The herpetofauna of Sonora, Mexico, with comparisons to adjoining states

Situated in the topographically complex transition between the Neotropics and the temperate biomes of North America, the state of Sonora, Mexico, has an extraordinarily diverse herpetofauna. Surprisingly little research has been conducted on the state’s amphibians and reptiles and many systematic and biogeographic questions remain unanswered. To facilitate future research, we provide a checklist of Sonora’s herpetofauna, documenting species presence based on museum specimens, our fieldwork, and published research. Sonora’s herpetofauna is placed in a regional biogeographic perspective via a checklist for the six adjoining states together with faunal analyses. A total of 402 species of amphibians and reptiles are recorded from these seven states. Sonora has the greatest species richness (187 species), followed by Chihuahua (169 species), and Sinaloa (146 species). Sonora's herpetofauna is most similar to that of Chihuahua, with which it shares a long border. Eleven biogeographic affinity-based faunal groups are recognized. Of these, three are dominant in Sonora: a core group classified as "Sonoran" demonstrates strong affinity to Sonoran Desertscrub and Sinaloan Thornscrub communities; a Tropical group - with many species reaching their northern distributional limits in the state; and a Madrean group consisting largely of montane species. Our state-level faunal analysis provides some evidence of peninsular depauperization of the herpetofauna on the Baja California peninsula due in part to the small number of Neotropical species present in Baja California Sur. Our faunal analysis points toward distinctive mainland and peninsular Sonoran Desert herpetofaunas centered on Sonora and the Baja California Peninsula, respectively, each with about 50 non-insular species, and each with species-level endemism nearing 50%.


Introduction
Sonora is the second-largest state in Mexico with a mainland territory spanning an area of 185,430 km 2 (Felger et al. 2001) between latitudes 26°16'49" N and 31°19'8" N ( Figure 1; Table 1). Chihuahua (Mexico's largest state), the continental divide, and the Sierra Madre Occidental are situated near the eastern border. The Gulf of California and Sonora's 14 islands (Appendix 1) lie westward and comprise the state's western terminus.
Arizona and New Mexico border on the north, while Sinaloa lies to the south. The Tropic of Cancer crosses the 108th meridian ca. 318 km south of the Sonoran border near the eastern tip of Baja California Sur. Van Denburgh (1922) and Slevin (1928) provided the first published lists of Sonora's herpetofauna. Legendary herpetologist Edward H. Taylor (1938) was the first to summarize the amphibians and reptiles. The best-known treatment of the herpetology of Sonora is that of Charles Bogert and James Oliver (1945). The most recent summary of Sonora's herpetofauna comes from Rorabaugh (2008). The present paper differs from Rorabaugh in that it (1) is regional in scope covering Sonora and all adjoining states; (2) uses the revised nomenclature published in Crother (2008) and Liner and Casas-Andreu (2008); (3) is inclusive of marine species and insular endemics; and (4) presents detailed discussion of species richness, faunal similarity, endemism and distributional limits among the bordering states in Mexico and U.S.

Materials and Methods
Scientific and standard English names used in this publication are based on the taxonomic lists published by Crother (2008) and Liner and Casas-Andreu (2008). Where these two checklists differed (e.g., in the recognition of Syrrhophus, Sceloporus vandenburgianus, and Holbrookia approximans) we followed the arrangement in the former. Taxa not included in the above lists, but added here are Sceloporus albiventris (Smith 1939), Lampropeltis webbi (Bryson et al. 2005), and Tropidodipsas repleta (Smith et al. 2005), as are species recognized in a recent revision of the Trimorphodon biscutatus complex (Devitt et al. 2008) and the Xantusia vigilis complex (Bezy et al. 2008), both appearing subsequent to the above two checklists.
It has been nearly four decades since the publication of the seminal work on the amphibians and reptiles of Sinaloa by Hardy and McDiarmid (1969), and we made a concerted effort to update the herpetofaunal list for this state, including a complete search of distributional records in Herpetological Review and an examination of all relevant species accounts in the Catalogue of American Amphibians and Reptiles.
We obtained data for Sonoran specimens of amphibians and reptiles in the following collections from curatorial staff (see acknowledgements) or via the HerpNET data portal (http://www.herpnet.org) on 13 May 2006 -occurrence of the center and majority of its distribution within a particular biotic region. Biotic communities follow Brown et al. (2007), except in the Sonoran and Chihuahuan regions where we include adjoining or contained semidesert grasslands, chaparral, and thornscrub. We note that although Sinaloan Thornscrub (86 % of which occurs in Sonora) is in some respects transitional between Sonoran Desertscrub and Sinaloan Dry Deciduous Forest, most of the characteristically Sonoran species occur in it. To avoid complicating the analysis, therefore, we have retained such species as "Sonoran". Rosen (2007) found little evidence for a Mohave Desert herpetofauna distinct from that of the Lower Colorado River Valley subdivision of the Sonoran Desert, and for simplicity, we have included species with Mohavean affinities as "Sonoran" (McLaughlin 1992). Species that do not demonstrate a particular biotic region association and occur over a wide distributional range are labeled "generalist." Species that occur widely in two or more desert regions (particularly in both the Chihuahuan and Sonoran deserts) are classified as "North American Desert generalist".
Species that enter the region from the east, with the great bulk of their ranges in the eastern U.S., are classified as "eastern temperate" without further specification. Although these affinity definitions and assignments are not quantitative, we found them to be a useful tool for characterizing the herpetofauna. To minimize such problems, some species that could have been designated as North American Desert generalist (such as Sonora semiannulata) or Madrean (such as Hyla arenicolor), which utilize divergent environments (such as woodland and forest, and semi-desert grassland and desertscrub, in these two examples) were entered in the "Generalist" category (Table 2).
We used Jaccard's cluster analysis to measure biotic relationships per state and faunal similarities between Sonora and bordering states. Clustering procedures are based on the nearestneighbor criteria. Principal Coordinate Analysis (PCO) of biotic affinity per state (as defined above) was performed using the Gower General Similarity Coefficient with data expressed as number of species per biotic affinity. All the analyses were carried out using the program MVSP Version 3.1 (Kovach 1999). ----------------

Results and Discussion
Species richness A total of 402 native species of amphibians and reptiles arrayed in 36 families are found in the region (Tables 2, 3; Appendix 2); six additional species are present but exotic to the area. Overall native herpetofaunal diversity (Table 3) is highest in Sonora (187 species), Chihuahua (169 species), and Sinaloa (146 species). Sonora has the highest number of species of turtles (15), lizards (64), and snakes (72), and the second highest number of frogs (32 in Sonora, 35 in Sinaloa) and salamanders (3 in Sonora, 4 in Chihuahua).
Sinaloa lacks a recent treatment of the herpetofauna, and we suspect it supports greater diversity than we indicate; however, it is likely less than that of Sonora and Chihuahua due to its lower physiographic diversity and smaller area (Table 1). Recent analyses that were restricted to Mexico's Pacific lowlands and interior valleys found that the tropical dry forests of Sinaloa had higher diversity than those of Sonora, consistent with a general and expected decline in species richness from south to north (Garcia 2006, Garcia et al. 2007).
Total species richness continues to diminish to the north, away from the high diversity of the tropics (Table 3; Figure 2), dropping to 135 and 128 species in the well-studied faunas of Arizona There is a correlation between number of species and state area ( Figure 2), but species richness in both Baja California and Baja California Sur are substantially lower than expected on the basis of area. This may be explained by peninsular species depauperization, which consists of a decrease in species diversity from the base to the tip due the spatial effects of peninsulas on immigration and extinction (Taylor and Regal 1978;Rosenzweig 1995). Application of this model to the Baja California peninsula has been contentious (Taylor and Regal 1978, Seib 1980, Busak and Hedges 1984, Wiggins 1999. Such disagreement is not surprising considering the peninsula's complex geological history (involving rifting from the mainland and hypothesized trans-peninsular seaways; summaries in Grismer 1994, Riddle and Honeycutt 1990, Riddle et al. 2000, Murphy and Aguirre Leon 2002, Riddle and Hafner 2006, Devitt et al. 2008, Mulcahy 2009) and its ecological and topographic diversity (e.g., the Sierra San Pedro Martir and Mediterranean climate in the north and San Lucan Dry Deciduous Forest and Sierra La Laguna on the cape). Table 3. Total number of amphibian and reptile species in each state arranged according to taxonomic order. Caudata  1  3  0  4  3  1  3  Anura  24  11  3  30  22  35  32  Crocodilia  0  0  0  0  0  1  1  Testudinata  6  6  7  13  10  12  15  Sauria  52  49  49  51  44  35  64  Serpentes  52  39  32  71  49  62  72  TOTAL  135 108  91 169  128 146  187   Table 4. Insular and non-insular endemic amphibians and reptiles by state. Insular  0  15  22  0  0  0  11  48  Non-insular  4  4  14  1  3  1  5  29  TOTAL  4  19  34  1  3  1 15 77 Figure 2. Logarithmic regression analysis of species/area relationship. Solid diamonds and continuous regression line represent all species of amphibians and reptiles; hollow dots and dotted regression line represent squamate species numbers. Species numbers do not include island endemics.

----------------
However, when considering only squamate reptiles (as in Seib 1980), we found evidence supporting a peninsular depauperization effect ( Figure 2). This effect is almost entirely accounted for by the scarcity of tropical species (11 mostly endemic vicariant species in Baja California Sur), compared to 28 Californian-affiliated species in Baja California and 39 and 49 tropical species in Chihuahua and Sonora, respectively. The near absence of Neotropical species on the Baja California peninsula could reflect decreased migration from the mainland due to the arid conditions at the base of the peninsula and the large over-water dispersal distance. It may also reflect the peninsula's modest area of tropical vegetation and species extinction since separation from the mainland. These latter two factors, rather than the distance from source populations, most likely account for the apparent effect seen in Figure 2. However, considering the relatively low number of Sonoran-affiliated species (37 species in 4.4 million ha of Sonoran desertscrub for Baja California; 31 species in about 5.5 million ha of Sonoran desertscrub in Baja California Sur), and that Baja California Sur is at a lower latitude and would be expected to have higher species diversity for that reason, we cannot exclude the possibility of some form of peninsula depauperization. This is surprising in view of the extensive evolution of vicariant endemic species in Baja California Sur, and suggests further analysis at varied scales is required.

Faunal similarities
Similarity indices based on species occurrences indicate Sonora is closest to its adjoining continental states, particularly to Chihuahua. However, it should be noted that the phenograms (Figures 3-7) are visual representations of descriptive biogeography and do not necessarily reflect historic biogeographical processess that are better inferred using reconstruction of phylogenies for species endemic to the sub-regions of the study area. Limiting our similarity analysis to species occurrence data, rather simply reflects geographic proximity. This is clearly shown by the two minimum spanning trees (Figures 5,6), in which the relationships of the states closely mirror their geographical positions. Our analyses for species biogeographic affinities (Figures 4,7) group Sonora with Sinaloa, Arizona with New Mexico, and Arizona-New Mexico with Chihuahua. These arrangements seem closer to underlying historical and ecological processes detailed below (see discussion under 'Biotic affinities'). Faunal similarity based on species presence-absence suggests Sonora is closest to Chihuahua for (a) total herpetofauna, (b) amphibians, and (c) snakes (Figures 3,5). This is likely due to the long political border (ca. 800 km) shared between these two states. The Sierra Madre Occidental is near this border and many Madrean species occur in both states. In addition, many widespread North American Desert species are shared by both states, and several Chihuahuan Desert species reach northeastern Sonora (e.g., Anaxyrus debilis, Phrynosoma cornutum, P. modestum, Sceloporus cowlesi, Aspidoscelis exsanguis, A. uniparens, Gyalopion canum). Together these influences result in high faunal similarity values between Sonora and Chihuahua.
For turtles, with a very small species pool, Sonora clusters with Sinaloa and together these two ally with Baja California and Baja California Sur ( Figure 3). Because the species pool is so small here, faunal similarity values can be explained by the distribution of a few species. These include tropical-affiliated species linking Sonora and Sinaloa: Kinosternon alamosae, K. integrum, Rhinoclemmys pulcherrima and Terrapene nelsoni. The anomalous similarity of Sonora to Baja California Sur is due to the presence of the shared Gopherus agassizii and Trachemys nebulosus, both of which might represent introductions to the southern peninsula (Grismer 2002), although, alternatively, turtle distributions may preserve evidence of the ancient attachment of the peninsula to the west coast of Mexico.      Our results (Figures 3-7) consistently indicate that the two states of peninsular Baja California group together and are distinctive from all the other states in the region. This pattern is consistent with regional biogeographical effects demonstrated by Riddle et al. (2000). It is not surprising considering the peninsula's high degree of isolation from the rest of the study area. Moreover, the entire peninsula is highly depauperate in anuran amphibians, its ecology being strongly dominated by arid environments supporting a notable array of unique reptilian endemics. Additionally, the boundary between the states crosses the ranges of many peninsular species. These include wide-ranging peninsular species as well as some formerly viewed as cape endemics, such as Bipes biporus, which range into the southern edge of the state of Baja California, and some former Baja California endemics, such Aspidoscelis labialis, are now known to reach Baja California Sur (summary in Grismer 2002).
Regardless, the herpetofauna of Baja California and Baja California Sur are quite distinctive as is reflected in the Jaccard's coefficient dendrograms (Figures 3, 4). Undoubtedly, the California affiliated faunal elements in Baja California, and the tropical affiliated species in the San Lucan Thornscrub and San Lucan Dry Deciduous Forest in Baja California Sur contribute directly to the distinctiveness of these two closely allied states.

Biotic affinity
Our interpretation of biogeographic affinity is largely based on a qualitative-discrete assignation per our regional experience and known species distributions within the floristic provinces defined in Brown et al. (2007). Of course, some species showed several biogeographic affinities that placed them in the category of generalists. The generalist assignation could be interpreted to a greater degree if we were to take into consideration the complete distribution of each species and its phylogenetic relationships.
Thus, the lack of distributional, phylogenetic, and geographical studies in the region further obscures biogeographic placement of certain species. Future studies accounting for historical-phylogenetic relationships may help refine biogeographic affinity and may also clarify the relationships we suggest here.
Baja California and Baja California Sur (excluding insular species) support the highest proportion of species with Sonoran affinities (each with 34 %; Figure 1;  (47) in both the proportion and number of Sonoranaffinity species (Figure 1; Table 2). Despite the unique and distinctive elements in the species assemblage of the Baja Californian desert herpetofauna, this most likely represents nothing more than the strong area-dominance of Sonoran Desert in both peninsular states.
The similarity phenograms based on biotic affinity ( Figure 4) produced strikingly different results from those based strictly on shared species ( Figure 3). Analysis of biotic affinity produced a close association of Sonora and Sinaloa, whereas Arizona and New Mexico were closely associated and linked to Chihuahua. We believe these examples, based on our assignment of species biogeographic affinities, come closer to reflecting evolutionary and biogeographic processes underlying the structure of this regional assemblage of herpetofaunas, and suggest phylogenetic reconstructions to test this hypothesis.
Although Arizona shares many Sonoran species with the state of Sonora, and New Mexico has grassland and Chihuahuan affinities that connect it with Chihuahua, these two U. S. states share much with each other. They share high proportions of wide-ranging (Generalist) species, North American Desert Generalists, Great Basin, and Great Plains species, and they have relatively low proportions of Tropical and Madrean species. Chihuahua, like Sonora, has many Madrean and Tropical species, but is dominated by Chihuahuan Desert (Figure 1; Table 2). The Chihuahuan Biome shares species with the Great Plains (e.g., Lithobates blairi, Thamnophis radix, Tropidoclonion lineatum), and this tends to ally Chihuahua with the temperate zone and, thus, with Arizona and New Mexico (Figure 4).
Chihuahua and Durango, rather than Sonora, include most of the higher elevations of the Sierra Madre Occidental, producing an additional Chihuahua-to-New Mexico connection (e.g., Opheodrys vernalis and Thamnophis sirtalis) and to the more temperate U.S. states ( Figure 3) via the shared biota of the U.S. Rocky Mountains and the highest elevations of the Sierra Madre. Although Sinaloa has a greater representation of tropical species than Sonora (54 % versus 26 %), these are nonetheless the two most tropical states in our study area, and they are closely similar in all other affinity categories ( Figure 1; Table 2). Jaccard's dendrogram for affinity ( Figure 4) and PCO using Gower's coefficient (Figure 7) both indicate a close association between Sinaloa and Sonora. This result is not surprising considering the principally subtropical evolutionary derivation of Sonoran Desertscrub and Sinaloan Thornscrub biota.
There is general paleobotanical agreement (Brown 1994) that Madrean Woodland, Sinaloan Dry Deciduous Forest, Sinaloan Thornscrub, Sonoran Desertscrub, and Mohave Desertscrub all share an ancient origin in the Madro-Tertiary Geoflora, a semi-arid formation that arose between Mesophytic (more tropical) and Arcto-Tertiary (more boreal) geofloras. There is also fairly general agreement that Sonoran Desertscrub is derived from Sinaloan Thornscrub, which in turn, is derived from Sinaloan Dry Deciduous Forest. Chihuahua and Sonora share an ancient Madrean heritage, but are in other important respects herpetofaunally and biogeographically divergent. We suspect that the derivation of species and gene clades of amphibians and reptiles may have followed along similar lines in this region, and that phylogenetic and phylogeographic analyses would yield tree topologies more similar to those of the Jaccard dendrograms for affinity ( Figure 4) than comparisons based on shared species ( Figure  3).

Limits of geographic distribution
The global latitudinal range of 66 species terminates in Sonora ( Figure 8, Appendices 3, 4). The map illustrates that the distributional limits of species centered in the tropics (1) are concentrated in areas with greatest accessibility to herpetologists (Álamos and Yécora regions) and (2) diverge from the coast at higher latitudes. Species of tropical affinity reaching their northern limits in Sonora do so largely in moderate elevation zones of the Sierra Madre Occidental Archipelago. At higher latitudes, these species generally do not descend into warm, lower elevations (personal observations), presumably due to moisture requirements.
Eight species (7 %) of the Sonoran Desertassociated herpetofauna reach their southern range limits in Sonora, mainly near the coast (Figure 8), while the largest number (50 species, 27 %) do so in Sinaloa (Hardy and McDiarmid 1969), primarily in a progressively narrowing band of subtropical Sonoran Desertscrub and Sinaloan Thornscrub. Another large component of the herpetofaunal diversity is associated with the Sierra Madre Occidental (31 species, 17 %) and continues southward though Chihuahua and Durango (Webb 1984;McCranie and Wilson 1987;Lemos Espinal and Smith 2007). Three notable exceptions are Aspidoscelis sonorae, Lithobates yavapaiensis, and Kinosternon arizonense, which reach their southern limits on the western slopes of the Sierra Madre Occidental Archipelago in Sonora.
The transitions from Sonoran Desertscrub to Sinaloan Thornscrub, and thence to Sinaloan Dry Deciduous Forest, are suggested by these patterns in the latitudinal range limits of amphibians and reptiles. A more refined analysis of these transitions would be of great interest for understanding the ecological structure of the regional biota, its evolution, and its conservation.

Endemism
Despite the species richness of Sonora, it does not have an unusually high number of state endemics (Table 4). There are only five mainland endemics (four lizards: Aspidoscelis opatae, Crotaphytus dickersonae, Phrynosoma ditmarsi, Xantusia jaycolei and one turtle: Trachemys yaquia), similar to the figures for Arizona (4), Baja California (4), New Mexico (3), Chihuahua (1), and Sinaloa (1), but far less than that of Baja California Sur (14). The high number of endemics in Baja California Sur (Table 4) is a reflection primarily of the unique tropical and semi-tropical herpetofauna of the Cape region, the existence of the isolated Sierra La Laguna, and the historical separation of the cape region from the rest of the peninsula (Murphy and Aguirre- Leon 2002, Riddle et al. 2000, Grismer 1994. It is worth noting, however, that this state-by-state comparison may obscure the rich endemism comprising the region's core Sonoran Desert herpetofauna, which is centered on Sonora but importantly includes Arizona, Sinaloa, and the Baja California peninsula. The Sonora herpetofauna includes 11 island endemics (9 lizards: Aspidoscelis bacata, Aspidoscelis estebanensis, Aspidoscelis martyris, Ctenosaura conspicuosa, Ctenosaura nolascensis, Sauromalus hispidus, Sauromalus varius, Uta nolascensis, and Uta palmeri; and 2 snakes: Coluber slevini, Crotalus estebanensis), notably fewer than Baja California (15) and Baja California Sur (22). This most likely reflects the greater number of islands included in Baja California and Baja California Sur compared to Sonora. Overall, it is clear that the greatest numbers of endemics in the region covered by this checklist involve the islands of the Gulf of California (Table 4). In addition to numerous recently derived species found on "land bridge" islands, these endemics include some remarkable ancient relicts: Aspidoscelis ceralbensis, Sceloporus (Sator) angustus, and Sceloporus (Sator) grandaevus.
The question thus arises whether the unique history of the Baja California peninsula, including its rifting from the mainland and its hypothesized trans-peninsular seaways has led to enhanced species diversity compared to the mainland. Here, we restrict this comparison to species we classify as having Sonoran and/or North American Desert Generalist affinities and exclude those classified as having Madrean, Tropical, Californian, and other non-desert affinities. In the count for Sonoran Desert endemics on the peninsula (Baja California + Baja California Sur) we include species that extend into southernmost California (south of San Gorgonio Pass), and for the mainland count (Arizona + Sonora + Sinaloa), we include species that enter extreme southwest New Mexico or extreme western Chihuahua. As thus defined, the Sonoran Desert peninsular herpetofauna consists of 52 total species of which 26 are endemic, whereas the mainland Sonoran Desert herpetofauna consists of 54 species of which 26 are endemic. These data indicate that both endemism and overall species number in the Sonoran Desert herpetofauna of the Baja California peninsula is roughly comparable to that of the mainland (Arizona + Sonora + Sinaloa).
With about 9.9 million ha of Sonoran Desertscrub on the peninsula and about 17.9 million ha on the mainland, this result confirms both the uniqueness of the peninsular herpetofauna within the Sonoran Desert context and the richness apparently connected to its strong vicariant history. It again suggests that the Gulf of California and the Sonoran Desert -isolated from the mainland's tropical influence -form dispersal barriers for Neotropical species and are key factors driving a possible peninsular depauperization effect.

Conclusions
Based on our similarity and affinity analyses, we consider the Sonoran assemblage (species occurring principally in Sinaloan Thornscrub and Sonoran Desertscrub) to represent the unique and regionally endemic core of Sonora's herpetofauna. Madrean, tropical, and northern temperate elements have played a strong role interacting with core faunal elements. This accords well with the early dispersal model of Bogert and Oliver (1945), which hypothesized five dispersal routes that influenced overall species richness in Sonora: from the south, (1) along the narrow coastal plain and (2) along the Sierra Madre Occidental; (3) from the eastern plateau or plains via valleys dissected into the north end of the Sierra Madre; (4) from the north, that is, from the mountains in Arizona; and (5) from the northwest, the Colorado-Mojave Desert. Our data support these historic connections, but are not adequate to distinguish this dispersal concept from vicariant hypotheses involving in situ endemism. Recent taxonomic changes and species discoveries suggest that the core Sonoran assemblage may have provided conditions for in situ adaptive speciation of some species. Determining the historical basis and importance of Sonora in speciation events within the Sonoran Desert and the topographically complex Sierra Madre Occidental and its associated archipelago and isolated valleys will require thorough distributional and phylogenetic studies, which are virtually nonexistent for the region. To assist and perhaps stimulate such research, we offer the following conclusions: 1. Sonora has the highest overall herpetofaunal richness of any state in the region. This is due to its situation in the topographically complex transition between the Neotropical and Nearctic zones, and the presence of the richest desert formation in North America and the highly diverse, semi-temperate Madrean biota with its deep southerly penetration from the north into tropical latitudes.
2. For total herpetofauna, Sonora is most similar to Chihuahua, with which it shares a long portion of the Sierra Madre Occidental as well as many wide-ranging North American desert species.
3. The herpetofauna of Sonora is especially notable as the core and evolutionary center of the unique, subtropical Sonoran Desertscrub-Sinaloan Thornscrub species assemblage. 4. Our data on Sonora's lizard fauna clusters it with Arizona, New Mexico, and Chihuahua, reflecting the large number of shared species with eco-physiological adaptations to the hot, arid climates of the Sonoran and Chihuahuan Deserts. 5. Many Sonoran species with tropical affinities reach their northern latitudinal limits at moderate elevations in the Sierra Madre Occidental region, where they are driven upward in minimum elevation away from desert aridity and downward in maximum elevation by high-latitude cold. Thus, there is a bio-climatically determined habitat wedge for tropical and subtropical species distributions with a north-pointing tip.

Species with Sonoran Desertscrub and Sinaloan
Thornscrub affinity have distributions that constrict south toward the coast. Most of these species range into Sinaloa, albeit usually rather sparingly. Desert-like bio-climatic conditions associated with high temperature and low rainfall at low elevation define this south-pointing habitat wedge in the distribution of the core desertscrubthornscrub species assemblage.
7. At the state-level of analysis, the number of mainland endemics in Sonora is low and comparable to that in Arizona, New Mexico, and Baja California, but far less than Baja California Sur, where high endemism may be due in part to the historical separation of the cape region from the remainder of the peninsula and, to some extent, from mainland Sonoran influences.
8. Sonora has fewer islands and insular endemics than Baja California and Baja California Sur. 9. Peninsular Baja California, with a rich, unique herpetofauna marked by notable endemism is proportionately the most "Sonoran" in herpetofaunal affiliation. This is due to the paucity of temperate and, especially, tropical species resulting in peninsular depauperization, which is apparent at our state level of analysis.

----------------
Ken Tighe (U.S. National Museum of Natural History); Jonathan Campbell, Carl Franklin (University of Texas at Arlington). We owe a special debt of gratitude to George Bradley at the University of Arizona for providing frequent access (often on short notice) to the collection of amphibians and reptiles at the University of Arizona.