Lizards on newly created islands independently and rapidly adapt in morphology and diet

Contributed by Thomas W. Schoener, June 21, 2017 (sent for review December 31, 2016; reviewed by Raymond B. Huey and Dolph Schluter)
July 31, 2017
114 (33) 8812-8816


We report for island populations of the termite-eating common gecko species Gymnodactylus amarali rapid parallel morphological and ecological change in response to human-caused environmental disturbance. The islands were formerly part of an extensive terrestrial ecosystem; in 1997, the area was flooded to construct a reservoir, fragmenting the higher portions into separate islands. Populations on all five islands studied have proportionally larger heads than populations at five nearby mainland sites. The new island morphology is accompanied by an increase in dietary niche breadth, mainly via expansion toward larger prey. This expansion is likely due to the greater availability of such prey on the newly formed islands after the extinction there of four larger lizard species that typically also included termites in their diets.


Rapid adaptive changes can result from the drastic alterations humans impose on ecosystems. For example, flooding large areas for hydroelectric dams converts mountaintops into islands and leaves surviving populations in a new environment. We report differences in morphology and diet of the termite-eating gecko Gymnodactylus amarali between five such newly created islands and five nearby mainland sites located in the Brazilian Cerrado, a biodiversity hotspot. Mean prey size and dietary prey-size breadth were larger on islands than mainlands, expected because four larger lizard species that also consume termites, but presumably prefer larger prey, went extinct on the islands. In addition, island populations had larger heads relative to their body length than mainland populations; larger heads are more suited to the larger prey taken, and disproportionately larger heads allow that functional advantage without an increase in energetic requirements resulting from larger body size. Parallel morphological evolution is strongly suggested, because there are indications that, before flooding, relative head size did not differ between future island and future mainland sites. Females and males showed the same trend of relatively larger heads on islands, so the difference between island and mainland sites is unlikely to be due to greater male–male competition for mates on islands. We thus discovered a very fast (at most 15 y) case of independent parallel adaptive change in response to catastrophic human disturbance.
Rapid evolution has been recorded recently for several taxa (1, 2), with rates of phenotypic change approaching, but not quite matching, rates of ecological change (3). Fast evolution is often driven by sudden anthropogenic environmental alteration, such as the flooding of large areas after the construction of hydroelectric dams. This inundation can convert mountaintops into islands, drastically shrinking continuous ranges of populations (4). The MacArthur–Wilson Species Equilibrium Model (5, 6) predicts that isolation and reduction of area will lead to species loss. In particular, a guild of competing species may lose some members, allowing those remaining to expand their niches (i.e., “ecological release”) (7). In this and other ways, island species may systematically differ from their mainland counterparts, phenomena that have been explored in various taxa (8), including lizards (912).
An example of rapidly human-created islands is at the lake resulting from the Serra da Mesa Hydroelectric Plant in central Brazil (Fig. S1). The reservoir is located within the Cerrado hotspot, a region of great conservation importance due to its combination of unique species and major human disturbance (13). The filling of the Serra da Mesa reservoir began in 1996 and flooded an area of 170,000 ha, comprising several valleys, forming ∼290 islands (14). Periodic monitoring at island and mainland sites showed that lizard communities suffered significant impacts, including the extinction of most large-sized lizard species on many islands (refs. 15 and 16; Table S1 and SI Methods). We selected the termite-specializing gecko Gymnodactylus amarali (Gekkonidae) as the focus of this study, because it was the most common lizard species in the area at the time of the field study.
Fig. S1.
Serra da Mesa hydroelectric power plant reservoir study sites. M1, M2, M3, M4, and M5 are the mainland sites, and I34, I35, I37, I38, and IX are the island sites.
Table S1.
Lizard species sampled in 1996, 2001, and 2011 at Sector 1 of the Serra da Mesa reservoir, municipality of Minacu, Goias, Brazil
Species that were not registered on the islands, but were found at Sector 1 of the Serra da Mesa reservoir area.
We evaluated the effects of isolation (actually, insularization) on diet and morphology of G. amarali populations on islands formed by the Serra da Mesa reservoir. We collected data on lizard diet and morphology on five islands, as well as five nearby mainland areas, to evaluate the changes that occurred as a result of insularization. One of our studied island sites has been periodically connected to the mainland (Island IX), but the other four islands (I34, I35, I37, and I38) became and remained isolated since 1998, when the reservoir was filled (SI Methods). We proposed the following hypotheses:
Island populations of G. amarali have greater food niche breadth, based on prey size, than mainland populations. Reduction of species richness on islands likely reduces interspecific competition, allowing a niche expansion of the remaining species. Because G. amarali specializes in termites (17), this expansion would probably not occur through the addition of nontermite taxa to their diet. However, increased availability of larger termites could increase niche breadth along the food-size dimension. This hypothesis was driven by the fact that four species of termite-eating lizards that became extinct on the islands, but not at the mainland sites, were all larger than G. amarali (SI Methods). Because these larger lizards are able to ingest larger termites, their extinction would reduce competition for this resource, increasing the availability of large-sized termites to G. amarali. Termite-nest mound number and volume showed no difference since 1996 between islands and the mainland (15, 16), suggesting that termite abundance has not changed significantly as a result of insularization.
Island populations consume larger termite sizes than mainland populations. Because diet expansion is mainly expected toward larger prey for energetic reasons (ref. 18 and below), the increase in niche breadth hypothesized under hypothesis 1 would result in larger mean prey sizes being consumed on islands.
Island lizards have larger head lengths than mainland lizards. We expect that island individuals would have larger head lengths than those at mainland sites for the same body size (i.e., would have disproportionately larger heads) for the following reasons. Lizards consuming larger prey should have a larger trophic apparatus. However, the energetic advantage of consuming larger prey would diminish if total energy requirements were to increase as well due to a larger body size. This disadvantage would be mitigated were island lizards to have larger heads for the same body size, allowing an increase in the size of prey taken without substantially increasing energy requirements.
All island populations had food-niche breadths (inverse of Simpson’s diversity index; Methods) that were greater than all mainland populations (means 3.74 vs. 2.38 respectively, t = 2.511, df = 4, one-tailed P = 0.033; Table 1). Thus, lizards on islands ate prey with a broader distribution of body sizes than lizards of the nonisolated, mainland areas. These results confirmed that populations of G. amarali in Serra da Mesa did indeed increase their dietary niche breadth once isolated on islands, as hypothesized (hypothesis i above).
Table 1.
Diet niche breadth (B) for populations of G. amarali in each island and mainland sites of the Serra da Mesa reservoir, using the inverse of Simpson’s diversity index (34)
Mainland B012.917
Mainland B022.259
Mainland B041.964
Island I343.399
Island I353.160
Island I384.645
Mean mainland2.380
Mean island3.735
As hypothesized (hypothesis ii above), because of differential expansion along the food-size dimension toward larger prey, G. amarali on islands ate larger termites than individuals with the same body length on mainland areas (Fig. 1A; mixed model analysis: adjusted means 4.93 vs. 4.23 mm, respectively, χ2 = 4.2, df = 1, two-tailed P = 0.041). Two tests gave no indication that male and female lizards differed in prey size (sex effect: χ2 = 2.2, df = 1, two-tailed P = 0.13; sex × location interaction: χ2 = 1.2, df = 1, two-tailed P = 0.27).
Fig. 1.
Mean prey size (A) and head length (B) as a function of body size (SVL) for G. amarali on islands and mainland sites of the Serra da Mesa reservoir. Each point is a separate individual.
Lizards had a greater head length on islands than on the mainland for a given body size (adjusted means 10.2 vs. 9.8 mm, respectively, χ2 = 11.0, df = 1, two-tailed P < 0.0001; Fig. 1B; details are in Methods). Head length was correlated with mean termite length in stomachs (Pearson product-moment correlation coefficient = 0.32, df = 48, two-tailed P = 0.025). Analyses distinguishing the sexes gave no significant difference in relative head length or body size between males and females [sex effect on relative head length: χ2 = 0.02, df = 1, two-tailed P = 0.88; on snout-vent length (SVL): χ2 = 0.54, df = 1, two-tailed P = 0.46], nor was there a significant interaction between sex and location (relative head length: χ2 = 0.00, df = 1, two-tailed P = 0.95; SVL: χ2 = 1.5, df = 1, two-tailed P = 0.22). These results were bolstered by the analysis of residuals that considered each site as a separate population; it showed that island populations have a larger mean head length relative to body length than mainland populations (t = 4.216, df = 6, two-tailed, P = 0.006)—in fact, there is a perfect ranking (Fig. 2).
Fig. 2.
Mean of the residual head lengths from linear regression of head length vs. SVL of G. amarali from five island and five mainland sites at the Serra da Mesa reservoir.
According to the ecological release hypothesis, the loss of species on newly formed islands should result in reduced interspecific competition and, consequently, less consumption of certain resources, allowing the remaining species to eat a larger spectrum of prey (i.e., have larger dietary niche breadths) (1921). In addition, island populations might be expected to evolve appropriate morphological changes, allowing them to use a greater resource availability. Our findings support these predictions. Populations of G. amarali in the Serra da Mesa islands increased their food-niche breadth by adding larger termites to their diet. They also exhibited larger heads, relative to body size, than their mainland counterparts. These two results are directly related, because lizards with larger trophic apparatuses are able to ingest larger prey, while still being able to ingest smaller prey (18), thereby adding new items to their diet and expanding their niche along the prey-size dimension.
That lizards of a given size on islands consumed larger termites on average than did those on mainland may also be because, once lizards have a morphology that allows them to consume a variety of sizes of prey, they will prefer those that give them more energy per handling and swallowing time—in this case, larger prey (18, 22). Indeed, Cohen et al. (23) showed a positive correlation between prey and predator size for a wide variety of organisms. However, larger predators may not always consume their preferred prey. In particular, because of lower abundance of large food items, larger predators may be forced to eat small items if large food has not been available for a period. Moreover, larger predators should take small prey if they have to make only a minimum effort, such as when small prey occurs near a perched lizard. Note that this scheme implies much variation in the relation of predator to prey size calculated from short feeding bouts (as when stomach contents are analyzed). Such a “snapshot” could differ greatly from the long-term diet (e.g., the long term prey-size distribution). In particular, some larger predators in the present study seem to have consumed inappropriately small prey. Fig. 1A illustrates this variability: Whereas the smallest lizards never take large prey (presumably because they cannot consume it efficiently), the largest lizards frequently take small prey—notice how the cluster of points is both quite diffuse and expands outward on the right-hand side.
Perhaps the most interesting result of this study is the disproportionately larger heads on islands than on the adjacent mainland. In fact, the relation is a perfect separation: All of the island populations have disproportionally larger heads than all of the mainland populations (Fig. 2). Before the creation of the reservoir, gecko populations at future island and mainland sites likely had similar characteristics, because they belonged to the same continuous population before the rise in water level. In fact, although data are not abundant, we did not detect differences in relative head length between future island (13 individuals) and mainland (10 individuals) sites before flooding (1996–1998; U1,23 = 69.0; P = 0.68). Considering the island-formation process, the astonishing aspect of this morphological shift is that the five island populations developed larger heads independently of one another, because the islands quickly separated before populations had much time to undergo change (SI Methods). Thus, each new population of G. amarali developed the same new traits independently on each island, driven by apparently similar changes in community structure in the newly isolated areas, such as the extinction of larger lizards on all studied islands. Specifically, the reduction in interspecific competition, and the consequent increased availability of certain resources, may have led to the observed dietary niche expansion in which larger prey were selected. Foraging for larger prey should have favored lizards with a disproportionately large head size, making them more efficient in the consumption of this resource. Note from Fig. 1 that the body-size range for island and mainland individuals is similar—the increase in relative head size is the key result. This set of results strongly indicates phenotypic parallelism.
An alternative hypothesis might be male–male competition for mates, with males hypothetically being sexually selected to have relatively larger heads on islands because of possible higher intraspecific densities there. Although Colli et al. (17) did find slightly larger heads for males than females in their study of the same species (but wide sexual overlap in overall morphological space), we found no significant sexual differences in our study for relative head length: First, in the mixed-model analyses with head length as the dependent variable, neither sex nor the sex × location interaction was significant, although there was a tendency for males to have longer heads than females (adjusted means 10.03 vs. 9.92 mm, respectively; P = 0.13); second, when we performed the mixed-model analyses separately for each sex, both males and females had significantly greater relative head lengths on islands (males: χ2 = 4.8, df = 1, two-tailed P = 0.03; females: χ2 = 11.4, df = 1, two-tailed P < 0.001). If some sort of social hypothesis such as greater efficacy in aggression on islands were correct, it would have to apply to both males and females.
Of the five sampled islands, four (I38, I35, I37, and I34; Fig. S1 and Fig. 2) were created immediately after the start of the reservoir filling. The fifth island sampled, island IX (Fig. 2), was isolated from other islands since the beginning of flooding, but for a while remained connected to the mainland. According to satellite images, it separated only when the reservoir reached its maximum capacity, which occurred a few times during the entire period. Hence, isolation time is shorter for this island than the others. Therefore, island IX may receive more immigration from the mainland in those periods when they are connected. For these reasons, changes in traits on IX should not be as extreme as on the other islands. In fact, that is the case: Fig. 2 shows that, of all of the islands, IX is the most similar in relative head size to the mainland.
Classic studies of natural, nonexperimental character change show some similarity to our discoveries. A major example of predictable and repeatable morphological change are Fenchel’s (24) Hydrobia gastropods in Denmark. The ranges of two species of these deposit-feeders have come together repeatedly and independently over approximately a 150-y period after the collapse of a sea wall. In each area of sympatry, the species have evolved about the same body-size difference (always with the same species being larger), in turn correlated with food-size difference, a striking example of parallel character displacement. A second well-known such example is the lake-inhabiting sticklebacks studied by Schluter (25): Solitary species are intermediate in gill-raker length and other morphological traits compared with coexisting pairs of species, traits related to habitat and correlated dietary differences. On a much larger scale, Caribbean Anolis lizards exhibit independent parallel evolution: the four largest islands (the Greater Antilles) have much the same set of “ecomorphs”—species specialized to use particular perch heights and diameters—and these evolved largely independently and repeatedly on each island (2628). A perhaps even more diverse system is the cichlids of African lakes, which evolved often repeated morphologies associated with feeding mode and/or habitat (29). Finally, Huey et al. (30) showed predictable character displacement in the termite-specializing fossorial lizard Acontias (Typhlosaurus in ref. 30), including prey-size changes precisely corresponding to the morphological changes. It should be recalled that what we have shown statistically for geckos is character change in head size relative to body size.
The above studies are all ones in which the character change of a given species in response to other similar species was known to be slower than in our gecko system, or not precisely known, but inferred to be slower. Two studies, however, show character change over comparable periods of time. One is the classic case reported by Grant and Grant (31) of Geospiza on Daphne Island, in which the invasion of a larger-beaked second species caused the resident species to shift toward smaller beak sizes in a few months. This study is, of course, a single instance, rather than a parallel and replicated change over several islands. A second study that is more comparable with ours focuses on Anolis carolinensis on small islands in Florida invaded by the conspecific Anolis sagrei. Using a combination of experimentally and naturally invaded islands, Stuart et al. (32) showed that the resident, more arboreal A. carolinensis shifted its height upward and evolved larger toe pads after 20 y.
Our results provide strong evidence that particular circumstances predictably and precisely shape traits of species in ecological communities and illustrate that populations can respond both rapidly, and in parallel, to ecological change—results of basic scientific interest. In addition, they may have applied significance, because understanding rapid evolution in fragmented populations is important for conservation purposes (e.g., ref. 33). Indeed, whereas the subject gecko of this study persisted and adapted in newly created habitat fragments, a number of other lizard species were extirpated, illustrating the potential consequences of insufficient responses to rapid environmental changes.

SI Methods

Study Area.

The study was conducted at Sector 1 of the Serra da Mesa reservoir, located in the upper Tocantins river basin, north of the state of Goias, in midwestern Brazil (13°50′S, 48°20′W) (Fig. S1). In October 1996, the dam was closed, flooding an area of ∼178,000 ha in a filling process that lasted for ∼2 y, creating the largest reservoir by water volume in Brazil. The area has an irregular topography, composed of several valleys that were flooded, leaving only the tops of hills above the water, which became isolated islands. The vegetation of these islands consists of cerrado sensu stricto and some patches of rocky cerrado.
Herpetofaunal monitoring (Table S1) has been conducted since 1996 in the area (14). Over a 5-y monitoring period, lizard richness decreased due to the rapid extinction of one species, Ameiva ameiva. Ten years after the complete filling of the reservoir, five other species become apparently extirpated from the islands of Serra da Mesa—Cercosaura ocellata, Colobosaura modesta, Norops meridionalis, Norops brasiliensis, and Tropidurus montanus — and two of these species—C. ocellata and C. modesta—were also not found on the sampled mainland sites (16).
Except for C. ocellata (a known spider eater; ref. 41) and C. modesta (feeding largely on spiders and orthopterans; ref. 42), all of the extinct species from the islands are larger than the study subject, G. amarali, and have termites in their diet: A. ameiva ranges from 34.2 to 126.0 mm SVL, and 75.41% of its diet is composed by Isoptera (43); N. meridionalis maximum SVL is 56.0 mm (44), and 16.83% of its diet was composed of Isoptera, with an importance value index (IVI) of 11.64 (45); N. brasiliensis maximum SVL is 69.0 mm, and 16.81% of its diet is Isoptera, with an IVI of 7.76 (46); T. montanus ranges from 30.5 to 98.0 mm in size (47), and 4.8% of its diet is Isoptera, present in 34.8% of the stomachs analyzed in the study (48). It is probable that these large lizards consumed larger termites, so their extinction may have increased the availability of such termites.

Study Subjects.

G. amarali is a Cerrado endemic gecko, characteristic of rocky fields, but often found in termite nests, because termites comprise most of their diet (49). Individuals range from 20 to 55 mm in size. They are thermoconformers (49) and breed in the dry season, perhaps because it is the period of greatest termite availability (17).

Field Procedures.

We sampled five islands, as well as five mainland areas (Fig. S1), to test our hypotheses. Sampling was carried out for a total of 2 mo spread throughout July–October 2011. All sites had similar structural and microhabitat features for lizards—vegetation, termite nests, burrows, and other typical shelters. For lizard collection, we used the “exhaustive sampling squares” methodology, developed and described by R.A.B. In each of the 10 sites, we sampled a 2,500-m2 plot. The plot perimeters were enclosed by vertical barriers of plastic sheeting to prevent lizard escape and was surrounded by firebreaks. After enclosure, the area inside the plot was burned to remove the undergrowth and facilitate the search for lizards. All precautions were taken to prevent fire from spreading to areas outside the plot. It is important to note that the flora and fauna of the Cerrado biome are adapted to the natural regime of fires (50). In previous situations in which this method was used (15, 16), no animal was found burned, and fire did not spread to other areas. The study area was inspected thoroughly in search of lizards just after fire, and all shelters were carefully checked. All lizards found were hand-caught, counted, weighed, and measured before being euthanized. After 2 h of inspection without finding additional lizards, we assumed that all individuals in the area were captured, and the enclosure was taken down. All of the animals collected were deposited in the Herpetological Collection of the University of Brasília, Department of Zoology.


Laboratory Procedures.

We recorded SVL and head length (anterior edge of tympanum to the nose tip) to the nearest 0.01 mm, using Mitutoyo digital calipers. All measurements were performed by the same person (A.C.R.L.). To determine diets, we analyzed stomach contents of each individual under a stereoscopic microscope, identifying prey items to order. G. amarali is highly specialized on termites (17), and 90% percent of all items in this study were termites. Therefore, we used only that taxon in our analysis. We recorded prey length (millimeters) of intact items with graph paper.

Statistical Analyses.

Diet niche breadth was calculated by using prey size as the dimension. We analyzed the contents of 50 G. amarali stomachs with three or more prey items—24 from island and 26 from mainland sites. Each study area was considered a distinct population. We sampled 10 areas (5 islands and 5 mainland sites), but 2 mainland and 2 island sites were omitted from niche-breadth analysis due to small sample size (<50 prey items), leaving a total of 3 island and 3 mainland sites. Animals were collected under Brazilian Permanent Permit 28190-1; our study follows the Brazilian law on ethical use of animals for research and education (Federal Law 6899/2009), being supported by permanent license from the Brazilian System of Access to Biological Information (SISBIO) and approved by the University of Brasília institutional review board [Commission on Animal Ethical Use (CEUA), Process #69/2010].
Seven size categories of termite length (in millimeters) were defined: 2–2.9; 3–3.9; 4–4.9; 5–5.9; 6–6.9; 7–7.9; and 8–8.9. We computed niche breadths (B) of each population using the inverse of Simpson’s diversity index (34):
where pi is the proportion of individuals of a given size (length) i found in the population diet, and n is the number of categories. Calculation of niche breadth using this index generates values ranging between 1 and n; values close to 1 have a narrower niche (the population commonly consumes few sizes of prey), whereas values close to n represent a broader niche (the population commonly consumes many sizes of prey).
A one-tailed t test for independent samples was performed on the mean island vs. mainland B values to detect differences in niche breadth between the two kinds of sites; the test is one-tailed because a change in the opposite direction is not predicted. All other analyses in this work use two-tailed tests.
We chose head length as the focal morphological variable, because it is functionally expected to relate to prey size and, in fact, was found to do so in many previous studies (22, 3537). Individuals with larger heads can eat larger prey, and this capability also results in a potentially wider range of prey sizes for larger-headed lizards (18, 23). We used linear mixed models to evaluate differences in head length of G. amarali between islands and mainland sites. Head length was the dependent variable; location (island vs. mainland) and SVL were fixed predictor variables (the latter was included to account for the fact that larger lizards have longer heads); and study site (each of the five islands and five mainland areas) was included as a random effect to account for the nonindependence of lizards collected at the same study site. The analysis of head length featured 49 individuals from the five islands and 43 individuals from the five mainland sites.
We ran a second linear mixed model to test whether lizards of the same body size were consuming different prey sizes on islands vs. mainland. Mean termite length in an individual’s diet was the dependent variable, and the independent variables were the same as those in the analysis of head length. Examination of residual plots from preliminary analyses of termite length suggested that differences in residual variation between islands and mainland may violate assumptions of homoscedasticity. This pattern is not surprising, given that our analysis of niche breadth showed a greater breadth of prey size used on islands than on mainland, consistent with hypothesis i (Table 1; the difference in the variation of termite length between islands and mainland is also apparent in Fig. 1A). To account for this heteroscedasticity, we included an additional parameter in our model that allowed islands and mainland to have different variance; the model with the additional parameter accounting for heteroscedasticity had a lower Akaike information criterion than the model without the adjustment for heteroscedasticity (1,775.4 vs. 177.2). The analysis of termite length featured 24 individuals from islands and 26 individuals from mainland sites.
We used likelihood ratio tests to assess the significance of fixed effects in linear mixed models. Preliminary analyses of both head length and termite length showed no significant interactions between the location (mainland vs. island) and SVL (P > 0.99 in both cases), and no effect of sex (P > 0.13 in both cases) or location × sex interaction (P > 0.27 in both cases), so these terms were dropped from the final model.
To further explore our data, we performed an analysis based on residuals: We calculated from the general regression (all data combined including both sexes because no significant difference between males and females was found) the residual head lengths for each individual and then took the average (signed) residual for the individuals at each site, giving five mainland and five island values of mean residuals. We then performed a t test on those means. Because of inadequate prey sample sizes for some individuals at some of the sites, a similar analysis could not be performed using residual prey sizes. We performed the same analysis separately for males and females, using only sites having at least five individuals (three each for island and mainland).
Only sexually mature individuals were used in all analyses—minimum 27-mm SVL (17) for both sexes. Three outliers were damaged during collection to the point of increasing the likelihood of measurement error; these lizards were not included in our analyses.
We carried out statistical analyses of diet breadth using SPSS Statistics 21 for Macintosh (38); mixed modeling and analysis of residuals were conducted by using the nlme package in R (39, 40).
Data are available at

Data Availability

Data deposition: Data have been deposited in the Dryad Digital Repository, (doi:10.5061/dryad.3nk78).


We thank David Spiller, Chris Martin, Derek Young, Maria Cristina, Heury Ferr, and those who helped with the fieldwork, especially Leonardo Gomes and Tayna Oliveira. This work was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior Foundation; Ministry of Education of Brazil Grant 11493/13-5; and Infraero Project Fauna Infraero/UnB.

Supporting Information

Supporting Information (PDF)


TW Schoener, The newest synthesis: Understanding the interplay of evolutionary and ecological dynamics. Science 331, 426–429 (2011).
TW Schoener, J Moya-Larano, J Rowntree, Preface. Adv Ecol Res 50, XIII–XX (2014).
JP DeLong, et al., How fast is fast? Eco-evolutionary dynamics and rates of change in populations and phenotypes. Ecol Evol 6, 573–581 (2016).
J Terborgh, K Feeley, Propagation of trophic cascades in tropical forests. Trophic Cascades, eds J Terborgh, JA Estes (Island, Washington, DC), pp. 91–106 (2010).
RH MacArthur, EO Wilson The Theory of Island Biogeography (Princeton Univ Press, Princeton, 1967).
TW Schoener, The MacArthur-Wilson equilibrium model: What it said and how it was tested. The Theory of Island Biogeography Revisited, eds JB Losos, RE Ricklefs (Princeton Univ Press, Princeton), pp. 52–87 (2009).
DO Mesquita, GR Colli, LJ Vitt, Ecological release in lizard assemblages of neotropical savannas. Oecologia 153, 185–195 (2007).
MV Lomolino, Body size evolution in insular vertebrates: Generality of the island rule. J Biogeogr 32, 1683–1699 (2005).
JB Losos, DJ Irschick, TW Schoener, Adaptation and constraint in the evolution of specialization of Bahamian Anolis lizards. Evolution 48, 1786–1798 (1994).
S Meiri, Size evolution in island lizards. Glob Ecol Biogeogr 16, 702–708 (2007).
P Raia, et al., The blue lizard spandrel and the island syndrome. BMC Evol Biol 10, 289 (2010).
K Sagonas, et al., Insularity affects head morphology, bite force and diet in a Mediterranean lizard. Biol J Linn Soc Lond 112, 469–484 (2014).
N Myers, RA Mittermeier, CG Mittermeier, GAB Fonseca, J Kent Hotspots (Cemex Conservation International, Arlington, VA, 2000).
RA Brandão, AFB Araújo, Changes in anuran species richness and abundance resulting from hydroelectric dam flooding in Central Brazil. Biotropica 40, 263–266 (2008).
RA Brandão, Monitoramento das populações de lagartos no Aproveitamento Hidroelétrico de Serra da Mesa, Minaçu, GO. Thesis (Universidade de Brasília, Brasilia, Brazil). (2002).
GRCC Santoro, Mudanças temporais, após 13 anos de insularização, em comunidades de lagartos (Squamata) em ilhas formadas por um grande reservatório no Brasil central. Thesis (Universidade de Brasília, Brasilia, Brazil). (2012).
GR Colli, DO Mesquita, PVV Rodrigues, K Kitayama, Ecology of the gecko Gymnodactylus geckoides amarali in a Neotropical Savanna. J Herpetol 37, 694–706 (2003).
TW Schoener, Models of optimal size for solitary predators. Am Nat 103, 277–313 (1969).
RH MacArthur, JM Diamond, JR Karr, Density compensation on island faunas. Ecology 53, 330–342 (1972).
ER Pianka Evolutionary Ecology (Harper and Row, 6th Ed, New York, 2000).
DI Bolnick, et al., Ecological release from interspecific competition leads to decoupled changes in population and individual niche width. Proc Biol Sci 277, 1789–1797 (2010).
TW Schoener, GC Gorman, Some niche differences in three Lesser Antillean lizards of the genus Anolis. Ecology 49, 819–830 (1968).
JE Cohen, SL Pimm, P Yodzis, J Saldana, Body sizes of animal predators and animal prey in food webs. J Anim Ecol 62, 67–78 (1993).
T Fenchel, Character displacement and coexistence in mud snails (Hydrobiidae). Oecologia 20, 19–32 (1975).
D Schluter The Ecology of Adaptive Radiation (Oxford Univ Press, Oxford, 2000).
EE Williams, Ecomorphs, faunas, island size, and diverse end points in island radiations of anoles (Sauria, Iguanidae). Lizard Ecology: Studies of a Model Organism, eds RB Huey, ER Pianka, TW Schoener (Harvard Univ Press, Cambridge, MA), pp. 326–370 (1983).
JB Losos, The evolution of convergent structure in Caribbean Anolis communities. Syst Biol 41, 403–420 (1992).
JB Losos Lizards in an Evolutionary Tree (Univ of California Press, Berkeley, CA, 2009).
TD Kocher, Adaptive evolution and explosive speciation: The cichlid fish model. Nat Rev Genet 5, 288–298 (2004).
RB Huey, ER Pianka, ME Egan, LW Coons, Ecological shifts in sympatry: Kalahari fossorial lizards (Typhlosaurus). Ecology 55, 304–316 (1974).
PR Grant, BR Grant, Evolution of character displacement in Darwin’s finches. Science 313, 224–226 (2006).
YE Stuart, et al., Rapid evolution of a native species following invasion by a congener. Science 346, 463–466 (2014).
SM Carlson, CJ Cunningham, PAH Westley, Evolutionary rescue in a changing world. Trends Ecol Evol 29, 521–530 (2014).
EH Simpson, Measurement of diversity. Nature 163, 688 (1949).
ER Pianka, HD Pianka, Comparative ecology of twelve species of nocturnal lizards (Gekkonidae) in the Western Australian desert. Copeia 1976, 125–142 (1976).
TW Schoener, The ecological significance of sexual dimorphism in size in the lizard Anolis conspersus. Science 155, 474–477 (1967).
B Herrel, B Van Damme, B Vanhooydonck, F De Vree, The implications of bite performance for diet in two species of lacertid lizards. Can J Zool 79, 662–670 (2001).
; IBM SPSS, Statistics for Macintosh (IBM Corp Armonk, NY), Version 21.0. (2012).
J Pinheiro, D Bates, S DebRoy, D Sarkar; R Core Team, nlme: Linear and Nonlinear Mixed Effects Models. R package, Version 3. Available at (2015).
; R Core Team, R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna) Available at (2015).
LJ Vitt, PA Zani, MA Espósito, Historical ecology of Amazonian lizards: Implications for community ecology. Oikos 87, 286–294 (1999).
DO Mesquita, GR Colli, FGR França, LJ Vitt, Ecology of Cerrado lizard assemblage in the Jalapão region of Brazil. Copeia 2006, 460–471 (2006).
TF Silva, BFE Andrade, RF Teixeira, M Giovanelli, Ecologia de Ameiva ameiva (Sauria, Teiidae) na Restinga de Guriri, São Mateus, Espírito Santo, sudeste do Brasil. Boletim do Museu de Biologia Mello Leitão 15, 5–15 (2003).
AB D’angiolella, et al., Anolis chrysolepis Duméril and Bibron, 1837 (Squamata: Iguanidae), revisited: Molecular phylogeny and taxonomy of the Anolis chrysolepis species group. Bull Mus Comp Zool 160, 35–63 (2011).
LBA Veludo, Ecologia de Anolis meridionalis (Squamata, Polychrotidae) no Cerrado brasileiro. Thesis (Universidade de Brasília, Brasilia, Brazil). (2011).
DO Mesquita, et al., The autoecology of Anolis brasiliensis (Squamata, Teiidae) in a Neotropical Savanna. Herpetol J 25, 233–244 (2015).
M Van Sluys, HMA Mendes, VB Assis, MC Kiefer, Reproduction of Tropidurus montanus Rodrigues, 1987 (Tropiduridae), a lizard from a seasonal habitat of south-eastern Brazil, and a comparison with other Tropidurus species. Herpetol J 1, 89–97 (2002).
M Van Sluys, CFD Rocha, D Vrcibradic, CAB Galdino, AF Fontes, Diet, activity, and microhabitat use of two syntopic Tropidurus species (Lacertilia: Tropiduridae) in Minas Gerais, Brazil. J Herpetol 38, 606–611 (2004).
LJ Vitt, et al., Uma atualização do guia fotográfico de répteis e anfíbios da região do Jalapão no Cerrado brasileiro. Occasional paper (Sam Noble Oklahoma Museum of Natural History, Norman, OK). (2005).
MM Cole The Savannas: Biogeography and Geobotany (Academic, London, 1986).

Information & Authors


Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 114 | No. 33
August 15, 2017
PubMed: 28760959


Data Availability

Data deposition: Data have been deposited in the Dryad Digital Repository, (doi:10.5061/dryad.3nk78).

Submission history

Published online: July 31, 2017
Published in issue: August 15, 2017


  1. rapid character change
  2. islands
  3. dietary shift
  4. Brazilian Cerrado
  5. lizards


We thank David Spiller, Chris Martin, Derek Young, Maria Cristina, Heury Ferr, and those who helped with the fieldwork, especially Leonardo Gomes and Tayna Oliveira. This work was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior Foundation; Ministry of Education of Brazil Grant 11493/13-5; and Infraero Project Fauna Infraero/UnB.



Mariana Eloy de Amorim1 [email protected]
Laboratório de Fauna e Unidades de Conservação, Departamento de Engenharia Florestal, Universidade de Brasília, Brasilia DF, Brazil CEP 70910-900;
Evolution and Ecology Department, University of California, Davis, CA 95616;
Thomas W. Schoener1 [email protected]
Evolution and Ecology Department, University of California, Davis, CA 95616;
Guilherme Ramalho Chagas Cataldi Santoro
Departamento de Pós-Graduação em Zoologia, Instituto de Biologia, Universidade de Brasília, Brasilia DF, Brazil CEP 70910-900;
Anna Carolina Ramalho Lins
Laboratório de Fauna e Unidades de Conservação, Departamento de Engenharia Florestal, Universidade de Brasília, Brasilia DF, Brazil CEP 70910-900;
Jonah Piovia-Scott
School of Biological Sciences, Washington State University, Vancouver, WA 98686-9600
Reuber Albuquerque Brandão
Laboratório de Fauna e Unidades de Conservação, Departamento de Engenharia Florestal, Universidade de Brasília, Brasilia DF, Brazil CEP 70910-900;


To whom correspondence may be addressed. Email: [email protected] or [email protected].
Author contributions: M.E.A., T.W.S., G.R.C.C.S., A.C.R.L., and R.A.B. designed research; M.E.A., G.R.C.C.S., A.C.R.L., and R.A.B. performed research; M.E.A., T.W.S., J.P.-S., and R.A.B. analyzed data; and M.E.A., T.W.S., G.R.C.C.S., A.C.R.L., J.P.-S., and R.A.B. wrote the paper.
Reviewers: R.B.H., University of Washington; and D.S., University of British Columbia.

Competing Interests

The authors declare no conflict of interest.

Metrics & Citations


Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.

Citation statements



If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by


    View Options

    View options

    PDF format

    Download this article as a PDF file


    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    Lizards on newly created islands independently and rapidly adapt in morphology and diet
    Proceedings of the National Academy of Sciences
    • Vol. 114
    • No. 33
    • pp. 8661-E7031







    Share article link

    Share on social media