The insular herpetofauna of Mexico: Composition, conservation, and biogeographic patterns

Abstract We compile a Mexican insular herpetofaunal checklist to estimate endemism, conservation status, island threats, net taxonomic turnover among six biogeographic provinces belonging to the Nearctic and Neotropical regions, and the relationships between island area and mainland distance versus species richness. We compile a checklist of insular herpetofaunal through performing a literature and collection review. We define the conservation status according to conservation Mexican law, the Red List of International Union for Conservation of Nature, and Environmental Vulnerability Scores. We determine threat percentages on islands according to the 11 major classes of threats to biodiversity. We estimate the net taxonomic turnover with beta diversity analysis between the Nearctic and Neotropical provinces. The Mexican insular herpetofauna is composed of 18 amphibian species, 204 species with 101 subspecies of reptiles, and 263 taxa in total. Endemism levels are 11.76% in amphibians, 53.57% in reptiles, and 27.91% being insular endemic taxa. Two conservation status systems classify the species at high extinction risk, while the remaining system suggests less concern. However, all systems indicate species lacking assessment. Human activities and exotic alien species are present on 60% of 131 islands. The taxonomic turnover value is high (0.89), with a clear herpetofaunal differentiation between the two biogeographic regions. The species–area and species–mainland distance relationships are positive. Insular herpetofauna faces a high percentage of threats, with the Neotropical provinces more heavily impacted. It is urgent to explore the remaining islands (3,079 islands) and better incorporate insular populations and species in ecological, evolutionary, and systematic studies. In the face of the biodiversity crisis, islands will play a leading role as a model to apply restoration and conservation strategies.


| INTRODUC TI ON
The islands have served as natural laboratories for the theoretical and empirical study of ecology, evolution, and conservation. Charles Darwin and Alfred R. Wallace were inspired by their observations in the Galapagos Islands and Malayan Archipelago to formulate the theory of evolution by natural selection (Darwin, 1859;Wallace, 1869).
Furthermore, the island's systematic studies led from the ecology and biogeography of descriptive work to more analytical approaches through the fundamental book "The Theory of Island Biogeography" of Robert MacArthur and Edward Wilson in 1967(Losos & Parent, 2009).
One of the most transcendental conclusions of this book is that the contribution of colonization, extinction, and speciation depends on the island area and the degree of isolation, which together explain species richness (Valente et al., 2020). Thus, island research has provided insights that have fundamentally transformed our view of biogeography, ecology, and evolution (Lomolino et al., 2009).
The oceanic islands are located in all latitudes and contain a considerable proportion of the planet's biodiversity. Due to their geographic isolation, islands harbor a disproportionately high percentage of endemic species, but moderate species richness compared with the mainland (Whittaker & Fernández-Palacios, 2007).
Moreover, islands have been involved in the evolution of both exceptional evolutionary patterns (e.g., adaptive radiations, evolutionary convergences) and unique phenotypic and functional traits (Kier et al., 2009;Russell & Kueffer, 2019). Currently, in the face of the biodiversity crisis, island ecosystems are the most threatened (Leclerc et al., 2020;Russell & Kueffer, 2019). Globally, islands are recognized as the epicenter of biodiversity loss (Spatz et al., 2017), with almost 40% of the species extinctions occurring on these systems, directly linked to human activities.
Both natural phenomena and anthropogenic activities can devastate entire island ecosystems. Hurricanes or cyclones have an intense effect on island biota, but these are agents of natural selection that shape the dynamics of colonization and extinction, and to which the species have been exposed and presumably adapted (Donihue et al., 2018;Losos & Ricklefs, 2009). In contrast, human activities are the more critical changing agent threatening the unique island biodiversity due to its short-time action (Donlan & Wilcox, 2008;Leclerc et al., 2018). Biological invasions, wildlife exploitation, and cultivation have been linked to the majority of insular extinctions and remain as the main threats to extant island species (Donlan & Wilcox, 2008;Leclerc et al., 2018). Additional synergistic factors such as pollution, urbanization, and climate change could accelerate the extinction of insular populations or species (Leclerc et al., 2020).
The risk of extinction of island species depends on both exposure and the interactions between threats (Leclerc et al., 2018).
According to the Red List of the International Union for Conservation of Nature (IUCN), amphibians and reptiles are the most underassessed groups within the terrestrial vertebrates, with 73% and 87% of evaluated species, respectively, compared with birds (100%) and mammals (91%; IUCN, 2021). The relative scarcity of extinction risk assessments could be due to comparatively limited geographic distribution, ecology, life history, and taxonomic studies (Böhm et al., 2013;Meiri & Chapple, 2016;da Silva et al., 2020;Tonini et al., 2016;Winter et al., 2016). Therefore, meetings continue to complete assessments to more accurately determine the category of risk of extinction (Gumbs et al., 2020;Tonini et al., 2016;Winter et al., 2016). Furthermore, of the amphibian and reptile species recorded as extinct by the IUCN, 62.86% and 90%, respectively, inhabited islands. Because island biodiversity has been the epicenter of global extinctions (Spatz et al., 2017), these patterns are likely to be reflected at local or regional scales. Here, we compile the distribution of amphibians and reptiles on islands of Mexico to determine the conservation status following three classification systems, endemicity level, threats associated with human activities, and the presence of invasive alien species on islands. Through a literature and scientific collections review, we compile a herpetofaunal checklist, resulting in the first attempt that integrates all the amphibian and reptile records for the Mexican islands (even among the terrestrial vertebrates). Further, with this inventory, we determine the percentage of different threats on islands, threat differences between biogeographic regions, the net taxonomic turnover associated with the Mexican biogeographic provinces, and the relationships between island-area and islandmainland distance against species richness.

| ME THODS
The islands in Mexico are located in six large oceanic ecoregions (Wilkinson et al., 2009). However, because most amphibian and reptile species have colonized islands from the mainland (to the exception of five marine turtles and one marine snake), we used the Mexican biogeographic province classification to explore the species presence by region and for taxonomic turnover. Thus, we grouped the islands within six biogeographic provinces associated with the Pacific and Atlantic versants. Moreover, this biogeographic classification includes the largest and most isolated islands in Mexico, according to Morrone et al. (2017). The following provinces were used: Californian (CP), Baja Californian (BCP), Sonoran (SP), Pacific Lowlands (PLP), Veracruzan (VP), and Yucatan Peninsula (YPP) provinces ( Figure 1). The Tamaulipas province has islands, but no record of amphibians or reptiles. The CP, BCP, and SP were included in the Nearctic region, and PLP, VP, and YPP correspond to the Neotropical region  Figure 1). Taxonomic position. We followed the AmphibiaWeb portal (AmphibiaWeb, 2020; http://amphi biaweb.org) for amphibians and F I G U R E 1 Map of Mexico showing the geographic location of the provinces in the Nearctic (blue colors) and Neotropical (yellow, red, and cream) regions, the number of species endemic to Mexico with a presence on the mainland (MX), island endemic species (ISE), island endemic subspecies (ISB), and nonendemic species (NME). The frog and lizard figures represent total amphibian and reptile species by province. Abbreviations: CP: Californian, BCP: Baja Californian, SP: Sonoran, PLP: Pacific Lowlands, VP: Veracruzan, and YPP: Yucatan Peninsula provinces. The map was based on Morrone et al. (2017) Uetz and Hošek (2020) (http://www.repti le-datab ase.org) for reptile names. For reptiles, Uetz and Hošek (2020) integrated into the database the subspecies category. We included subspecies because some of them are island endemics; therefore, we considered their recognition important. The taxonomic list is presented in Table S1.

| Data collection
We performed a literature review to search all possible "checklist," "inventories," and "new records" for "amphibians" and "reptiles," as well as other common names (e.g., frogs, toads, lizards, turtles, tortoise, and snakes) of the Mexican islands. We collected information from the literature by executing searches on Google Scholar and Web of Science (https://webof knowl edge.com) using the terms mentioned above plus "island," "islands," "cay," or "reef" with "Mexico." We located 16 studies resulting in a list of 111 islands with herpetofaunal records.
Additionally, we used the online platforms of Global Biodiversity Information Facility (GBIF; htpp://www.gbif.org) and Sistema Nacional de Información Sobre Biodiversidad de Mexico (CONABIO, 2020; http://www.snib.mx) to obtain additional records for museum specimens, as well as literature not published in scientific journals but that may contain regional lists or species list reports. Only those records that provided the deposit collection, voucher number (or photograph), coordinates, and the presence in marine or coastal islands were kept. Some coordinates did not coincide with any island despite the locality names, so these were excluded. With this revision, we compiled a total of 131 islands with herpetofaunal records (Table S2).

| Endemicity and conservation status
We classified the species as endemic to Mexico with a mainland presence, endemic island species, endemic island subspecies, and nonendemic species. We determined the conservation status of List is a classification system widely used in scientific research (e.g., Leclerc et al., 2018;Spatz et al., 2017). The categories include not evaluated (NE), data deficient (DD), least concern (LC), near threatened (NT), vulnerable (VU), endangered (EN), critically endangered (CR), extinct in the wild (EW), and extinct (EX). The last two categories were not recorded in this study. The most threatening categories for nonextinct species include VU, EN, and CR.
The EVS was initially proposed to assess the conservation status of amphibians and reptiles in Mesoamerica and later fitted to the Mexican herpetofauna. It consists of a series of ecological attributes, geographic distribution, reproductive biology (amphibians only), and human persecution level (reptiles only) to determine any of three risk categories: low, medium, and high (see Johnson et al., 2015;Wilson et al., 2013aWilson et al., , 2013b, with subsequent updates (García-Padilla et al., 2020).
For the three classificatory systems, we determined the status at the species level. The NOM059 is the only system that considers some subspecies under distinct threat categories. Only Aspidoscelis hyperythrus schmidti differed with respect to the nominal species, so it was recognized in the analysis. We excluded all invasive alien herpetofaunal species for the percentage estimation.

| Islands threats
We determined the presence of the 11 major classes of most significant threats to biodiversity according to the IUCN Red List, based on Leclerc et al. (2018). The definition is presented in Table S3. We determined threats on islands, rather than species, because some herpetofaunal taxa have not been assessed by IUCN, and insular populations could be exposed to different selective pressures (natural and anthropogenic) than continental populations. Furthermore, determining the presence of human activities and invasive alien species on the islands provides a risk assessment for any biological group present on these islands. Of the 11 threats, we considered climate change and geological events as present on all islands. Climate change can generate adverse effects on a global scale (e.g., generating drought, increasing the magnitude of climatological phenomena, or the complete sinking of islands), which could threaten the permanence of amphibians and reptiles on islands (Bellard et al., 2014;Winter et al., 2016). In the case of geological events, Mexico is located in the Ring of Fire, an area of high seismic and volcanic activity (García Acosta, 2004), which may increase the probability of catastrophic geological events on islands. For wildlife exploitation, we assume that any island with endemic insular species (in our case, all were reptiles) is subject to this threat, based on the fact that reptiles are the most trafficked legally and illegally worldwide (D'Cruze & Macdonald, 2016). Because these first three threats can obscure other human activities that endanger the insular herpetofauna or could be speculative (as in wildlife exploitation), we model two scenarios. Scenario 1 includes the climate change, geological events, and wildlife exploitation active. Therefore, in Scenario 2 we do not consider these threats.
For the remaining threats, we conducted a search for human settlements and activities (e.g., cultivation, permanent human population, tourism, seasonal fishing, mining/natural resources extraction), as well as the presence of invasive alien species. For human settlements and activities, we obtained information from the Atlas of the Inhabited Insular Territory of Mexico (INEGI, 1994). We used the invasive species diagnosis in protected natural areas of Mexico by the National Commission for the Knowledge and Use of Biodiversity Although some groups of invasive alien species (e.g., vertebrates as rats or cats) may pose more significant threats than others (e.g., plants), we consider any introduced plant or animal as invasive, because their presence could alter the island ecosystem dynamics  Table S4.

Island threats between biogeographic regions.
To determine differences in the threats between Nearctic and Neotropical regions, we used a Mann-Whitney U test, using the regions as category and the threat numbers recorded by island as dependent variable. We selected this test because normality and homoskedasticity failed even when different transformations were used. Also, because the number of islands for some provinces is low (VP: eight islands; PC: seven islands), it was not possible to determine differences between provinces using other statistic methods. The analysis was carried out with the R program version 4.0.1 (R Core Team, 2020), considering an alpha value of 0.05.

| Taxonomic turnover
The taxonomic turnover between biogeographic provinces was assessed with the Simpson dissimilarity index (βsim). We used this index because it allowed us to identify the role of the unshared biota size components in β diversity analyses. Otherwise, it would enable estimating the net taxonomic turnover between biotas (Baselga, 2010;Baselga & Orme, 2012). Analyses were performed with R version 4.0.0 (R Development Core Team, 2020) with the "betapart" package (Baselga & Orme, 2012). We emphasize that results from this analysis may be an underestimate because, for some islands, there are no systematic studies to determine taxonomic diversity.

| Area and continental distance versus species richness
Several mathematical models have been used to infer the relationship between area and species richness (Triantis et al., 2012). We used the power model since it is the simplest and widely utilized (Triantis et al., 2012). We performed a natural logarithm (Ln) transformation of island area, continental distance, and species richness.
This transformation enabled us to estimate the model parameters with linear regression analysis (Preston, 1962). Island area and mainland distance (the minimum distance from the closest end of the island to the mainland) were the independent variables, and species richness was the dependent variable. The analysis was carried out with the R program version 4.0.1 (R Core Team, 2020), considering an alpha value of 0.05.

| RE SULTS
We found that the insular herpetofauna was composed of 222 species with 101 subspecies among 131 islands. The number of subspecies ranged from one to six, with 263 total taxa recognized (Table 1; Table S3). Species richness was markedly unequal between amphibians and reptiles. For amphibians, we recorded only 18 species, belonging to two orders with two salamanders and 16 frogs on TA B L E 1 Taxonomic summary of the Mexican insular herpetofauna by biogeographic province and region

| Endemicity and conservation status
Of the 17 native amphibian species, only two (11.76%) are endemic to Mexico, and no species was an island endemic. For reptiles, 53.75% to 12 islands), and 71.93% inhabited only one island (41 taxa). The average island area was 76.53 ± 32.28 km 2 (range 0.004 to 1,198.75 km 2 ), and the mean mainland distance was 13.60 ± 1.55 km (range 1.33 to 51.00 km).
For the Neotropical provinces, the endemic insular taxa were grouped in PLP (four species and five subspecies) and YPP (one species), inhabiting 11 islands with one to three taxa per island (1.83 ± 0.26) (Figure 1) and P: 4%), and 44% were not listed (88 species; Figure 2). For the IUCN Red List, the majority of species for both groups were recorded as LC (amphibians: 94.12%; and reptiles: 70.85%), followed by DD (5.88%, one amphibian species) and NE (12.56%, 25 reptile species; Figure 2). We did not record amphibian species within some high-risk categories (VU, EN, or CR). The high-risk categories for reptiles represented 11.56% (23 species), and DD (3 species) represented 5.03% (Figure 2). The EVS showed the lowest unlisted values among the three systems for amphibians (all assigned to some category) and reptiles (7.00%, 14 species; Figure 2). For amphibians, most species (76.47%, 13 species) were in the low vulnerability category. For reptiles, most species were grouped in the high category (38.50%, 77 species), followed by medium (31.50%, 63 species) and low categories (23.00%, 46 species) ( Figure 2). When we explored the conservation status only for insular endemic species, the trends were similar. The NOM059 and EVS categorized most island species in the highest risk categories (Figure 2). On the IUCN Red List, 48% of the insular species were classified as LC, 30% in the three highrisk categories, and the remaining 22% as DD and NE (Figure 2).

| Island threats
With the first scenario, all islands are threatened by at least two processes, climate change (100%) and geological events (100%). The third greatest threat recorded for all the islands together was biological invasions (40.46%; Figure 3a). In this scenario, wildlife exploitation

| Taxonomic turnover
The average regional taxonomic turnover (β sim ) value was 0.89 ± 0.16 (mean ± SD). The BCP and PLP showed the highest number of unique species and subspecies (58.45%; Table 2). The lowest dissimilarity

| Area and continent distance versus species richness
According to theory, we found a significantly positive relationship between island area and species richness (F 1,128 = 88.69, p <.0001).

F I G U R E 3
The cumulative percentage of the nine threats associated with human activities (excluding climate change and geological events) by province, region, and total islands. If the bar is larger, the percentage of threats is higher. The model explained 40.28% of the variance (Figure 5a). Contrary to what we expected, we found a significant increase in species number with the mainland distance (F 1,128 = 5.61, p =.019), although the percentage of explained variance was lower (3.43%; Figure 5b).

| Composition and conservation status
The Mexican herpetofauna is the most diverse in the Mesoamerican region, which is the region that occupies the second place for biodiversity hot spots in the world (Johnson et al., 2017;Myers et al., 2000).
Our results suggest that the islands of Mexico harbor 17.18% of amphibian and reptile species recorded for the country (1,292 species; Johnson et al., 2017), with this percentage contained in a small and restricted area (0.0002% of Mexico territory). To commensurate diversity, it must be considered in comparative terms with other highly diverse areas, such as the rainforest of the Biosphere Reserve Los Tuxtlas and the Lacandona region, which harbor 15% and 9.67%, respectively, of Mexican herpetofauna (Hernández-Ordóñez et al., 2015;López-Luna, 2017). Considering that the sampled insular area represents 65.14% of the entire Mexican insular territory, with 3,079 islands (considering only true islands) left to study, the species richness on islands is likely to increase. Continuing exploration and documentation of taxonomic and functional diversity on islands is a promising avenue of research, which is necessary given the current crisis of biodiversity loss.
The conservation status among the three systems yields conflicting recommendations, especially in reptiles. First, the percentage of unlisted species in the NOM059 (82.35% and 44.00% for amphibians and reptiles, respectively) suggests that these taxa are not at risk of extinction or assessments have not been performed. Both claims are not mutually exclusive, although the second option is likely the most plausible. Currently, the NOM059 is the only legal instrument that establishes the conservation status category for wildlife use, management, and exploitation in Mexico (SEMARNAT, 2010). The species inclusion into some category requires exhaustive reviews by experts from each group, under a lengthy bureaucratic process given its impact in the Mexican legal framework, which could delay species or subspecies incorporation, and therefore its legal protection.
Second, the IUCN Red List categorizes most species as least concern (amphibians: 94.12%; and reptiles: 70.85%), even for endemic island species (48%). This could be interpreted as that the majority of species inhabiting islands are not at risk of extinction, despite that 60% of islands register more than three significant threats linked to vertebrate extinctions (Leclerc et al., 2018). It is likely that the high percentage of species categorized as least concern is due to several species having broadly distributed continental populations. The IUCN Red List also does not consider subspecies (although only one subspecies is considered in the NOM059 system) or island populations, which potentially masks the unique threats facing insular herpetofauna. It is also likely that many species harbor distinct insular lineages and evolutionarily significant units (ESUs). Thus, phylogenetic and systematic studies will be vital to appropriately characterize the uniqueness of island populations. We propose to prioritize those species not listed, or at least the 22 island endemic species missing in the NOM059 and the 25 species (seven island endemics) listed as NE by the Red List. As stated above, we also need to establish a distinction between island and mainland populations, which could trigger specific conservation programs for island populations or species. The EVS system groups the insular reptiles into a more top risk category (if the medium and high categories are considered).
However, its application is restricted to the herpetological community, and it suffers from the same limitation of not recognizing island populations. It is important to reiterate that in this study, we determine the main threats on islands and not on species. The extinction risk assessments used in classification systems are focused on species, and therefore some elements (e.g., ecology, population size) need to be considered in future research of island populations and increase the risk of extinction (Slavenko et al., 2016).
The information bias on taxonomy, ecology, and distribution of amphibians and reptiles is also recognized (Meiri & Chapple, 2016;da Silva et al., 2020;Winter et al., 2016), and in developing countries such as Mexico, it can generate an imprecise conservation status evaluation (Koleff et al., 2009). Our results are limited to records on islands that do not result from long-term studies to determine patterns of diversity; rather, information is derived from historical collections and fortuitous encounters. Our example focuses on Mexico, which is located in one of the most diverse areas in the world. It is possible that the Central and South American countries show similar patterns of information bias for their island biodiversity, as well as possible conflicts between status conservation systems and their specific laws.

| Threats to the insular herpetofauna
An interesting aspect of our regional approach is that it allows us to identify differential threat scenarios between the Nearctic and Neotropical regions. The geographic location and the species F I G U R E 5 The positive linear relationship between the natural logarithm (Ln) of the A) island area and B) mainland distance (MD) against species richness (SR) composition could be associated with these differences. The Nearctic region is drier and colder than the Neotropical region; thus, cultivation is limited, given the extremely low rainfall regimes (Grismer, 2002). This implies that human activities and settlements in Nearctic islands are less frequent and mainly associated with lighthouses, fishing camps, Mexican armed forces, and scientific research stations (INEGI, 1994;Samaniego-Herrera et al., 2007) We also wish to highlight that the interaction between threats (e.g., with climate change, pollution) can generate scenarios of considerable adversity for conservation (Leclerc et al., 2018), which requires further investigation.
Overall, biological invasions, human intrusions and disturbance, and habitat modifications are the main threats for the insular herpetofauna. Biological invasions are the only threat associated with species extinctions or population declines that have also been identified on a global scale (Leclerc et al., 2018), and this threat has been suggested as a major cause of insular vertebrate extinctions (Donlan & Wilcox, 2008). We analyze the presence of the threats on islands among all the species, rather than in high-risk species in isolation (Leclerc et al., 2018;Spatz et al., 2017), which may explain the differences in the relative importance of the threats. Also, differences in scale and methodological approaches lead to differences in the results. Similar scale approximations can provide information for the development of regional-or country-specific conservation strategies. For highly diverse nations, usually developing countries, it is essential for the future conservation of island herpetofauna and biodiversity in general.  Wilson, 1967); although other factors, such as island age, may be important (see Emerson & Oromi, 2005;Gillespie et al., 2008;Losos & Ricklefs, 2009). In contrast, the positive relationship between the distance to the continent and the number of species is not consistent with what is theoretically expected. Although the percentage of variance explained in our model is low, the significant relationship may be due to islands farthest from the mainland being larger.

| Biogeographic patterns
However, we did not find a significant relationship between area and distance from the continent (data not shown), at least with our database. Also, this trend could suggest that the remote islands are better studied. An increase in sampling effort and systematic studies on several islands are required to improve diversity estimates.
Recent phylogenetic studies on Mexican islands led to the description of two insular endemic rattlesnakes (Meik et al., 2018), the elevation to species of synonymous insular snake populations (Cox et al., 2018), and the proposal to elevate three subspecies of island  Table 3, we present the ten islands with the greatest recorded threats, which can guide specific conservation efforts. Biological invasions, wildlife exploitation, and habitat modifications are the main threats to the Mexican insular herpetofauna, with Nearctic and Neotropical islands facing different situations, ultimately endangering the flora and fauna that inhabits them. The islands have and will continue to be a fundamental study model in biology, and, in the face of the biodiversity crisis, will play a leading role in the development of restoration and conservation strategies.

ACK N OWLED G M ENTS
We thank Ruben Castro, Aquiles Argote, Guadalupe Bustos, Samuel Arechaga, and Yirdael Muñiz for their invaluable comments during the initial stages of this project. Special thanks to Marisol Gómez for their invaluable help in the elaboration of the figures. We thank André S. Fuentes for providing reptiles' information from the Roqueta Island. We appreciate the invaluable comments of two anonymous reviewers who significantly improved the quality of this work.

CO N FLI C T O F I NTE R E S T
There are no sources, relationship, financial, or any potential sources of conflict of interest that might be perceived as influencing the objectivity of this work.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available as appendices.