994; Spotila et al. 1994; Du and Ji
2003). We attempted to select temperatures that straddled
the pivotal temperature for G. sulcata, but limited the range
to that expected to produce high hatching rates (Richard
Fife, Riparian Farms, personal communication).
Fewer studies have examined Tinc effects on MR in rep-
tiles, but there is still wide variation in observed results.
The common snapping turtle (Chelydra serpentina (L.,
1758)) has been the subject of two such studies. Measure-
ments conducted shortly after hatching found a negative cor-
relation between Tinc and resting metabolic rate (O’Steen
and Janzen 1999), whereas the effect of Tinc on MR pro-
duced inconsistent patterns among 6-month-old juveniles
(Steyermark and Spotila 2000). We share Steyermark and
Spotila’s (2000) view that Tinc may act in a similar fashion
to acclimation temperature so that Tinc effects can be inter-
preted as acclimation effects. Therefore, it is not surprising
that in one study hatchling C. serpentina exhibited a MR
pattern consistent with positive compensation (O’Steen and
Janzen 1999), but that juveniles 6 months removed from the
acclimation period did not (Steyermark and Spotila 2000).
Geochelone sulcata showed a predictable positive correla-
tion between MR and acute body temperature in this study,
but there was no evidence of metabolic compensation to Tinc
immediately after hatching, or 40–100 d posthatching. Q10
values at all ages fell within the generally predicted 2–3
range, and was consistent at 40 and 100 d, but lower at
Fig. 4. (A) Growth of juvenile African spurred tortoise (Geoche-
lone sulcata) incubated at different temperatures. (B) Growth rate
calculated over 10 d intervals. Lower growth rates at 40 and 100 d
correspond with fasting periods prior to measuring metabolic rate.
Error bars are ±1 SE.
Fig. 5. Relationship between incubation temperature and sex ratio,
presented as percent male African spurred tortoise (Geochelone
sulcata). Values in parentheses are sample sizes; broken lines indi-
cate the interpolated pivotal temperature.
Fig. 6. Changes in resting metabolic rates (O2consumption) with
temperature of hatchling African spurred tortoise (Geochelone sul-
cata) incubated at five different temperatures. Error bars are ±1 SE.
Table 2. Incubation temperature and clutch effects on preferred body temperature of juvenile African spurred
tortoise (Geochelone sulcata).
Tinc Clutch Tinc clutch
Light cycle F df PFdf PFdf P
Daytime
Tpref (mean) 0.61 4, 21 0.657 9.23 1, 21 0.006 1.08 4, 21 0.390
Tpref (precision) 2.56 4, 21 0.069 4.57 1, 21 0.045 1.70 4, 21 0.187
Nighttime
Tpref (mean) 1.26 4, 21 0.317 8.90 1, 21 0.007 1.36 4, 21 0.279
Tpref (precision) 1.78 4, 21 0.168 7.85 1, 21 0.010 0.85 4, 21 0.510
Note: Standard deviations calculated from data acquired at 15 min intervals were used as measures of the precision with
which tortoises thermoregulated. Values in boldfaced type indicate statistically significant relationships.
Ligon et al. 69
Published by NRC Research Press
both later stages compared with that measured at hatching.
This temporal difference may stem from differences in di-
gestive state: hatchling tortoises were subsisting on stored
yolk reserves at the time of initial MR measurements,
whereas they were in a postabsorptive state following a pe-
riod of fasting during subsequent measurements. These re-
sults suggest that biochemical processes associated with
digestion may be affected by temperature to a greater degree
than those contributing to maintenance metabolism, thereby
producing elevated whole-animal MR.
Differences owing to maternity were more prevalent than
Tinc effects. Hatching success was low for two out of six
clutches, and egg and hatchling mass differed significantly
among clutches. There was a consistent correlation among
clutches between egg and hatchling mass, and clutch 3,
which was composed of the largest eggs, produced the larg-
est hatchlings. However, clutch 3 did not exhibit the greatest
egg mass-to-hatchling mass conversion efficiency, indicating
that, in addition to the quantity of material allocated to each
egg, females may vary the ratios or quality of egg compo-
nents and thereby influence hatchling size.
Finally, clutch, but not Tinc, strongly influenced Tpref. The
magnitude of difference between clutches 5 and 6 were dra-
matic and consistent during both daytime and nighttime
measurement intervals. It is surprising that, despite a nearly
78C difference in Tpref, tortoises in the two clutches for
which this endpoint was measured did not differ in other re-
spects such as growth rate or temperature-specific metabolic
rate. Unfortunately, without data from additional clutches it
is impossible to know whether tortoises from clutch 6 exhib-
ited preferences dramatically higher than typical, or if tor-
toises from clutch 5 selected unusually low temperatures. It
is possible that tortoises from clutch 6 had a mild infection
and were therefore maintaining chronic fever. However, this
seems unlikely for several reasons. First, steps were taken to
ensure that individual tortoises from different incubation
temperatures and clutches were treated equally; tortoises
were distributed randomly among enclosures, and cage
groups were redistributed frequently, thereby minimizing
differences owing to extrinsic factors. Second, interclutch
Tpref differences were consistent; therefore, there would
have to have been an extremely high clutch-infection corre-
lation. Finally, no other endpoints suggested differences in
the health status of tortoises in different clutches.
An alternative — and perhaps more likely — explanation
for the Tpref differences is that clutches 5 and 6 could have
been produced by dams that originated from different eco-
types of G. sulcata. Based on what little is known about the
natural history of G. sulcata and the climate to which it is
presently accustomed, it seems likely that daytime tempera-
tures of 30 8C are more commonly encountered in situ than
temperatures in the low to mid-20 8C range. There is no
published evidence of population-level variation in physiol-
ogy in the species, but this may be more indicative of the
general paucity of data from natural populations of G. sul-
cata than an absence of such differences, and investigations
of interpopulation variation are certainly warranted. Given
the rapid climatic and vegetational changes that have oc-
curred in the Sahel region to which this species is native
(Nicholson 2000, 2001), the possible presence of genetically
based differences in thermal optima may benefit the long-
term survival of the species. However, given the long gener-
ation time of G. sulcata and the rapid pace at which climatic
changes have occurred in the Sahel (Nicholson 2000), the
apparent resilience to changes in incubation temperature
that this species exhibited in this study may turn out to be
of greater importance than variation in posthatching Tpref to
the long-term survival of the species.
Our results do not support the hypothesis that TSD is ac-
tively maintained in this species because of males and fe-
males benefiting differently from development at different
temperatures. Recent investigations into the evolutionary
adaptiveness of TSD in a short-lived lizard revealed differ-
ential fitness owing to incubation temperature (Warner and
Shine 2008). However, similarly conclusive results remain
elusive for turtles, owing in part to the difficulty of measur-
ing lifetime fitness of long-lived organisms. Alternatively, it
is possible that more extreme incubation temperatures would
have produced measurable temperature differences likely to
correspond with fitness; however, the range that we used ex-
tended 2–3 8C on either side of the pivotal temperature for
the species, and thus should have been sufficiently broad to
reveal any ecologically relevant effects that existed. This
leaves two possibilities: (1) that TSD in G. sulcata is no lon-
ger evolutionarily adaptive but persists because selection
against it and in favor of other SDM’s is weak, or (2) that
TSD is an adaptive trait but for reasons not elucidated by
this study. Differentiating between these two competing hy-
potheses remains the challenge for investigators of TSD in
this and other chelonian species.
Acknowledgements
All procedures for this research were approved by the
Oklahoma State University Institutional Animal Care and
Use Committee (protocol #AS023), and were in compliance
with animal care guidelines described in the 7th edition of
the Guide for the Care and Use of Laboratory Animals (In-
stitute of Laboratory Animal Research, Commission on Life
Sciences, National Research Council, Washington, D.C.).
We thank Richard Fife at Riparian Farms for generously
contributing eggs for our research, and Gage Gregory, Elicia
Ligon, David Roper, and Shana Watkins for assisting with
animal care and data collection. Dr. Kay Backues, DVM,
volunteered a great deal of time and expertise to sex the ju-
venile tortoises. Earlier drafts of the manuscript benefited
greatly from comments by Stanley Fox, Mark Payton, and
two anonymous reviewers. Finally, financial support was
provided by the Oklahoma State University Environmental
Institute.
References
Bobyn, M.L., and Brooks, R.J. 1994. Interclutch and interpopula-
tion variation in the effects of incubation conditions in sex, sur-
vival and growth of hatchling turtles (Chelydra serpentina). J.
Zool. (Lond.), 233: 233–257.
Bonin, F., Devaux, B., and Dupre
´, A. 2006. Turtles of the World.
The Johns Hopkins University Press, Baltimore, Md.
Brooks, R.J., Bobyn, M.L., Galbraith, D.A., Layfield, J.A., and
Nancekivell, E.G. 1991. Maternal and environmental influences
on growth and survival of embryonic and hatchling snapping
turtles (Chelydra serpentina). Can. J. Zool. 69: 2667–2676.
doi:10.1139/z91-375.
70 Can. J. Zool. Vol. 87, 2009
Published by NRC Research Press
Bull, J.J. 1980. Sex determination in reptiles. Q. Rev. Biol. 55:3–
21. doi:10.1086/411613.
Bull, J.J., and Charnov, E.L. 1989. Enigmatic reptilian sex ratios.
Evolution, 43: 1561–1566. doi:10.2307/2409470.
Bull, J.J., and Vogt, R.C. 1979. Temperature-dependent sex deter-
mination in turtles. Science (Washington, D.C.), 206: 1186–
1188. doi:10.1126/science.505003. PMID:505003.
Burke, R.L., Ewert, M.A., McLemore, J.B., and Jackson, D.R.
1996. Temperature-dependent sex determination and hatching
success in the gopher tortoise (Gopherus polyphemus). Chelo-
nian Conserv. Biol. 2: 86–88.
Ciofi, C., and Swingland, I.R. 1997. Environmental sex determina-
tion in reptiles. Appl. Anim. Behav. Sci. 51: 251–265. doi:10.
1016/S0168-1591(96)01108-2.
CITES. 2007. Appendices I, II and III [online]. Convention on In-
ternational Trade in Endangered Species of Wild Fauna and
Flora (CITES) Secretariat, Geneva, Switzerland. Available from
http://www.cites.org/eng/app/appendices.shtml [accessed 12 Feb-
ruary 2008].
Conover, D.O. 1984. Adaptive significance of temperature-dependent
sex determination in a fish. Am. Nat. 123: 297–313. doi:10.
1086/284205.
Conover, D.O., and Kynard, B.E. 1981. Environmental sex deter-
mination: interaction of temperature and genotype in fish.
Science (Washington, D.C.), 213: 577–579. doi:10.1126/science.
213.4507.577. PMID:17794845.
Demuth, J.P. 2001. The effects of constant and fluctuating incuba-
tion temperatures on sex determination, growth, and perfor-
mance in the tortoise Gopherus polyphemus. Can. J. Zool. 79:
1609–1620. doi:10.1139/cjz-79-9-1609.
Du, W., and Ji, X. 2003. The effects of incubation thermal environ-
ments on size, locomotor performance and early growth of
hatchling soft-shelled turtles, Pelodiscus sinensis. J. Therm.
Biol. 28: 279–286. doi:10.1016/S0306-4565(03)00003-2.
Eendeback, B.T. 1995. Incubation period and sex ratio of Herman’s
Tortoise, Testudo hermanni boettgeri. Chelonian Conserv. Biol.
1: 227–231.
Ernst, C.H., and Barbour, R.W. 1989. Turtles of the World. Smith-
sonian Institution Press, Washington, D.C.
Ewert, M.A., and Nelson, C.E. 1991. Sex determination in turtles:
diverse patterns and some possible adaptive values. Copeia,
1991: 50–69. doi:10.2307/1446248.
Ewert, M.A., Jackson, D.R., and Nelson, C.E. 1994. Patterns of
temperature-dependent sex determination in turtles. J. Exp.
Zool. 270: 3–15. doi:10.1002/jez.1402700103.
Ewert, M.A., Hatcher, R.E., and Goode, J.M. 2004. Sex determina-
tion and ontogeny in Malacochersus tornieri, the pancake tor-
toise. J. Herpetol. 38: 291–295. doi:10.1670/149-03N.
Gettinger, R.D., Paukstis, G.L., and Gutzke, W.H.N. 1984. Influ-
ence of hydric environment on oxygen consumption by embryo-
nic turtles Chelydra serpentina and Trionyx spiniferus. Physiol.
Zool. 57: 468–473.
Gutzke, W.H.N., and Packard, G.C. 1987. The influence of
temperature on eggs and hatchlings of Blanding’s turtles,
Emydoidea blandingii. J. Herpetol. 21: 161–163. doi:10.2307/
1564476.
Gutzke, W.H.N., and Paukstis, G.L. 1983. Influence of the hydric
environment on sexual differentiation in turtles. J. Exp. Zool.
226: 467–469. doi:10.1002/jez.1402260317. PMID:6886668.
Gutzke, W.H.N., Packard, G.C., Packard, M.J., and Boardman, T.J.
1987. Influence of the hydric and thermal environments on eggs
and hatchlings of painted turtles (Chrysemys picta). Herpetolo-
gica, 43: 393–404.
Janzen, F.J. 1995. Experimental evidence for the evolutionary sig-
nificance of temperature-dependent sex determination. Evolu-
tion, 49: 864–873. doi:10.2307/2410409.
Janzen, F.J., and Krenz, J.G. 2004. Phylogenetics: which was first,
TSD or GSD? In Temperature-dependent sex determination in
vertebrates. Edited by N. Valenzuela and V. Lance. Smithsonian
Books, Washington, D.C. pp. 121–130.
Janzen, F.J., and Paukstis, G.L. 1991. Environmental sex determi-
nation in reptiles: ecology, evolution, and experimental design.
Q. Rev. Biol. 66: 149–177. doi:10.1086/417143. PMID:1891591.
McKnight, C.M., and Gutzke, W.H.N. 1993. Effects of embryonic
environment and of hatchling housing conditions on growth of
young snapping turtles (Chelydra serpentina). Copeia, 1993:
475–482. doi:10.2307/1447148.
Miller, K., Packard, G.C., and Packard, M.J. 1987. Hydric condi-
tions during incubation influence locomotor performance of
hatchling snapping turtles. J. Exp. Biol. 127: 401–412.
Morris, K.A., Packard, G.C., Boardman, T.J., Paukstis, G.L., and
Packard, M.J. 1983. Effect of the hydric environment on growth
of embryonic snapping turtles (Chelydra serpentina). Herpetolo-
gica, 39: 272–285.
Naylor, C., Adams, J., and Greenwood, P.J. 1988. Variation in sex
determination in natural populations of a shrimp. J. Evol. Biol.
1: 355–368. doi:10.1046/j.1420-9101.1988.1040355.x.
Nicholson, S. 2000. Land surface processes and Sahel climate. Rev.
Geophys. 38: 117–139. doi:10.1029/1999RG900014.
Nicholson, S.E. 2001. Climatic and environmental change in Africa
during the last two centuries. Clim. Res. 17: 123–144. doi:10.
3354/cr017123.
O’Steen, S., and Janzen, F.J. 1999. Embryonic temperature affects
metabolic compensation and thyroid hormones in hatchling
snapping turtles. Physiol. Biochem. Zool. 72: 520–533. doi:10.
1086/316690. PMID:10521320.
Packard, G.C., and Packard, M.J. 1988. Water relations of embryo-
nic snapping turtles (Chelydra serpentina) exposed to wet or dry
environments at different times in incubation. Physiol. Zool. 61:
95–106.
Packard, G.C., Packard, M.J., Miller, K., and Boardman, T.J. 1987.
Influence of moisture, temperature, and substrate on snapping
turtle eggs and embryos. Ecology, 68: 983–993. doi:10.2307/
1938369.
Peterson, C.C. 1990. Paradoxically low metabolic rate of the diur-
nal gecko Rhoptropus afer. Copeia, 1990: 233–237. doi:10.2307/
1445841.
Rhen, T., and Lang, J.W. 1995. Phenotypic plasticity for growth in
the common snapping turtle: effects of incubation temperature,
clutch, and their interaction. Am. Nat. 146: 726–747. doi:10.
1086/285822.
Rhen, T., and Lang, J.W. 2004. Phenotypic effects of incubation
temperature in reptiles. In Temperature-dependent sex determi-
nation in vertebrates. Edited by N. Valenzuela and V. Lance.
Smithsonian Books, Washington, D.C. pp. 90–98.
Rostal, D.C., Grumbles, J.S., Lance, V.A., and Spotila, J.R. 1994.
Non-lethal sexing techniques for hatchling and immature desert
tortoises (Gopherus agassizii). Herpetol. Monogr. 8: 83–87.
doi:10.2307/1467072.
Shine, R. 1999. Why is sex determined by temperature in many
reptiles? Trends Ecol. Evol. 14: 186–188. doi:10.1016/S0169-
5347(98)01575-4. PMID:10322531.
Spotila, J.R., Zimmerman, L.C., Binkley, C.A., Grumbles, J.S.,
Rostal, D.C., List, A., Breyer, E.C., Phillips, K.M., and Kemp,
S.J. 1994. Effects of incubation conditions on sex determination,
hatching success, and growth of hatchling desert tortoises, Go-
pherus agassizii. Herpetol. Monogr. 8: 103–116. doi:10.2307/
1467074.
Ligon et al. 71
Published by NRC Research Press
Steyermark, A.C., and Spotila, J.R. 2000. Effects of maternal iden-
tity and incubation temperature on snapping turtle (Chelydra
serpentina) metabolism. Physiol. Biochem. Zool. 73: 298–306.
doi:10.1086/316743. PMID:10893169.
Steyermark, A.C., and Spotila, J.R. 2001. Effects of maternal iden-
tity and incubation temperature on snapping turtle (Chelydra
serpentina) growth. Funct. Ecol. 15: 624–632. doi:10.1046/j.
0269-8463.2001.00564.x.
Vleck, D. 1987. Measurement of O2consumption, CO2production,
and water vapor production in a closed system. J. Appl. Physiol.
62: 2103–2106. doi:10.1063/1.339528. PMID:3110127.
Warner, D.A., and Shine, R. 2008. The adaptive significance
of temperature-dependent sex determination in a reptile.
Nature (London), 451: 566–568. doi:10.1038/nature06519.
PMID:18204437.
Wilhoft, D.C., Hotaling, E., and Franks, P. 1983. Effects of
temperature on sex determination in embryos of the snapping
turtle, Chelydra serpentina. J. Herpetol. 17: 38–42. doi:10.2307/
1563778.
Yntema, C.L. 1976. Effects of incubation temperatures on sexual
differentiation in a turtle, Chelydra serpentina. J. Morphol. 150:
453–461. doi:10.1002/jmor.1051500212.
Yntema, C.L. 1978. Incubation times for eggs of turtle Chelydra
serpentina (Testudines: Chelydridae) at various temperatures.
Herpetologica, 34: 274–277.
72 Can. J. Zool. Vol. 87, 2009
Published by NRC Research Press
