SEARCH

SEARCH BY CITATION

Keywords:

  • behavioural thermoregulation;
  • biologging;
  • energetics;
  • terrestrial turtle

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    It is widely accepted that reptiles are able to regulate behaviourally their body temperature (Tb), but this generalization is primarily based on studies of lizards and snakes in the temperate zone. Because the precision of Tb regulation may vary considerably between taxa and over geographical ranges, studies of semi-terrestrial turtles in climatic extremes are relevant to the understanding of reptilian thermoregulation.
  • 2
    We studied thermoregulation in 21 free-ranging wood turtles (Glyptemys insculpta) at the northern limit of their range in Québec, using miniature data loggers to measure their internal Tb and external temperature (Text) continuously. We simultaneously recorded the available operative environmental temperature (Te) using 23 physical models randomly moved within each habitat type, and we located turtles using radiotelemetry.
  • 3
    The habitat used by wood turtles was thermally constraining and the target temperature (Tset) was only achievable by basking during a short 5-h time window on sunny days. Wood turtles did show thermoregulatory abilities, as determined by the difference between turtle Tb distribution and the null distribution of Te that resulted in Tb closer to Tset. Although most individuals regulated their Tb between 09.00 h and 16.00 h on sunny days, regulation was imprecise, as indicated by an index of thermoregulation precision (| Tb – Tset |).
  • 4
    The comparison of habitat use to availability indicated selection of open habitats. The hourly mean shuttling index (| Text – Tb |) suggested that turtles used sun/shade shuttling from 09.00 to 16.00 h to elevate their Tb above mean Te.
  • 5
    Based on laboratory respirometry data, turtles increased their metabolic rate by 20–26% over thermoconformity, and thus likely increased their energy gain which is assumed to be constrained by processing rate at climatic extremes.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Temperature reflects the average kinetic energy of a biochemical system, so changes in temperature dramatically alter the rate of all physiological functions (Beitinger & Fitzpatrick 1979). For this reason, body temperature (Tb) regulation is inextricably linked to the ecology of ectotherms (Heatwole & Taylor 1987Berman & Quinn 1991). Moreover, understanding how living organisms respond to temperature is a key step to predict the consequences of climate change on populations (Parmesan & Yohe 2003Kearney, Shine & Porter 2009). Thermoregulation should aim to maintain Tb as close as possible to an optimal temperature (ToHuey & Slatkin 1976) where fitness is maximized through improvements in performance (Huey & Kingsolver 1989). Because reptiles generate negligible amounts of body heat and have limited physiological control over Tb, they regulate their Tb largely through habitat selection and behaviour. The exception is one of the largest (> 500 kg) reptile, the leatherback turtle, which can remain warm in near-freezing water (James, Davenport & Hays 2006). Thus, for most reptiles we expect a strong link between habitat use and thermoregulation (Blouin-Demers & Weatherhead 2002), especially at climatic extremes where the environment is constraining (Blouin-Demers & Weatherhead 2001). Here our general goal is to explore the relationship between thermoregulation and habitat use in wood turtles, Glyptemys insculpta (Le Conte, 1830[1829]), near the northern limit of their range.

Reptiles are able to regulate their Tb to some extent (Avery 1982Huey 1982Hertz, Huey & Stevenson 1993), but the precision of Tb regulation varies considerably (Adolph 1990Rummery et al. 1995Dorcas & Peterson 1998). Since the ecological costs and physiological benefits associated with thermoregulation are largely independent (Huey & Slatkin 1976), the evolutionarily optimal temperature may be different from the physiological optimum. Thus, variation in the cost–benefit ratio is assumed to underlie variation in thermoregulatory precision. The cost–benefit model of thermoregulation (Huey & Slatkin 1976) predicts that thermoregulation should cease when the costs outweigh the benefits. Two costs of thermoregulation are predation risk (Pianka & Pianka 1970Huey 1982) and the reduction in time for other nonthermoregulatory activities such as foraging (Huey & Slatkin 1976Gregory, Crampton & Skebo 1999). The primary benefit of thermoregulation is the improvement in physiological performances, which can be grouped in two broad categories: locomotory (Huey 1982Stevenson, Peterson & Tsuji 1985Blouin-Demers, Weatherhead & McCracken 2003) and energy intake (Beaupre, Dunham & Overall 1993van Marken Lichtenbelt, Vogel & Wesselingh 1997Niu, Zhang & Sun 1999).

Energy intake is often more sensitive to Tb than locomotion (Huey 1982Stevenson et al. 1985Angilletta Jr, Hill & Robson 2002) and the benefit of enhanced locomotion is likely to be trivial for turtles that do not use rapid locomotion to capture food or to escape predators (Harding & Bloomer 1979Ernst 2001). Moreover, turtles are long-lived iteroparous organisms with low and unpredictable annual reproductive success and high adult survival (Congdon, Dunham & Sels 1993Ernst, Barbour & Lovich 1994). Fitness is thus likely to be tied more closely to energy gain than to locomotion in turtles. Therefore, the cost–benefit relationship of thermoregulation likely differs for turtles compared to lizards and snakes from which most of our understanding of reptilian thermoregulation is drawn.

The main benefit of thermoregulation for freshwater turtles, increased energy gain, should be higher in climatic extremes. At high latitudes, temperature can be viewed as an ecological resource (Magnuson, Crowder & Medvick 1979Tracy & Christian 1986) that may limit growth and reproduction of ectotherms more than food availability because food processing rate is constrained by low temperatures and short growing seasons (Congdon 1989Grant & Porter 1992Niewiarowski 2001). Indeed, studies have shown that growth of turtles (Koper & Brooks 2000) and snakes (Brown & Weatherhead 2000), and clutch frequency in turtles (Rollinson & Brooks 2007), increase with air temperature (Ta) in northern climates. An increase in energy gain through careful regulation of Tb, metabolism, and food processing capacity should increase allocation to growth and reproduction, allowing for earlier age at maturity and higher fecundity and clutch frequency in turtles. In southern Canada, only 7% of female wood turtles reproduce annually (Foscarini 1994).

To demonstrate thermoregulation, two criteria must be met. First, the distribution of Tb must differ significantly from a null or random distribution which represents Tb of hypothetical thermoconforming animals (Hertz et al. 1993Wills & Beaupre 2000). Second, Tb of the animals must deviate directionally from the null distribution such that Tb approaches the target temperature (Hertz et al. 1993). For instance, behavioural thermoregulation has recently been demonstrated in loggerhead sea turtles at the limit of their breeding range with turtles actively selecting patches of warm water (Schofield et al. 2009). Thermoregulation might similarly be expected in freshwater turtles at climatic extremes. In this study, we used miniature data loggers to measure available operative environmental temperatures (Te), internal Tb, and external carapace temperatures (Text) in wood turtles. We assume that the Tb that turtles seek to maintain (Tset) is 30 °C, based on their choice in a laboratory thermal gradient (Dubois, Blouin-Demers & Thomas 2008). Based on cost–benefit arguments, we predicted that wood turtles would be precise thermoregulators. In addition, there is a strong link between habitat use and thermoregulation (Adolph 1990Grover 1996Blouin-Demers & Weatherhead 2002). Wood turtles prefer open habitats such as alder stands (Kaufmann 1992Compton, Rhymer & McCollough 2002Arvisais et al. 2004) which may be driven by thermoregulation needs (Tuttle 1996Arvisais 2000Compton et al. 2002). We predicted that wood turtles would select open habitats and use sun/shade shuttling to regulate Tb close to Tset.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

study site

We conducted this study during summer 2004 in Brome-Missisquoi County in southern Québec. The study area was a 7-km section of meandering river surrounded by flood plains and higher nonflooded terraces. Half of the study site was open habitat: active or abandoned hay fields, successional fields with herbaceous and shrub vegetation, wetlands, and lawns. The other half was closed woodland habitat dominated by alder (Alnus rugosa) stands or mixed deciduous forest.

capture, radiotelemetry, and body temperature measurement

Turtles were captured by hand and marked with a unique code by carapace notching (Cagle 1939). Adult males were recognized by their concave plastron and longer, thicker pre-anal tail (Lovich, Ernst & McBreen 1990Brooks et al. 1992). Individuals smaller than 160 mm carapace length were considered juveniles.

Seven males, females, and juveniles (21 turtles; 330–1200 g; 132–215 mm carapace length) were equipped with radiotransmitters and temperature data loggers. Turtles were injected with ketamine intramuscularly (40 mg kg−1) 30 min before surgery and were maintained at a surgical plane of anaesthesia with isoflurane. A temperature data logger (Dallas iButton thermochron (Sunnyvale, California, USA) model DS1921L; 3 g) was implanted in the coelomic cavity through an incision made just anterior to the hind leg. After surgery, turtles were kept under observation for 5 to 7 days before release at their capture site. A radiotransmitter (Holohil Systems, Carp, Ontario, Canada, model AI-2F for adults or model SI-2F for juveniles) was affixed with stainless steel bolts to the posterior margin of the carapace for adult males and juveniles, and to the anterior margin for adult females to minimize interference during copulation. A second temperature data logger was mounted on top of the radiotransmitter. The implanted logger was used to measure internal Tb and the second logger to measure Text. All data loggers used in the study were calibrated in a water bath from 5 °C to 50 °C. The protocol was approved by the Animal Care Committee at the Université de Sherbrooke (DT14) and permits were provided by the Société Faune et Parcs du Québec.

telemetry and temperature recording

Turtles were located two to four times weekly from 1 June to 2 August for a total of 478 locations. Each day, the tracking sequence was chosen randomly to ensure that individuals were located at different times from 08.00 to 19.00 h. The implanted data loggers recorded Tb at 30-min intervals from 24 June to 4 August. The data loggers mounted on the radiotransmitters recorded temperature at 10-min intervals and were downloaded ~ every 15 days because of their limited storage capacity (2048 readings).

habitat use

Turtle locations were mapped with a global positioning system (GPS) with 3–10 m accuracy (mean = 8·2 m). During a pilot study, we noted vegetation cover at 452 turtle locations and paired random points within a 50-m radius (see below) of the turtles’ locations to define cover type classes as forest, alder, shrub (< 2 m tall), herbaceous, graminoid (hayfield, flood plain, and wetlands), bare ground (lawn and bare soil), and river. In 2004, we characterized 478 turtle locations (21 to 25 characterizations per individual, mean = 23·5).

We characterized all available habitats within the study site with digital orthophotos (1:5000 scale) in arcview 3·2 (ESRI, Redlands, California, USA). Turtles were located at most 250 m from the river. We superimposed a grid, composed of 25 × 25 m squares, on a 250-m buffer area bordering the river. We determined the proportion of each of the 2838 grid squares covered by closed habitat (alder and forest), open habitat (shrub, herbaceous, graminoid, and bare ground), and river to the nearest 20%.

To determine the habitats available to turtles, we developed an index that we term truly available habitat area (TAHA). Because 75% of movements between consecutive locations were ≤ 50 m, we considered that all habitats within a 50-m radius of a turtle's location were available. Therefore, we used the 100-m wide corridor linking successive locations as the TAHA. All 1822 grid squares encompassed in the TAHA were ground truthed: we recorded the proportion of each cover type (forest, alder, shrub, herbaceous and graminoid) to the nearest 20%.

We compared habitat use to habitat availability at two scales: the entire study site and the TAHA. The first scale was to test whether turtles’ TAHA were randomly dispersed in the study site, while the second scale was to test whether turtles were randomly located in their TAHA. At the study site scale, habitat availability was the sum of the grid squares in the study site and habitat use was the sum of grid squares in TAHA of all turtles. At the TAHA scale, habitat availability was the sum of grid squares in turtles’ TAHA and habitat use was the sum of turtles’ locations. For this analysis, the five terrestrial cover types were classified as closed habitat (alder and forest) and open habitat (shrub, herbaceous, graminoid, and bare ground) and χ2 tests were used to compare availability to use at each scale.

operative environmental temperatures

Operative environmental temperature (Te) integrates the effects of Ta, wind, and solar radiation to infer the temperature that an ectotherm would attain at a given point and time (Bakken & Gates 1975). We used 23 rectangular water-filled plastic containers (180 × 10 × 60 mm) painted to approximate the colour and reflectance of a wood turtle carapace to act as ‘Te thermometers’ (Tracy 1982Bakken 1992). These models did not reproduce the exact size and shape of wood turtles, but studies have demonstrated that model shape, size, and colour have minor effects on Te (< 1 °C on mean TeVitt & Sartorius 1999Shine & Kearney 2001). A data logger placed in each model recorded temperature at 30-min intervals. We confirmed the accuracy of Te measures by recording simultaneously the temperatures of four models and four turtles (carapace length 145–182 mm, mass 452–1050 g) on bare ground in an outdoor enclosure.

The expected null distribution of Tb for thermoconforming animals can be obtained by randomly sampling the thermal environment (Hertz et al. 1993). We reduced the number of models required to sample adequately spatial and temporal variation in Te over the entire study site by using a combination of stratified and random sampling. We placed four models in each of the five cover types, one model on bare ground to measure the maximal available Te, and two data loggers in the river. Models were randomly moved within their cover type once a day (5 days week−1) to capture spatial heterogeneity encountered by a randomly moving turtle within a given cover type.

The null Te distribution of a thermoconforming control for each turtle was obtained by first randomly picking one of the 23 models where the probability that a particular model was picked reflected the proportional vegetation cover type available in the TAHA of a given turtle. Once a model was chosen, the entire model's temperature data set (2048 readings) was used. Loggers in models and in turtles recorded temperatures during identical time periods, so picking the entire model's temperature data set allows us to control for temporal variation in model and turtle samples, revealing the spatial variation only. This procedure was repeated 150 times with replacement to generate the null Te distribution for a given turtle.

thermoregulation analyses

Thermoregulation indices are often based on differences between an average animal's Tb and an average control's Te (Hertz et al. 1993Brown & Weatherhead 2000Blouin-Demers & Weatherhead 2001). Hertz et al. (1993) argued that mean values can be misleading because they lose information about underlying distributions and variability. This is especially relevant for Tb and Te distributions which are rarely normally distributed (Dewitt & Friedman 1979Wills & Beaupre 2000). We tested for differences between Tb and null Te distributions with Kolmogorov–Smirnov tests (Blouin-Demers & Weatherhead 2008) and used the D-statistic as a thermal selection index (TSI) to quantify the extent of departure from thermoconformity.

We then verified that turtles’ Tb deviated directionally from thermoconformity such that Tb approached Tset. Tset should approximate To and is often defined as the central 50% range (25–75% quartiles) of preferred Tb of post-absorptive animals in a thermal gradient (Hertz et al. 1993Christian & Weavers 1996Blouin-Demers & Weatherhead 2001). Due to low Ta and a short growing season at our northern latitude, the ability to digest and pass food rather than to find and ingest food is likely to limit energy acquisition (Congdon 1989Koper & Brooks 2000Niewiarowski 2001), so we expect Tset to maximize food processing rate. As argued elsewhere (Dubois et al. 2008), energy intake from processed food increases much faster than metabolic rate, thus increasing net energy retention at Tb below To. We defined Tset (30 °C) based on the upper selected Tb (95% percentile) of fed animals in laboratory settings, which more closely approximates To for energy acquisition compared to a range of preferred Tb that assumes a performance plateau (Dubois et al. 2008). Our approach is thus similar to that used by Huey and Slatkin (1976). We calculated the accuracy of Tb (db) by the deviation of Tb from Tset (| Tb – Tset |; Hertz et al. 1993).

To analyse the use of sun/shade shuttling as a behavioural adjustment for thermoregulation, we created a shuttling index based on our continuous temperature recordings. The shuttling index was the absolute difference between a turtle's external temperature (Text) and its internal Tb (| Text – Tb |). Without solar radiation, Text should track Tb such that Text ≈ Tb. Text and Tb should deviate when a turtle moves into the sun (Text > Tb) or when a turtle moves into the shade after basking (Text < Tb).

statistical analyses

Sunny and cloudy days (> 5 h and ≤ 5 h of sun, respectively) were analysed separately because they represent different thermal conditions. Thermoregulation analyses were restricted to temperature data from six each for males, females, and juveniles, and from 22 physical models. We lost one juvenile turtle and one physical model and two implanted data loggers failed. Text data were incomplete for most individuals (4761 to 5709 Text/turtle) because of data logger loss or impossibility of downloading before the memory was full.

To control for repeated measures on individuals and some missing values, we used nested linear mixed models with individual intercepts as a random factor, thus separating individual and treatment effects. Models were fitted by the lme function with restricted maximum log-likelihood (REML) in software r (R project version 1·9·1). The significance of each independent variable was tested with type III sums of squares. Normality of residuals was checked with quantile–quantile normal plots and distribution histograms. Data are presented as means ± 1 SD.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

operative environmental temperatures

In the outdoor enclosure, Tb of turtles were highly correlated with the temperatures of physical models recorded simultaneously (r = 0·99, F1,24 = 1346·3, P < 0·001). Although the mean temperature of models was significantly higher than that of turtles (0·44 ± 0·28 °C, paired t = 3·126, P = 0·002), this difference was less than the measurement error of the data loggers (±0·5 °C). Therefore, we conclude that our physical models adequately predict available Te for wood turtles.

habitat thermal quality and heterogeneity

We recorded 49 152 Te from 22 physical models. We first calculated the hourly mean Te of each cover type during cloudy (n = 18) and sunny (n = 24) days (Fig. 1). Terrestrial cover types were thermally homogeneous from 20.00 to 08.00 h and offered lower Te than the river (15·6 ± 1·9 °C and 19·5 ± 0·9 °C, respectively). From 09.00 to 19.00 h, the hourly mean Te in terrestrial cover types exceeded Te in the river only for shrub, herbaceous, and graminoid covers from 12.00 to 17.00 h on sunny days. The hourly mean Te exceeded Tset only on bare ground on sunny days from 11.00 to 18.00 h. During that time, 86% of bare ground Te exceeded Tset, while only 3% of shrub and graminoid Te exceeded Tset and Te never exceeded Tset in other cover types. Turtles can attain Tset only during sunny days by basking on bare ground.

image

Figure 1. Hourly mean operative environmental temperature (Te) available to wood turtles in Québec (a) on cloudy days and (b) on sunny days. The dashed line represents Tset for wood turtles.

Download figure to PowerPoint

The thermal heterogeneity of terrestrial cover types from 09.00 to 19.00 h was evaluated with a mixed model nested ancova where Te was the dependent variable, time of day, cover type (forest, alder, shrub, herbaceous, and graminoid), and their interaction were predictor variables, and individual physical model intercept was a random factor. We conducted separate analyses for sunny and cloudy days. For both sunny and cloudy days, mean Te was significantly affected by time of day (P < 0·001), but not by cover type (P = 0·21 and P = 0·41, respectively). The interaction between time of day and cover type was significant for sunny days (P < 0·001), but not for cloudy days (P = 0·92; Table 1). Thus, not surprisingly, cover types were thermally heterogeneous only on sunny days, and thermal heterogeneity varied with time of day, with a peak at 14.00 h (Fig. 1).

Table 1.  Mixed model nested ancova for the effects of terrestrial cover type and time of day on mean environmental temperature available for wood turtles in Québec on sunny and cloudy days (09.00–19.00 h)
Random effectsFixed effects
LevelResidual variancePredictord.f.FP
Cloudy daysCloudy days
 Among models0·25 (12·0%) Cover type 4, 14 1·084 0·42
 Within models1·84 (88·0%) Time of day10, 171511·562< 0·0001
 Cover type × time of day40, 1715 0·719 0·91
Sunny daysSunny days
 Among models0·46 (15·2%) Cover type 4, 14 1·840 0·18
 Within models2·56 (84·8%) Time of day10, 554852·201< 0·0001
 Cover type × time of day40, 5548 2·088< 0·0001

thermoregulation analyses

We recorded 36 864 Tb and 92 741 Text from 18 turtles. We first compared the distribution of Tb to the null Te distribution representing the expected Tb distribution for a thermoconforming turtle randomly moving within its TAHA. The Tb and null Te distributions of the 18 turtles were pooled three ways: (i) complete data set (00.00 to 23.00 h), (ii) cloudy days (09.00 to 18.00 h), and (iii) sunny days (09.00 to 18.00 h). For the three data sets, the null Te distribution encompassed turtles’ Tb distribution, indicating that the null Te distribution adequately reflected the temperatures that were available to turtles (Fig. 2). The Kolmogorov–Smirnov tests indicated that the distribution of Tb was significantly different from the null Te distribution for the complete data set (D = 0·22, P < 0·001), cloudy days (D = 0·17, P < 0·001), and sunny days (D = 0·25, P < 0·001).

image

Figure 2. Distribution of Tb and null distribution of Te for wood turtles from June to August in Québec: (a) complete data set (00.00–23.00 h), (b) cloudy days (09.00–18.00 h), and (c) sunny days (09.00–18.00 h). Black bars represent Tb distributions and grey bars represent available Te for thermoconforming controls. The dashed line represents Tset for wood turtles.

Download figure to PowerPoint

To compare the hourly variation in Tb and available Te, we calculated the hourly mean Tb for turtles with the same three data sets. These hourly mean Tb were superimposed on the hourly mean Te for bare ground, river, and terrestrial cover type weighted by availability of the cover types in turtles’ TAHA (Fig. 3). Turtle Tb was always higher than terrestrial cover type Te. From 20.00 to 08.00 h, turtles maintained Tb intermediate between terrestrial cover types and river Te, suggesting that they partitioned their time between the river and terrestrial habitats. During daytime, turtle Tb exceeded river Te from 10.00 to 17.00 h on sunny days, whereas Tb only marginally exceeded river Te on cloudy days. After 15.00 h on sunny days, turtle Tb decreased more or less in concert with bare ground Te and approached available Te in terrestrial cover types and river around 18.00 h.

image

Figure 3. Hourly mean Tb of wood turtles and Te on bare ground, in river, and in terrestrial habitats from June to August in Québec: (a) complete data set (00.00–23.00 h), (b) cloudy days (09.00–18.00 h), and (c) sunny days (09.00–18.00 h). The dashed line represents Tset for wood turtles.

Download figure to PowerPoint

To gain information on interindividual and hourly variation in thermal selection, we compared the hourly Tb and null Te distributions separately for each of the 18 turtles with Kolmogorov–Smirnov tests. We used data for sunny days (09.00 to 18.00 h) when opportunities for thermoregulation are present. Figure 4 presents the number of turtles with an hourly Tb distribution significantly different from their hourly null Te distribution (D > 0·262, P < 0·05), which suggests active thermoregulation. From 09.00 to 16.00 h, the number of thermoregulating turtles varied between 12 and 15 (67–83% of individuals), while it was much lower late in the afternoon when the sun was lower [7 (39% of individuals) at 17.00 h and 6 (33% of individuals) at 18.00 h].

image

Figure 4. Number of individuals with an hourly distribution of body temperatures different from thermoconformity (Kolmogorov–Smirnov test D > 0·262, P < 0·05) for wood turtles in Québec on sunny days (09.00–18.00 h).

Download figure to PowerPoint

We verified that the departure from thermoconformity was towards Tset with the D-statistic (our TSI). We used TSI as an explanatory variable, along with time of day and their interaction, to explain variation in individual hourly mean db. The mixed model nested ancova, with individual turtle intercept as a random factor, indicated that TSI and time of day both had an effect on db (P < 0·001), and there was no interaction between TSI and time of day (P = 0·75; Table 2). db was negatively correlated with TSI (–6·12 ± 0·44). Mean db decreased with time of day from 09.00 to 13.00 h, was stable between 13.00 h and 16.00 h, and increased from 17.00 to 18.00 h (Fig. 5). Mean db for all turtles on sunny days (09.00 to 18.00 h) was 7·1, and it decreased to 5·9 between 11.00 h and 16.00 h, when it was possible for turtles to attain their Tset (Fig. 5).

Table 2.  Mixed model nested ancova for the effects of thermal selection index (TSI) and time of day on mean deviation of body temperatures from the Tset (db) of wood turtles in Québec on sunny days (09.00–18.00h)
Random effectsFixed effects
LevelResidual variancePredictord.f.FP
Among individuals0·85 (63·0%)TSI1, 14327·68< 0·0001
Time of day9, 14320·35< 0·0001
Within individuals0·50 (37·0%)TSI × time of day9, 143 0·660·75
image

Figure 5. (a) Hourly mean deviations of Tb from Tset (db) on sunny days (09.00–18.00 h) in Québec. (b) Individual mean db as a function of mean thermal selection index (TSI) from 11.00 to 16.00 h on sunny days. The dashed line indicates the threshold above which TSI values represent a significant departure from thermoconformity. The black triangles indicate the mean db on both figures.

Download figure to PowerPoint

habitat selection and sun/shade shuttling

Turtles’ TAHAs were nonrandomly located within the study site (χ2 = 1584·9, d.f. = 2, P < 0·001), with a preference for open habitats and river (Fig. 6). At the TAHA scale, we compared habitat use on sunny (n = 380 locations) and cloudy days (n = 98 locations). Turtles were more often located in terrestrial habitat on sunny days (75%) than on cloudy days (62%; χ2 = 6·97, d.f. = 1, P < 0·01; Fig. 6). When turtles were on land, the proportion of locations in closed habitats compared to open habitats was not different on sunny and cloudy days (χ2 = 0·30, d.f. = 1, P = 0·61; Fig. 6). We thus compared terrestrial habitat use to availability by pooling cloudy and sunny days. Turtles showed a strong preference for open habitats (χ2 = 68·60, d.f. = 1, P < 0·001; Fig. 6). We observed basking behaviour in 8% (36/478) of telemetry locations: turtles flattened the vegetation or used already flattened vegetation or bare ground to bask.

image

Figure 6. (a) Relative availability of closed and open habitats and river in the study site (black bars) and in pooled wood turtle truly available habitat areas (TAHA; dark grey bars) from June to August in Québec. (b) Relative frequency of turtle locations in river and terrestrial habitats on cloudy days (hatched bars) and on sunny days (light grey bars). (c) Relative availability of closed and open terrestrial habitats in pooled turtle TAHAs (dark grey bars) in comparison with the relative frequency of turtle terrestrial locations in both habitat types on cloudy days (hatched bars) and on sunny days (light grey bars).

Download figure to PowerPoint

We also tested the use of sun/shade shuttling as a behavioural adjustment for thermoregulation on sunny days (09.00 to 18.00 h). We used the hourly mean shuttling index as a dependent variable in a mixed model nested anova, with time of day as an explanatory variable and individual turtle intercept as a random factor. Time of day had an effect on the shuttling index (F9,134 = 14·1986, P < 0·001; Fig. 7). The homogeneous grouping based on 95 % confidence intervals of the model coefficients of each hour indicated that the shuttling index was higher from 10.00 to 12.00 h when turtles were warming up, intermediate at 09.00 h and between 13.00 h and 16.00 h, and lower from 17.00 to 18.00 h when the sun was low.

image

Figure 7. Hourly mean shuttling index (| Text – Tb |) as a function of time of day for wood turtles on sunny days (09.00–18.00 h) in Québec. The letters indicate significant differences in hourly mean shuttling index based on the 95% confident intervals of the mixed model nested anova coefficients.

Download figure to PowerPoint

internal body temperature and external carapace temperature

For the 18 free-ranging turtles during sunny days (09.00 to 18.00 h), the internal body temperature (Tb) and external carapace temperature (Text) were correlated (r = 0·93, F1,23853 = 15·5900, P < 0·001), but the residuals ranged from –20 to +8 °C. Text (20·3 ± 5·95 °C) was significantly higher than Tb [18·7 ± 6·00 °C; paired t test : t(23854) = 40·7154, P < 0·001). The distribution of Tb (9·8 to 34·5 °C) was significantly different from the distribution of Text (11·0 to 44·0 °C; Kolmogorov–Smirnov test: D = 0·17, P < 0·001).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The low Te generally available under different vegetation cover types indicates that habitat thermal quality is relatively low for wood turtles at our study site and that turtles must exert careful habitat selection if they are to achieve Tb close to Tset. This suggests that while regulation of Tb close to Tset will bring a benefit over thermoconformity, primarily through an increase in energy gain, it should incur a cost that is expressed as a constraint on a turtle's time-activity budget and as predation risk while exposed on bare ground for basking (Pianka & Pianka 1970Huey 1982Shine & Madsen 1996).

As previously reported (Tuttle 1996Arvisais 2000Compton et al. 2002), we observed a strong link between habitat use and thermoregulation. Turtles selected open habitats that allow access to solar radiation. Turtles were more often in terrestrial habitats on sunny days than on cloudy days, suggesting that turtles took advantage of higher available Te and greater basking opportunities offered by terrestrial habitats. The shuttling index indicated that basking was most pronounced in the morning as turtles sought to warm up, intermittent in early afternoon, and ceased in late afternoon, probably due to the scarcity of sun patches as the sun started to set. Tb profiles suggest that turtles regularly used the river as a thermal refuge at night, just as snakes and lizards use hot rocks or crevices (Huey & Porter 1989Webb & Shine 1998Blouin-Demers & Weatherhead 2001). This could explain why wood turtles stayed within 250 m of the river during the summer active period in this and in another northern population (Arvisais et al. 2002). It appears that the requirements for thermoregulation impose constraints on the spatial distribution of these turtles.

We calculated how much thermoregulation would increase metabolic rate in turtles based on Dubois et al. (2008), excluding temperatures above Tset, which represented less than 2% of the recorded Tb and Te. We assume that habitats with Te > Tset did not impose a measurable cost or constraint on turtle movements and feel this is a reasonable assumption given the scarcity of spatial and temporal combinations that yield Tb above Tset in our study area. When we consider the complete and sunny days data sets, turtles exhibited a 20% and a 26% increase in metabolic rate compared with metabolic rates predicted for thermoconforming individuals. If we assume that turtles were generally absorptive when they selected higher Tb, a 20% to 26% increase in metabolic rate will be accompanied by a substantial increase in net energy retention (see Appendix 1 in Dubois et al. 2008). Somewhat surprisingly, turtles did not exploit the benefits of thermoregulation to the maximum. A hypothetical turtle that maintained the highest possible Tb without exceeding Tset would have been able to increase its metabolic rate by 72% and 93%, respectively, for complete and sunny days.

Wood turtles are the only semi-terrestrial turtles at the latitude of our study. Using the same approach as above, we calculated the consequences of terrestriality for metabolic rate. For the complete and sunny days data sets respectively, a hypothetical fully aquatic turtle (i.e. one where Tb equals water temperature) would have a metabolic rate 21% and 4% greater than that of a thermoconforming fully terrestrial turtle. Because food processing rate is tightly tied to Tb and metabolic rate (Stevenson et al. 1985Dorcas, Peterson & Flint 1997Niu et al. 1999), we argue that even a semi-terrestrial lifestyle is only profitable in northern populations when it is accompanied by thermoregulation. A completely terrestrial lifestyle is not as energetically profitable as a fully aquatic lifestyle at our high latitude because terrestrial temperatures fall below that of water bodies at night. By using the river as a thermal refuge at night, turtles were able to increase metabolic rate by 21 % compared with a turtle that remained terrestrial at night. Thus, there is a clear benefit for turtles to stay close to the river and to use it as a thermal refuge at night. Our analyses indicate that the spatial distribution, temporal activity patterns, and semi-terrestrial lifestyle of wood turtles are shaped by their thermal biology. We also note that advanced data loggers now allow a host of new information to be obtained about free-living animals, including activity patterns, foraging, and even metabolic rate (Wilson et al. 2006). This new generation of devices might shed more light on the links between thermoregulation and physiological ecology.

For all ectotherms, regulating Tb near Tset to increase net energy gain may be critical for fitness by increasing growth and reproductive rates, but this may be particularly true at northern latitudes due to the low Ta (Congdon 1989Dunham, Grant & Overall 1989Angilletta 2001). Because turtles forage continuously on small prey items (Harding & Bloomer 1979Ernst et al. 1994), they may derive some benefit from higher Tb and enhanced processing rate almost continuously. The imprecise Tb regulation observed, however, suggests that the benefit of maintaining Tb close to the physiological optimum may be lower than the ecological cost, so that trade-offs are necessary to achieve the greatest net benefit (Huey & Slatkin 1976). Predation risk associated with thermoregulation (basking, movements; Pianka & Pianka 1970Huey 1982) may represent one important ecological cost for turtles. For long-lived organisms with a high adult survival rate and low and variable reproductive success (Congdon et al. 1993Congdon et al. 2003), selection may favour a risk-averse strategy aimed at increasing reproductive life span at the expense of high physiological performance and reproductive effort (Rollinson & Brooks 2007).

It is noteworthy that the thermal sensitivity of metabolism (Q10 = 1·92 ± 0·10; Dubois et al. 2008) over the 16 to 32 °C range where turtles are active (Ernst 1986Tuttle 1996) is at the low end of the range reported for reptiles (Q10 = 2 to 3; Bennett & Dawson 1976). A low Q10 testifies to a comparatively low decline in metabolic rate (and we presume food processing capacity) as temperatures fall below the physiological optimum. If the limited opportunities for thermoregulation at high latitudes where climates are cool translate to high thermoregulatory costs, turtles with a low thermal sensitivity of metabolism may have greater overall fitness than those with a high thermal sensitivity. In other words, the former will maintain high physiological performance despite relatively cool Tb, while the latter will see their metabolic and processing rates fall sharply and for long periods when turtles are unable to regulate Tb close to Tset. Measuring how thermal sensitivity of metabolic rate varies with latitude within a given species may shine light on the physiological adaptations that permit ectotherms to perform in climatically challenging high latitudes.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank P.A. Bernier, N. Delelis, W. Bertacchi, M. Gauthier, C. Daigle, J. Jutras and many volunteers for their help in the field and the analysis. We also thank the Dr. Jacques Dancosse of the Biodôme de Montréal for the surgical implantation of data loggers. This study was supported by a grant (D.T.) and a scholarship (Y.D.) from the National Sciences and Engineering Research Council of Canada and by a grant to the Biodôme de Montréal from the Canada Trust Friend of the Environment Fund. We are grateful to M. Angilletta, J. Roe, and 2 anonymous reviewers for insightful comments that improved this manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Adolph, S. (1990) Influence of behavioral thermoregulation on microhabitat use by two Sceloporus lizards. Ecology, 71, 315327.
  • Angilletta, M.J. (2001) Thermal and physiological constraints on energy assimilation in a widespread lizard (Sceloporus undulatus). Ecology, 82, 30443056.
  • Angilletta Jr, M.J., Hill, T. & Robson, M.A. (2002) Is physiological performance optimized by thermoregulatory behavior? A case study of the eastern fence lizard, Sceloporus undulatus. Journal of Thermal Biology, 27, 199204.
  • Arvisais, M. (2000) Caractérisation et sélection d’habitats à l’intérieur des domaines vitaux chez la tortue des bois (Clemmys insculpta) au nord de son aire de répartition, Québec, Canada. M.Sc. thesis, Université du Québec à Trois-Rivière, Trois-Rivière, Québec, Canada.
  • Arvisais, M., Bourgeois, J.C., Levesque, E., Daigle, C., Masse, D. & Jutras, J. (2002) Home range and movements of a wood turtle (Clemmys insculpta) population at the northern limit of its range. Canadian Journal of Zoology, 80, 402408.
  • Arvisais, M., Levesque, E., Bourgeois, J.C., Daigle, C., Masse, D. & Jutras, J. (2004) Habitat selection by the wood turtle (Clemmys insculpta) at the northern limit of its range. Canadian Journal of Zoology, 82, 391398.
  • Avery, R.A. (1982) Field studies of body temperatures and thermoregulation. Biology of the Reptilia (ed. C.Gans), Vol. 12, pp. 93166. Academic Press, New York.
  • Bakken, G.S. (1992) Measurment and application of operative and standard operative temperature in ecology. American Zoologist, 32, 194216.
  • Bakken, G.S. & Gates, D.M. (1975). Heat-transfer analysis of animals: some implications for field ecology, physiology, and evolution. Perspectives in Biophysical Ecology (eds D.M.Gates & R.B.Schmerl), pp. 255290. Springer-Verlag, New York.
  • Beaupre, S.J., Dunham, A.E. & Overall, K.L. (1993) The effects of consumption rate and temperature on apparent digestibility coefficient, urate production, metabolizable energy coefficient and passage time in canyon lizards (Sceloporus merriami) from 2 populations. Functional Ecology, 7, 273280.
  • Beitinger, T.L. & Fitzpatrick, L.C. (1979) Physiological and ecological correlates of preferred temperature in fish. American Zoologist, 19, 319329.
  • Bennett, A.F. & Dawson, W.R. (1976) Metabolism. Biology of the Reptilia (eds C.Gans & W.R.Dawson), Vol. 5, pp. 127223. Academic Press, New York.
  • Berman, C.H. & Quinn, T.P. (1991) Behavioural thermoregulation and homing by spring chinooksalmon, Oncorhynchus tshawytscha (Walbaum), in the Yakima River. Journal of Fish Biology, 39, 301312.
  • Blouin-Demers, G. & Weatherhead, P.J. (2001) Thermal ecology of black rat snakes (Elaphe obsoleta) in a thermally challenging environment. Ecology, 82, 30253043.
  • Blouin-Demers, G. & Weatherhead, P.J. (2002) Habitat-specific behavioural thermoregulation by black rat snakes (Elaphe obsoleta obsoleta). Oikos, 97, 5968.
  • Blouin-Demers, G. & Weatherhead, P.J. (2008) Habitat use is linked to components of fitness through the temperature-dependence of performance in ratsnakes (Elaphe obsoleta). Israel Journal of Ecology & Evolution, 54, 361372.
  • Blouin-Demers, G., Weatherhead, P.J. & McCracken, H.A. (2003) A test of the thermal coadaptation hypothesis with black rat snakes (Elaphe obsoleta) and northern water snakes (Nerodia sipedon). Journal of Thermal Biology, 28, 331340.
  • Brooks, R.J., Shilton, C.M., Brown, G.P. & Quinn, N.W.S. (1992) Body size, age distribution, and reproduction in a northern population of wood turtles (Clemmys insculpta). Canadian Journal Zoology, 70, 462469.
  • Brown, G.P. & Weatherhead, P.J. (2000) Thermal ecology and sexual size dimorphism in northern water snakes, Nerodia sipedon. Ecological monographs, 70, 311330.
  • Cagle, F.R. (1939) A system of marking turtles for future identification. Copeia, 1939, 170172.
  • Christian, K.A. & Weavers, B.W. (1996) Thermoregulation of monitor lizards in Australia – an evaluation of methods in thermal biology. Ecological monographs, 66, 139157.
  • Compton, B.W., Rhymer, J.M. & McCollough, M. (2002) Habitat selection by wood turtles (Clemmys insculpta): An application of paired logistic regression. Ecology, 83, 833843.
  • Congdon, J.D. (1989) Proximate and evolutionary constraints on energy relations of reptiles. Physiological Zoology, 62, 356373.
  • Congdon, J.D., Dunham, A.E. & Sels, R.C.V. (1993) Delayed sexual maturity and demographics of blanding turtles (Emydoidea blandingii): implications for conservation and management of long-lived organisms. Conservation Biology, 7, 826833.
  • Congdon, J.D., Nagle, R.D., Kinney, O.M., Van Loben Sels, R.C., Quinter, T. & Tinkle, D.W. (2003) Testing hypotheses of aging in long-lived painted turtles (Chrysemys picta). Experimental Gerontology, 38, 765772.
  • Dewitt, C.B. & Friedman, R.M. (1979) Significance of skewness in ectotherm thermoregulation. American Zoologist, 19, 195209.
  • Dorcas, M.E. & Peterson, C.R. (1998) Daily body temperature variation in free-ranging rubber boas. Herpetologica, 54, 88103.
  • Dorcas, M.E., Peterson, C.R. & Flint, M.E.T. (1997) The thermal biology of digestion in rubber boas (Charina bottae): physiology, behavior, and environmental constraints. Physiological Zoology, 70, 292300.
  • Dubois, Y., Blouin-Demers, G. & Thomas, D. (2008) Temperature selection in wood turtles (Glyptemys insculpta) and its implications for energetics. Ecoscience, 15, 398406.
  • Dunham, A.E., Grant, B.W. & Overall, K.L. (1989) Interfaces between biophysical and physiological ecology and population ecology of terrestrial vertebrate ectotherms. Physiological Zoology, 62, 335355.
  • Ernst, C.H. (1986) Environmental temperatures and activities in the wood turtle, Clemmys insculpta. Journal of Herpetology, 20, 222229.
  • Ernst, C.H. (2001) Some ecological parameters of the wood turtle, Clemmys insculpta, in southeastern Pennsylvania. Chelonian Conservation and Biology, 4, 9499.
  • Ernst, C.H., Barbour, R.W. & Lovich, J.E. (1994) Turtles of the United-States and Canada. Smithsonian institution Press, Washington, DC.
  • Foscarini, D.A. (1994) Demography of the wood turtle (Clemmys insculpta) and habitat selection in the Maitland river valley. M.Sc. thesis, University of Guelph, Guelph, Ontario, Canada.
  • Grant, B.W. & Porter, W.P. (1992) Modeling global macroclimatic constraints on ectocherm energy budgets. American Zoology, 32, 154178.
  • Gregory, P.T., Crampton, L.H. & Skebo, K.M. (1999) Conflicts and interactions among reproduction, thermoregulation and feeding in viviparous reptiles: Are gravid snakes anorexic? Journal of Zoology, 248, 231241.
  • Grover, M.C. (1996) Microhabitat use and thermal ecology of two narrowly sympatric Sceloporus (Phrynosomatidae) lizards. Journal of Herpetology, 30, 152160.
  • Harding, J.H. & Bloomer, T.J. (1979) The wood turtle, Clemmys insculpta, a natural history. Bulletin of the New York Herpetological Society, 15, 926.
  • Heatwole, H.F. & Taylor, J. (1987) Ecology of Reptiles. Surrey Beatty & Sons Pty Limited, Chipping Norton, Australia.
  • Hertz, P.E., Huey, R.B. & Stevenson, R.D. (1993) Evaluating temperature regulation by field-active ectotherms: The fallacy of the inappropriate question. American Naturalist, 142, 796818.
  • Huey, R.B. (1982) Temperature, physiology, and the ecology of reptiles. Biology of the Reptilia (ed. C. Gans), Vol. 12, pp. 2592. Academic Press, New York.
  • Huey, R.B. & Kingsolver, J.G. (1989) Evolution of thermal sensitivity of ectotherm performance. Trends in Ecology & Evolution, 4, 131135.
  • Huey, R.B. & Porter, W.P. (1989) Hot rocks and not-so-hot rocks: retreat-site selection by garter snakes and its thermal consequences. Ecology, 70, 931944.
  • Huey, R.B. & Slatkin, M. (1976) Cost and benefits of lizard thermoregulation. Quarterly Review of Biology, 51, 363384.
  • James, M.C., Davenport, J. & Hays, G.C. (2006) Expanded thermal niche for a diving vertebrate: a leatherback turtle diving into near-freezing water. Journal of Experimental Marine Biology and Ecology, 335, 221226.
  • Kaufmann, J.H. (1992) Habitat use by wood turtles in central Pennsylvania. Journal of Herpetology, 26, 315321.
  • Kearney, M., Shine, R. & Porter, W.P. (2009) The potential for behavioral thermoregulation to buffer ‘cold-blooded’ animals against climate warming. Proceedings of the National Academy of Sciences, USA, eFIRST date: 21 Feb 2009.
  • Koper, N. & Brooks, R.J. (2000) Environmental constraints on growth of painted turtles (Chrysemys picta) in northern climates. Herpetologica, 56, 421432.
  • Lovich, J.E., Ernst, C.H. & McBreen, J.F. (1990) Growth, maturity, and sexual dimorphism in the wood turtle, Clemmys insculpta. Canadian Journal of Zoology, 68, 672677.
  • Magnuson, J.J., Crowder, L.B. & Medvick, P.A. (1979) Temperature as an ecological resource. American Zoologist, 19, 331343.
  • Niewiarowski, P.H. (2001) Energy budgets, growth rates, and thermal constraints: toward an integrative approach to the study of life-history variation. American Naturalist, 157, 421433.
  • Niu, C., Zhang, T. & Sun, R. (1999) Food consumption and growth of juvenile Chinese soft-shelled turtles (Pelodiscus sinensis) in relation to body weight and water temperature. Asiatic Herpetological Research, 8, 8184.
  • Parmesan, C. & Yohe, G. (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, 3742.
  • Pianka, E.R. & Pianka, H.D. (1970) The ecology of Moloch horridus (Lacertilia: Agamidae) in western Australia. Copeia, 1970, 90103.
  • Rollinson, N. & Brooks, R.J. (2007) Proximate constraints on reproductive output in a northern population of painted turtles: an emperical test of the bet-hedging paradigm. Canadian Journal of Zoology, 85, 177184.
  • Rummery, C., Shine, R., Houston, D.L. & Thompson, M.B. (1995) Thermal biology of the Australian forest dragon, Hypsilurus spinipes (Agamidae). Copeia, 1995, 818827.
  • Schofield, G., Bishop, C.M., Katselidis, K.A., Dimopoulos, P., Pantis, J.D. & Hays, G.C. (2009) Microhabitat selection by sea turtles in a dynamic thermal marine environment. Journal of Animal Ecology, 78, 1421.
  • Shine, R. & Kearney, M. (2001) Field studies of reptile thermoregulation: how well do physical models predict operative temperatures? Functional Ecology, 15, 282288.
  • Shine, R. & Madsen, T. (1996) Is thermoregulation unimportant for most reptiles: an example using water pythons (Liasis fuscus) in tropical Australia. Physiological Zoology, 69, 252269.
  • Stevenson, R.D., Peterson, C.R. & Tsuji, J.S. (1985) The thermal-dependence of locomotion, tongue flicking, digestion, and oxygen-consumption in the wandering garter snake. Physiological Zoology, 58, 4657.
  • Tracy, C.R. (1982) Biophysical modeling in reptilian physiology and ecology. Biology of the Reptilia (eds C.Gans & F.H.Pough), Vol. 12, pp. 275315. Academic Press, New York.
  • Tracy, C.R. & Christian, K.A. (1986) Ecological relations among space, time, and thermal niche axes. Ecology, 67, 609615.
  • Tuttle, S.E. (1996) Ecology and natural history of the wood turtle (Clemmys insculpta) in southern New Hampshire. M.Sc. thesis, Antioch University, Keene, New Hampshire.
  • Van Marken Lichtenbelt, W.D., Vogel, J.T. & Wesselingh, R.A. (1997) Energetic consequences of field body temperatures in green iguana. Ecology, 78, 297307.
  • Vitt, L.J. & Sartorius, S.S. (1999) HOBOs, tidbits and lizard models: the utility of electronic devices in field studies of ectotherm thermoregulation. Functional Ecology, 13, 670674.
  • Webb, J.K. & Shine, R. (1998) Using thermal ecology to predict retreat-site selection by an endangered snake species. Biological Conservation, 86, 233242.
  • Wills, C.A. & Beaupre, S.J. (2000) An application of randomization for detecting evidence of thermoregulation in timber rattlesnakes (Crotalus horridus) from northwest Arkansas. Physiological and Biochemical Zoology, 73, 325334.
  • Wilson, R.P., White, C.R., Quintana, F., Halsey, L.G., Liebsch, N., Martin, G.R. & Butler, P.J. (2006) Moving towards acceleration for estimates of activity-specific metabolic rate in free-living animals: the case of the cormorant. Journal of Animal Ecology, 75, 10811090.