Thermal ecology and thermoregulatory behavior of Coleodactylus natalensis ( Squamata : Sphaerodactylidae ) , in a fragment of the Atlantic Forest of Northeastern , Brazil

We studied the thermal ecology and thermoregulatory behavior of Coleodactylus natalensis Freire, 1999 in a remnant of a northern coastal patch of the Brazilian Atlantic Forest. Data were collected during four 20-day field excursions over the course of one year. We assessed the importance of substrate and air temperatures, in addition to time of exposure to sunlight, as relevant factors for the regulation of body temperature in this species. After each specimen was captured, body (Tb), substrate (Ts) and air (Ta) temperature were measured 10 cm above the ground, using a temperature sensor coupled to a fast response thermo-hygrometer. Ad libitum and focal animal methods were used to describe thermoregulatory behavior. The mean body temperature of C. natalensis was 31.3 ± 3°C (amplitude of 26.9 and 38.4°C, n = 20). A positive relationship was found between Tb and environmental temperatures; further, substrate temperature explained the additional variability of temperature variations in this species. With respect to environmental observations, individuals of C. natalensis did not expose themselves directly to the sun, moving equally between full and filtered sun. Our results indicate that C. natalensis is umbrophylic and a passive thermoregulator.

Like other ectothermal animals, lizards depend on environmental sources to absorb heat (PIANKA & VITT 2003).They use environmental mechanisms to adjust their body temperature and maintain it within a range that meets their physiological and ecological needs (COWLES & BOGERT 1944, HUEY 1982).The thermoregulatory behavior of a lizard involves costs and associated benefits, which reflect individual ecological and thermoregulatory priorities (HUEY & SLATKIN 1976, DOWNES & SHINE 1998).
Behavioral regulation of body temperature can be achieved through different mechanisms of sun exposure (COWLES & BOGERT 1944, BOGERT 1959, PIANKA 1971, HUEY et al. 1977, ROCHA 1988, ROCHA & BERGALLO 1990).These mechanisms play an important role in the active temperature regulation of individual animals (HUEY & SLATKIN 1976).In passive thermoregulation, body temperatures generally reflect environmental temperatures.Active and passive thermoregulation strategies can be considered as the two extremes of a continuum of thermoregulatory options (HUEY & SLATKIN 1976).Body temperature regulation is therefore a complex process, influenced not only by environmental heat sources, but also by the ecological and life history characteristics of the species (ROCHA 1994, KIEFER et al. 2005, 2007, ROCHA et al. 2009).
Coleodactylus natalensis Freire, 1999 (Fig. 1) was initially described as endemic to the Parque Estadual das Dunas de Natal, and is considered to be endemic to the Rio Grande do Norte Atlantic Forest.It is a diurnal lizard that inhabits leaf litter in shady areas (FREIRE 1999, CAPISTRANO & FREIRE 2009), making it a forest species restricted to the Atlantic Forest biome.Other than a description of the species (FREIRE 1999), few ecological studies have been conducted involving reproduction, habitat use, predation, activity period and diet (LISBOA et al. 2008, CAPISTRANO & FREIRE 2009, SOUSA & FREIRE 2010, SOUSA et al. 2010).
The aims of our study were to: (I) evaluate mean body temperature of active C. natalensis in an Atlantic Forest rem-

MATERIAL AND METHODS
Data were collected in a remnant of the Brazilian Atlantic Forest, the Parque Estadual Mata da Pipa (PEMP -6°14'S and 35°03'W, Fig. 2).It is located in the Municipality of Tibau do Sul, state of Rio Grande do Norte, covering an area of approximately 290 ha (RIO GRANDE DO NORTE 2006) at 63 m above sea level.
The climate of the region is classified as sub-humid with a mean temperature of 26.5°C (maximum 32°C/minimum 21°C), mean annual relative humidity of 74% and highest rainfall levels in April and June.The soil is predominantly composed of dystrophic quartz sands, which are relatively flat and excessively drained and deep with extremely low natural fertility and a sandy texture (EMPARN 2010).
We conducted four 20-day field excursions (two in the rainy season: October 29 to November 18 2008 and October 8-28 2009 and two in the dry season: March 4-24 2008 and July 1-21 2008), totaling 80 days of daytime (between 0700 and 1800).Individuals were located by means of active searches in the high forest of PEMP.The percale area used during field work contains large trees (above 15 m) with diameters breast high   (DBH) of over one meter and a canopy that blocks out most of the sunlight, except for some areas with clearings resulting from numerous tree falls.The litter is thick and abundant with extensive resprouting of the underbrush.We explored four 500 m transects in the habitat.Those were located 50 m apart and crossed the area in a mainlandsea direction.Transects were traversed in a linear fashion, deviating 5 m to the right or left every 10 m in order to better cover the microhabitats used by these specimens.Individual animals were collected manually.
When studying C. natalensis specimens, behavioral observations were made ad libitum (ALTMANN 1974), to record behavioral activities at regular time intervals.The focal animal method (ALTMANN 1974), by which a single individual is observed for a given length of time to record behavior frequency and duration, was also used.We recorded the length of time the animals were exposed to three categories of sunlight: shade, filtered sun and direct sunlight (Figs 3, 4, and 5 respectively), as well as the length of time during which animals were motionless or moving.In both methods, 5-minutes observation intervals were established, followed by an additional five minutes for recordings.
At the moment of capture we measured cloacal temperature (considered as body temperature -Tb), as well as substrate (Ts) and air (Ta) temperature (°C)10 cm above ground.This procedure was performed with the help of a temperature sensor (Instruterm ® model S-02K) coupled to a digital thermo-hygrometer (Instruterm ® model HTR-160 -accurate to 0.1°C and 1-s response time).Only body temperatures obtained up to ten seconds after capture were considered.The collected specimens were deposited in the Coleção Herpetólogica do Departamento de Botânica, Ecologia e Zoologia (CHBEZ), Universidade Federal do Rio Grande do Norte (UFRN).
We calculated mean body temperature of active C. natalensis as the arithmetic mean of cloacal temperatures recorded for all the lizards collected.The paired t-test was used to evaluate whether there was a significant difference between air and substrate temperatures recorded for the species.Body temperature dependency (Tb, dependent variable) in relation to that recorded in the microhabitat (Ts and Ta, independent variables) was determined by simple linear regression.If a significant correlation was found, multiple linear regression was carried out using the stepwise method to evaluate if either of the two environmental variables (Ta and Ts) explains an additional variation in the lizard's body temperature (DANCEY & REIDY 2006).
To estimate the degree of behavioral thermoregulation (passive or active thermoregulator), we used the absolute values of the differences between Tb and Ta (⌬TA) and between Tb and Ts (⌬TS) in the module (VRCIBRADIC & ROCHA 1998, KIEFER et al. 2007).The Wilcoxon non-parametric test was used to compare the ⌬TA and ⌬TS values (DANCEY & REIDY 2006).All statistical procedures were calculated with the aid of the software SPSS 15.0 for Windows.
A significant, positive correlation was found between body (Tb) and air temperatures (Ta, r 2 = 0.588, p < 0.001, Tab.I, Fig. 6).The same relationship was found between body (Tb) and substrate temperatures (Ts, r 2 = 0.591, p < 0.001, Tab.I, Fig. 7).However, after disregarding the effect of air temperature, substrate temperature explains part of the additional variation in body temperature during activity.With respect to the degree of behavioral thermoregulation (Fig. 8), median ⌬TA (M = 3.6°C, n = 20, Tab.I) and ⌬TS (3.6°C, n = 20, Tab.I) values were equal, with no significant difference among the remaining values (paired Wilcoxon, T = 15.17,z = -0.525,p = 0.600).The highest percentage of negative values was recorded for ⌬TA (15%), while for ⌬TS it was only 5%.
Our behavioral observations totaled 16 minutes and 20 seconds (n = 9; Tab.II).Coleodactylus natalensis individuals did not expose themselves directly to sunlight in the clearings and were primarily found moving between shade (72.4%) and filtered sun (27.6%, Tab.II).These observations were supplemented by the fact that 53.9% (n = 41) of the total number of Table I.Mean body temperature (Tb) of active Coleodactylus natalensis (n = 20), mean air (Ta) and substrate (Ts) temperatures and median values of the module of differences between Tb and Ta (⌬Ta) and between Tb and Ts (⌬Ts) recorded in the lizard microhabitats, including a summary of paired t-test and simple and multiple linear regressions between the temperatures, as well as the paired Wilcoxon test between ⌬TA and ⌬TS.⌬TA vs ⌬TS T = 15.17;Z = -0.525;p = 0.600 individuals of C. natalensis observed (n = 76) were found in shaded areas.The remaining animals (46.1%; n = 35) were spotted under filtered sun (Fig. 9).
temperature (PIANKA 1977, JAKSIC & SCHWENK 1983, MAGNUSSON 1993).In some cases, a species living in habitats with lower environmental temperatures may also have lower body temperatures than conspecifics in habitats where higher temperatures prevail (KIEFER et al. 2005, KOHLSDORF & NAVAS 2006).The body temperature recorded for C. natalensis, an umbrophile species, may be due to the fact that minute species such as the Sphaerodactylidae exhibit high rates of water loss in relation to their small size and, consequently, greater capacity to absorb heat (MACLEAN 1985, STEINBERG et al. 2007).Furthermore, this species shows tolerance to less shaded areas, as a pre-adaptive trait to warmer environments (FREIRE 1999), which might explain the high body temperatures recorded in two individuals who had body temperatures that possibly fled to the standard population (Figs 6 and 7).However, they were collected in a poorly shaded spot, near the forest edge, where the ambient temperatures were high, confirming that this species is tolerant to warmer environments.
To date, body temperature records for Coleodactylus are nonexistent, since species of this genus are the smallest among all lizards.Their small size hinders the use of methods routinely employed for lizards (VITT et al. 2005).However, the mean temperature recorded for the substrate used by C. natalensis (Tab.I) did not differ from that recorded for the substrate of Amazonian congeners (C.amazonicus and C. septentrionalis - VITT et al. 2005), or from the population of C. natalensis studied in its type-locality (CAPISTRANO & FREIRE 2009).
The body temperature of C. natalensis showed a significant relationship with the temperatures recorded at the capture site (Figs 6 and 7), which suggests that the individuals use behavior to maintain a certain degree of control over their temperature (VITT & CARVALHO 1992).Since substrate temperature offers the best explanation for body temperature variations, we suggest that this species has a thigmothermal behavior.The high proportion of individuals with Tb above Ts, along with the fact that no specimens were seen directly exposed to sunlight (Fig. 9), suggests that this species exhibits a certain degree
The use of habitat by lizards has been mainly studied in the context of interspecific competition and niche partitioning, without considering the thermoregulatory restrictions in microhabitat use (GROVER 1996).However, some characteristics of habitats and (shade level and presence of water bodies) microhabitats used by lizards have an important influence on the food and thermal ecology of these animals (GANDOLFI & ROCHA 1998, HUEY 1982, ROCHA 1994, VAN SLUYS 1992, VRCIBRADIC & ROCHA 1998).
The thermal patterns of lizards may be influenced by phylogenetic and foraging factors, body size, activity period and habitat (HUEY & PIANKA 1983, PIANKA 1986, ROCHA 1994, ROCHA et al. 2009).The low mean temperature recorded during activity for lizard species from forest areas may be due to the lower availability of thermal sources for thermoregulation, as well as lower microhabitat temperatures, compared to open environments (HOWLAND et al. 1990, VITT 1991, VITT et al. 1997, 1998).Our data, therefore, corroborate findings that the thermal characteristics of species are a result of environmental factors and are regulated by behavioral strategies.

Figure 2 .
Figure 2. Location of the Parque Estadual Mata da Pipa (PEMP), municipality of Tibau do Sul, state of Rio Grande do Norte, northeastern Brazil.

Figure 8 .
Figure 8. Absolute values of the differences between Tb and Ta (⌬TA) and between Tb and Ts (⌬TS) in the module for Coleodactylus natalensis (N = 20) in Parque Estadual Mata da Pipa, Tibau do Sul, Rio Grande do Norte, Brazil.

Figure 9 .
Figure 9. Frequency distribution of active Coleodactylus natalensis (n = 76) under different sunlight exposure conditions in Mata da Pipa State Park, Tibau do Sul, Rio Grande do Norte, Brazil.Numbers on top of bars indicate sample sizes.

Thermal ecology and thermoregulatory behavior of Coleodactylus natalensis (Squamata: Sphaerodactylidae), in a fragment of the Atlantic Forest of Northeastern, Brazil Pablo A. G. de Sousa 1, 2 & Eliza M. X. Freire 1
nant; (II) identify environmental factors involved in the thermoregulation process of this species, and (III) quantify exposure time to sunlight, as determinant factors for regulating body temperature.

Table II .
Length of exposure to light conditions and movement for Coleodactylus natalensis (n = 9).