Aquatic Ecology

, Volume 44, Issue 4, pp 723–729

Giant spiny-frog (Paa spinosa) from different populations differ in thermal preference but not in thermal tolerance


    • Institute of EcologyZhejiang Normal University
  • Chun-Tao Liu
    • Institute of EcologyZhejiang Normal University

DOI: 10.1007/s10452-009-9310-3

Cite this article as:
Zheng, R. & Liu, C. Aquat Ecol (2010) 44: 723. doi:10.1007/s10452-009-9310-3


To examine whether different thermal environments have induced a change in thermal characteristics, we have conducted a between-population comparison on broad geographic patterns of preferred body temperature and critical thermal maximum in a giant spiny-frog Paa spinosa. We found a bimodal pattern of preferred body temperature during the day, with high preferred body temperature during the inactive diurnal period and low temperature during the active nocturnal period. There were significant differences among six populations of P. spinosa in preferred body temperatures, which decreased along a south to north gradient. Unlike preferred body temperatures, critical thermal maximum did not differ between frogs from the six localities. Although not all characteristics of thermal physiology in P. spinosa underwent parallel changes between the populations, the shift of preferred body temperatures suggests that the features of thermal physiology in the frog may change along a latitudinal gradient in response to different thermal environments.


AnuranThermal physiologyLatitudeTemperature preferenceBehavior


Body temperature has an important influence on physiological and behavioral performance of ectothermic animals. To maintain body temperatures within the optimal range for whole-animal performance, ectotherms may cope with environmental temperature fluctuations by behavioral adjustment or by changes of thermal sensitivity of physiological function through acclimation or adaptation (Huey and Bennett 1990), or both. Means of attaining optimum body temperatures by thermoregulation are well documented in reptiles (e.g. Hertz et al. 1993; Bauwens et al. 1996; Castilla et al. 1999). In general, amphibians are regarded as poor thermoregulators compared with other terrestrial vertebrates (Brattstrom 1962, 1963, 1979; Hutchison and Dupré 1992). Some habits and habitats offer limited opportunities for thermoregulation, at least during the portion of the day when the animals are active. Nocturnal species, for example, cannot bask in the sun, and the thermal mosaic of the environment is more uniform at night than it is during the day. Furthermore, amphibians face a unique challenge among vertebrate ectotherms because the skin of many species offers little or no resistance to water loss, so frogs must have a trade-off between body temperature and water balance (e.g. Brattstrom 1959; Lillywhite 1970; Tracy 1975; also see review in Navas et al. 2008). Possibly because of this, data on amphibian thermal preferences and body temperatures are more variable than those reported for many heliothermic lizards.

Comparative studies on thermal physiology of ectotherms provide a good opportunity to explore the ecological and evolutionary consequences of variation in physiological performance (Kingsolver and Watt 1983; Bauwens et al. 1995; Marden et al. 1996). In particular, intraspecific studies let us to understand the scope for physiological adjustments within species, which is a valid and important goal in evolutionary physiology, ecology and physiological ecology.

Here we conducted a between-population comparison on broad geographic patterns of preferred body temperature and critical thermal maximum to reveal local adaptation in thermal physiology in the giant spiny-frog Paa spinosa. Because of biotic and abiotic constraints, field observations of body temperature alone do not reveal the thermal preferences of ectotherms (Huey and Slatkin 1976), and a controlled laboratory environment investigation can provide information on the optimum physiological temperature for given performance functions, the preferred temperature (Tp) and thermal tolerance (Hutchison and Dupré 1992). Given thermoregulatory constraints, adjustments across ecological gradients of temperature may rely on physiological adjustment, and we studied behavior and physiology to see what seems more relevant to thermoregulation of P. spinosa. The existence of a circadian rhythm of body temperature in reptiles is well documented (for review, see Refinetti and Menaker 1992). Specifically, we measured body temperatures of the frog during artificial light/dark (LD) cycles and examined whether P. spinosa can show circadian rhythms of body temperature with amplitude similar to those recorded in many ectotherms. Obviously, understanding geographical variations in preferred body temperature and thermal tolerance has important implications for husbandry practice as well as effective conservation of this species.

Materials and methods

Study animal

The giant spiny-frog, P. spinosa, is most distinctly characterized with keratinized skin spines on chest. Its current distribution is limited to ten provinces of China and the northern Vietnam ( This species inhabits rocky streams in evergreen forest and open countryside on hills and mountains from 500 to 1,500 m above sea level (Zhao 1998). Due to its high values of nutrition and medicinal importance, this species is highly priced in the market in China. Since 1980s, artificial breeding and domestication efforts have been launched, and have successfully reproduced a second generation in captive conditions for this species (Liu 2004).

Collection and animal care

During June and early July of 2008, we collected a sample of 312 adult giant spiny-frogs (137 male, 175 female; snout-vent length 90–128 mm) from six latitudes along a south–north transect that spanned about 4°, which the maximum distance separating populations is about 1,300 km: 28.7° (Pingjiang, PJ); 28.3° (Lishui, LS); 27.3° (Wuyishan, WYS); 26.3° (Jinggangshan, JGS); 25.8° (Longsheng, GLS); 24.4° (Yangshan, YS) (Table 1; Fig. 1). Frogs from each locality were transported to the laboratory immediately after capture and kept in an air-conditioned room; the air temperature was 19.0 ± 1.0°C and the light cycle was light:dark (LD) 12:12 (07:00 to 19:00 G + 8).
Table 1

The geographical position and mean temperature (°C) of six localities (courtesy of National Climatic Data Center)

Collected localities

Longitude and latitude

Average minimum temperature in January (°C)

Average maximum temperature in June (°C)

Average temperature in year


E 112.63°, N 24.48°





E 109.98°, N 25.82°





E 114.10°, N 26.34°





E 118.01°, N 27.27°





E 119.54°, N 28.27°





E 113.58°, N 28.72°



Fig. 1

Study map showing the geographic location of the collecting sites of P. Spinosa

The frogs from each locality were maintained separately under identical conditions, ca. 10 in each terrarium, which was filled with soil and grasses to mimic the natural habitats where the frogs were captured. Frogs were fed mealworms (larvae of Tenebrio molitor) and water enriched with vitamins and minerals. Live mealworms were confined by a plastic plate (30 cm diameter) with smooth sloping sides placed at the opposite end of the box to the water tray. Uneaten mealworms were removed, the trays and boxes were washed, and boxes relined with moistened absorbent paper toweling.

Preferred body temperature

We measured 206 adult giant spiny-frogs (102 male, 104 female; snout-vent length 95–102 mm) from six populations that a week prior to our work had been selected from 312 frogs. Frogs were acclimated under identical conditions (19.0 ± 1.0°C; 12L:12D) for 1 week, followed by Tp measurements. Tp was determined in a thermal gradient chamber, 1,000 mm × 500 mm × 400 mm (length × width × height). A porcelain heater was suspended 15 cm above the terrarium floor to provide supplementary heating which created a thermal gradient approximately ranging from 19 to 50°C. Petri dishes filled with tap water were placed throughout the chamber in order to allow the frog to choose between dry or wet areas at all temperatures. The water temperature in the Petri dishes was virtually identical to the floor temperature. One frog at a time was placed in the middle of the temperature gradient and left undisturbed for 10 h. Body temperatures were determined at 2 h intervals during the light cycle (12L:12D). We determined the body temperature of each frog three times. Body temperatures (cloacae) were taken using a DM6801A electronic thermometer (±0.1°C, Shenzhen Meter Instruments, China). The mean body temperature of all measured frogs was used as Tp for separate populations (Ji et al. 1996; Du et al. 2000). The substrate temperature (where the frogs were sighted) was also recorded (±0.1°C).

Critical thermal maximum

Forty-seven adult giant spiny-frogs (24 male, 23 female) from six populations were used in this experiment. We conducted all trials between 12:00 and 15:00 to minimize possible circadian effects on thermal tolerance. Critical thermal maximum (CTMax) was determined in a LRH-250A incubator (Ningbo Medical Instruments). Frogs were heated from 28°C at the rate of 0.5°C per min.

We used the “loss of righting response (LRR)” as our criteria of CTMax as recommended by Brattstrom (1968). We defined LRR as the temperature at which frogs lost muscular coordination and were unable to right themselves in a coordinated manner when placed on their dorsum. One of us monitored and regulated each frog’s body temperature gain to 0.5°C/min, while the other independently monitored the behavioral and physiological status of each frog until LRR was reached. Each frog was heated until LRR occurred; Immediately after LRR, we removed frogs from the heat source and placed them in water at room temperature until recovery. Recovery was usually rapid and frogs appeared to function and behave normally after the trials. With the exception of those retained as voucher specimens, frogs were released after processing.

Statistical analyses

Regression, variance analysis (ANOVA) and independent sample T-test were used. Normality and variance-homogeneity assumptions were tested using Kolmogorov–Smirnov test and Levene test, respectively. Descriptive statistics are presented as mean ± SD and range of data. Pearson’s correlation coefficients were calculated and tested for significance of linear relationship among continuous variables. P < 0.05 was considered statistically significant. All calculations were performed using SPSS12. 0.


The preference body temperatures of P. spinosa in each populations during the photophase (7:00–19:00) was higher than during the scotophase of light/dark (LD) cycles (t = 3.89, P < 0.001; Fig. 2). The mean Tps of six P spinosa populations varied from 22.1 ± 2.5°C in WYS to 24.4 ± 2.5°C in YS. The broadest range of Tp was recorded for YS population, where it varied from 19.4 to 30.3°C (Table 2). Pearson correlations showed there was a significantly positive correlation between the preferred body temperature and substrate temperature in each population after Bonferroni correction (PJ r = 0.871; LS r = 0.905; JGS r = 0.820; YS r = 0.883; GLS r = 0.910; WYS r = 0.861; All P < 0.001).
Fig. 2

Daily rhythm of mean preferred body temperature of P. Spinosa held on thermal gradients under artificial light/dark (LD) cycles. Black bars represent the scotophase of the LD cycle. The white bar represents the photophase of the LD cycle

Table 2

The analysis of the CTMax (°C) and the mean preferred body temperature (°C) of P. spinosa in six populations (mean ± SD)










42.1 ± 1.9



24.4 ± 2.5a




43.4 ± 2.0



23.9 ± 1.4a




42.5 ± 1.3



23.0 ± 2.0ab




43.0 ± 2.6



22.1 ± 2.5b




43.7 ± 1.0



22.3 ± 2.3b




41.6 ± 3.4



22.6 ± 1.6b




F5,35 = 0.911, P = 0.483


F5,194 = 5.882, P < 0.001


F1,35 = 0.431, P = 0.375


F1,194 = 0.376, P = 0.652


F5,35 = 0.522, P = 0.211


F5,194 = 0.616, P = 0.571

The Bonferroni significant difference test is used for multiple comparisons. Means with different superscripts on each column are statistically different

Two-way ANOVAs with sex and locality as the factors revealed that CTMax did not differ between sexes and among localities in frogs, and that the interaction effects between sex and locality on these measures were not significant. Sex and interaction between sex and locality did not also have significant effects in preferred temperatures (Tps). But the Tps significantly differed among six pairs of populations (Table 2). Four P. spinosa populations inhabiting the north (PJ, JGS, WYS and LS) had significantly lower mean preferred body temperatures than conspecific populations in the south (YS and GLS; Table 2). And there was a negative linear relationship between Tps and the latitude of each population (R2 = 0.74, P < 0.001, n = 6; Fig. 3).
Fig. 3

Relationship between mean preferred body temperature (Tp) and latitude (R2 = 0.74, P < 0.001, n = 6) among six populations of P. spinosa along the south of China


Ectotherm, such as lizards held in continuously operating laboratory thermal gradients display distinctive daily rhythms of behavioral thermoregulation whether they are diurnal (Rismiller and Heldmaier 1988; Firth and Belan 1998; Ellis et al. 2007) or nocturnal (Refinetti and Susalka 1997). Generally, this daily rhythm of behavioral thermoregulation is characterized by the selection of higher photophase and lower scotophase temperatures (i.e. voluntary hypothermia) under artificial light: dark (LD) cycles (Regal 1967). Although numerous studies have demonstrated the existence of daily rhythms of behavioral thermoregulation of many reptile species (for review, see Refinetti and Menaker 1992), very few studies have demonstrated this in amphibians. In a field study, P. spinosa showed bimodal pattern of temperature selection in a thermal gradient during the day (Yu et al. 2008). Interestingly, a similar bimodal pattern of body temperatures was also found for P. spinosa in our laboratory study where there were no thermal restrictions (Fig. 2). This may reflect not only changes in the thermal environment but also a voluntary selection pattern. P. spinosa were active at lower body temperatures during the night than during daytime, as indicated by the results from the field study (Yu et al. 2008) and by our laboratory study. Although reasons for such a pattern remain to be examined, one possible explanation might be related to that high daytime Tps may facilitate predator-avoidance by inactive P. spinosa, and low nighttime Tps may reduce locomotory performance, but they may also confer advantages to amphibian, including reduced evaporative water loss (review in Navas et al. 2008) and energetic cost of locomotion and maintenance (Lima et al. 2005).

We found geographical differences in Tps among populations of P. spinosa, with decrease along a south to north gradient, suggesting the Tp may be a result of adaptation to the local thermal environment. Historical weather data on average yearly temperature, average minimum temperature and average maximum temperature show differences in thermal environment among the six localities (Table 1). The data on anuran thermal preferences are scarce (review in Navas et al. 2008), however, some lizards are demonstrating geographic variation in Tp for different populations (e.g. Van Berkum 1986; Bauwens et al. 1995; Grant and Dunham 1990; Andrews 1998; Rock et al. 2000; Du 2006), and others are not (e.g. Yang et al. 2008; Sepúlveda et al. 2008). For long-term fluctuation in climate, as encountered along altitudinal or latitudinal gradients, natural selection may favor evolutionary change of thermal sensitivity (Huey and Bennett, 1990). This may comprise adjustments that affect survival at extreme temperatures (resistance adaptations) and adjustments that affect physiological capacities at non-extreme body temperatures (capacity adaptations; Precht et al. 1973). An example of capacity adaptations are shifts in the position (or shape) of performance functions (Huey and Bennett 1987), indicating shifts in optimal temperature of some physiological or whole-animal performance function. Optimal temperature for many performance functions is often within the range of preferred body temperature (Dawson 1975; Huey 1982). Although many factors are known to affect thermal preference (e.g., nutritional or reproductive state), the Tp is a true characteristic of some ectotherms which can provide information on temperature preferences of field-active (Licht 1968; Angilletta and Werner 1998). However, the intraspecific variation in thermal biology in P. spinosa among our populations under laboratory thermal gradient suggests that the features of thermal physiology in the frog may change along a longitudinal gradient in response to different thermal environments.

The preferred body temperature of P. spinosa showed a strong correlation with substrate temperature, thus, it might initially be considered that P. spinosa demonstrate a certain degree of thermoconformism. The permeability of amphibian skin is often referred to as a morphological constraint limiting the efficiency of behavioral thermoregulation, but evaporative cooling is essential for a number of anuran species, particularly in some specific ecological settings (Navas et al. 2008). Following the pioneering findings by Lillywhite (e.g. 1970, 1971, 1975), we learnt that Australian tree-frogs (Litoria) can set an upper limit to the match between body and air temperatures, and avoid reaching body temperatures above 38°C through evaporative cooling (Buttemer 1990). In contrast to our prediction, we found no statistically significant difference in the ability to endure the extreme high temperature in six P. spinosa populations. This implies that not all features of thermal physiology undergo parallel changes between the populations. Some etotherms could show such unparallel changes of thermal physiology (e.g. the northern grass lizard Takydromus septentrionalis, Du 2006). We suggest that frogs could show such unparallel changes of thermal physiology for three reasons. First, differences in thermal environment along a south–north transect that spanned 4° are too small to induce changes of thermal physiology in this species. The variation in high ambient temperatures of six localities seems to be quite small, maybe explaining why no differences in CTMax are found. Second, the frogs have not experienced strict stresses from extremely high or low ambient temperatures (monthly ambient temperatures are located below the CTMax (Tables 1, 2)). Third, it is generally known that behavioral thermoregulation may considerably buffer the intensity of selection on thermal physiology traits (Huey and Kingsolver 1989). For example, P. spinosa is a nocturnal frog, daytime retreat sites have been shown to be important for overall daily temperature regulation, as demonstrated by other nocturnal ectotherm (Tracey and Christian 2005). And behavior is seemingly more plastic than physiology. Thus, the intensity of selection in these populations was reduced, and not to force evolutionary change of the thermal tolerance. Certainly, these speculations need more corroborative empirical evidence and deserve further studies.

Our results showed acclimation-induced differences in thermal physiology, which may be entirely derived from phenotypic plasticity. It is an interesting possibility. This could be asked in future experiments rising in the laboratory egg masses from amplexed couples of the two extreme-most latitudes.


We are grateful to Baogen Yu for his assistance in the laboratory and Weiguo Du for his comments. We also thank two anonymous reviewers for their constructive comments on a previous version of this manuscript. This research was supported by the Science Technology Commission of Zhejiang Province of China (No. 2006C22031).

Copyright information

© Springer Science+Business Media B.V. 2010