Journal of Biogeography (1976) 3, 1-18 Originsand affinitiesof vertebratesof the North American SonoranDesert and the Monte Desert of NorthwesternArgentina W. FRANK BLAIR Desert of northern Chile and southern Peru, which ARTHUR C. HULSE is one of the driest areas on earth, to dry habitats Department of Zoology, University of Texas at within the tropics, such as those found in northern Austin, Austin, Texas 78712, U.S.A. Venezuela. Evolutionary studies of desert organisms on each continent are particularly interesting because MICHAEL A. MARES the continents were isolated from one another from Biology Department, University of Pittsburgh, the Triassic to the late Pliocene, when the formation Pittsburgh, Pennsylvania 15260, U.S.A. of the Central American land bridge united both continents and allowed extensive faunal interchange Abstract (Dietz & Holden, 1970; Patterson & Pascual, 1972). Thus one can use the geological record and palae- The Monte Desert of Argentina and the Sonoran ontological data to determine when particular Desert of North America are very similar in most groups of organisms may have had access to desert aspects of their physical environments. The origins regions and compare their degree of adaptation and and the degree of affinity of the faunas of three classes adaptational strategies to arid environments. of vertebrates in the two deserts are discussed. The Monte Desert of northwestern Argentina and Heterotherms show high endemism, with 72 0 % of the the Sonoran Desert of North America are very Monte species either endemic to the Monte or with similar to one another in climate and geomorphology Monte-Chaco distribution and with 61-4% of the (Morello, 1958; Simpson & Vervoorst, 1976). In Sonoran Desert heterotherms endemic to deserts of addition, these deserts are of about the same age and the southwest. Mammals show low endemism and support plant communities of very similar physiog- their desert inhabitantsgenerallybelong to wide-ranging nomies (Solbrig, 1972, 1976, Figs 1, 2). It is largely species. Mammals alone among these vertebratesshow because of these similarities, particularly the abiotic wide disparity between the two deserts in number of ones, that the Origin and Structure of Ecosystems species, with more than twice as many in the Sonoran Subprogram of the U.S. component of the Inter- Desert as in the Monte. national Biological Program chose the two deserts as Desert adaptation seems to have been occurring study sites for an investigation into patterns and over much of western North America throughout processes of convergent evolution of ecosystems. much of the Cenozoic and in the present Monte- Since convergence requires that floras and faunas be Chaco-Patagonia region since at least the Miocene. phylogenetically distinct, it was necessary to compare The only species and most of the genera shared the origins and affinities of the organisms occurring between the two deserts are wide ranging or volant in both deserts. This report reviews the faunal make- mammals. Only one of seven families of anurans is up, origins, and taxonomic relationships of three common to both deserts. Reptiles have many families classes of terrestrial vertebrates of the disjunct in common, while mammals of the two deserts have deserts. relatively few families in common. Methods Introduction The North American desert system is well studied as North and South America contain a number of arid far as vertebrate distribution patterns are concerned. and semi-arid regions, ranging from the Atacama We have consulted a number of references in 1 1 2 W. Frank Blair, Arthur C. Hulse and Michael A. Mares .a Fig. 1. Hillside in Sonoran Desert near Tucson, Arizona. Columnar cactus is Cereiusgiganteus. Fig. 2. Hillside in Monte Desert near Andalgala, Catamarca Province, Argentina. Columnar cactus is Trichocereiusterschecki. Origins of desert vertebrates 3 compiling faunal lists for the various arid and semi- tional species, would be found in this general area. arid areas of North America (principally Lowe. We therefore feel that the exclusion of this area will 1964; Stebbins, 1951, 1954 for reptiles and amphi- not greatly affect discernible patterns of colonization bians; Hall & Kelson, 1959 for mammals). The or taxonomic relationships of the faunas of the fauna of the Monte Desert and adjacent areas is disjunct deserts. For both the Monte and the much more poorly known, and we have relied Sonoran Desert our discussions concern the lowland heavily on our own field experience in the area tc desert faunas and exclude those of the mountain compile species lists. We have also used various ranges that are found in both of these regions. published references to supplement our observations (Barrio, 1964a, b; 1965a, b; 1968; Cei, 1955a, b, 1959, 1962; Freiberg, 1942; Gallardo, 1966; Peters Categories for vertebrate distribution patterns (see & Orejas-Miranda, 1970; Peters & Donoso-Barros, Figs 3 and 4) 1970; Reig & Cei, 1963 for reptiles and amphibians; Cabrera, 1957, 1960 for mammals). Various divisions of the Sonoran and Monte (1) Sonoran-Endemic: Sonoran biotic province as deserts and adjacent areas were used in describing delimited by Dice (1939), which includes the Mojave vertebrate distribution patterns. We realize that any Desert. such subdivisions of areas which may not in actuality (2) Sonoran-Chihuahuan: Category 1 plus arid and be sharply delineated one from another is artificial to semi-arid regions of New Mexico, Texas, Chihuahua a certain extent. Nevertheless, since a species (Mexico) south of Zacatecas, San Luis Potosi. inhabiting a desert may possess a geographic (3) Sonoran-Chihuahuan-Grasslands: Category 2 distribution extending well beyond the desert's plus the Great Plains Grasslands. boundaries, it is necessary to examine its overall (4) Sonoran-Great Basin: Category 1 plus the Great habitat requirements, and those of its genus (or Basin Desert. family) in order to determine how that particular (5) Sonoran-Great Basin-Chihuahuan: Categories species managed to invade the desert in the first 2 and 4 combined. place. Naturally there is additional error in trans- (6) Sonoran-Thorn Scrub: Category 1 plus adjacent ferring records of occurrence from distribution maps thorn scrub habitat along the western Mexican to habitat maps. Since we are dealing with extensive Lowlands. areas, however (e.g. the Chihuahuan or Sonoran (7) All deserts-Grasslands: Category 5 plus Great deserts), and since we have some knowledge of the Plains Grasslands. ecological requirements of the great majority of (8) Widespread: Ranges well beyond deserts and vertebrates discussed, we feel that errors of this type grasslands. are minimized. We have not included Baja California as part of (1) Monte-Endemic: Morello's (1958) northern, the Sonoran Desert, nor do we discuss the origins of western, and eastern limits of the Monte, south to its fauna. We have deleted this area in order to avoid southern Mendoza and northeastern La Pampa. the confouh4ing effects of evolution on a peninsula (2) Monte-Chaco: western lImits of Monte east to and its offshore islands. No comparable area exists San Luis, northern Cordoba, to Corrientes through in or near the Monte Desert, and since we hope to Santa Fe, Chaco, Formoas and Paraguay northward. determine whether the origins and overall distri- (3) Monte-Pampas-Patagonia: Category 3 minus bution patterns of vertebrates in both deserts are the Chaco. similar, the inclusion of the northern peninsula (4) Monte-Chaco-Pampas-Patagonia: Category 2 would only seem to add noise to our analyses. then south and east through La Pampa; Buenos We have also excluded the vertebrates of the Aires through Neuquen, Rio Negro, Chubut and southern Monte (south of southern Mendoza) from Santa Cruz. the Monte endemic category. This is primarily (5) Monte-Chaco-Cordillera: Category 2 plus the because we are not personally familiar with the slopes of the Andean or pre-Andean chain. patterns of faunal occurrence in the southern portion (6) Monte-Cordillera: Category 1 plus slopes of of the desert, particularly as far as distributional Andean or pre-Andean chain. limits are concerned. Available references indicate, (7) Widespread: ranges well beyond the preceding however, that few endemics, and indeed few addi- areas. 4 W. Frank Blair, Arthur C. Hulse and Michael A. Mares Geographic origins of the desert vertebrates deserts represent an extreme in aridity within much broader regions of deficient but less severely deficient The evolution of desert-adapted species will be moisture. Transitions from these less arid areas to conditioned by a variety of factors, in addition to the the deserts are typically gradual (e.g. from arid Chaco physical characteristics of the desert environment. to Monte) and desert endemics may have distri- Among terrestrial vertebrates, one set of factors is butions that deviate from map lines delineating purely phyletic, with the various classes expectedly actual desert boundaries (Kuchler, 1964; Cabrera, showing very different patterns of adaptation simply 1971). because of their different physiologies and morpho- logies. Another set of factors is biogeographical. What is the nature of source areas where selection Anurans may have produced attributes that are pre-adaptive for the evolution of desert species? Biogeographic Endemic species of anurans are few in both deserts factors of ecological and physical barriers and of (Tables 1 and 2). The Monte has no endemic anuran long-range dispersal assume importance in com- among