Accessibility Page Navigation
Style sheets must be enabled to view this page as it was intended.
For tortoise, terrapin and turtle care and conservation



Department of Zoology, University of Western Australia, Perth, W.A. 6907, Australia.

Article originally published in Herpetologica 54(l): 103-112 and reproduced with permission from the Herpetologists' League, Inc.


Pseudemydura umbrina is active only during winter and spring when environmental temperatures are relatively low and the ephemeral swamps that it inhabits contain water. We used operative environmental temperature thermometers to define the microclimates available to P umbrina in its natural habitat during late winter and early spring, and we monitored the use of these microclimates by seven adult tortoises using radio-telemetry. Although air and water temperatures may be as low as 12-14°C during the day, we found that microclimates available to the animals have operative temperatures of up to 35-44°C out of water and 27°C in water. All animals increased in body mass during this study, and females increased in body mass faster than mates. Tortoises that spent more time in warm water increased their body mass faster than those that spent more time in cold water. During the cool time of the year, P. umbrina enhances its physiological condition by using warmer microclimates in the range of potential available microclimates.

The western swamp tortoise, Pseudemydura umbrina (Chelidae: Pleurodira), is the only extant member of the sub-family Pseudemydurinae and the world's rarest chelonian. Its distribution is restricted to the Swan coastal plain in Western Australia. The climate here is Mediterranean with cool, wet winters and hot, dry summers (Burbidge, 1981 ). At the time of this study, the only viable wild population was in the Ellen Brook Nature Reserve.

Pseudemydura umbrina inhabits ephemeral swamps that are filled with water only during winter and spring (i.e., June-November). When the swamps dry or water temperatures rise above 28°C in late spring, tortoises leave the water and move to aestivation sites in underground holes or beneath leaf litter (Burbidge, 1981). They remain in aestivation until the following autumn (May-June), when the swamps begin to fill again. This species therefore has an activity period limited to the cooler time of the year; feeding and mating occur only in water and are restricted to periods when the swamps contain water. Previous studies investigating climatic constraints on chelonian activity have mainly involved species that are active during the warm seasons and hibernate throughout winter: e.g., Pseudemys scripta (Crawford et al., 1983; Spotila et al., 1989), Chelydra serpentine (Brown et al., 1990; Williamson et al., 1989), and Chrysemys picta (St. Clair and Gregory, 1990; Taylor and Nol, 1989).

Being a semi-aquatic ectothermic vertebrate, the body temperature of P umbrina is largely determined by the temperature of the environment in which the animal is situated. Basking is the primary thermoregulatory behaviour adopted by freshwater turtles to elevate their body temperature, and occurs usually when the ambient temperature is low (Avery, 1982). To determine whether temperature selection and basking are important to these animals, some aspect of their short-term physiological performance must he measured. The rate of increase in body mass is related to short-term performance

Earlier investigations (Burbridge, 1967; Lucas et al., 1963) found that P umbrina is active and feeds at water temperatures between 14 and 28°C. During much of winter and early spring, water temperatures may be < 14°C, and therefore below the optimal range associated with activity in this species. In this study, we measure the operative environmental temperature of microclimates which are potentially available to P umbrina in its natural habitat by using water-filled, hollow, copper models of the same dimensions as an average-sized adult tortoise. The internal temperature of these models records the operative environmental temperature (Tc) of the microclimate in which the model is situated. The use of these potential microclimates by free-ranging adults of P. umbrina is determined by temperature-sensitive radiotransmitters glued to the carapaces of the tortoises.

The hypothesis is that during the cool time of the year, P umbrina exploits the warmer microclimates that are available to it, either in the water column or by basking out of the water, in order to elevate body temperature. By selecting warm micro climates within the range of potential available microclimates, animals will enhance their temperature-dependent physiological performance which will he reflected by an increase in body mass.

Materials and Methods

Study Area

The study site is a fenced area of 29 ha at Ellen Brook Nature Reserve and has shallow (420 mm maximal depth), ephemeral, winter-wet swamps on clay soil between slightly higher areas of bushland. The swamps are filled by rain and receive little or no surface drainage from outside the fenced area. In most swamps, bushes (Melaleuca lateritia) form an open to dense spreading canopy about 1-2 m above the water level. Various aquatic plants and clumps of sedges occur without forming dense associations. Throughout the study area, the low bush canopy over the flooded areas provides a mosaic of shaded and sunny patches, and potential basking sites (low branches, clay humps which form small islands and sedge tussocks) are common.

Environmental Temperature (Tc)


We constructed 16 hollow, copper models with the same dimensions as an average-sized adult tortoise except that the models were simplified (i.e., they were - basically the shape of the shell of the tortoise without a head or legs). The models were painted matt black to replicate as closely as possible the spectral emissivity of the tortoises themselves, whose carapace colours varied from brown to almost black. We filled each model with water so that the mass of the model was similar to that of an animal. A thermistor was inserted into the centre of the model, and the point of insertion was made water-tight using a cork and silastic, water-proof sealant. The thermistors had previously been calibrated. We had 25 thermistors available to use with the models. We inserted 16 into the centre of the models and glued the remaining thermistors to the upper surface toward the rear of nine models. Thermistors were linked to UNIDATA data loggers and temperatures were recorded every 15 minutes for the period between 26 August and 28 September 1993.

The internal temperature of the models recorded the operative environmental temperature (Tc of the microclimate in which the model was situated. We placed each of the 16 models in a different microclimate. The microclimates selected covered almost all of the range of microclimates potentially available to tortoises at Ellen Brook. The models were either exposed to full sun or full shade (under bushes) and positioned either on the ground or suspended on branches in the water. Using a combination of these factors, 16 different microclimates were defined (Table 1). Shallow water models had their base at a depth of < 10 cm, mid-depth models were at 25 cm, and deep models were at 40 cm or more

Table 1. Table of average and minimum internal temperatures (C) and thermal category for all models between 26th August and 9th September 1993.

Model position/microclimatemax. temp. (°C)Min. Temp. (°C)Thermal category
Dry, on soil, sun35.07.1Bask
Dry, on log, sun27.46.7Bask
Dry, on soil, shade15.511.9(No category)
Protruding, on soil, sun21.011.1Warm
Protruding, on soil, shade16.611.1Cold
protruding, suspended, sun18.312.2Warm
protruding, suspended, shade17.112.6Warm
Shallow, on soil, sun20.311.7Warm
Shallow, on soil, shade16.411.6Cold
Shallow, suspended, sun17.012.3Warm
Mid-depth, on soil, sun15.812.4Cold
Mid-depth, on soil, shade15.013.0Cold
Ad-depth, suspended, sun15.412.5Cold
Deep, on soil, sun14.612.6Cold
Deep, on soil, shade14.813.6Cold

The range of potential available microclimates was determined for the whole period of the study and for the specific days on which animals were observed and monitored. To define selected microhabitats in the range of potential available temperatures, the models were ranked into three categories according to their internal temperatures: basking out of water in the sun (B), warm water (W), and cold water (C). This ranking scheme allowed us to identify the microclimates selected by animals on the days of observation in relation to the different thermal categories.

To give an indication of the change in thermal climate over the period of the study, we plotted the mean temperatures of the three thermal categories, between 0600 h and 1900 h, for each of the 5 weeks of the study.


We individually calibrated temperature- sensitive radio-transmitters (AVM Co. Transmitter type: SM1 Temp. B) with a mercury thermometer in a water bath from 0°C-40°C prior to their use, and also at the end of the study. A transmitter was then fixed to the carapace of each tortoise using Epiglue epoxy resin. We glued the transmitter onto the back of the carapace and fixed the antenna around the edge of the carapace on the marginal scutes. Using a directional antenna and radio-receiver, we recorded the pulse interval of the signal emitted from the transmitter with an AVM electronic pulse-interval timer. The pulse interval was then converted to a temperature reading using the regression equation obtained from the calibration of the specific transmitter being used.

We monitored four female (F 140, F 173, F 191, F 137) and three male tortoises (M I0, M 151, M 144) during this study. The animals were observed in the field over a 3 week period during September 1993. A pulse frequency recording was made for each animal before it was approached and became visible. This ensured that a recording of carapace temperature was made before the animal had been disturbed. The animal was then monitored for several hours, or for the rest of the day, from a position close enough to observe the animal without disturbing it. In practice, this was very difficult because the dense vegetation and murky water in many of the swamps made viewing impossible at distances > I m. In these situations, the pulse frequency from the transmitter was recorded every I0- 15 minutes for several hours and the exact location of the animal was determined every hour. In swamps where an animal could be seen easily, its carapace temperature, position in the water, and any activity was recorded every I0- 15 min. All observations and carapace temperature recordings were made between about 0900 h and 1800 h. Graphs of carapace temperatures were plotted for all animals on the days of observation in relation to the available thermal categories determined from the models on these days. We calculated the time spent by each animal in each of the three thermal categories.

We measured the body mass of each animal at the beginning of the observation period, when it were first fitted with radio transmitters, and again at the end of the observation period. For four animals, this was a I0 week period and for the other three animals this period of time was 6, 5 and 3 weeks. Because we observed animals over different periods of time, the absolute increase in body mass could not be compared; instead, we used the rate of increase in body mass (g/wk) for comparisons. The weekly increase in body mass for each animal was then correlated with temperature selection by that animal.


Potential Available Microclimates

The mean maximum and minimum internal temperatures for each model and their ranking into thermal categories are given in Table 1. Maximum temperatures for deep and mid-depth models ranged from 14.6-15.8°C. The fluctuation in temperature over a 24-hour period for these microclimates was between 1.3°C and 3.4°C; i.e., the deep and mid-water temperatures remained relatively constant. The maximum temperature of models in shallow water in the shade, both on the soil and suspended, was 16.4°C. Maximum temperatures of models with the top protruding both in the sun and the shade varied between 16.6°C and 21.0°C. Temperatures of models situated in the sun in shallow water with the top protruding and out of water all showed a large fluctuation over a 24-hour period. The average maximum temperatures reached in these microclimates were between 17.1°C and 35.0°C.

Graphs of the mean temperatures available in the three categories, for each of the 5 wk of the study, show the increase in thermal climate between the end of August and the end of September (Figure la - e). The upper limit of the range of temperatures available in these microclimates is determined by internal temperatures from the "Dry, on soil, sun" model, and the lower limit defined by the " Deep, on soil, shade" model. The upper limit of the warm water category is determined by the "Protruding, on soil, sun' model internal temperature, and the upper limit of the cold water category defined by the 'Protruding, on soil, shade' model temperature.

Use of Available Temperatures by Animals

On average, each animal was observed and temperature recordings taken for a total of 10 h. All animals were observed both in the morning and afternoon on at least three different days, and some animals were observed over a whole day. Over all days of observations, the range of temperatures that were recorded in the different categories between 0900 hours and 1800 hours were: basking, 16-44°C; warm water, 13-27.5°C; and cold water, 12-19°C. We plotted the carapace temperatures, recorded from the radio-transmitters, on a graph along with the three thermal categories recorded from the models for the specific day of observation The percentage of time spent in each of the three thermal categories was calculated for each animal, using carapace temperature

Body Mass

All the study animals increased their body mass over the period of observation (paired t-value = 4.08, P = 0.0065, df = 6). The average weekly increase in body mass for females and males was 3.8 ± 0.58 and 1.6 ± 0.56 g/wk respectively (Figure 3). The rate of increase in body mass for females was significantly greater than for males (t-value = 2.64, P = 0.0461, df = 5).

All animals used in the study were adults, and none of the animals increased their carapace length over the course of the study. Males had a significantly larger initial carapace length than females (t-value = 7.86, P = 0.0005, df = 5). The regression of initial carapace length against change in body mass was significant (r2 = 0.7 1, P = 0.01 73, df = 6) (Figure 4). Females were smaller and had a greater rate of increase in body mass than males.

Physiological Performance in Relation to Temperature Selection

Stepwise, multiple regression was used to examine the relationship between sex, initial carapace length, and time spent in warm water. Sex was included in the model at step one (F-value 60.74) and time spent in warm water was entered at step two (F-value 29.65); initial carapace length was not included in the model. Multiple regression of sex and the percentage of time spent in warm water against change in body mass was significant (ANOVA F-value = 38.25, P = 0.0025, df = 6). Both sex and time spent in warm water had a significant effect on change in body mass (t-value = 7.793, P = 0.00 1 5; and t-value = 5.445, P = 0.0055, respectively).

Separate regressions of percentage of time spent in warm water against increase in body mass were determined for males and females.The correlation coefficients of both regressions were high (males r2 = 0.963; females r2 = 0.925); however the regression was only significant for females (P = 0.0383, df = 3) and not for males (P = 0. 1 236, df = 2) because of the small number of males used in this study.


The western swamp tortoise is active only for 5-6 months of the year during winter and spring. During this limited period of annual activity, the animals must feed and assimilate enough energy for growth, reproduction, and storage for the time of aestivation. A major part of the annual activity period of P umbrina falls into the cold, wet time of the year, because the ephemeral swamps that it inhabits contain water only from about June-November. During winter, minimum air temperatures in the early morning are sometimes close to 0°C and the lowest maximum daily air temperatures between June and September 1993 were between 13 C and 15 C (Perth Bureau of Meteorology).

Pseudemydura umbrina shows reduced activity at temperatures below about 14°C ( Burbidge, 1966; Lucas et al., 1963).

During winter, P. umbrina spends most time in the water. The internal temperatures of the T c models demonstrate that a wide range of microclimates are available to P. umbrina. The operative environmental temperatures recorded show that should animals choose to use these different microclimates during the day they would be exposed to temperatures ranging from about 12°C up to 44°C, even when air temperatures are low (average maximum shade temperature over the study period was 17.5°C). By exploiting the temperature gradient in the water, animals could expose themselves to temperatures up to 27°C. By basking out of water in the middle of the day, animals could elevate their body temperatures to well above those that they may reach in the water, but our results suggest that basking out of the water is infrequent. Aquatic basking, where animals select appropriate thermal environments in the water, may be more common because animals are able to forage whilst also maintaining an optimal body temperature. Similar results were found for snapping turtles, Chelydra serpentine (Brown et al., 1990). Crawford et al. (1983) suggested that turtles could reduce the amount of time spent thermoregulating by selecting microclimates with high operative environmental temperatures, and that selected microhabitats may represent a compromise between demands, benefits, and costs associated with different activities such as basking and foraging.

Our results generally support this hypothesis. However, the few direct observations of atmospheric basking of P. umbrina in our study may be due to the presence of the observer rather than because this behaviour is rare. The animals were very wary of the observer, and on the few occasions when they were seen basking out of the water, they moved back into the water as soon as the observer was detected. On other occasions, animals probably moved back into the water before they were sighted.

Previous studies have found a strong relationship between transmitter temperature, on the surface of the turtle's carapace, and cloacal temperature when the animals are in water, thus the transmitter temperature seems to be an accurate estimate of body temperature, T b (Brown and Brooks, 1991). However, when the animals are either basking out of water, or in the water with their carapace exposed, the transmitter may register changes in temperature faster than the animal's actual body temperature changes. In these situations, the temperature reported by the transmitter is likely to be an overestimate of body temperature. The body temperature of a turtle with its carapace out of water will be close to or equal to the surface water temperature, because this imposes the upper limit on T b (Chessman, 1987). The main aim of our study was to observe utilization of potential available microclimates by the animals rather than to obtain accurate measurements of body temperatures. Carapace temperatures provide accurate estimates of the temperatures of the surrounding microclimate. Temperatures recorded by the radio-transmitters showed that P umbrina in shallow, sunny water had carapace temperatures within the warm water category as defined by the model temperatures.

All animals used in the study were adults. None of the animals showed an increase in carapace length, however all gained in body mass over the period of observations. Animals were replenishing energy stores that would have been depleted over the previous summer and autumn. When comparing the weekly increase in body mass, all females gained mass at a significantly higher rate than males. Captive females of this species show a marked increase in body mass preceding ovulation, which occurs between late September and early November (Kuchling and Bradshaw, 1993). During late winter and spring, females must assimilate enough energy for growth, storage, and reproduction. Vitellogenesis, the energy allocation to prospective eggs, commences in summer during aestivation. When aestivation ends in late autumn, the ovarian follicles already contain more than half of the egg yolk mass, presumably originating from stored body lipids. If females of P umbrina do not dramatically increase their body mass in early spring, they arrest ovulation and do not reproduce in that year (Kuchling and Bradshaw, 1993). The males show a less dramatic increase in body mass, possibly because the production of sperm requires lower energy investment than the production of eggs. The males were also observed to move over larger distances than females during this time, possibly in search of mates. During winter and early spring searching for mates may be more important for males than feeding and assimilating energy, because mating activity in P umbrina is largely restricted to winter and comes to an end during the second half of September (Kuchling and DeJose, 1989).

Williamson et al. (1989) investigated the effects of temperature on physiological performance in hatchlings and juveniles of C. serpentina. They found rapid growth in young animals kept at 25 C which was related to a high food intake, but no growth occurred at 15°C. This was attributed to a reduced rate of digestive processes at the lower temperature. The present study supports the hypothesis that P umbrina increases physiological condition (reflected by increased body mass) by using warmer microclimates, in the range of potential available microclimates. Animals that spent more time in warm water increased in body mass faster than those that spent more time in cold water.

The increase in body mass of an animal reflects short-term physiological performance, because the overall fitness of an animal increases as a function of the net rate of energy intake (Spotila and Standora, 1985). The greater increase in mass seen in animals selecting the warmer microclimates is most likely related to the increased rate of basic physiological processes, and hence energy assimilation. A Potential problem with using body mass as a factor to reflect physiological performance in chelonians is that the mass of individuals may fluctuate considerably over a short period of time if urine in the bladder or stored water in the lateral bladders is voided (Burbidge, 1967). In our study, all animals were weighed only once at the beginning and once at the end of the observation period, and we handled them carefully. The turtles were weighed immediately after capture and before any water or urine was discharged (often a reaction to handling). In this way, the probability of an error in body mass should be minimal and the same for each individual. We assume that the pattern of change in body mass observed in this study is an adequate indirect measure of the physiological performance of the individuals.

Statistical analysis of much of the data is limited by the small sample size, a factor of the rarity of these animals. Generalizations about temperature selection and physiological performance in this species must therefore be made with caution, because individual variations of behaviour may be inherent rather than a result of the thermal environment to which an individual is exposed. Having stated this, the results do show that P. umbrina exploits the warm microclimates that are available in the natural habitat during winter and early spring. Animals are able to increase the rate of energy assimilation and body mass by exploiting warmer microclimates. The study also shows that during the cool time of the year females assimilate energy faster than males.


The radio-tracking equipment was funded by the Bundesverband far fachgerechten Natur-und Artenschuz, Germany and the British Chelonia Group. This work was partially supported by the World Wide Fund for Nature Australia, the National Endangered Species Program of the Australian Conservation Agency, and the Western Autralian Department of Conservation and Land Management (CALM). We thank A. A. Burbidge, P. Fuller and J. Kinal (CALM) for providing expertise and access to long-term monitoring data of P umbrina. All procedures used on the animals were approved by The University of Western Australia Animal Welfare Committee.

Llterature Cited

Avery, R. A. 1982. Field studies of body temperature and thermoregulation. pp93-166. In C. Gans and F. H. Pough (Eds), Biology of the Reptilia, Vol. 12, Academic Press, London, UK.

Brown, G. P. and Brooks, R. J. 1991. Thermal and behavioural responses to feeding in free-ranging snapping turtles, Chelydra serpentine. J. Herpetol, 25: 273-278.

Brown, G. P., Brooks, R. J. and Layfield, J.A. 1990. Radiotelemetry of body temperatures of free-ranging snapping turtles (Chelydra serpentine)during summer. Can. J. Zool. 68:1659-1663.

Burbidge, A. A. 1967. The Biology of South Western Australian Tortoises. Ph.D. Thesis, University of Western Australia,Perth, Australia. -. 1981. The ecology of the western swamp tortoise, Pseudemydura umbrina (Testudines: Chelidae). Aust. Wildl. Res. 8:203-223.

Chessman, B. C. 1987. Atmospheric and aquatic basking of Australian fresh-water turtle Emydura maquarii (Gray). Herpetologica 43:301-306.

Crawford, K. M., Spotila, J. R. and Standora, E. A. 1983. Operative environmental temperaures and basking behavior of the turtle Pseudemys scripta. Ecology 64:989-999.

Kuchling, G. and Bradshaw, S. D. 1993. Ovarian cycle and egg production of the western swamp tortoise, Pseudemydura umbrina (Testudines: Chelidae) in the wild and in captivity. J. Zool. 229:405-419.

Kuchling, G. and DeJose, J. P. 1989. A captive breeding operation to rescue the critically endangered western swamp turtle Pseudemydura umbrina from extinction. lnt. Zoo Yb. 28:103-109.

Lucas, J. S., Burbidge, A. A., Graham., M. T., Leung, S. C., Lindgren, E. and Philips, B. F. 1963. Ecology and Conservation of Pseudemydura umbrina Siebenrock 1901. Honours Thesis, University of Western Australia, Perth, Australia.

Spotila, J. R., and Standora, E. A. 1985- Energy budgets of ectothermic vertebrates- Am. Zool. 25:973-986.

Testudo Volume Four Number Five 1998