Is it like night and day? Nocturnal versus diurnal perch use by dwarf chameleons (Bradypodion pumilum)

ABSTRACT Arboreal chameleons must navigate a complex, three-dimensional landscape consisting of trees, bushes and/or grasses of various sizes. This complexity equates to the microhabitat of chameleons, that is, the branches upon which they perch and through which they move. Therefore, chameleons rely on their ability to grip perches available to them, and this is evident by their specialised adaptations such as their prehensile tail and their grasping hands and feet. To date, ecological studies of chameleons have relied on gathering data on perch use only at night because locating chameleons during the daytime is extremely difficult. However, the night-time perch represents the sleeping perch of individuals, and this is not necessarily reflective of overall diel perch use. Many other arboreal reptiles are constrained to using thin perches at night, but by day their microhabitat is more variable, using thicker branches, limbs or tree trunks. To assess whether this well-entrenched paradigm (night-time perch use being constrained to thin perches) also extends to fully arboreal chameleons, daytime and night-time perch diameters were measured for the Cape Dwarf Chameleon, Bradypodion pumilum, and were compared using analyses of variance statistical approaches. Neither perch diameters nor variance in perch diameters differed between the two photoperiods, and the variance in diameters of perches used was remarkably high. Despite this variance, there was a significant, positive relationship between body size and the diameter of perches used. Unlike other reptiles, the results show that for B. pumilum the night-time perch use is not different than diurnal perch use, and there appears to be a fairly wide range of perches used by chameleons of a given body size over both photoperiods. Nevertheless, the positive trend for chameleons of larger body sizes to choose larger perches than smaller chameleons implies that there are some upper and lower constraints for the diameters of perches used, possibly relating to their ability to effectively grip a range of perch sizes.

Introduction effectively navigate the complex three-dimensional substrate which they use (e.g. Tolley et al. 2008Tolley et al. , 2013Hopkins and Tolley 2011;da Silva and Tolley 2013;Higham et al. 2015). The main selection pressure for these morphological adaptations is assumed to relate to microhabitat differences, i.e. the diameter of the grasses, twigs or branches through which they move (Hopkins and Tolley 2008;da Silva and Tolley 2013;Tolley et al. 2013). Gripping therefore, is an important ecological trait relating to microhabitat substrate, and thus the size of the perches available to chameleons is considered a driver of morphological adaptation (Herrel et al. 2013). Adaptations for tail, hands and feet allow chameleons to grip the available substrate with the highest possible force in order to maintain a steady purchase (Herrel et al. 2013;Higham et al. 2015). The diameter of perches used, therefore, is considered an important component of their ecology (da Silva et al. 2014;Measey et al. 2014).
Several studies have investigated adaptation of morphological and performance traits in chameleons, particularly for the Bradypodion, a genus of chameleons that is nearendemic to South Africa. The 20 species of Bradypodion occur in a variety of different habitats, ranging from closed-canopy Afrotemperate forests to open canopy grassland and heathlands. These habitats are fundamentally different in structure, i.e. the diameter of available perches (see Higham et al. 2015), and this difference in substrate is an important factor in driving the morphological disparity between species that occur in these diverse habitats, but also for driving convergence and/or conservatism for species that occur in similar habitats (da Silva et al. 2014).
While inferences regarding morphological and performance adaptations relating to microhabitat substrate appear to hold across all species studied to date, a fundamental assumption that underlies these inferences is that the night-time perch size use reflects the daytime perch use. This is because all studies to date have relied upon measurements of perches that were recorded during the night. Their extreme crypsis makes finding chameleons during the day exceptionally difficult. While chameleons can sometimes be observed during daytime, these are rare events, and therefore daytime observations are too few to gather data in the quantity needed for biological studies. Thus, data from chameleons are almost ubiquitously gathered at night when they are easily visible in torchlight. However, chameleons are not active at night and so data on perching site reflects the place at which the individual has chosen to sleep. To date, there has been no comparison of perch sites used during these two photoperiods, so it is unknown whether the night-time perch use correlates to the size of perches that would be used during the daytime.
Given the reliance on night-time perch site data to make inferences on overall chameleon ecology, habitat use and adaptation, it is necessary to establish whether perch data collected at night is an adequate proxy for diel perch use or whether night-time perch measurements represent perch use specifically relating to the sleeping perch. For some arboreal or semi-arboreal reptiles (e.g. Anolis, agamids, geckos, snakes), sleep sites can be very different in structure from the substrates used during daytime, often with thin branches used at night as opposed to thick branches, trunks or terrestrial habitats used during the day (Chandler and Tolson 1990;Reaney and Whiting 2003;Vitt et al. 2002Vitt et al. , 2003Ikeuchi et al. 2012;McCranie and Kohler 2015;Cabrera-Guzmán and Reynoso 2010;Mohanty et al. 2022). There is even variation within the family Chamaeleonidae, with the species of some genera selecting arboreal perch sites at night but moving to the ground to forage in the leaf litter by day (e.g. Raxworthy 1991;Razafimahatratra et al. 2008;Miller 2017;Villeneuve 2017). Much of the literature on sleep ecology of reptiles therefore points to a disparity of perch sites between diurnal and nocturnal periods, casting doubt on the use of night-time perches as a proxy for overall diel perch use.
Given the differences between diurnal and nocturnal perch use in reptiles investigated to date, it is possible that there are different selective pressures operating at night than in the day, and therefore individuals carefully select night-time perches based on specific criteria. These sleep sites must serve as a resting place and safe refuge for approximately 12 hours, during which time, an individual is not able to actively evade predators (see Mohanty et al. 2022). For chameleons, and other reptiles that sleep in the arboreal setting, sleep sites comprise unstable, thin branches that presumably offer some measure of safety from predators (Mohanty et al. 2016(Mohanty et al. , 2022. The weight of a predator probably would not be supported on a thin branch (Chander and Tolson 1990;Storks and Leal 2022), and/or the movement by predators would amplify down the branch and be felt strongly toward the tips of the branch alerting the sleeping individual and allowing fast escape by dropping to the ground (Raxworthy 1991;Measey et al. 2014;Mohanty et al. 2016Mohanty et al. , 2022. Therefore, there is general consensus that perches used for sleep tend to be thin and unstable, chosen to minimise predation risk. However, for chameleons there has been no comparison of perches used during the night-time to those used during the daytime, so the paradigm that chameleons constrain themselves to using thinner perches at night than during the day is unchallenged. Whether the diameters of perches used differs between the two photoperiods has never been explicitly tested for chameleons because of the difficulty in gathering data during the daytime. However, during previous studies on the Cape Dwarf Chameleon (Bradypodion pumilum), daytime perch measurements were recorded from radiotracked individuals (Rebelo et al. 2022). Night-time perches were recorded from the same radio-tracked individuals and also during capture-mark-recapture studies (Tolley et al. 2010;Katz et al. 2013Katz et al. , 2014Rebelo et al. 2022). The availability of these yet unpublished perch use data makes it possible to examine whether daytime and night-time perch use is dissimilar for this population of chameleons, as has been recorded for other arboreal reptiles. If perch diameters are dissimilar between photoperiods, this would support the assumption the night-time perches are fundamentally different from those used during the daytime, and that the night-time sleeping perch is not representative of diurnal and/or overall diel perch use. In addition, if the variation in perch use is lower for the night-time perches, then perch use may be more constrained for sleeping sites than for daytime perch use. To test these assumptions, perch diameters collected from radio-tracked and marked chameleons were compared between photoperiods for B. pumilum.

Materials and methods
Chameleon perches were measured at a single study site in Cape Town, South Africa (34°06'43" S; 18°22'44" E, Figure 1) during field work carried out in 2009, 2010 and 2014 (Tolley et al. 2010;Katz et al. 2013Katz et al. , 2014Rebelo et al. 2022). The vegetation at the site is moderately transformed (originally Sand Fynbos vegetation) with a mixture of indigenous and non-native bushes, trees, reeds and grasses. The vegetated area is bordered by suburban housing, sports fields and urban green space (i.e. mowed grassy patches). There is no artificial lighting at the site, and the suburban houses in the area are >300 m distant and project minimal light pollution toward the site. Thus, it was assumed the sleep sites of chameleons are not biased by the lighting in this peri-urban setting (see Mohanty et al. 2021). Perches used by chameleons were measured during two photoperiods i.e. daytime and night-time. All perch diameters were measured to the nearest 0.1 mm using a digital calliper, at a point on the branch beneath the body of the chameleon as close to the front legs as possible without disturbing the animal.
The daytime perches were measured during a radio-tracking study of 13 adult chameleons in 2010 and 2014 (2010: females = 4; males = 4, 2014: females = 2; males = 3). The radio-tracking study was designed to collect data on chameleon movements (see Rebelo et al. 2022), but ad hoc perch diameter measurements were also made from these individuals. The animals were located through triangulation of radio transmitter signal (Rebelo et al. 2022). Sex and body size (snout-vent length; SVL) of each individual was recorded upon the first capture event. Daytime perch measurements were taken after the chameleons became active (e.g. after 06:00) and ceased once the chameleons were no longer active (e.g. usually by 17:00). For the daytime perch data collection, several measurements per day (usually hourly) were recorded for each individual (total females: n = 161, total males: n = 218). If chameleons were in transit when located, the perch measurement was only taken after they had settled on a branch (i.e. did not move for several minutes).
The night-time perch data were collected using two methods. Firstly, during the radiotracking study, the last perch measurements of the day (usually after 17:00), or the first measurement of the morning (usually as of 06:00) were considered representative of night-time perches given that this characterised the times when chameleons were not active and were settled on their sleeping perch. This premise was confirmed given that each chameleon was on the same perch in the morning that it had been on or after 17:00 the evening before (A Rebelo, pers. obs.). Of the 13 radio-tracked chameleons, 178 of these night-time perch measurements were taken over the course of about two weeks (females = 88, males = 90). Additional night-time measurements were taken at the same study site during monthly capture-mark-recapture (CMR) surveys beginning in late 2009 and concluding in 2010. Sleeping chameleons were located by torchlight after full darkness (see Tolley et al. 2010;Katz et al. 2013Katz et al. , 2014. CMR data were originally collected for studies on survival, abundance and population genetics (see Tolley et al. 2010;Katz et al. 2013Katz et al. , 2014, so perch data were collected ad hoc and were not taken for all individuals encountered. Of those measured, the perch diameter was measured only once, at its first encounter during the CMR study (i.e. one measurement per individual, 61 females, 71 males).
A total of 689 perch diameter measurements were taken for chameleons from the study site during daytime and night-time (females = 310; males = 379). Of these, perch measurements taken for 37 females and 26 males were considered outliers (ranging from 6.5-21.9 mm for females, 6.9-17.4 mm for males), identified by being more than 1.5 times the 75th and 25th interquartile range of values recorded (i.e. Tukey's Hinges). Outliers were removed from the dataset, for a final dataset of 626 perches (273 females, 353 males; Table 1). The data were checked for heteroscedasticity and were found to be homoscedastic (Breusch-Pagan test: females: c 2 = 0.66, p = 0.80; males: c 2 = 0.98, p = 0.32). Therefore, parametric approaches were considered appropriate to use (see Schmider et al. 2010;Blanca Mena et al. 2017).
To assess whether daytime perch diameters were significantly different from night-time, the diameters for these two photoperiods were compared. Perch height in Bradypodion has been shown to be influenced by sex, with males usually perching higher than females (see Hopkins and Tolley 2008;da Silva and Tolley 2013;Rebelo et al. 2022). To minimise any bias that perch height might have on diameter, the sexes were analysed separately. All perch diameters available for daytime and night-time were compared using a mixed model analysis of variance (ANOVA) in SPSS v26 (IBM Statistics) that incorporated photoperiod (day/night) as a fixed factor and sampling year (2010/2014) as a random factor. In addition, the ANOVA was run using a subset of data from only the radiotracking study to assess whether the method of locating chameleons (i.e. radio-tracking or CMR study) might have introduced bias. Only the radio-tracking subset was run for this additional analysis, as the CMR data did not consist of daytime and night-time measurements. As with the analysis of the combined dataset, photoperiod was the fixed factor and sampling year was included as a random factor. Furthermore, given that perch measurements of radio-tracked chameleons were measured for the same individuals over the course of a few weeks, potential dependence in the dataset was examined using a mixed model ANOVA with photoperiod as the fixed factor and chameleon individual (ID) as a random factor. Given the small sample size of radio-tracked chameleons, sampling year was not included as an additional random factor in this latter analysis.
While hind foot size has been shown to correlate with the size of perches used for some species of Bradypodion (da Silva et al. 2014), the hindfeet had not been measured for the present study. Therefore, body size (SVL) was assessed as a potential covariate for the diameters of perches used, as SVL strongly correlates with both hand and feet size in Bradypodion (K.A. Tolley unpubl. data) and has been shown to be positively correlated with perch diameter for the chameleon Brookesia decaryi (Razafimahatratra et al. 2008). Linear regressions for SVL and perch diameter were run separately by sex and showed a positive significant relationship (see Results). Therefore, to correct for any potential bias in body size, a mixed model analysis of covariance (ANCOVA) was also run on the combined dataset (both radio-tracking and CMR studies) with SVL as the covariate, photoperiod as the fixed factor and sampling year as a random factor.
To assess whether the variance in perch diameter was different between photoperiods, Levene's test of equality of variances was used. For this test, the assumption is that the two photoperiods have equal population variance, and this can be tested through examination of the sample variances. If the sample variances differ, the null hypothesis of equal population variance can be rejected.

Results
The mixed model ANOVA did not show any significant differences in perch diameter between the photoperiods for either sex for the combined dataset of perch measurements from the CMR and the radio-tracking studies (Table 2). In addition, sampling year had no effect. When the subset of data from only radio-tracked chameleons was run through the same analysis, there was no significant difference between photoperiods for females or for males (Table 3). When chameleon ID was incorporated as a random factor, there was no significant difference in photoperiod for females (Table 3). For males there was no significant difference for photoperiod or the interaction term, although there was a significant difference among chameleons individuals (photoperiod: F = 0.36, p = 0.55; chameleon ID: F = 8.92, p < 0.01; interaction term: F = 0.57, p = 0.72).
Post hoc examination of the residuals for males shows that this effect is due to a single individual (N522: Supplementary Material, Figure S1) who utilised larger perches than three other radio-tracked males (ANOVA: F = 2.40, p < 0.05, Tukey's post hoc test for multiple comparisons p < 0.05, Supplementary Material Table S1). There were only five perch measurements made for this individual (4 daytime, 1 night-time; see Supplementary Material, Figure S1), and therefore these might not have been a good random sample of measurements. Excluding these five measurements from N522 and taking body size into account, there was no significant effect for chameleon ID for males (Table 3). There was a significant, positive regression between SVL and perch diameter for both females (r 2 = 0.09, p < 0.001) and males (r 2 = 0.12, p < 0.001). The r 2 values were not high, but this is probably a result of the notable variation in perch diameter across the range of SVL values (Figure 2). For example, female chameleons larger than 80 mm SVL used perches ranging from approximately 2.0-5.8 mm, whereas female chameleons between 50-55 mm SVL were recorded on perches from about 0.8 to 6.1 mm (Figures 2 & 3). In addition, the ANCOVA using SVL as the covariate did not show any significant differences in perch diameter between the daytime and night-photoperiods for either sex for the combined dataset of perch measurements from the CMR and the radio-tracking studies (Table 2).
Levene's test indicated that the variation in perch diameters used was not significantly different between photoperiods for either females (W = 0.01, p = 0.93) or males (W = 0.45, Table 2. Results of the mixed model analyses of variance between daytime and night-time perch diameters for Bradypodion pumilum for data collected during radio-tracking and capture-markrecapture. The results are given for the analyses of variance (ANOVA) and the analyses of covariance (ANCOVA) using body size as the covariate. F values and significance values (sig) are shown for the fixed factor (photoperiod), random factor (year) and the interaction (photoperiod + year).

Discussion
Unlike what has been found for other arboreal reptiles, the diameters of perches used by Bradypodion pumilum did not differ between photoperiod for either sex, regardless of whether perch diameter was corrected for chameleon body size. In addition, the method of locating chameleons (radio-tracking or CMR study) did not appear to introduce any bias in the result. Perch diameter was remarkably variable for both photoperiods, with a large range of sizes being used by all chameleons of all size classes. The diameter variance was not different between photoperiods, suggesting that perch use is not more constrained during the night-time than the daytime, contrary to the prediction based on other arboreal reptiles. Overall, the results suggest that photoperiod does not differentially affect the diameters of perches used, and night-time perch measurements can be viewed as representative of diel perches. Thus, it is reasonable to assume that the incorporation of night-time perch measurements as an ecological variable to examine and explain morphological adaptations is appropriate (e.g. Hopkins and Tolley 2008;da Silva and Tolley 2013;Herrel et al. 2013;Higham et al. 2015) and future studies that require habitat data for chameleons can indeed rely on measurements that are gathered during the night-time. Unlike other reptiles that shelter in holes, in rock cracks, under stones or other debris at night, arboreal reptiles, including chameleons, are assumed to sleep on exposed, thin branches (Mohanty et al. 2022) presumably to evade predation by nocturnal predators, primarily snakes (Chandler and Tolsen 1990;Tolley and Burger 2007;Mohanty et al. 2016Mohanty et al. , 2022Bors et al. 2020). However, the premise that night-time perches used are thinner, as compared to daytime substrates, had not been explicitly tested. In contrast to most arboreal reptiles that shelter in trees or bushes to sleep, fully arboreal chameleons are not known to move out of the matrix of branches during the day onto the ground, tree trunks or limbs. Thus, it is logical to reason that fully arboreal chameleons are entirely dissimilar from other arboreal reptiles in this respect, and there is no real expectation that the diameters of perches used during daytime should differ from those used at night, particularly as the range of perches used are all fairly thin and unlikely to support the weight of a predator. Indeed, night-time perch choice may be instead heavily influenced by the position of the branch rather than the thinness of the branch, such that arboreal predators find it difficult to reach their potential prey (Yorks et al. 2003). Along with assessing the original concept that perches used at night are thinner than those used during the day was a corollary that the diameters would be constrained during the night-time (i.e. less variable), particularly as it has been suggested that chameleons use thin perches to reduce predation risk at night (Raxworthy 1991;Measey et al. 2014;Mohanty et al. 2016;Mohanty et al. 2022). Implicit in that assumption is that during the daytime, fully arboreal chameleons should be more malleable, using a variety of different sized perches in order to utilise the highly complex three-dimensional microhabitat. This hypothesis was not supported however, with the variance in perch diameter being equivalent across photoperiods. Therefore, chameleons were not constrained to thin branches at night, but used a wide range of perch diameters. Again, chameleons do not conform to the expected paradigm that seems to apply to most arboreal reptiles of being constrained to thin perches at night, but the explanation may lie in the fundamental differences between how most arboreal reptiles and fully arboreal chameleons interact with their environment. On the foraging spectrum (Cooper 2005), chameleons fall within the range of being ambush foragers (Keren-Rotem et al. 2006), requiring prey to actively come within striking range of their ballistic tongue. Dietary studies support this notion, with B. pumilum electing to consume active prey over sedentary prey , although chameleons appear to be labile in their foraging strategy depending on prey availability Carne and Measey 2013;Dollion et al. 2017). Nevertheless, chameleons spend most of the day motionless making only brief, infrequent movements and this could make them vulnerable to predators, much as they are at night. Although they are alert during the day, they rely on crypsis instead of speed or refuge to evade predators making the mode of escape similar in the two photoperiods. Furthermore, snakes can be active diurnal predators of chameleons (e.g. Dispholodus typus) and have been observed preying upon chameleons during the day (A Turner, pers. comm.). Thus, perching on thin branches could be similarly important during the day as at night for arboreal chameleons, which may not be the case for most other arboreal reptiles that utilise larger branches, limbs, trunks or the ground during the day. What has not been typically emphasised in studies of arboreal reptiles, certainly not for chameleons, is that other factors besides predator avoidance could influence the diameters of perches used (although see Mohanty et al. 2022). For example, a steady grip on a slightly larger perch would be advantageous during adverse conditions (e.g. wind and storms), it may be beneficial to be positioned deeper into the vegetation or closer to the ground during cold or rainy periods such that thermoregulation is maximised, and/or that the relative position of conspecifics is exploited to the individual's advantage. It may even be important to consider body mass and tensile strength of the different substrate types to assess how this influences perch diameter used. While positioning on thin perches most certainly reduces predation risk, it is more likely that multiple factors determine where an individual positions itself for sleep and by day, and that the diameters of perches used is likely to be highly variable depending on the combination of intrinsic and extrinsic factors.
The results for B. pumilum also show a considerable degree of variation in perch sizes used for chameleons of all body sizes over both photoperiods. It should be noted that body size may not be the best predictor of perch use, and that other morphological traits, such as hand/feet size, tail length or even limb length might be better predictors (e.g. Herrel et al. 2011;da Silva et al. 2014), and these could show a stronger relationship for a narrower range of perches used relative to those morphological traits. Nevertheless, there is a significant positive relationship between body size and perch diameter for B. pumilum, indicating that there is probably an optimal range of perch diameters for an individual of a given size. While beyond the scope of this study, it can be assumed that if perches used are too small or too large, gripping would be less effective and an individual could lose its purchase during aggressive encounters with conspecifics (e.g. Stuart-Fox and Whiting 2005;Stuart-Fox 2006;Stuart-Fox et al. 2006b), could be dislodged by predators (see Stuart-Fox et al. 2006a), or could simply fall off during adverse weather. Despite the apparent flexibility for perch use, the correlation with body size suggests that there are some limitations to the range of perches used, outside of which gripping performance probably becomes ineffectual. For the population of B. pumilum investigated, perches smaller than 1 mm and larger than 6 mm were not commonly used, regardless of body size, reinforcing the notion that gripping ability is reduced on perches outside this range. Indeed, reduced grip strength on large perches (10 mm) has been recorded for Bradypodion (e.g. Herrel et al. 2011;da Silva et al. 2014). It should also be noted however, that lower limit of perches used may also be compounded by the effect of body mass in relation to the tensile strength of the perch, and this can be deduced given the positive correlation between perch diameter and body mass in arboreal reptiles (Irschick et al. 1997;Sawant et al. 2013).
While these results demonstrate that this population of chameleons uses perches of a similar size range regardless of whether active in the day or sleeping at night, this study also raises some interesting new questions. For example, is perch use more variable in species that occur in habitats with intrinsically diverse substrates than those that occur in habitats with more homogeneous substrates? Chameleon species that occur in homogeneous habitats (e.g. heathlands or grasslands) might be comparatively more constrained for perch use than species that occur in habitats with heterogeneous substrates, such as forests. Assuming that perch use is a driver of adaptation for chameleon morphology, species in homogeneous habitats should be under stabilising selection to maintain the morphological trait states (e.g. small body size, small hand/foot/tail sizes) best suited to a small range of perch sizes available to them (e.g. Herrel et al. 2013). These species could potentially become stuck on adaptive peaks for their hand/foot/tail morphology, able to grip only a small range of perch sizes. In contrast, species that occur in more structurally complex habitats (e.g. forests) may be more able to cope with a range of conditions, and this might be a trait that inherently allows them to be 'pre-adapted' to changing conditions. In addition, are chameleons actively choosing certain perches while avoiding others? Comparison to perches available versus those chosen does suggest this is the case for chameleons from some microhabitat types but not others (Higham et al. 2015). Do chameleons use different perches (whether by day or night) in anthropogenically transformed habitats as compared to natural habitats? Clearly, these are avenues of research that could help to provide knowledge on which species are labile enough to rapidly adapt to the novel habitats that are arising in the Anthropocene (e.g. Winchell et al. 2016).

Conclusions
For some arboreal reptiles studied to date, daytime perch use seems to differ fundamentally from those used at night (Chandler and Tolson 1990;Raxworthy 1991;Reaney and Whiting 2003;Vitt et al. 2002Vitt et al. , 2003Razafimahatratra et al. 2008;Cabrera-Guzmán and Reynoso 2010;Ikeuchi et al. 2012;McCranie and Kohler 2015;Mohanty et al. 2022). However, this paradigm does not easily extend to fully arboreal chameleons, given that B. pumilum uses similarly sized branches throughout the diurnal and nocturnal periods. In addition, while sleeping sites have been quantified and described for a number of arboreal or semi-arboreal reptiles supporting the notion that sleeping sites are constrained to thin perches (e.g. Cabrera-Guzmán and Reynoso 2010; Mohanty et al. 2016;Bors et al. 2020;Storks and Leal 2020;Kaiser and Kaiser 2021), for many species there has been no comparative framework that includes diurnal perch diameters used. Consequently, it can be challenging to make inferences regarding whether there are constraints on the sizes of perches used at night unless these are compared to those used during the day. As found for B. pumilum, the two may not be fundamentally different, and this could have important ramifications for interpreting the factors dictating sleep ecology. It should be noted that inferences from this study stem from measurements made from a single species from one study site. While it is likely that the interpretations made here also apply to fully arboreal chameleons in general, it could be useful to corroborate these findings by investigating diurnal perch use for additional chameleon species to establish whether these conclusions can be considered universal for fully arboreal chameleons. Finally, the disparate results found for this population of chameleons suggest that it would be useful to broaden studies on the sleep ecology of arboreal reptiles by including comparisons to daytime microhabitat substrates as often as possible.
helpful discussions and feedback. Two anonymous reviewers provided very helpful comments that improved the manuscript. This work was supported by the National Research Foundation of South Africa (NRF) though grants to the author (grant# 120332 (Dimensions of Biodiversity), 72890 and 85413) and to A. Rebelo (NRF Honours Innovation Scholarship). Data collected for this study was carried out under research permits from CapeNature (0035-AAA007-00056 and AAA004-00322-0035, 056-AAA08-0027) and under research agreements with South African National Parks. Ethics approvals were issued by South African National Biodiversity Institute (002/10 and 001/2014) and University of Cape Town (2010/V21/EK and 2014/V4/RA).