The island syndrome in lizards

bs_bs_banner Global Ecology and Biogeography, (Global Ecol. Biogeogr.) (2013) 22, 184 191 RESEARCH PAPER The island syndrome in lizards Maria Novosolov 1 *, Pasquale Raia 2 and Shai Meiri 1 1 Department of Zoology, Tel Aviv University, Tel Aviv, Israel, 2 Dipartimento di Scienzedella Terra, Università Federico II, Naples, Italy *Correspondence: Shai Meiri, Department of Zoology, Tel Aviv University, 69978, Tel Aviv, Israel. E-mail: uncshai@post.tau.ac.il ABSTRACT Aim Islands are thought to promote correlated ecological and life-history shifts in species, including increased population density, and an infrequent production of few, large, offspring. These patterns are collectively termed the island syndrome. We present here the first, phylogenetically informed, global test of the island syndrome hypothesis, using lizards as our model organisms. Location World-wide. Methods We assembled a database containing 641 lizard species, their phylogenetic relationships, geographic ranges and the following life-history traits: female mass, clutch size, brood frequency, hatchling body mass and population density. We tested for life-history differences between insular and mainland forms in light of the island syndrome, controlling for mass and latitude, and for phylogenetic nonindependence. We also examined the effects of population density and, in insular endemics, of island area, on lizard reproductive traits. Results We found that insular endemic lizards lay smaller clutches of larger hatchlings than closely related mainland lizards of similar size, as was expected by the island syndrome. In general, however, insular endemics lay more frequently than mainland ones. Species endemic to small islands lay as frequently as mainland species. Continental and insular lizards have similar productivity rates overall. Island area had little effect on lizard reproductive traits. No trait showed association with population density. Main conclusions Island endemic lizards mainly follow the island syndrome. We hypothesize that large offspring are favoured on islands because of increased intraspecific aggression and cannibalism by adults. Stable populations on islands lacking predators may likewise lead to increased intra-specific competition, and hence select for larger hatchlings that will quickly grow to adult size. This view is supported by the fact that lizard populations are denser on islands although population density per se was uncorrelated with any of the traits we examined. Keywords Clutch size, island biogeography, island syndrome, life history, lizards, population density, reproduction, reversed island syndrome. INTRODUCTION Animal populations confined on islands frequently show several substantial morphological and behavioural differences from related mainland forms (Adler & Levins, 1994; Blondel, 2000; Raia et al., 2010). Small insular vertebrates often grow larger than their mainland counterparts do, whereas larger species sometimes dwarf on islands (Van Valen, 1973; Lomolino, 2005). This pattern, known as the island rule, was considered universal. Recent data and analyses, however, have challenged this view (Meiri et al., 2008, 2012). Insular species likewise often evolve substantial differences from the mainland phenotype in their locomotor and feeding apparatuses (Grant, 1965; Sondaar, 1977), behaviour and coloration, among others. For example, on DOI: 10.1111/j.1466-8238.2012.00791.x 184 2012 Blackwell Publishing Ltd http://wileyonlinelibrary.com/journal/geb

Life history evolution in island lizards islands many birds (especially rails) became flightless (Roff, 1994), dull-coloured (Omland, 1997) large-billed (Grant, 1999; Clegg & Owens, 2002) and relatively tame (Blondel, 2000). Insular rodents are often larger, less aggressive, and less fecund than their mainland relatives are, and live in denser (and more stable) populations, a phenomenon termed the island syndrome (Adler & Levins, 1994). The island syndrome thus predicts changes in population dynamics, body size (Sondaar, 1977; Raia & Meiri, 2006), anti-predatory behaviour (Schoener et al., 2005) and life history (Raia et al., 2003, 2010). Aspects of this syndrome have been erratically assigned to such different types of organisms as vertebrates, arthropods, molluscs and plants (Whittaker & Fernandez-Palacios, 2007). MacArthur & Wilson (1967, pp. 149 153) predicted that, in general, established insular populations would shift towards K selection as the population density of insular populations increases. Therefore, clutch sizes would be lower on islands, especially in temperate regions. Many models and data describe single aspects of life-history evolution on islands. Consistent, taxon-wide evidence for the full range of phenomena predicted under the island syndrome, however, is currently limited to rodents (Adler & Levins, 1994) and passerine birds (Blondel, 2000), and was never formally tested even in these groups. In reptiles, trait shifts consistent with the island syndrome are common. Many insular lizards are melanistic (Fulgione et al., 2008; Runemark et al., 2010), have modified limb lengths and head shapes (Herrel et al., 2008; Raia et al., 2010), altered body sizes (Case, 1978; Meiri, 2007; Pafilis et al., 2011), small clutch sizes (Huang, 2007, cf. Case, 1975), and low growth rates (Andrews, 1976). Shifts to herbivorous diet (Van Damme, 1999; Meiri, 2008) and either reduced or increased aggressiveness (Stamps & Buechner, 1985; Pafilis et al., 2009; Raia et al., 2010) are also common on islands. Insular faunas often have few predators and competitors. This can lead to increased population densities ( density compensation and density overcompensation ). Such increased population density prompts the emergence of the island syndrome according to Adler and Levins model (Adler & Levins, 1994; see also Raia et al., 2010; Pafilis et al., 2011). Increased population densities are thought to select for few, large-sized, offspring (MacArthur & Wilson, 1967; Andrews, 1979; Adler & Levins, 1994; Pafilis et al., 2011), which grow into large adults (Stamps & Buechner, 1985; Sinervo et al., 2000). Raia et al. (2010) predicted that insular populations facing extreme environmental unpredictability would display an opposite array of trait shifts (a reversed island syndrome ). They predicted that such highly uncertain conditions, which may be common on very small islands, will keep population density low and drive the production of frequent, large clutches of small hatchlings. Theory predicts that evolutionary divergence should intensify as islands become more insular smaller and more isolated (Whittaker & Fernandez-Palacios, 2007). The larger and closer to the mainland an island is the more it is assumed to resemble the mainland in important ecological attributes (e.g. in species richness), and hence in the traits of organisms inhabiting it (Heaney, 1978; Melton, 1982; Lomolino, 2005; cf. Meiri et al., 2005). We hypothesize that insular lizards differ from continental species in their population densities and in their clutch size, hatchling mass, brood frequency and productivity (progeny biomass produced per unit time). Specifically, we predict the following: 1. Because of lower predation and competition on islands, insular populations will be denser than mainland populations (Case, 1975; Bennett & Gorman, 1979; Rodda & Dean-Bradley, 2002; Buckley & Jetz, 2007). 2. Insular lizards shift towards K strategy in response to low predation, high population density and increased intra-specific competition. Thus, we predict that insular lizards lay smaller clutches of larger hatchlings, and lay infrequently and thus their overallproductivityrateislow. 3. Island mainland differences in life-history traits will be strongest on small islands, while on large islands trait values will be similar to those observed on the mainland. 4. Alternatively, a shift to K strategy on medium-sized islands, predicted by the island syndrome, may be followed by a shift towards r strategy on very small islands, as predicted under the reversed island syndrome. METHODS We gathered data on mean female and hatchling body length, clutch size, brood frequency and productivity data for 641 species of lizards. We collected data from the primary literature, field guides, our own observations and museum records. A complete list of the 1794 sources for the different trait values is provided in Appendix S1 in Supporting Information. Of the 641 species, 100 are insular endemics. Population density estimates were available for 220 species (Appendix S1). We determined whether lizard species were insular endemics or mainland inhabitants using the reptile database (http://reptiledatabase.reptarium.cz), field guides and the primary literature. Island endemic species are those inhabiting only islands, whereas we treated all species that are found on the mainland as mainland species regardless of whether they also occur on islands. We made sure, however, that the trait data we use for these species originated from mainland populations. We recorded the area (in km 2 ) of the largest island inhabited by insular endemics for testing the correlation between island area and reproductive traits. Island areas were obtained from the UN island directory (http://islands.unep.ch/iindex.htm). Data that were unavailable in the directory were obtained from the National Imagery and Mapping Agency (NIMA, 1997). We estimated female and hatchling masses from snout vent lengths (SVLs) using family-specific equations for legged species, and different equations for legless and leg-reduced lizards from Meiri (2010) except for the following: for Liolaemus and Phymaturus we used equations from Pincheira-Donoso et al. (2011); for different gecko clades and for Anolis we used specific equations using data gathered by S.M. (unpublished). Appendix S2 contains the allometric equations for converting SVLs to mass in these clades. For 22 species, we had no data for female SVL and we therefore used species-specific SVLs, or actual female mass (for Global Ecology and Biogeography, 22, 184 191, 2012 Blackwell Publishing Ltd 185

M. Novosolov et al. Phrynosoma blainvillii). For seven species, we had no data on hatchling SVL, and we used hatchling masses instead. We used mean clutch/litter sizes and frequencies where possible. Where more than one mean was reported for a species we used the midpoint of the range of means. Where means were unavailable, we used the midpoint of the reported trait range. Meiri et al. (2012) have recently shown that productivity is best quantified as a rate biomass produced per unit time. We thus define productivity as the product of brood frequency, clutch size and hatchling mass, in units of g per year. Mean population densities (mean number of adult lizards per hectare) were recorded from the literature (Appendix S1). Where means were unavailable, we used the mean of log-transformed maximum and minimum population density values. To correct for possible phylogenetic effects in the data, we assembled a composite species-level phylogeny from the literature, following the broad-scale squamate phylogenetic relationship reported by Wiens et al. (2010). We used the taxonomy of the reptile database (http://reptile-database.reptarium.cz). In assembling the tree we gave priority to recently published phylogenies that are based on nuclear DNA, then on mitochondrial DNA sequences. For species where no molecular phylogeny was available, we relied on phylogenies based on morphological data. Where phylogenetic data were unresolved, we sunk species into a polytomy within their genus. The phylogenetic relationships between the species and the sources of phylogenetic data for each are depicted in Appendix S3. We did not account for branch lengths. Instead, we scaled branches to make the tree ultrametric using the cladogram transform in FigTree (Rambaut, 2010). Arbitrary branch lengths computed this way give a necessarily imprecise measure of the expected phenotypic covariation between species, yet they are still much better than neglecting phylogenetic effect altogether. To account for possible effects of climate we mapped the geographic ranges of the different species in ArcGIS using published data on lizard distribution (Appendix S1). A full analysis of the climatic drivers of lizard life history is beyond the scope of the current study. We thus used only a simplistic measure of climate: the absolute value of latitudinal centroid for each species. This variable was used to correct for, for example, the tendency of tropical species to produce more clutches than temperate species (Meiri et al., 2012). Linear models and phylogenetic generalized least square All the data except latitude were log 10-transformed in all analyses. For each species we used the clutch size, number of yearly broods, hatchling mass and productivity (= clutch size number of yearly broods hatchling mass) as response variables and regressed them against female mass (g) and latitude, using insularity as a main effect in ANCOVA. After analysing the entire dataset we repeated all analyses with only insular endemics inhabiting islands smaller than 1000 km 2. To examine the relationship between island area and reproductive traits we ran similar analyses with all insular endemics, and used island area as an additional covariate. We test two predictions with regard to island area: (1) a shift towards K strategy as islands grow smaller as predicted by the island syndrome; (2) a shift towards K strategy on medium-sized islands; but a shift towards r strategy on very small islands (the reversed island syndrome). We therefore also use the quadratic term of area in these models. Additionally, we used population density as a predictor for species for which we had these data. All four sets of analyses were duplicated to account for phylogenetic non-independence by using phylogenetic generalized least square (PGLS) regression. PGLS fits the regression of a given predictor variable on a given response variable via a GLS procedure, by using a phylogenetically informed hypothesis for the distribution of residuals around the response. The analysis is based on the shared evolutionary history of different species drawn from the phylogenetic tree. PGLS assumes that trait evolution proceeds according to a Brownian motion model (Freckleton et al., 2002). We adjusted the strength of phylogenetic non-independence using the maximum likelihood value of the scaling parameter l (Freckleton et al., 2002) implemented in the R package caper (Orme et al., in press). Pagel s l is a multiplier of the off-diagonal elements of the variance covariance matrix, which provides the best fit of the Brownian motion model to the tip data. To further examine whether our results stem from comparisons of lizards belonging to very different clades, we compared hatchling size, clutch size, brood frequency and productivity in a paired design across the 17 genera for which we have both island and mainland representatives. The results of these analyses (Appendix S4) were similar to those of the phylogenetic analysis and are therefore not explored further. In all the analyses, we selected models with a backwards elimination procedure based on P-values. RESULTS All the results presented below are for models that include female body mass as a covariate (mass was significantly correlated with the response variable in all models). The effects of mass and latitude on the various response variables are presented in Appendix S5. Island endemic lizards populations are, on average, nearly four times as dense as mainland ones (138.0 0.16 for islands and 36.3 0.08 lizards per hectare, for mainland, t = 3.11, P < 0.001). A difference remains after correcting for body mass (intercept: islands = 3.08 0.15; mainland = 2.22 0.11, t = 5.86, P < 0.001). Significant, albeit smaller, differences remain after phylogeny and mass are both accounted for (l =0.62, intercept: islands = 2.61 0.16, mainland = 2.11 0.28, t = 3.20, P = 0.001). Population density and island area Population density shows no association with any of the response variables. Clutch size, brood frequency and hatchling mass are likewise unassociated with island area. Productivity, 186 Global Ecology and Biogeography, 22, 184 191, 2012 Blackwell Publishing Ltd

Life history evolution in island lizards however, is negatively correlated with island area only in the non-phylogenetic model (slope =-0.08 0.02, P = 0.001). The results of all regressions of the various response variables on population density and island area are shown in Appendix S6. No quadratic terms of island area were significant in any analysis (Appendix S7). when examining all insular endemic species (non-phylogenetic island intercept: 0.04 0.03; mainland: 0.22 0.03, slope = 0.27 0.01, t =-6.72, P < 0.001; phylogenetic island intercept: 0.12 0.02; mainland: 0.24 0.08, slope = 0.23 0.02, t =-5.23, P < 0.001, l=0.84). These results are in agreement with the predictions of the island syndrome. Clutch size Clutch sizes of lizards endemic to small (< 1000 km 2 ) islands are smaller than those of mainland species (female mass and latitude-corrected) (Fig. 1a), in both the non-phylogenetic (island intercept: 0.08 0.06; mainland: 0.21 0.04, slope = 0.28 0.02, t =-2.21, P = 0.03) and phylogenetic models (island intercept: 0.15 0.04; mainland: 0.23 0.08, slope = 0.25 0.02, t =-2.29, P = 0.02, l=0.87). Similar results are obtained Brood frequency Brood frequency of small-island endemics is similar to that of mainland species (corrected for female mass and latitude) in both the non-phylogenetic (intercept islands: 0.64 0.06; mainland: 0.58 0.03, slope =-0.14 0.02, t = 1.07, P = 0.29) and phylogenetic models (island intercept: 0.41 0.05; mainland: 0.45 0.08, slope: -0.07 0.02, t =-0.99, P = 0.32, l=0.75) (Fig. 1b). However, across all insular endemic species Log Clutch Size 1.5 1.0 0.5 Clutch size on islands vs. mainland Brood frequency on islands vs. mainland a) b) Log Brood Frequency (broods/year) 1.5 1.0 0.5 0.0 0.0 0 1 2 3 4-0.5 0 1 2 3 4 Log Female Mass (g) Log Female Mass (g) Hatchling mass on islands vs. mainland Productivity on islands vs. mainland Log Hatchling Mass (g) 2.0 1.5 1.0 0.5 0.0-0.5-1.0-1.5 c) d) Log Productivity (g/year) -1 0 1 2 3 0 1 2 3 4 0 1 2 3 4 Log Female Mass (g) Log Female Mass (g) Figure 1 Relationship of (a) (log-transformed) clutch size and (b) (log-transformed) brood frequency (broods per year) with (log-transformed) female mass (in g) on islands (white, solid line) and the mainland (black, dashed line). Relationship of (c) (log-transformed) hatchling mass (g) and (d) (log-transformed) productivity (g per year) (log transformed) with female body mass (g) on islands (white) and the mainland (black). For (c) and (d) the lines show the best fit model regression. Only one regression line is shown where the difference between island and mainland was not significant. Global Ecology and Biogeography, 22, 184 191, 2012 Blackwell Publishing Ltd 187

M. Novosolov et al. brood frequency is higher on islands than on the mainland (corrected for latitude and female mass; island intercept: 0.69 0.03; mainland: 0.62 0.03, slope =-0.15 0.01, t = 2.7, P = 0.007). Phylogenetic models show no significant correlation between insularity and brood frequency (island intercept: 0.44 0.02; mainland: 0.45 0.08, slope =-0.07 0.02, t =-0.24, P = 0.8, l=0.84). Hatchling mass Small-island endemic species have larger hatchlings than those of mainland species of comparable size (non-phylogenetic model; island intercept: -0.87 0.05; mainland: -0.98 0.03, slope = 0.69 0.01, t = 2.04, P = 0.04; phylogenetic model; island intercept: -0.73 0.05; mainland: -0.86 0.07, slope = 0.65 0.02, t = 2.91, P = 0.004, l=0.7) (Fig. 1c). Across all insular endemics we obtain similar results in a phylogenetic model (island intercept: -0.78 0.03; mainland: -0.86 0.07, slope = 0.64 0.02, t = 3.05, P = 0.002, l=0.72), but not in the non-phylogenetic model (island intercept: -0.96 0.03; mainland: -0.99 0.03, slope = 0.71 0.01, t = 1.25, P = 0.21). There is an interaction between insularity and female mass in the non-phylogenetic model (insular slope 0.11 0.05 higher than mainland slope, P = 0.04). These results are, in general, consistent with the island syndrome. Productivity There are no significant differences in productivity rates between insular and continental species. Mass-corrected intercepts of small-island endemics and continental species are not significantly different (non-phylogenetic model, islands: -0.21 0.07; mainland: -0.25 0.02, slope = 0.85 0.02, t = 0.53, P = 0.6; phylogenetic model, islands: -0.21 0.07; mainland: -0.20 0.09, slope = 0.82 0.03, t =-0.06, P = 0.95, l=0.56, results corrected for latitude) (Fig. 1d). Similar results are obtained with a phylogenetic model across all insular endemic species (mass- and latitude-corrected; islands intercept: -0.23 0.04; mainland: -0.18 0.09, slope = 0.80 0.02, t =-1.44, P = 0.15, l=0.59). Only the non-phylogenetic model across all species show a significant difference: productivity rates are lower on islands (corrected for female mass and latitude; islands intercept: -0.21 0.04; mainland: -0.15 0.04, slope = 0.83 0.02, t =-2.14, P = 0.03). DISCUSSION As expected under the island syndrome (Adler & Levins, 1994; Blondel, 2000), insular lizards lay smaller clutches of larger hatchlings than do mainland lizards of comparable size. Brooding rates and productivity rates of mainland and insular lizards are similar. Thus in general we see a shift towards K selection in island taxa, in accordance with island biogeography theory (MacArthur & Wilson, 1967) and the island syndrome (Adler & Levins, 1994). The results of the phylogenetic and non-phylogenetic analyses are mostly congruent. However, the non-phylogenetic analysis of all islands revealed higher brood frequencies, similar-sized hatchlings and low productivity on islands. In both the nonphylogenetic analyses and in the phylogenetic analysis of small islands we found similar brood frequency and productivity on islands and on the mainland and larger hatchlings on islands. Higher brood frequency on islands was found only in the nonphylogenetic analysis that included large islands but not in the phylogenetic analyses, and in analyses of only small islands. We therefore suggest that lizards belonging to lineages that lay frequently (e.g. anoles and geckos), while having similar laying rates on islands and continents, are relatively more common on islands than on the mainland. We suspect that members of these lineages are good colonizers of islands (i.e. have good dispersal ability and/or low extinction probability on islands). A nonmutually exclusive alternative is that lizards belonging to such lineages frequently radiate on large islands. The first hypothesis is supported by data on the geographic distribution of insular lizards (M.N., unpublished), and the second is supported by findings of lizard radiations on large islands (Losos & Schluter, 2000). Overall, we detected no differences in productivity rates between islands and continental areas. To us this suggests that insularity is neither a panacea of unlimited opportunity in the absence of predators and competitors on the one hand, nor is it the epitome of harsh environments, subjected to chronic food shortages that is often advanced as the cause of insular dwarfing (e.g. Köhler & Moyà-Solà, 2010, cf. Meiri & Raia, 2010). It appears that insular lizards invest in fewer, larger offspring, as predicted by the island syndrome. At least on large islands lizards brood more frequently than on the mainland. Overall there is no association between productivity rates and insularity. The rodent model of Adler & Levins (1994) predicts smaller broods of large offspring and high population density on islands. This model applies to lizards as well. In fact, ours are the first quantitative comparative and phylogenetic analyses to present the island syndrome in any clade. Raia et al. (2010) experimentally studied a population of the Italian wall lizard, Podarcis sicula, from the tiny islet of Licosa, off the western coast of southern Italy. They pointed out that under the very unpredictable environmental conditions of Licosa, the uncertain mortality schedule favours a great investment in the reproduction of numerous, small offspring. Such perturbations also make population density highly variable, and usually low. Raia et al. (2010) also showed that the Licosa P. sicula is very aggressive toward conspecifics. We found a single result supporting the reversed island syndrome (Raia et al., 2010): a negative relationship between island area and productivity. We suspect that if such a pattern generalizes it is more likely to manifest itself at intra-specific levels, on smaller islands than most of the ones we examined here. Island biogeography theory usually predicts more pronounced evolutionary shifts on small islands, because their faunas are more different from those of the mainland and larger islands (e.g. Thomas et al., 2009). Perhaps the lack of relationship between island area and reproductive 188 Global Ecology and Biogeography, 22, 184 191, 2012 Blackwell Publishing Ltd

Life history evolution in island lizards traits implies that both the island syndrome and its reverse occur on (different) small islands and together they therefore cancel each other out. Surprisingly, although we found some relationship between insularity and reproductive traits (e.g. smaller clutches) neither population density nor island area (or its quadratic) were correlated with them. Because predator and competitor richness increase with island area, population density is predicted to decrease with island area (Case, 1975; Rodda & Dean-Bradley, 2002; Lomolino et al., 2010). However, across the 56 insular endemic species from which we have both population density and area data, density and area are uncorrelated (correcting for mass, slope =-0.15 0.08, t =-1.71, P = 0.09; uncorrected slope =-0.06 0.12, t =-0.57, P = 0.57). This is surprising given that the islands with endemic lizards in our sample range in area from 0.13 km 2 (Columbretes Islands, Podarcis liolepis) tothec. 786,000 km 2 island of New Guinea. On small islands, such as the Columbretes, lizards are sometimes the only terrestrial vertebrates, yet the larger islands are inhabited by a panoply of lizards, snakes, birds and mammals (as well as much greater diversity of arthropods) similar to the conditions on continental areas. Population density was uncorrelated with any of the traits we have examined. This is intriguing, as population density is often advocated as a major determinant of the intensity of intraspecific competition, and thus the evolution of life history towards larger offspring that are assumed to be better competitors. The large size of insular offspring, in turn, is hypothesized to result in longer embryonic development, and thus in smaller and less frequent clutches (Melton, 1982; Adler & Levins, 1994). Population density data are extremely noisy. For many of the species in our database for which we have more than one population density estimate the variation spans at least one, and sometime even two or three orders of magnitude. Within species, densities may often vary between habitats, between years, across climatic gradients etc. (e.g. Schoener & Schoener, 1980). To a large extent variation in reported values of population density is greatly influenced by the spatial extent of the area over which population density was quantified (Blackburn & Gaston, 1996). Because population density estimates are always for areas in which lizards are found (i.e. areas with zero population density are excluded) smaller areas usually only encompass the best habitat for a lizard, but larger areas do often contain habitats where no individuals dwell. In our data the area over which population density was measured was strongly and negatively correlated with population density in fact, this factor explains more of the variation in population density than body size does (M.N. & S.M., unpublished). Island populations are denser than mainland ones, and this pattern hold regardless of whether we correct for mass, for phylogeny, or for both. This is in line with established knowledge regarding population density of mainland and insular lizards (Andrews, 1979; Bennett & Gorman, 1979; Buckley & Jetz, 2007). It may be that the lack of relationship between population density and reproductive characteristics stems from the noise in our data (we were unable to fully correct for the effect of sampling area, as it is often not reported). Alternatively, increased population density need not always result in increased intra-specific competition. It is possible, for example, that population density is controlled by the amount of available food, and thus if dense areas also hold more resources the intensity of competition need not increase. This may partially depend on the frequent shift to herbivory which many insular lizards show (Van Damme, 1999; Meiri, 2008), and certainly was the case, for instance, with the hugely dense population of Bonaire whiptail Cnemidophorus murinus (Dearing, 1993). Although we think this hypothesis merits some quantitative treatment, it is beyond the scope of this work. Overall, lizards follow most of the predictions of the island syndrome: they lay fewer eggs, and occur at greater population densities on islands than on the mainland. They also produce larger offspring than closely related mainland forms. We found no relationship between population density and life history, and island area is likewise unrelated to either population density or to the life-history traits we examined. Productivity rates are similar across islands and the mainland suggesting a common constraint on the amount of energy available for reproduction. Predator-induced mortality is probably lower on many islands (Adler & Levins, 1994). This is likely to result in increased intraspecific competition, selecting for larger hatchlings. Intraspecific predation on juveniles may also impose a selection pressure for large hatchling size (Pafilis et al., 2009). Together with the often high abundance of food on some islands these factors may explain most of the variation in population density and life-history characteristics that is unaccounted for by body size and latitude. The increase in hatchling size will result in smaller broods, thus explaining the island syndrome. ACKNOWLEDGEMENTS We thank Lital Dabool for valuable discussion. Erez Maza has been instrumental in obtaining data on lizard distributions. We thank Panayiotis Pafilis and two anonymous referees for constructive comments on an earlier version of this work. Shai Meiri is supported by an Alon Fellowship. REFERENCES Adler, G.H. & Levins, R. (1994) The island syndrome in rodent populations. Quarterly Review of Biology, 69, 473 490. Andrews, R.M. (1976) Growth rate in island and mainland anoline lizards. Copeia, 1976, 477 482. Andrews, R.M. (1979) Evolution of life histories: a comparison of Anolis lizards from matched island and mainland habitats. Breviora, 454, 1 51. Bennett, A.F. & Gorman, G.C. (1979) Population density and energetics of lizards on a tropical island. Oecologia, 42, 339 358. Blackburn, T.M. & Gaston, K.J. (1996) Abundance body size relationships: the area you census tells you more. Oikos, 75, 303 309. Blondel, J. (2000) Evolution and ecology of birds on islands: trends and prospects. Vie et Milieu, 50, 205 220. Global Ecology and Biogeography, 22, 184 191, 2012 Blackwell Publishing Ltd 189

M. Novosolov et al. Buckley, L.B. & Jetz, W. (2007) Insularity and the determinants of lizard population density. Ecology Letters, 10, 481 489. Case, T.J. (1975) Species numbers, density compensation, and colonizing ability of lizards on islands in the Gulf of California. Ecology, 56, 3 18. Case, T.J. (1978) On the evolution and adaptive significance of postnatal growth rates in the terrestrial vertebrates. Quarterly Review of Biology, 53, 243 282. Clegg, S.M. & Owens, I.P.F. (2002) The island rule in birds: medium body size and its ecological explanation. Proceedings of the Royal Society B: Biological Sciences, 269, 1359 1365. Dearing, M.D. (1993) An alimentary specialization for herbivory in the tropical whiptail lizard, Cnemidophorus murinus. Journal of Herpetology, 27, 111 114. Freckleton, R.P., Harvey, P.H. & Pagel, M. (2002) Phylogenetic analysis and comparative data: a test and review of evidence. The American Naturalist, 160, 712 726. Fulgione, D., Guglielmi, S., Odierna, G., Rippa, D., Caliendo, M.F. & Rastogi, R.K. (2008) Morphological differentiation and genetic structure in island lizard populations. Zoological Science, 25, 465 474. Grant, P.R. (1965) The adaptive significance of some size trends in island birds. Evolution, 19, 355 367. Grant, P.R. (1999) Ecology and evolution of Darwin s finches. Princeton University Press, Princeton, NJ. Heaney, L.R. (1978) Island area and body size of insular mammals: evidence from the tri-colored squirrel (Callosciurus prevosti) of Southeast Asia. Evolution, 32, 29 44. Herrel, A., Huyghe, K., Vanhooydonck, B., Backeljau, T., Breugelmans, K., Grbac, I., Van Damme, R. & Irschick, D.J. (2008) Rapid large-scale evolutionary divergence in morphology and performance associated with exploitation of a different dietary resource. Proceedings of the National Academy of Sciences USA, 105, 4792 4795. Huang, W.-S. (2007) Ecology and reproductive patterns of the agamid lizard Japalura swinhonis on an East Asian island, with comments on the small clutch sizes of island lizards. Zoological Science, 24, 181 188. Köhler, M. & Moyà-Solà, S. (2010) Physiological and life history strategies of a fossil large mammal in a resource-limited environment. Proceedings of the National Academy of Sciences USA, 106, 20354 20358. Lomolino, M.V. (2005) Body size evolution in insular vertebrates: generality of the island rule. Journal of Biogeography, 32, 1683 1699. Lomolino, M.V., Riddle, B.R., Whittaker, R.J. & Brown, J.H. (2010) Biogeography, 4th edn. Sinauer Associates Inc., Sunderland, MA. Losos, J.B. & Schluter, D. (2000) Analysis of an evolutionary species area relationship. Nature, 408, 847 850. MacArthur, R.H. & Wilson, E.O. (1967) The theory of island biogeography. Princeton University Press, Princeton, NJ. Meiri, S. (2007) Size evolution in island lizards. Global Ecology and Biogeography, 16, 702 708. Meiri, S. (2008) Evolution and ecology of lizard body sizes. Global Ecology and Biogeography, 17, 724 734. Meiri, S. (2010) Length weight allometries in lizards. Journal of Zoology, 281, 218 226. Meiri, S. & Raia, P. (2010) Reptilian all the way? Proceedings of the National Academy of Sciences USA, 107, E27. Meiri, S., Dayan, T. & Simberloff, D. (2005) Area, isolation, and size evolution in insular carnivores. Ecology Letters, 8, 1211 1217. Meiri, S., Cooper, N. & Purvis, A. (2008) The island rule: made to be broken? Proceedings of the Royal Society B: Biological Sciences, 275, 141 148. Meiri, S., Brown, J.H. & Sibly, R.M. (2012) The ecology of lizard reproductive output. Global Ecology and Biogeography, 21, 592 602. Melton, R.H. (1982) Body size and island Peromyscus: a pattern and a hypothesis. Evolutionary Theory, 6, 113 126. NIMA (1997) National Imagery and Mapping Agency (NIMA), vector map level 0 (VMAP0), 4th edn. NIMA, Washington, DC. Omland, K.E. (1997) Examining two standard assumptions of ancestral reconstructions: repeated loss of dichromatism in dabbling ducks (Anatini). Evolution, 51, 1636 1646. Orme, C.D.L., Freckleton, R.P., Thomas, G.H., Petzoldt, T., Fritz, S.A. & Isaac, N.J.B. (in press) Caper: comparative analyses of phylogenetics and evolution in R. Methods in Ecology and Evolution. Pafilis, P., Meiri, S., Foufopoulos, J. & Valakos, E. (2009) Intraspecific competition and high food availability are associated with insular gigantism in a lizard. Die Naturwissenschaften, 96, 1107 1113. Pafilis, P., Foufopoulos, J., Sagonas, K., Runemark, A., Svensson, E. & Valakos, E.D. (2011) Reproductive biology of insular reptiles: marine subsidies modulate expression of the island syndrome. Copeia, 2011, 545 552. Pincheira-Donoso, D., Fox, S.F., Scolaro, J.A., Ibargüengoytía, N., Acosta, J.C., Corbalan, V., Medina, M., Boretto, J., Villavicencio, H.J. & Hodgson, D.J. (2011) Body size dimensions in lizard ecological and evolutionary research: exploring the predictive power of mass estimation equations in two Liolaemidae radiations. Herpetological Journal, 21, 35 42. Raia, P. & Meiri, S. (2006) The island rule in large mammals: paleontology meets ecology. Evolution, 60, 1731 1742. Raia, P., Barbera, C. & Conte, M. (2003) The fast life of a dwarfed giant. Evolutionary Ecology, 17, 293 312. Raia, P., Guarino, F.M., Turano, M., Polese, G., Rippa, D., Carotenuto, F., Monti, D.M., Cardi, M. & Fulgione, D. (2010) The blue lizard spandrel and the island syndrome. BMC Evolutionary Biology, 10, 289. doi: 10.1186/1471-2148- 10-289. Rambaut, A. (2010) Figtree, version 1.3.1. Institute of Evolutionary Biology, University of Edinburgh. Rodda, G.H. & Dean-Bradley, K. (2002) Excess density compensation of island herpetofaunal assemblages. Journal of Biogeography, 29, 623 632. 190 Global Ecology and Biogeography, 22, 184 191, 2012 Blackwell Publishing Ltd

Life history evolution in island lizards Roff, D.A. (1994) The evolution of flightlessness: is history important? Evolutionary Ecology, 8, 639 657. Runemark, A., Hansson, B., Pafilis, P., Valakos, E.D. & Svensson, E.I. (2010) Island biology and morphological divergence of the Skyros wall lizard Podarcis gaigeae: a combined role for local selection and genetic drift on color morph frequency divergence? BMC Evolutionary Biology, 10, 269. doi: 10.1186/ 1471-2148-10-269. Schoener, T.W. & Schoener, A. (1980) Densities, sex ratios, and population structure in four species of Bahamian Anolis lizards. Journal of Animal Ecology, 49, 19 53. Schoener, T.W., Losos, J.B. & Spiller, D.A. (2005) Island biogeography of populations: an introduced species transforms survival patterns. Science, 310, 1807 1809. Sinervo, B., Svensson, E. & Comendant, T. (2000) Density cycles and an offspring quantity and quality game driven by natural selection. Nature, 406, 985 988. Sondaar, P.Y. (1977) Insularity and its effects on mammal evolution. Major patterns of vertebrate evolution (ed. by M.K. Hecht, P.C. Goody and B.M. Hecht), pp. 671 707. Plenum Press, New York. Stamps, J.A. & Buechner, M. (1985) The territorial defense hypothesis and the ecology of insular vertebrates. Quarterly Review of Biology, 60, 155 181. Thomas, G.H., Meiri, S. & Phillimore, A.B. (2009) Body size diversification in Anolis: novel environment and island effects. Evolution, 63, 2017 2030. Van Damme, R. (1999) Evolution of herbivory in lacertid lizards: effects of insularity and body size. Journal of Herpetology, 33, 663 674. Van Valen, L.M. (1973) A new evolutionary law. Evolutionary Theory, 1, 1 30. Whittaker, R.J. & Fernandez-Palacios, J.M. (2007) Island biogeography. Ecology, evolution, and conservation, 2nd edn. Oxford University Press, Oxford. Wiens, J.J., Kuczynski, C.A., Townsend, T., Reeder, T.W., Mulcahy, D.G. & Sites, J.W. (2010) Combining phylogenomics and fossils in higher-level squamate reptile phylogeny: molecular data change the placement of fossil taxa. Systematic Biology, 59, 674 688. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Appendix S1 The dataset and metadata (literature source). Appendix S2 Equations used to estimate mass from snout vent lengths in clades not reported in Meiri (2010) and the data used to generate them. Appendix S3 Phylogenetic relationships between the lizard species in our dataset and the literature sources used to reconstruct them. Appendix S4 Results of comparisons between life-history traits of congeneric insular and mainland species. Appendix S5 The relationship between female mass and latitude in the best models for different life-history traits. Appendix S6 Results of regressions of the four response variables on population density and island area. Appendix S7 Results of the quadratic regressions. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. BIOSKETCHES Maria Novosolov is an MSc student studying the evolution of life-history traits in island endemic reptiles. Pasquale Raia studies palaeobiology of insular vertebrates, evolutionary responses to insularity, and the evolutionary biology of Cenozoic mammals. Shai Meiri studies the biogeography of animal traits in different vertebrate clades, evolutionary responses to insularity, and the patterns, drivers and consequences of the global distribution of reptiles Editor: Miguel Olalla-Tárraga Global Ecology and Biogeography, 22, 184 191, 2012 Blackwell Publishing Ltd 191