The evolution of climatic niches in squamate reptiles

Transcription

1 Downloaded from on July, 18 rspb.royalsocietypublishing.org Research Cite this article: Pie MR, Campos LLF, Meyer ALS, Duran A. 17 The evolution of climatic niches in squamate reptiles. Proc. R. Soc. B 8: Received: 9 February 17 Accepted: 1 June 17 Subject Category: Evolution Subject Areas: evolution, ecology Keywords: squamata, diversification, climatic niche evolution, heterotachy, MOTMOT, ARBUTUS Author for correspondence: Marcio R. Pie marcio.pie@gmail.com Electronic supplementary material is available online at figshare.c The evolution of climatic niches in squamate reptiles Marcio R. Pie 1,,3, Leonardo L. F. Campos 1, Andreas L. S. Meyer 1, and Andressa Duran 1,3 1 Departamento de Zoologia, Universidade Federal do Paraná, Curitiba, Paraná, Brazil Programa de Pós-Graduação em Zoologia, and 3 Programa de Pós-Graduação em Ecologia e Conservação, Universidade Federal do Paraná, CEP Curitiba, Paraná, Brazil MRP, Despite the remarkable diversity found in squamate reptiles, most of their species tend to be found in warm/dry environments, suggesting that climatic requirements played a crucial role in their diversification, yet little is known about the evolution of their climatic niches. In this study, we integrate climatic information associated with the geographical distribution of 188 squamate species and their phylogenetic relationships to investigate the tempo and mode of climatic niche evolution in squamates, both over time and among lineages. We found that changes in climatic niche dynamics were pronounced over their recent squamate evolutionary history, and we identified extensive evidence for rate heterogeneity in squamate climatic niche evolution. Most rate shifts involved accelerations, particularly over the past Myr. Most squamates occupy similar regions of the climatic niche space, with only a few lineages diversifying into colder and humid climatic conditions. The changes from arid to mesic conditions in some regions of the globe may have provided opportunities for climatic niche evolution, although most lineages tended to remain near their ancestral niche. Variation in rates of climatic niche evolution seems common, particularly in response to the availability of new climatic conditions over evolutionary time. 1. Introduction An intriguing feature of many ecological phenomena is that they are fairly obvious to observe, but remarkably difficult to explain. One such phenomenon is the uneven distribution of species across the planet. Although many potential correlates of species richness have been identified (e.g. temperature, precipitation and primary productivity [1 3]), these correlates often fall short when similar conditions are associated with substantially different levels of species richness [,]. These discrepancies are not surprising, given that the location and extent of the distribution of a given species is the complex outcome of a variety of mechanisms operating at different ecological and evolutionary scales [, 1]. As a consequence, a complete theory of species distribution patterns must necessarily be framed in terms of the mechanisms that actually modify species numbers, namely speciation, extinction and dispersal [11 13]. Interestingly, all three of these mechanisms are influenced to some extent by the ecological niche of a species. The ecological niche has been defined as multidimensional hypervolume space within which a species can persist [1]. More recently, niche definitions have been separated into two main classes: Grinnellian and Eltonian [1]. While the former is related to non-interactive broad-scale environmental variables, the later focuses on biotic interactions [1]. To date, most studies that addressed the influence of ecological niches on speciation, extinction and dispersal have focused on the Grinnellian class, particularly on the climatic niche of a species (here defined as the set of climatic conditions associated with the occurrence and persistence of a species; e.g. [1 18]). Although relevant progress has been made, our current understanding of how climatic niches evolve is still incipient. & 17 The Author(s) Published by the Royal Society. All rights reserved.

2 Downloaded from on July, 18 Studies on closely related species and populations often find rapid evolution in climatic niches (e.g. [19,]), which could suggest that climatic niches are evolutionarily labile. When viewed at broader evolutionary timescales, however, niche evolution appears to be much more constrained, with most changes occurring along a few major dimensions. For instance, although closely related, Primates exhibit high diversity in their climatic niches [1], nearly % of the climatic niche evolution in Primates involves changes along a temperature axis, particularly during the winter months, and half of the remaining variation is distributed on a precipitation axis, especially during the driest months [,3]. Few studies have investigated climatic niche evolution using methods that explicitly test for variation in evolutionary rates among lineages. Duran & Pie [] used a recently developed method based on a reversible-jump Markov chain Monte Carlo approach to test for variation in rates of climatic niche evolution in different primate clades (i.e. Platyrrhini, Strepsirrhini and Catarrhini) and found evidence for multiple independent rate shifts. Interestingly, these shifts were concentrated on the past 1 Myr, which is corresponds to a period of overall world climatic cooling []. Indeed, most of the observed shifts involved rate accelerations associated with the adaptation of lineages from an ancestral warm/wet climatic niche to novel colder climatic conditions. These observations suggest a scenario in which lineages tend to maintain their ancestral climatic niches, but also evolve into novel areas of climatic niche space as they become available during the evolutionary history of a clade. A test of such a scenario for the evolution of climatic niches could be attained by investigating another clade with comparable age but with a distinct ancestral climatic niche, which therefore would respond differently to changing world climatic conditions. One such clade is Squamata, the largest extant order of reptiles and the second-largest order of vertebrates, with over 99 species worldwide []. Squamates are found on every continent outside Antarctica, thus being exposed to a variety of climatic conditions during their evolutionary history []. Given that squamates are ectotherms, they might experience different constraints in relation to environmental conditions than endotherms [,7]. For example, the extent to which climate influences species distribution is greater in ectotherms than in endotherms [8], which may have led to a stronger effect of historical climate change on squamate climatic niche evolution. More importantly, squamates are ancestrally associated with warm/dry environments that were already present prior to the period of accelerated climatic niche evolution in Primates. Therefore, one could predict that climatic niche evolution would be more protracted than that of Primates, even though particular clades exhibit particular life history traits (e.g. dispersal ability and generation time) that might allow them to independently accelerate the evolution of their climatic niches to occupy novel habitats. In this study, we test this scenario with a large-scale analyses of climatic niche evolution in several squamate clades. The specific objectives of our study are (i) to describe the general structure of squamate climatic niche space; (ii) to test for variation in rates of climatic niche evolution, both among clades and over time, and (iii) to assess the extent to which niche evolution corresponded to large-scale climatic changes over the course of their history. We found extensive evidence for rate heterogeneity (heterotachy) in squamate climatic niche evolution, both over time and among lineages.. Material and methods We obtained information on the geographical distribution of squamate species using shapefiles from the International Union for Conservation of Nature database [9] (available at Climatic information was obtained from WorldClim v. 1. (available at which comprises data on 19 temperature and precipitation variables at a spatial resolution of arc-min (1 km at the Equator) [3]. For each distribution map we extracted the climatic information from all grid cells within the species distribution polygons and calculated the mean values for each variable for each species using the packages RASTER.-31 [31], RGDAL.9-1 [3], MAPTOOLS.8-3 [33] and PLYR [3] (see electronic supplementary material, table S1 for the raw dataset). When climatic niches are measured from the environment associated with species distribution ranges, what is measured is an approximation of the realized climatic niche of the species (i.e. a part of the fundamental climatic niche occupied by the species, after accounting for factors such as dispersal [3]). Information on species fundamental niches would probably provide a more thorough evaluation of the climatic niche evolution in squamate reptiles [3]. However, obtaining such information for hundreds of species is challenging, given that it involves the use of experimental methods. Therefore, the term climatic niche hereafter refers to an approximation of the realized climatic niche of the species. The phylogenetic relationships and the corresponding divergence times among squamate species were obtained from Zheng & Wiens [37]. We used in our analyses those clades for which we had simultaneously climatic and phylogenetic information, namely Gekkota (97 species), Scincoidea (8 species), Lacertoidea (179 species), Iguania ( species), Anguimorpha ( species) and Serpentes ( species), and an additional three species as a sister group to all squamates, accounting for 188 squamate species. This dataset comprises approximately 19% of currently recognized squamate species (lizards and snakes) [38], which, although not complete, should be a representative sample of the evolutionary patterns for their entire clades. We performed a phylogenetically corrected principal component analysis (ppca) [39], based on the covariance matrix of intraspecific mean bioclimatic variables, using the phyl.pca function from PHYTOOLS.- [], given that it has been recently suggested that using PC scores from standard PCA could bias the fit of alternative models of evolution [1]. We retained only those PC axes that explained the largest variation in the dataset, according to the broken-stick criterion [], thus characterizing the main axes of variation of squamate climatic niches. We transformed raw data into z-scores prior to the ppca to account for variation in the measurement units among variables. Scores of the selected PCs were used in the remaining analyses. Given that several recent studies have shown that the fit of simple evolutionary models can provide misleading results [3,], particularly in the case of large datasets [,], we used ARBUTUS [] to evaluate the adequacy of trait evolution models []. ARBUTUS first fits a model of evolution to a dataset and then simulates datasets using the estimated parameters (N ¼ 1 in this study). Test statistics are then calculated on each simulated dataset and contrasted to the original one, such that discrepancies between the simulated and the observed datasets indicate the ways in which the model does not capture specific properties of the original dataset []. The six test statistics implemented by ARBUTUS were: M SIG is the mean of the squared independent contrasts and detects potential violations regarding the overall rate of trait evolution; C VAR is the coefficient of variation of the absolute value of the independent contrasts and evaluates whether the evolutionary model is properly accounting for the rate heterogeneity across the phylogeny; S VAR, S ASR and S HGT are based the slope of a linear model fitted to the absolute value of the contrasts against their expected rspb.royalsocietypublishing.org Proc. R. Soc. B 8: 178

3 Downloaded from on July, 18 (a) PC (17%) Iguania 3 Anguimorpha Scincoidea Gekkota Lacertoidea Serpentes 1 PC1 (%) 1 (b) bio 1 bio 3 bio 19 bio 1 bio 17 bio 11 bio 1 bio 1 bio 13 bio bio 9 bio 18 bio 8 bio 1 bio bio 1 PC1 Figure 1. Climatic niche space of squamate species represented by the scores of the first two PC axes (a). Each point represents the mean score for each of the 188 species. High scores on PC1 indicate arid climatic conditions, particularly in drier months, whereas high scores on PC indicate warmer climatic conditions. Numbered points represent examples of species that occupy extreme regions the climatic niche space: (1) Hemidactylus albofasciatus, () Brachymeles minimus, (3) Enyalioides heterolepis, () Oligosoma acrinasum. The arrows in panel (b) indicate the correlations of climatic variables along the first two axes of the ppca (see also electronic supplementary material, table S): annual mean temperature (bio 1), mean diurnal range (bio ), isothermality (bio 3), temperature seasonality (bio ), maximum temperature of warmest month (bio ), minimum temperature of coldest month (bio ), temperature annual range (bio 7), mean temperature of wettest quarter (bio 8), mean temperature of driest quarter (bio 9), mean temperature of warmest quarter (bio 1), mean temperature of coldest quarter (bio 11), annual precipitation (bio 1), precipitation of wettest month (bio 13), precipitation of driest month (bio 1), precipitation seasonality (bio 1), precipitation of wettest quarter (bio 1), precipitation of driest quarter (bio 17), precipitation of warmest quarter (bio 18) and precipitation of coldest quarter (bio 19). Temperature and precipitation variables are represented as dashed and continuous arrows, respectively. (Online version in colour.) bio 1 bio bio 7 3 rspb.royalsocietypublishing.org Proc. R. Soc. B 8: 178 variances, their ancestral states inferred at the corresponding node and their node depth, respectively. Finally, D CDF tests if the independent contrast distributions follow the normal distribution, as expected for contrast evolving under Brownian motion. S VAR allows for testing whether the expected variances in the contrasts are proportional to the branch lengths, S ASR is used to test if the variation in evolutionary rates are related to the trait values, S HGT evaluates if the variances in the contrasts are related to time []. (The complete descriptions of the test statistics can be assessed in table 1 from Pennell et al. [].) We used ARBUTUS based on the Brownian motion model [7], which describes the scenario where traits evolve as a random walk process, with traits changing with a constant rate, non-directional and with their magnitude independent of current or past states [7]. This model was chosen given that it represents a simple starting point for the understanding of the evolutionary dynamics and thus would more efficiently indicate the departures from a constant-rate model. General patterns of temporal variation in climatic niche evolution was studied using two complementary methods. First, we used disparity-through-time plots (DTT) [8], as implemented in GEIGER V...3 [9]. This metric is calculated based on the Euclidean distances between PC scores, with the disparity of subclades being calculated by dividing the average disparity of the entire tree by the average of each clade. The observed data are then compared with a null expectation following 999 simulations under an unconstrained Brownian motion model. The second approach was the phenogram function from PHYTOOLS 3.1. [], in which the phylogeny is plotted in a space defined by the climatic trait (PC axes) in the Y-axis and the time on the X-axis, such that the position of each node corresponds to the maximum-likelihood estimate of the corresponding ancestral state for each of the selected PC scores. Variation in climatic niche evolution among lineages was assessed using the approach called models of trait macroevolution on trees (MOTMOT) [], as implemented in MOTMOT package 1.1. []. This method uses maximum-likelihood evaluation, using phylogenetically independent contrasts, to infer the number and the position on the phylogeny where changes in evolutionary rates occurred, showing also the direction of change of these evolutionary rates. MOTMOT analyses were run using the transformphylo.ml function, using both algorithms (tm1 and tm), with the maximum number of rate shifts set to 1 and without estimating the minimum clade size. The MOTMOT analyses were repeated twice for each selected PC to ensure convergence. All analyses in this study were carried out in R [1]. 3. Results The ppca efficiently reduced the dimensionality of the dataset, with the first three PC axes accounting for 83% of the variance in the data (see electronic supplementary material, table S for ppca loadings). The first PC reflected variation in precipitation, especially during the driest and coldest months (e.g. precipitation of driest months), whereas the second PC corresponded to variation in mean temperature ( particularly mean annual and during the coldest quarter) and precipitation during the wettest month. Finally, the third PC reflected mostly variation in temperature, especially during the warmest months (electronic supplementary material, table S). The distribution of squamate species on the ppca ordination plot showed that most species can be found in arid climatic conditions, with relatively low precipitation and high temperatures (figure 1). Iguania, the second largest squamate clade has the broadest occupation of

4 Downloaded from on July, 18 PC1 PC rspb.royalsocietypublishing.org Proc. R. Soc. B 8: 178 PC M SIG C VAR S VAR S ASR S HGT D CDF Figure. Distribution of simulated test statistics on the first three PC scores based on posterior predictive simulations []. See texts for definitions of each tested statistic. The dashed lines correspond to the empirical data, whereas the histograms were based on simulations (N ¼ 1 ). Significant departures from expectation ( p,.1) were detected in all statistics, except for M SIG ( p ¼.97.98). climatic niche space, whereas Anguimorpha, with only species, has its climatic niche nested within the climatic niche distributions of other clades (figure 1). Interestingly, few squamate species seem to occupy extreme regions of the climatic niche space, such as cold, humid regions (figure 1). Posterior predictive simulations based on the ARBUTUS method indicated that the BM model showed a poor fit to the dataset for all three PCs (figure ). Except for M SIG, all the remaining test statistics presented significant departures from the distribution of simulated statistics, with the observed values calculated from the original dataset being much higher or much lower than the expectations under the model. For instance, for all three PCs, the C VAR statistic suggests that the model is not accounting for variation in rates of climatic niche evolution across the phylogeny, whereas the S VAR indicates that more climatic niche evolution occurred on short branches (figure ). Interestingly, the S ASR statistic showed that both PCs mainly related to variation in precipitation conditions (PC1 and PC) showed lower observed values than the expectations under BM, whereas the temperature axis (PC3) showed higher observed values (figure ), suggesting that the variation in rates of climatic niche evolution may behave differently with respect to precipitation and temperature. In general, the ARBUTUS analyses indicate substantial heterotachy across the squamate phylogeny, suggesting that a simple and constant-rate evolutionary models, such BM, provide poor statistical fits to the evolution of squamate climatic niches. A consistent pattern was uncovered when variation in rates through time was assessed based on DTT plots, indicating higher levels of disparity than expected by chance (figure 3). In general, squamate subclades presented more climatic niche evolution, especially near the present. Nevertheless, these results indicate the presence of heterogeneity through time in rates of climatic niche evolution, with more disparity near the present (figure 3), as one would expect in the context of a radiation into new climatic conditions. Likewise, the ancestral state reconstructions were consistent with this interpretation, showing a remarkable temporal concordance, regardless of the niche axis and the clade, with an accentuated climatic niche diversification, particularly during the last Myr (figure ). Interestingly, this diversification was continuous through this time, and strongly asymmetric for both PC1 and PC, where there seems to have a directional change for low scores (figure ). The MOTMOT analyses detected frequent variation in rates of climatic niche evolution, both over time and among lineages (figure ). For PC1 and PC axes, MOTMOT identified the most pronounced shifts, with the majority being rate increases (1 of ). All the identified decelerations in rates of climatic niche evolution were observed in clades with more than three species, whereas the accelerations were more frequent in

5 Downloaded from on July, PC1 PC PC3 entire dataset Anguimorpha. 1. Gekkota 1.. rspb.royalsocietypublishing.org Proc. R. Soc. B 8: Iguania disparity Lacertoidea 1. Scincoidea Serpentes time before present (Myr) Figure 3. Relative disparity-through-time (DTT) of PC scores representing squamate climatic niches. Solid lines indicate the observed DTTs, whereas the dashed lines and the corresponding polygons represent the averages and 9% confidence intervals of the expectations given a constant accumulation of disparity over time based on 999 pseudoreplicates. recent individual branches, such as Thamnophis radix (Serpentes), Liolaemus multimaculatus (Iguania) and Phymaturus mallimaccii (Iguania; figure ). The age of rate shifts varied considerably, with lineages presenting changes that happened early in the origin of the clades, and fairly recently for several species (less than Myr). For instance, in Gekkota clade there were three rate decreases, one of which dating from Myr, which corresponds to the approximate age of the origin of the clade. The

6 Downloaded from on July, 18 PC1 PC entire dataset 1 rspb.royalsocietypublishing.org PC3 Anguimorpha Proc. R. Soc. B 8: Gekkota 8 PC scores Iguania 1 8 Lacertoidea 8 Scincoidea 1 Serpentes time before present (Myr) 1 1 Figure. Evolution along the first three climatic niche axes of each squamate clades based on maximum likelihood ancestral state reconstruction. X-axis corresponds to the approximate age of origin of squamate clades to the present, whereas nodes indicate the inferred climatic niches for the most recent common ancestor of the extant taxa defined by that node, in each of the three PC axes. Divergent climatic niche evolution is apparent in branches that deviate from most branches of the phylogeny. Note that divergent branches are more frequent near the present time. greatest number of rate shifts were identified in Serpentes, which is also the most species-rich clade in this study, whereas the smallest clade (Anguimorpha) showed one rate shift comprising the entire clade (PC1, figure ). It is also important to note that most of the single-species rate shifts involving accelerations ( of 9, figure ) involved species from Southeast Asia, suggesting an unusual condition for the evolution of climatic niches in that region. Conversely, all three rate shifts involving decelerations were associated with island species (i.e. Socotra, New Caledonia and Vanuatu).

7 Downloaded from on July, 18 PC1 1 PC 7 3 Serpentes Iguania rspb.royalsocietypublishing.org Proc. R. Soc. B 8: Anguimorpha 8 Lacertoidea Scincoidea 1* 9 17 Gekkota 18* 1* 19* * Figure. Shifts in rates of climatic niche evolution along the first two PCs based on the MOTMOT method. Branch lengths are scaled according to the relative rates of evolution. Deviations from the background rate are highlighted with numbers (1 1 in PC1; 11 in PC), showing that the accelerations in rates of climatic niche evolution (numbers in bold face) were more frequent than decelerations (numbers marked with asterisk). Silhouettes of species representing the six clades at the tips were drawn by the authors. (Online version in colour.). Discussion The integration of a variety of phylogenetic comparative methods and the use of large-scale datasets allowed us to provide a general scenario for the evolution of climatic niches in Squamata. For instance, the most important dimension of the squamate climatic niche is dominated by precipitation, particularly during the dry season, leaving temperature variables to an important but secondary role. In addition, most squamate lineages are found in warm/dry conditions,

8 Downloaded from on July, 18 which in general seem to be their ancestral state. Interestingly, the evolution away from those conditions was far from homogeneous, with departures from a single-rate model being apparent both over time and among lineages. The occupation of regions with higher precipitation was more pronounced over the past Myr as a result of the increasing availability of habitats with those climatic conditions worldwide [,]. Interestingly, that occupation was associated with several independent rate shifts in different lineages, with the vast majority of those shifts involving accelerations. The climatic niche disparity on the entire dataset showed a considerable increase since 8 Myr (Late Cretaceous) for all three climatic niche axes. Curiously, these changes may be uncoupled from squamate taxonomic diversification, given that extant genera diversified fairly recently in squamate evolutionary history (Miocene Pliocene) []. Therefore, in general, our results corroborate the tested scenario in which clades tend to remain near their ancestral climatic niche, with some lineages accelerating the evolution of their climatic niches to occupy novel conditions as they become available during the history of the clade. These conclusions are supported by data on the thermal physiology of lizards, which shows that several thermal tolerance metrics are conserved in their evolutionary history [3]. Given that these metrics provide a good approximation of the fundamental climatic niches of squamates, our conclusions might be extended to fundamental climatic niches. However, this issue deserves further investigation. Before addressing the potential implications of these results, it is important to have in mind a number of caveats. First, our analyses included 188 species, which accounts for approximately 19% of the total number of described squamate species. Therefore, it is likely that several additional rate shifts could be identified in the future, as both improved phylogenetic information and species distribution data become available. However, it is important to note that, by the overdispersed nature of taxonomic sampling for phylogenetic studies, additional terminals will tend involve new species and genera, with increasingly fewer tips involving deeper branches. As a consequence, new data are unlikely to change our conclusion of accelerated rates of evolution near the present. Second, some single-branch rate shifts (figure ) might have been caused by poor placements on the phylogeny, or errors in their associated geographical distribution data, and therefore should be interpreted with caution. Finally, because of differences in behaviour, phenology and morphology, broad scale climate conditions do not necessarily describe the thermal and hydric conditions that species experience (i.e. just because species occupy the same or similar climates does not mean they experience the same conditions), such that the patterns detected in the present study pertain mostly to broad-scale environmental characteristics present throughout the distribution of different squamate lineages. By contrasting the results of the present study with those of previous studies (e.g. [9,,,3, 7]), one might begin to infer some tentative general principles regarding niche evolution. First, climatic niche evolution seems to be lowdimensional, with a few axes accounting for the vast majority of the observed variation. Although one might be tempted to explain the relative simplicity in the way climatic niches are structured in terms of climatic tolerances or some other physiological constraint, we argue that this pattern might ultimately have to do with the way in which climatic conditions are distributed across the globe. In particular, variation in climatic conditions is not isotropic, such that many combinations of climatic conditions are exceedingly rare or even absent altogether (e.g. colder climates with little annual variation in temperature). As a consequence, climatic conditions and their respective covariances form a template over which different lineages adapt and diversify [7]. In other words, the low dimensionality of climatic niches might simply result from the low dimensionality of the climatic conditions themselves. Second, different climatic niche axes might evolve according to distinct evolutionary dynamics [,,,7]. Several studies have already documented varying degrees of phylogenetic signal in different climatic niche axes [,8,9], yet few explicit tests have been carried out to assess variation in rates of climatic niche evolution. The results of the present study, as well as a previous analysis of primate climatic niches [] indicate that variation in rates is common, particularly in response to the availability of new, non-analogous climatic conditions over evolutionary time. Third, the ancestral climatic niche condition plays a large role in later evolutionary trends in a given clade. For instance, the protracted acceleration of squamate niche evolution over the course of the past Myr that led to the occupation by some lineages of more warm/humid environments is qualitatively different from the more rapid evolution of Primates when they diversified from their warm/humid ancestral conditions into colder habitats over only 1 Myr []. The ancestral climatic niches have been recognized by several authors as an important driver of lineage diversification, as posited in the tropical conservatism hypothesis [1]. More recently, it has also been recognized that this phenomenon is not exclusive to ancestral tropical conditions. For instance, Lampropeltini have a temperate origin, leading to an inverse diversity gradient in the region as lineages were limited in their occupation of more tropical habitats [8]. However, our results challenge some of the expectations of classical notions of phylogenetic niche conservatism. In particular, although ancestral climatic niches provide the initial conditions for further niche diversification, the evolution into other regions of climatic space will depend not only on biological constraints in terms of novel adaptations for thermoregulation and desiccation, but also on the availability of areas with different climates that are biogeographically accessible to the clade in question. In addition, once a novel climate is available, there might be some lag time until a lineage is able to colonize it []. The extent to which these three principles are indeed general requires further investigation, including both endo- and ectothermic clades. Data accessibility. All data are accessible through public sources described in Material and methods. Authors contributions. M.R.P. designed the study, participated in data analysis and drafted the manuscript; L.L.F.C. participated in data collection and analyses; A.L.S.M. participated in data collection and analysis, and helped draft the manuscript; A.D. participated in data collection, carried out the statistical analyses and helped draft the manuscript. Competing interests. We have no competing interests. Funding. This study was funded through graduate scholarships to A.L.S.M. and A.D. by CAPES (Ministry of Education Brazilian Government) and a research fellowship to M.R.P. by CNPq (Ministry of Science and Technology Brazilian Government, 3897/1-). 8 rspb.royalsocietypublishing.org Proc. R. Soc. B 8: 178

9 Downloaded from on July, 18 References 9 1. Currie DJ Energy and large-scale patterns of animal- and plant-species richness. Am. Nat. 137, 7 9. (doi:1.18/81). Mittelbach GG, Steiner CF, Scheiner SM, Gross KL, Reynolds HL, Waide RB, Willig MR, Dodson SI, Gough L. 1 What is the observed relationship between species richness and productivity? Ecology 8, 381. (doi:1.37/799) 3. McCain CM. 7 Could temperature and water availability drive elevational species richness patterns? A global case study for bats. Glob. Ecol. Biogeogr. 1, (doi:1.1111/j x). Robert ER, Schluter D Species diversity in ecological communities: historical and geographical perspectives, nd edn. Chicago, IL: University of Chicago Press.. Jiménez I, Ricklefs RE. 1 Diversity anomalies and spatial climate heterogeneity. Glob. Ecol. Biogeogr. 3, (doi:1.1111/geb.1181). Wiens JJ, Parra-Olea G, García-París M, Wake DB. 7 Phylogenetic history underlies elevational biodiversity patterns in tropical salamanders. Proc. R. Soc. B 7, (doi:1.198/rspb..31) 7. Sexton JP, McIntyre PJ, Angert AL, Rice KJ. 9 Evolution and ecology of species range limits. Annu. Rev. Ecol. Evol. Syst., 1 3. (doi:1.11/ annurev.ecolsys ) 8. Pyron RA, Burbrink FT. 9 Can the tropical conservatism hypothesis explain temperate species richness patterns? An inverse latitudinal biodiversity gradient in the new world snake tribe Lampropeltini. Glob. Ecol. Biogeogr. 18, 1. (doi:1.1111/j x) 9. Smith SA, Beaulieu JM. 9 Life history influences rates of climatic niche evolution in flowering plants. Proc. R. Soc. B 7, 3 3. (doi:1.198/rspb ) 1. Morales-Castilla I, Olalla-Tárraga MÁ, Bini LM, De Marco Jr P, Hawkins BA, Rodrígues MÁ. 1 Niche conservatism and species richness patterns of squamate reptiles in eastern and southern Africa. Austral Ecol. 3, 8. (doi:1.1111/j x) 11. Ricklefs RE Community diversity: relative roles of local and regional processes. Science 3, (doi:1.11/science ) 1. Wiens JJ, Donoghue MJ. Historical biogeography, ecology and species richness. Trends Ecol. Evol. 19, 39. (doi:1.11/j.tree ) 13. Mittelbach GG et al. 7 Evolution and the latitudinal diversity gradient: speciation, extinction and biogeography. Ecol. Lett. 1, (doi: /j x) 1. Hutchinson GE. 197 Concluding remarks. Cold Spring Harb. Symp. Quant. Biol., 1 7. (doi:1.111/sqb ) 1. Soberón J. 7 Grinnellian and Eltonian niches and geographic distributions of species. Ecol. Lett. 1, (doi:1.1111/j x) 1. Hua X, Wiens JJ. 13 How does climate influence speciation? Am. Nat. 18, 1 1. (doi:1. 18/79) 17. Pyron RA, Costa GC, Patten MA, Burbrink FT. 1 Phylogenetic niche conservatism and the evolutionary basis of ecological speciation. Biol. Rev. 9, (doi:1.1111/brv.11) 18. Cooney CR, Seddon N, Tobias JA. 1 Widespread correlations between climatic niche evolution and species diversification in birds. J. Anim. Ecol. 8, (doi:1.1111/13-.13) 19. Knouft JH, Losos JB, Glor RE, Kolbe JJ. Phylogenetic analysis of the evolution of the niche in lizards of the Anolis sagrei group. Ecology 87, S9 S38. (doi:1.189/1-98()87 [9:paoteo]..co;). Evans MEK, Smith SA, Flynn RS, Donoghue MJ. 9 Climate, niche evolution, and diversification of the bird-cage evening primroses (Oenothera, sections Anogra and Kleinia). Am. Nat. 173,. (doi:1.18/977) 1. Kamilar JM, Muldoon KM. 1 The climatic niche diversity of Malagasy primates: a phylogenetic perspective. PLoS ONE, e1173. (doi:1.1371/ journal.pone.1173). Duran A, Pie MR. 1 Tempo and mode of climatic niche evolution in Primates. Evolution 9, 9. (doi:1.1111/evo.173) 3. Duran A, Meyer ALS, Pie MR. 13 Climatic niche evolution in new world monkeys (Platyrrhini). PLoS ONE 8, e838. (doi:1.1371/journal.pone.838). Zachos J, Pagani M, Sloan L, Thomas E, Billups K. 1 Trends, rhythms, and aberrations in global climate Ma to present. Science 9, (doi:1.11/science.191). Vitt LJ, Caldwell JP. 1 Herpetology: an introductory biology of amphibians and reptiles, th edn. Amsterdam, The Netherlands: Elsevier Science.. Gillooly JF, Brown JH, West GB, Savage VM, Charnov EL. 1 Effects of size and temperature on metabolic rate. Science 93, 8 1. (doi:1. 11/science.1197) 7. Shine R. Life-history evolution in reptiles. Annu. Rev. Ecol. Evol. Syst. 3, 3. (doi:1. 11/annurev.ecolsys ) 8. Aragón P, Lobo JM, Olalla-Tárraga MÁ, Rodríguez MÁ. 1 The contribution of contemporary climate to ectothermic and endothermic vertebrate distributions in a glacial refuge. Glob. Ecol. Biogeogr. 19, 9. (doi:1.1111/j x) 9. IUCN. 1 The IUCN red list of threatened species, v. 1.. Spatial data download. 3. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol., (doi:1.1/joc.17) 31. Hijmans RJ. 1 raster: Geographic data analysis and modeling. R package version.-31. See CRAN.R-project.org/package=raster. 3. Bivand R, Keitt T, Rowlingson B. 1 rgdal: Bindings for the Geospatial Data Abstraction Library. R package version.9-1. See org/package=rgdal. 33. Bivand R, Lewin-Koh N. 1 maptools: Tools for reading and handling spatial objects. R package version.8-3. See package=maptools. 3. Wickham H. 1 plyr: Tools for splitting, applying and combining data. R package version See 3. Peterson AT, Soberón J, Pearson RG, Anderson RP, Martínez-Meyer E, Nakamura M, Araújo MB. 11 Ecological niches and geographic distributions. Princeton, NJ: Princeton University Press. 3. Kearney M, Porter W. 9 Mechanistic niche modelling: combining physiological and spatial data to predict species ranges. Ecol. Lett. 1, (doi:1.1111/j x) 37. Zheng Y, Wiens JJ. 1 Combining phylogenomic and supermatrix approaches, and a time-calibrated phylogeny for squamate reptiles (lizards and snakes) based on genes and 1 species. Mol. Phylogenet. Evol. 9, (doi:1.11/j. ympev.1.1.9) 38. Uetz P, Hošek J. 1 The reptile database. See www. reptile-database.org (accessed 9 January 1) 39. Revell LJ. 9 Size-correction and principal components for interspecific comparative studies. Evolution 3, (doi:1.1111/j x). Revell LJ. 1 Phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 3, (doi: /j.1-1x x) 1. Uyeda JC, Caetano DS, Pennell MW. 1 Comparative analysis of principal components can be misleading. Syst. Biol., (doi:1. 193/sysbio/syv19). Jackson DA Stopping rules in principal components analysis: a comparison of heuristical and statistical approaches. Ecology 7, 1. (doi:1.37/19397) 3. Ho LST, Ané C. 1 Intrinsic inference difficulties for trait evolution with Ornstein Uhlenbeck models. Methods Ecol. Evol., (doi: /1-1x.18). Thomas GH, Cooper N, Venditti C, Meade A, Freckleton RP. 1 Bias and measurement error in comparative analyses: a case study with the Ornstein Uhlenbeck model. biorxiv. (doi:1.111/3). Pennell MW, FitzJohn RG, Cornwell WK, Harmon LJ. 1 Model adequacy and the macroevolution of angiosperm functional traits. Am. Nat. 18, E33 E. (doi:1.18/8). Silvestro D, Kostikova A, Litsios G, Pearman PB, Salamin N. 1 Measurement errors should always rspb.royalsocietypublishing.org Proc. R. Soc. B 8: 178

10 Downloaded from on July, 18 be incorporated in phylogenetic comparative analysis. Methods Ecol. Evol., 3 3. (doi: /1-1x.1337) 7. Felsenstein, J Maximum-likelihood estimation of evolutionary trees from continuous characters. Am. J. Hum. Genet., Harmon LJ, Schulte JA, Larson A, Losos JB. 3 Tempo and mode of evolutionary radiation in iguanian lizards. Science 31, (doi:1. 11/science.1878) 9. Pennell MW, Eastman JM, Slater GJ, Brown JW, Uyeda JC, FitzJohn RG, Alfaro ME, Harmon LJ. 1 Geiger v.: an expanded suite of methods for fitting macroevolutionary models to phylogenetic trees. Bioinformatics 3, (doi:1.193/ bioinformatics/btu181). Thomas GH, Freckleton RP. 1 MOTMOT: models of trait macroevolution on trees. Methods Ecol. Evol. 3, (doi:1.1111/j.1-1x x) 1. R Development Core Team. 1 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.. Singh G History of aridland vegetation and climate: a global perspective. Biol. Rev. 3, (doi:1.1111/j.19-18x.1988.tb9.x) 3. Grigg JW, Buckley LB. 13 Conservatism of lizard thermal tolerances and body temperatures across evolutionary history and geography. Biol. Lett. 9, 11. (doi:1.198/rsbl.1.1). Cooper N, Freckleton RP, Jetz W. 11 Phylogenetic conservatism of environmental niches in mammals. Proc. R. Soc. B 78, (doi:1.198/rspb. 1.7). Schnitzler J, Graham CH, Dormann CF, Schiffers K, Peter Linder H. 1 Climatic niche evolution and species diversification in the Cape flora, South Africa. J. Biogeogr. 39, (doi:1.1111/jbi.18). Araújo MB, Ferri-Yáñez F, Bozinovic F, Marquet PA, Valladares F, Chown SL. 13 Heat freezes niche evolution. Ecol. Lett. 1, (doi:1.1111/ ele.11) 7. Pie MR. 1 The macroevolution of climatic niches and its role in ant diversification. Ecol. Entomol. 1, (doi:1.1111/een.13) 8. Luxbacher AM, Knouft JH. 9 Assessing concurrent patterns of environmental niche and morphological evolution among species of horned lizards (Phrynosoma). J. Evol. Biol., (doi:1.1111/j x) 9. Olalla-Tárraga MÁ et al. 11 Climatic niche conservatism and the evolutionary dynamics in species range boundaries: global congruence across mammals and amphibians. J. Biogeogr. 38, (doi:1.1111/j x). Guerrero PC, Rosas M, Arroyo MTK, Wiens JJ. 13 Evolutionary lag times and recent origin of the biota of an ancient desert (Atacama Sechura). Proc. Natl Acad. Sci. USA 11, (doi:1.173/pnas ) 1 rspb.royalsocietypublishing.org Proc. R. Soc. B 8: 178