MUS. COMP. ZOOL. an r LIBRARY 655 JNIVERSITY OF KANSAS) MAR 1 9 1980 pn te “43 WUSEUM OF NATURAL HISTORY = jarmvaro PUBLICATION LI er mere TY An Ecogeographic Analysis of the Herpetofauna of the Yucatan Peninsula By Julian C. Lee )*| oO UNIVERSITY OF KANSAS LAWRENCE 1980 AEC, J KH UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly published in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publications, Museum of Natural History, began with number 1 in 1946. Longer research papers are pub- lished in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subjected to critical review by intra- and extramural specialists; final acceptance is at the discretion of the Director. Institutional libraries interested in exchanging publications may obtain the Occa- sional Papers and Miscellaneous Publications by addressing the Exchange Librarian, University of Kansas Library, Lawrence, Kansas 66045. Individuals may purchase separate numbers of all series. Prices for all publications of the Museum may be obtained from the Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. THe UNIVERSITY OF KANSAS MusEUM OF NATURAL HIsTory MISCELLANEOUS PUBLICATION No. 67 February 29, 1980 An Ecogeographic Analysis of the Herpetofauna of the Yucatan Peninsula By Juutan C. LEE Museum of Natural History and Department of Systematics and Ecology University of Kansas Lawrence, Kansas 66045 PRESENT ApDRESS: DEPARTMENT OF FioLocy, UNIversIry OF MIAMI, CoraL GABLES, FLORIDA 33124 THE UNIVERSITY OF KANSAS LAWRENCE 1980 UNIVERSITY OF KANSAS PUBLICATIONS, MusEUM OF NATURAL HISTORY Editor: E. O. Wiley Miscellaneous Publication No. 67 pp. 1-48, 22 figures, 7 tables, 27 plates Published February 29, 1980 Museum or NATURAL HIsTORY Tue UNIVERSITY OF KANSAS LAWRENCE, KANSAS 66045 WESeAe PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS CONTENTS iTV IPTV BUH Crd RY CO ees i ee i ta ee ac eRe eee er oe it PACS RN OMEN NL Gia as nee ce Nae SS BES a 8 SI ae goa ae ea gle oe ee it IMT REN WATER ONIN TIEN lisence ee, eet Ee ce acti et at ot Lk ee 2 IPT nWSASIVO}2A HEY 6) ONY» Ia is 20 ek 9a non OE Ole ESR Rea NE cer eeee = 2 COT rani ae eee Sere Oe one ea She es en ee wt Ree 4 Ve rele Conte pega eked ee RAD 0s. 2s eee ee i iibe sere tate Boe ae vt 5 ech eR eS Ia TRY re Rie op ian ap re RE Ra Ce aie ar De 6 COMPOSITION OF THE HERPETOFAUNA ___. WW 7 SECTION I. PATTERNS OF DISTRIBUTION, SPECIES NGS TAY 4 ANOS INGIB SIVAN | 2p eee ee ee 8 INUETRELOD Stace os ee eee lala Nee ee Ee ee ON eae Ee 8 JEATOSOTEGGS cts ae ie a Wie WO RO cee ec ie oe ee SS 10 ID YRS LORSSTC ON Rg II Ie yee ace a dd eo 15 SECTION II. ENVIRONMENTAL CORRELATES OF SPECIES DENSITY 17 MirREODS) ete Srreeae Termeni SM AoA l, Wee eel Mire. 2. a. 2g 17 FEES WTR INS gan Sr Rs I WN KS LN ent LOO LET = oT apa ie et SER tcl 22, DISCUSSION ei 2:2) eee ee Se) A tee I Ue ee 24 SECTION III. EVOLUTION OF A NEOTROPICAL PENINSULAR eA Urea AN IN Aye See tek eh iat ad te a ED sa bls Sle eg 30 IW UTBESTTSCOYDISS wa at PN Ly ee ra RE oF Ge a 31 IRV ESTOS Sue ae ee ee Oe cE et (ARS OG St DRE ele OPO Renee ec mre 31 ID ISGUSSTO Ni mpe eee POE Se el at ee 35 SUMMARY AND CONCLUSIONS UE SWINE Ne oto Mee ee PRAY. . S eoielet 8 tS Loe oe Al LITERATURE CITED Pee ENIX” 2 ad Wa to Uae) ATE ROLE la rea eR Mae 22 ha cg 47 4 . s rap/ { ’ a | j $ : ‘ce I iin (a3 ‘ Fe 2 ~ i Y < ~ INTRODUCTION For several reasons the peninsula of Yucatan offers an excellent opportunity to study patterns of animal distribution and to assess the relative contributions of several factors thought to be important in setting distribution limits and thus in controlling the numbers of co-occurring species: (1) The area can be delineated objectively using natural features that constitute a barrier to dispersal of the terrestrial fauna; the peninsula is thus a relatively discrete and_ self-contained unit. (2) The area forms a cul-de-sac, with faunal interchange generally re- stricted to movement along a north-south axis; thus, although the point of origin for the various faunal elements may not be known, most of the fauna must have entered from the south and spread north- ward. (3) The northern end of the pen- insula is comparatively young geolog- ically, thus affording an opportunity to assess the effects of time in shaping pat- terns of distribution and species density. (4) Strong north-south gradients in pre- cipitation and vegetation structure exist within the peninsula and are not seri- ously confounded by elevational vari- ation. In this study I have sought to assem- ble and integrate into a coherent whole certain distributional, ecological, and his- torical data pertaining to the herpeto- fauna of this restricted portion of the Neotropics. I use these data to test vari- ous hypotheses that have been invoked to explain gradients in species density, and I formulate a series of hypotheses concerning the evolution of the penin- sular herpetofauna. In so doing I have found it expedient to organize the study into three sections. In section one I seek to identify recurring patterns of distribu- tion, endemism, and species density. In section two I attempt to relate the pat- terns of distribution and species density to environmental features in order to evaluate the importance of various fac- tors in setting distribution limits and in controlling the numbers of co-occurring species. Finally, in section three I treat the historical development of the herpe- tofauna and its patterns of distribution through ecological and evolutionary time. ACKNOWLEDGEMENTS It is a pleasure to acknowledge the many people who assisted in various phases of this project. For loan of speci- mens, answers to my written queries, and/or provision of working space, I am indebted to the following curators and their institutions: Richard G. Zweifel and Charles W. Myers, American Mu- seum of Natural History; Clarence J. McCoy, Carnegie Museum of Natural History; Shi-Kuei Wu, University of Col- orado Museum; Max A. Nickerson and Robert W. Henderson, Milwaukee Pub- lic Museum; Arnold G. Kluge and Ron- ald A. Nussbaum, Museum of Zoology, University of Michigan; Hymen Marx, Field Museum of Natural History; John W. Wright and Robert L. Bezy, Los An- geles County Museum; Ernest E. Wil- liams, Museum of Comparative Zoology; David B. Wake, Museum of Vertebrate Zoology; George R. Zug and W. Ronald Heyer, National Museum of Natural His- tory; Dorothy Smith, University of Ili- nois; Walter Auffenberg and John Iver- son, Florida State Museum. Ernest Liner and James Knight supplied locality data from their private collections, as did Rev. Lennard Dieckmann, S. J., St. John’s Col- lege, Belize City, Belize. Collecting permits were graciously issued by Mr. E. O. Bradley, Chief For- estry Officer, Ministry of Trade and In- dustry, Belize; and by Dr. Antonio Landazuri Ortiz, Direccién General de la Fauna Silvestre, México. I thank the authorities of Tikal National Park for permission to measure trees there. Field work was supported in part by a National Science Foundation grant for improving doctoral research in the field bo sciences (DEB 76-09303), and by grants from the William Saul Fund and the Watkins Fund, administered through the Museum of Natural History, University of Kansas. Transportation to museums was partly financed by monies from the Graduate School, and satellite imagery was purchased by the Department of Systematics and Ecology, both of The University of Kansas. During various phases of this project I have benefited from discussions with many talented biologists, including Rob- ert L. Bezy, Martha L. Crump, John D. Lynch, Clarence J. McCoy, Michael V. Plummer, Alan H. Savitzky, L. C. Stuart, Catherine A. Toft, Richard Wassersug, and John Wright. William E. Duellman, chairman of my doctoral committee, shared with me his wealth of knowledge of Neotropical biology. He also carefully read and criticized the manuscript, as did Richard F. Johnston, Norman A. Slade, and two anonymous reviewers. All errors are my responsibility. A host of Yucatecans, Belizians, and Guatemalans aided in specimen acquisi- tion and in other ways facilitated the field work. Richard Lacer and Michael V. Plummer provided companionship in the field. Janet M. Lee assisted during six months of field work, often in remote areas and under unpleasant conditions. In addition to performing more than her share of maintenance activities, she re- corded data, prepared voucher speci- mens, and identified plants. The project could not have been completed in its present form without her conscientious efforts. Finally, I owe an obvious debt to L. C. Stuart, who anticipated by 20 to 40 years many of my conclusions. In addi- tion to sharing with me his unparalleled knowledge of Middle American herpe- tology, “Don Pancho” provided accom- modations at Panajachel and smoothed the way for me in many other ways. He made available to me his considerable unpublished data on peninsular amphib- ians and reptiles, and was, from the out- MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY set, enthusiastic in his support of this project. THE ENVIRONMENT Many authors have treated casually or in detail various aspects of the en- vironment of the Yucatan Peninsula. Here my intent is to provide a brief overview of the physiography, climate, geology, and vegetation of the peninsula, emphasizing those features that are im- portant for an understanding of animal distribution. Readers interested in more thorough discussions of these topics are referred to the publications cited in the following sections. Physiography.—The Peninsula of Yu- catan is a broad, flat limestone shelf jutting north-northeast into the Gulf of México and the Caribbean Sea. Bounded on the north, east, and west by water, and to the south and southwest by the highlands of Alta Verapaz, Guatemala and the Mesa Central of Chiapas, Meéx- ico, the area is a relatively discrete nat- ural unit of approximately 240,000 km?, and spans nearly six degrees of tropical latitude. The peninsula contains all of the Department of El] Petén, Guatemala; the Republic of Belize (formerly British Honduras); and the Mexican states of Yucatan, Campeche, and Quintana Roo; as well as the eastern portion of Tabasco and the Lacandén region of Chiapas (Fig. 1). The northern third of the peninsula is devoid of major topographic relief. Only the Sierrita de Ticul (maximum elevation 270 m; Heilprin, 1891) breaks the monotony of the countryside. (See Figure 2 for the locations of many of the place names used in this discussion). The central portion of the peninsula rises gradually to a maximum of 350 m in southeastern Campeche (West, 1964; Paynter, 1955), and is continuous with the rolling uplands of northern E] Petén (Stuart, 1958). South of parallel 17°N, in central and southern E] Petén, a parallel series of folded limestone ridges runs east-west and thence northwest into Chi- YUCATAN HERPETOFAUNA 3 apas and Tabasco, producing a more varied topography. To the south and southwest these ridges give way to high- lands, the 600 m contour of which, for purposes of this work, is taken as the southern boundary of the peninsula. Commencing in northeastern El Petén and continuing through northern Belize and into southern Quintana Roo is a series of major faults which produce low limestone ridges and intervening swampy areas (West, 1964). The most conspicuous topographic feature of the peninsula is the uplifted south-central portion of Belize, termed variously the Cockscomb or Maya Moun- tains. These reach a maximum elevation of 1158 m ( Wadell, 1938). The surface of much of the peninsula consists of eroded and_ thoroughly karsted limestone. Caves, caverns, and subterranean waterways abound, espe- cially in the north. The porosity of the limestone precludes much accumulation of surface water; lakes are uncommon, and rivers are virtually absent from the northern third of the peninsula. Through- out much of this area natural wells (cenotes—from the Mayan dzonot), which result from collapse of the lime- YUCATAN ra ee a EL PETEN \ Fic. 1—Map of the Yucatan Peninsula show- ing political subdivisions and major topographic features. stone roofing of subterranian chambers, are important sources of fresh water and support a mesophilic biota. These caves and cenotes have been studied in detail by Cole (1910), Hatt et al. (1953), Mercer (1896), Pearse (1938), and Thompson (1897). Scattered throughout the peninsula are depressions (aquadas) which fill with water during the rainy months, but are frequently dry at other times. A belt of lakes extends across southern Campeche and through south- ern Quintana Roo. From west to east these are: Laguna Silvituc, Zoh Laguna, Laguna Chacanbacab, Laguna Om, and Lago Bacalar. Further south, at approxi- mately 17°N, a chain of lakes lies in a major east-west fault. Among these are Laguna Perdida, Lago Macanché, La- guna Yaxha, and Lago Petén Itza; the latter is the largest and deepest lake in the peninsula with a depth in excess of 32 m and a surface area of 567 km? (Covich, 1976). The northernmost river of any conse- quence is the Rio Champotdn, which drains portions of west-central Cam- peche and enters the Gulf of México at the town of Champotoén. In southwest- ern Campeche several rivers flow in a northerly direction into Laguna de Tér- minos, a large bay which is nearly cut off from the Gulf of México by Isla del Carmen. Among these is the Rio Can- delaria, which originates in northwestern El Petén, and the rios Champan and Palizada. By far the largest river is the Usumacinta, which originates in the De- partments of Huehuetenango and Alta Verapaz, Guatemala, and flows north- westward onto the Tabasco lowlands where it joins the Rio Grijalva before entering the Gulf of México. Two of its major tributaries, the Rio de la Pasion and the Rio San Pedro Martir, drain much of El Petén. Draining an esti- mated 102,828 km?, and with an average annual discharge of approximately 28,118,000,000 m?, the Usumacinta is the most important river in Middle America (Tamayo, 1964). The northernmost river of the Caribbean drainage is the Rio + MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Hondo, the headwaters of which drain northeastern E] Petén and southeastern Campeche, where the river is known as the Rio Azul. The Rio Hondo forms the international boundary between Belize and México as it courses northeastward, finally to enter Bahia Chetumal. Most of north-central Belize is drained by the Belize and Sibun rivers, while the south- ern third of the country is dissected by numerous small rivers and _ streams, among which are the Deep, Monkey, and Golden rivers. As it flows eastward into Bahia de Amatique, the Rio Sar- stoon, which originates in the depression between the Maya Mountains and the Sierra de Santa Cruz, forms the southern border between Belize and Guatemala. The west coast of the peninsula is essentially a sandy beach, occasionally interrupted by low cliffs and rocky areas, as in the vicinity of the town of Cam- peche. Paralleling much of the northern margin of the peninsula, from Celestun at the northwest corner, to the vicinity of Chiquila near the Yucatan-Quintana Roo border, is a sandy barrier beach, behind which lies a series of swamps, marshes, and shallow lagoons known as La Cienega. Along portions of the east coast of Quintana Roo, limestone out- crops form sea cliffs and headlands with which alternate small sandy beaches, as at Tulum. Halfway down the east coast of Quintana Roo are the large shallow bays known as Bahia Ascension and Bahia Espiritu Santo. Further south lies Bahia Chetumal, which marks the coastal boundary between Belize and México. To the north and west of the peninsula, the Campeche Banks extend up to 250 km from shore, in contrast to the east side of the peninsula where the conti- nental shelf is narrow. Immediately off the northeast coast of Quintana Roo lie several small, sandy islands, possibly the remnants of a barrier bar (Paynter, 1955). Among these are the Islas Contoy, Can- cun, and Mujeres. Beginning at the northeast corner of the peninsula and ex- tending discontinuously southward for roughly 650 km to the Gulf of Honduras, lies the longest coral barrier reef in the Atlantic Tropics (Edwards, 1957). Hun- dreds of tiny islets and atolls dot the reef, which lies approximately 40 to 60 km off shore. The protected shallow la- goon behind the reef contains numerous small mangrove islands. Climate.—Aspects of the climate of the Yucatan Peninsula have been essayed by Page (1933, 1938), Lundell (1937), Vivo Escoto (1964), and Garcia (1965), from whose works the following discus- sion is drawn. Owing to its tropical setting, low elevation, and to strong maritime influ- ences, the region enjoys a warm and homogeneous temperature regime, with only slight fluctuations in mean temper- ature from one locality to another, and from season to season. Mean annual temperatures for Progreso, Yucatan; Champotén, Campeche; and Paso de los Caballos, E] Petén are 24.9, 26.2, and 27.2 C, respectively (Page, 1933, 1938). The annual range of mean monthly tem- perature is 6.2 C at Champoton, 4.2 C at Progreso, and 6.1 C at Paso de los Caballos. However, within a_ single month temperature extremes can be con- siderable, especially during winter and spring when variations of 22° to 28° C have been recorded at Progreso, Mérida, and Valladolid. The monthly march of temperatures is similar throughout the peninsula, with January and May usually the coldest and warmest months, respec- tively. Frost and freezing temperatures are unknown. The lowest temperature reported by Page (1933) is 4.0° C for Champoton in January, 1926; the maxi- mum is 47.0° C for the same station in March of the same year. The amount and seasonality of rain- fall vary considerably throughout the peninsula, and from year to year at any one locality. In general rainfall is great- est at the base of the peninsula and de- creases to the north and, especially, to the northwest. Progreso, on the north- west coast, receives an average of 500 mm of rain per annum, whereas Paso de los Caballos, in northwestern E] Petén, YUCATAN HERPETOFAUNA 5 receives in excess of 1700 mm, and areas farther south receive somewhat more (Page, 1933, 1938). Complicating this general pattern is an area of unusually high rainfall in northern Quintana Roo, where 1200 to 1500 mm may fall in a year (Garcia, 1965). As elsewhere in Middle America, “summer” is the rainy season, with most of the rain falling from May through October. During these six months, rainfall is bimodal, generally with peaks in June and September sepa- rated by a relatively dry July. The per- centage of total annual rainfall occurring from May to October—a measure of sea- sonality—increases from south to north- west; in much of western Yucatan and northern Campeche 80 to 90% of the rain falls during this period. The correspond- ing figures for much of El Petén and eastern Quintana Roo are 60 to 70%. In summary, the climate of the Yuca- tan Peninsula may be characterized as thoroughly tropical, with uniformly high temperatures and seasonal rainfall. An- nual rainfall is greatest in the south and east portions of the peninsula and least at the northwest corner. Seasonality of rainfall exhibits an opposite pattern, and is greatest in the northwest portion. Vegetation._Several attempts have been made to describe and classify the vegetation of the Yucatan Peninsula Lundell (1934, 1937) combined floristic, climatic, and physiographic information to recognize six phytogeographic divi- sions in the peninsula, none of which i especially well defined. His Southerr Campeche Division includes roughly the southeast third of the state of Campeche. from about the latitude of Champoton south to the Campeche-E] Petén border, and from the Campeche-Quintana Roo border west for a distance of approxi- mately 85 km. According to Lundell, the area is a well-drained calcareous upland supporting a forest dominated by the zapote (Achras zapota) and the chaca (Bursera simaruba) both of which rarely exceed 20 m in height in this area. Palms, figs, (Ficus spp.), and mahogany (Swietenia macrophylla) are rare, and groves of ramon (Brosimum alicastrum) are widely scattered. The Southwestern Campeche Division encompasses the southwestern third of the state and is characterized by Lundell as a rainforest dominated by cedar (Cedrela mexicana), Swietenia macrophylla, Achras zapota, and Ficus spp. Approximately the north- ern third of Campeche, together with all of Yucatan and the northern tip of Quin- tana Roo comprise the Northern Division of Lundell’s classification. He considered the scrubby thorn forest of this area to be a subclimax resulting from centuries of shifting slash-burn agriculture prac- ticed by the Maya. He further supposed that the region once supported a climax vegetation similar to that of southern Campeche. Embracing nearly all of Quintana Roo and the northern third of Belize is Lundell’s East Coast Division, botanically a poorly known region at the time Lundell wrote. He characterized the southern two-thirds of this area, ex- clusive of Belize, as a vast forest domi- nated by Achras zapota and Swietenia macrophylla. The Northern Petén Divi- sion lies almost entirely within the De- partment of El Petén north of the 17th parallel. The botany of this region was treated in detail by Lundell (1937), who characterized the vegetation of these well-drained uplands as a_ luxuriant broadleaf evergreen quasi-rainforest, where forest giants such as Ceiba pen- tandra and Swietenia macrophylla may attain heights of 50 m. South and south- west of Lago Petén Itza lies the phyto- geographic division termed by Lundell (1937) the Central Petén Savanna Coun- try. The region is characterized by a series of disconnected grassy savannas upon which are scattered low, scrubby trees, especially the nanze (Byrsonima crassifolia). The boundaries of this sa- vanna country are said to coincide with the boundaries of a tongue of Cretaceous limestone (Lundell, 1937), thereby sug- gesting a possible edaphic explanation for the anomalous occurrence of savan- nas amidst the luxurious mesophytic for- est. However, Lundell (1937) favored 6 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY an anthropogenic origin for the savannas. Leopold (1950) presented a vegeta- tion map of México, in which he at- tempted to reconstruct the pre-human distribution of vegetation types. He rec- ognized five such types in the Yucatan Peninsula: rain forest, tropical evergreen forest, tropical deciduous forest, thorn forest, and savanna. Paynter (1955) de- vised a simple scheme involving only three vegetation zones: a scrub zone bordering the north coast and extending inland for perhaps 20 km; a deciduous forest zone extending over much of Yu- catan and northern Campeche; and a rainforest zone covering central and southern Campeche, Quintana Roo, northern El Petén, and northern Belize. Paynter’s scrub zone corresponds to the thorn forest type of Leopold, and his de- ciduous forest zone corresponds to Leo- pold’s tropical evergreen and _ tropical deciduous forest types combined. Their rainforest zones are essentially the same. Wagner (1964) utilized a structural classification devised by Beard (1955) in which plant associations are defined on the basis of floristic similarity; the asso- ciations are grouped into formations ac- cording to physiognomic similarity and are united to form formation series. Two formation series are depicted by Wagner as occurring in the Yucatan Peninsula: a dry evergreen formation series in north- ern and central Yucatan which also oc- curs as isolated patches in southwest Campeche, eastern Belize, and central Petén; and a tropical rain forest forma- tion series occurring elsewhere. Each of the above vegetation classi- fications has merit, yet no system of veg- etation classification nor vegetation map can accurately reflect the complex mo- saic that is the vegetation of the Yucatan Peninsula. There, vegetation types grade subtly and imperceptably into one an- other, or interdigitate in intricate pat- terns. Slope, aspect, elevation, drainage, and edaphic factors combine to produce a heterogeneous vegetation even within limited areas. Add to this the effects of climate and long-term human disturb- ance and the result is a vegetation so complex as to defy simple generalization. Nonetheless, it cannot be denied that both the height and luxuriousness of the forest diminish dramatically from south to north. From a structurally complex mesophytic forest in southern El Peten, the vegetation gradually gives way to a low scrubby xerophytic thorn forest at the north end of the peninsula. Nor can it be doubted that the nature of the veg- etation exerts a strong influence on the composition of the herpetofauna. Geology.—According to Sapper (1937), the Yucatan Peninsula (as defined here- in) consists of two distinct orographic units named the “Yucatan Peninsula” and the “South Petén and Maya Moun- tains.” The boundary between the two areas according to both Sapper (1937) and Wadell (1938) is the east-west fault that commences about 40 km south of Belize City on the Caribbean, skirts the northern slope of the Maya Mountains, and then passes just north of the Lakes Yaxha, Macanché, and Petén Itza. The boundary then swings northwest, forms the valley of the Rio San Pedro Martir, and then continues through the region of Tenosique to terminate immediately west of Laguna de Términos. North and northeast of the Sierrita de Ticul the geology is reasonably well understood. According to Sapper (1937) the area north of approximately the 21st parallel, including the barrier beach, is of Qua- ternary age. Southward, extending to the Sierrita lie marine limestones, mostly of Pliocene age, from which the overlying Pleistocene strata have been largely eroded, except in the vicinity of Mérida and Izamal. According to Galloway (in Hatt et al., 1953) the Sierrita is of Mio- cene age. Of special biogeographic im- portance is the conclusion of Hatt et al. (1953) that “There is indeed no geolog- ical evidence that any of the peninsula from the Serrania (= Sierrita) north- wards was available to land vertebrates until late Pleistocene-Recent time. . .” Extending southwest from the Sierrita to the vicinity of Laguna de Términos and YUCATAN HERPETOFAUNA a into northern El Petén is a vast area mapped by Sapper (1937) as Miocene limestone, but which, according to the profile drawn by Galloway (in Hatt et al., 1953) should include Oligocene de- posits in northern Campeche, and pos- sible Eocene deposits in northern El Petén. The entire Tabasco-Campeche alluvial plain in the vicinity of Laguna de Términos was considered by Sapper (1937) to consist of Quaternary sedi- ments. He considered the limestones of northern Belize and northeastern El Petén to be of Oligocene age, having been stripped of their Miocene covering (Wadell, 1938). The area of southern E] Petén, con- sidered by Wadell (1938) a geologic- orographic continuation of Chiapas and Tabasco, consists of dolomitic limestones resting conformably upon the Lower Cretaceous limestones of northern Alta Verapaz. Wadell (1938) thus considered them to be of Upper Cretaceous age. These are overlaid by Tertiary breccias and conglomerates, generally of Eocene to Oligocene age. Quaternary deposits of gravel, sand, and clay occur along rivers and lakes, and in topographic de- pressions. As described by Ower (1929), the Maya Mountains of south-central Belize are an uplifted block of Upper Carboniferous granite, surrounded by Cenozoic limestones. Ower (1929) be- lieved that the mountains arose as part of a general Pliocene orogeny, but Stuart (1966) indicated that they have been land positive since the Cretaceous. Thus, with only local exceptions, the Peninsula of Yucatan can be viewed as a continuous block of marine limestone of various ages, sloping upward toward the south. Emergence of this unit, which apparently commenced in the Miocene (Vinson and Brineman, 1963), proceeded from south to north such that the depos- _its become progressively younger to the north. Throughout the late Tertiary the main portion of the peninsula together with the Maya Mountains were probably land areas in firm connection with Nu- clear Central America, although marine transgressions in the form of the Chapa- yal Basin and the Amatique Embayment may, in the Upper Tertiary, have sev- ered this connection (Vinson and Brine- man, 1963). COMPOSITION OF THE HERPETOFAUNA As presently understood, the known herpetofauna of the Yucatan Peninsula, exclusive of marine turtles and strictly insular forms, consists of 164 species representing 25 families and 93 genera (Table 1). This does not include the faunas of the hundreds of islands and atolls adjacent to the peninsula. Their treatment is beyond the scope of this study. For completeness I have included several species which, although widely distributed elsewhere, barely enter the peninsula and can scarcely be considered integral elements of the herpetofauna. Among these is Natrix rhombifera, which reaches its southern distribution limit in eastern Tabasco and southwest Campe- che, and Geophis carinosus, which is generally restricted to situations at 1000 to 1500 m, but which has been taken at Palenque, Chiapas. I am aware of no specific peninsular localities for Storeria dekayi WHolbrook (=Storeria tropica Cope). The type locality for S. tropica is “Petén, Guatemala” (Cope, 1884). Stuart (1934, 1963) considered the spe- cies present in the Department of El Petén. I consider S. dekayi a valid mem- ber of the peninsular herpetofauna, but exclude the species from subsequent analyses. In comparison with other tropical TaBLE 1.—Taxonomic Composition of the Her- petofauna of the Yucatén Peninsula. Group Families Genera Species Salamanders 1 2 5 Anurans 7 15 30 Turtles 4 8 11 Crocodilians 1 1 2 Lizards 6 22, 43 Snakes 6 45 ie Total 95 93 164 8 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY areas, the peninsular herpetofauna is depauperate. For instance, the Mexican state of Michoacan, with about one- fourth the area, possesses only one less species than does the Yucatan Peninsula (Duellman, 1965b). Comparisons with Amazonia are even less favorable; a 3 km? area of Ecuadorian rainforest is known to support 173 species of amphib- ians and reptiles (Duellman, 1978). The numbers of species occurring in the pen- insula, and in various regions within the peninsula can be explained, at least in part, in terms of historical and ecological factors. These explanations are presented in the sections which follow. SECTION I: PATTERNS OF DISTRIBUTION, ENDEMISM, AND SPECIES DENSITY The task of documenting distribu- tions and searching for patterns is the concern of descriptive (as opposed to historical or ecological) biogeography, and forms the substance of the discus- sion which follows. Specifically, my pur- poses here are to (1) ascertain the peninsular distribution of each of the 164 species of amphibians and reptiles in the Yucatan Peninsula, (2) identify areas of concordance of distribution limits, (3) identify and delineate areas of faunal homogeneity, (4) identify areas of en- demism, and (5) document patterns of species density. METHODS Prerequisite to biogeographic analysis is accurate mapping of the geographic distributions of taxa. Minimally, such mapping requires locality records suffi- cient to infer distributions accurately, and some understanding of the phyletic relationships of the organism considered. At the very least one must know whether or not samples drawn from different lo- calities represent conspecifics. These kinds of information are not uniformly available for the herpetofauna of the Yucatan Peninsula. Figure 2 identifies those areas where important collections of amphibians and reptiles have been made. Because many archeological ex- peditions to the peninsula included biol- ogists among their personnel, biological investigation in the area tends, in part, to reflect the activities of Mayanists, and the biota in the vicinity of many impor- tant Mayan centers are comparatively well known. In general Yucatan, north- ern Quintana Roo, central E] Petén, and Belize have been well sampled, whereas portions of southern Campeche, southern Quintana Roo, and northern El Petén form an area where much remains to be learned concerning the herpetofauna. Problems of nomenclature and alpha taxonomy persist for possibly ten percent of the 164 species here considered. Espe- cially troublesome are members of the genera Sphaerodactylus, Eleutherodac- tylus, Elaphe, Micrurus, Tantilla, and Pliocercus. Additional collecting and tax- onomic study will resolve these questions and refine the emerging picture of ani- mal distribution in the peninsula. To what extent these additional data will modify the general conclusions here set forth remains to be seen, but I believe they will in no major way prove con- tradictory. I assembled locality records for each species considered by me to be a valid member of the peninsular herpetofauna. In so doing I accepted published records from reliable literature sources, and I examined all major and several minor collections of Yucatecan materials in the United States. I augmented these data with approximately 2,000 specimens rep- resenting 103 species obtained during nine months of field work. These are de- YUCATAN HERPETOFAUNA 9 71 40-0 032 O33 Fic. 2.—Important collecting stations in the Yucatan Peninsula. Closed circles indicate areas in which major collections have been made. Open circles represent minor collections. 1. Al- tun Ha, Belize; 2. Apazote, Campeche; 3. Augustine, Belize; 4. Balchacaj, Campeche; 5. Becan, Campeche; 6. Belize City, Belize; 7. Belmopan, Belize; 8. Calcehtok, Yucatan; 9. Campeche, Campeche; 10. Candelaria, Cam- peche; 11. Catmis, Yucatan; 12. Cayo, Belize; 13. Celesttin, Yucatan; 14. Central Farm, Be- lize; 15. Champotén, Campeche; 16. Chetumal, Quintana Roo; 17. Chichén Itza, Yucatan; 18. Chinaja, Alta Verapaz; 19. Chunyaxché, Quin- tana Roo; 20. Ciudad del Carmen, Campeche; 21. Cobaé, Quintana Roo; 22. Corozal, Belize; 23. Double Falls, Belize; 24. Dzibalchén, Campeche; 25. Dzibilchalttin, Yucatan; 26. Dzilam Bravo, Yucatan; 27. Dzitas Yucatan; 28. Dziuché, Yucatan; 29. El] Ceibal, El Petén; 30. El Desempefo, El Petén; 31. Emiliano Zapata, Tabasco; 32. Encarnacion, Campeche; 33. Escarcega, Campeche; 34. Es- meralda, Yucatan; 35. Felipe Carrillo Puerto, Quintana Roo; 36. Flores, E] Petén; 37. Fron- tera, Tabasco; 38. Gallon Jug, Belize; 39. Hopelchén, Campeche; 40. Isla Aguada, Cam- peche; 41. Kantunil, Yucatan; 42. Kikil, Yuca- tan; 43. Laguna Alvarado, Campeche; 44. La- guna Chacanbacab, Quintana Roo; 45. Laguna Chumpich, Campeche; 46. Laguna Silvituc, Campeche; 47. Laguna Yaxha, El Petén; 48. La Libertad, El] Petén; 49. Las Canas, El -Petén; 50. Libre Unién, Yucatan; 51. Limones, Quintana Roo. 52. Manatee, Belize; 53. Maya- pan, Yucatan; 54. Mérida, Yucatan; 55. Mid- dlesex, Belize; 56. Orange Walk, Belize; 57. Palenque, Chiapas; 58. Paso de los Caballos, El Petén; 59. Peto, Yucatan; 60. Piedras Ne- posited in the collections of the Museum of Natural History, University of Kansas. i cannot claim to have personally exam- ined every available museum specimen. Rather, I sought to verify peripheral or otherwise questionable records, and I accepted uncritically those records fall- ing well within known distributions. Concerning questions of nomenclature and taxonomy, I generally have accepted the conclusions of the most recent au- thority to have dealt with a group in a thorough and comprehensive manner. Occasionally taxonomic decisions are based upon my own investigations, the results of which will appear elsewhere. I summarized the locality records as spot maps—one for each species—and inferred from the maps the limits of dis- tribution for each species (see appendix for spot maps). Though I followed no particular rule for inferring limits, I was conservative and was guided solely by the locality records, rather than by con- siderations of habitat. My estimates of distribution are therefore probably min- imal ones. To the extent that the distri- bution maps reflect the distribution of the animals rather than the activity of collectors, they provide answers to a number of questions, the most funda- gras, El Petén; 61. Pisté, Yucatan; 62. Playa del Carmen, Quintana Roo; 63. Popolna, Yuca- tan; 64. Poptin, El Petén; 65. Progreso, Yuca- tan; 66. Pueblo Nuevo X-Can, Quintana Roo; 67. Puerto Juarez, Quintana Roo; 68. Puerto Morelos, Quintana Roo; 69. Ramate, El] Petén; 70. Rio Lagartos, Yucatan; 71. Sabancuy, Cam- peche; 72. San Andres, El Petén; 73. San Jose Carpizo, Campeche; 74. San Luis, El Petén; 75. San Pedro Columbia, Belize; 76; Sayaxché, E] Petén; 77. Silk Grass, Belize; 78. Sisal, Yuca- tan; 79. Sojio, El Petén; 80. Stann Creek, Be- lize; 81. Tekom, Yucatan; 82. Telchac, Yucatan; 83. Telchac Puerto, Yucatan; 84. Tenosique, Tabasco; 85. Tikal, E] Petén; 86. Tizimin, Yu- catan; 87. Toocog, El Petén; 88. Tower Hill, Belize; 89. Tres Brazos, Campeche; 90. Tulum, Quintana Roo; 91. Tuxpena, Campeche; 92. Uaxactin, El Petén; 93. Uxmal, Yucatan; 94. Valentin, Belize; 95. Vigia Chico, Quintana Roo; 96. Xcalak, Quintana Roo; 97. X-Can, Yucatan; 98. Xcopen, Quintana Roo; 99. Xpujil, Campeche; 100. Xunantunich, Belize; 101. Yokdzonot, Yucatan; 102. Zotz, El Petén. 10 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY mental of which is whether or not the limits of distribution occur at random in the peninsula. To answer this, I super- imposed over each map a_transpar- ent grid, the squares of which repre- sented 50 km on a side. Grid size is a compromise between the resolution with which one hopes to perceive patterns, and the accuracy with which one can plot localities and infer distributional limits. I tallied the number of distribu- tion limits that fell within each of the 108 grid squares, and cast the resultant values into a frequency distribution. Fol- lowing the suggestion of Hagmeier and Stults (1964), I then compared it to a Poisson distribution in order to detect departures from a random distribution. Because a distribution limit in one square is not likely to be independent of the occurrence of that same limit in an adjacent square the Poisson is not pre- cisely the expected distribution. How- ever, it seems a reasonable approxima- tion. I used the same grid to detect patterns of species density and ende- mism, and for a cluster analysis of the 108 grid squares, based upon presence or absence of species in each square. I considered a species present in a square if its distribution covered 50% or more of the land in the square. I calculated sim- ilarities for all pairwise comparisons of grid squares using the coefficient of Baroni-Urbani and Buser (1976) for binary data: VAD) Aen where A is the number of species com- mon to both squares, B is the number present in the first but not the second, C is the number present in the second but not the first, and D is the number absent from both but present in other squares. The coefficient ranges from 0 to 1 and allows negative matches. I used the similarity coefficients to perform cluster analyses using the unweighted pair group method with arithmetic av- erages (UPGMA). The UPGMA is an agglomerative hierarchical clustering technique which unites operational taxo- nomic units (OTU’s) or groups of OTU’s on the basis of some criterion of simi- larity or dissimilarity. To perform the calculations I used the TAXON program of the Numerical Taxonomy System of Multivariate Statistical Programs, version three, written by F. James Rohlf, John Kishpaugh, and David Kirk. The pro- gram also calculates cophenetic correla- tion coefficients which measure the dis- tortion introduced by the clustering process. Most workers have found that cophenetic correlation coefficients gener- ally range from 0.60 to 0.95 (Sneath and Sokal, 1973); high values indicate little distortion. Of several hierarchical clus- tering techniques, the UPGMA is said generally to introduce the least distortion (Rohlf, 1970). Sneath and Sokal (1973) give the algorithm for this clustering technique, and an example of its appli- cation to biogeographic data is given by Hagmeier and Stults (1964) and Hag- meier (1966). See Peters (1971) for a discussion of the limitations of this technique. RESULTS Distribution.—The distributions of amphibians and reptiles in the Yucatan Peninsula are summarized in the Appen- dix. Inspection of these figures reveals several general patterns. A number of species are restricted to the base of the peninsula, as for example: Bolitoglossa dofleini, B. mexicana, B. rufescens, Oedi- pina elongata, Eleutherodactylus lati- ceps, E. loki, E. rugulosus, Syrrhophus leprus, Centrolenella fleischmanni, Kino- sternon acutum, Anolis biporcatus, A. capito, A. uniformis, Sceloporus teapen- sis, Lepidophyma flavimaculatum, Ame- iva festiva, Celestus rozellae, Adelphicos quadrivirgatus, Clelia clelia, and Coni- ophanes fissidens. Other species range through the base of the peninsula and then northward along the east side, avoiding the north- YUCATAN HERPETOFAUNA 11 oononoaws & WR = 3 11 Fic. 3.—Cluster analysis of 108 grid squares on the basis of presence or absence of frog species. Squares clustered at the 0.95 level of similarity or higher are united, assigned a number, and cir- cumscribed by a dotted line. On the map a solid line encloses major areas of faunal homogeneity. The cophenetic correlation coefficient is 0.82. west corner. Conspicuous among these are: Agalychnis callidryas, Hyla ebrac- cata, H. loquax, H. microcephala, H. picta, Anolis tropidonotus, Corytophanes hernandezi, Eumeces sumichrasti, Den- drophidion vinitor, Imantodes cenchoa, Leptophis ahaetulla, and Xenodon rab- docephalus. Still others are restricted to the north end of the peninsula, such as: Bolito- glossa yucatana, Eleutherodactylus yu- catanensis, Kinosternon creaseri, Terra- pene mexicana, Sceloporus cozumelae, Leptotyphlops phenops, Coniophanes meridanus, Imantodes tenuissimus, Pli- ocercus andrewsi, Symphimus mayae, Tantilla cuniculator, and Bothrops yuca- tanicus. Finally there are those species which are pan-peninsular. These include: Lep- todactylus labialis, L. melanonotus, Bufo marinus, B. valliceps, Phrynohyas venu- losa, Smilisca baudinii, Hypopachus variolosus, Rana pipiens, Anolis rod- riguezi, A. sericeus, Basiliscus vittatus, Ameiva undulata, Boa constrictor, Coni- ophanes imperialis, Drymarchon corais, Drymobius margaritiferus, Leptodeira frenata, Leptophis mexicanus, Mastigo- dryas melanolomus, Ninia sebae, Spilotes pullatus, Tropidodipsas sartori, and Mi- crurus diastema. Statistical confirmation that the limits of distribution of amphibians and rep- tiles do not fall randomly through the peninsula is presented in Table 2, which compares frequency distributions of numbers of distribution limits per 50 x 50 km grid square with the expected fre- quencies assuming a Poisson distribution. For each major taxon, and for the entire herpetofauna, the approximate chi- square values substantially exceed the expected chi-square values for the appro- priate degrees of freedom at the 0.005 level. The null hypothesis that the limits of distribution follow a Poisson distri- bution, and thus are placed randomly, is decisively rejected in all cases. Table 2 also shows that in each case there is an 12 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Lizards oon om Oe Bae COL NI = 3 ~) My = tox oS ®$sGoaesy ant BR a : | | 1 0.90 088 0.81 0.73 0.67 060 0.53 Fic. 4.—Cluster analysis of 108 grid squares on the basis of presence or absence of lizard species. Squares clustered at the 0.90 level of similarity or higher are united, assigned a number, and cir- cumscribed by a dotted line. On the map a solid line encloses major areas of faunal homogeneity. The cophenetic correlation coefficient is 0.79. excess of squares with few distribution limits, and an excess of squares with many limits, indicating that the limits of distribution are contagious (clumped). The coefficients of dispersion (C.D.) in- dicate this also, for in all cases they sub- stantially exceed unity. Faunal areas.—Contagious distribu- tion limits indicate the existence of areas where distribution limits are concordant, i.e., areas of rapid faunal transition. These in turn imply the existence of areas of faunal homogeneity, the loca- tions of which are indicated in Figures 3, 4 and 5, which summarize the results of separate cluster analyses for frogs, lizards, and snakes. What constitutes a major cluster depends upon the level of similarity used to define it, and in this I have followed no particular rule; rather I have identified clusters, and the faunal areas they represent, by inspection of the phenograms. Some might disagree with my interpretations, but this is not a serious issue, for the clusters are usually easily recognizable. Thus, in Figure 3 four areas of faunal homogeneity are in- dicated for frogs: one in the northwest corner, one in the northern half of the peninsula exclusive of the northwest cor- ner, a central area, and a southern area. A similar pattern exists for lizards, al- though the picture is less clear. In Fig- ure 4 I recognize essentially the same four faunal areas identified for frogs, plus one small area in north-central Belize. A .somewhat different pattern emerges for snakes, where only three major areas are apparent: a northern, central, and southern area (Fig. 5). In contrast to frogs and lizards, the north- west corner of the peninsula does not YUCATAN HERPETOFAUNA 13 TaBLE 2.—Number of Distribution Limits per 50 X50 km Block Fitted to a Poisson Distribution. Anurans Lizards Snakes All Species #t limits observed expected # limits: observed expected # limits observed expected # limits: observed expected 0 35 19.5 Ov ey, 1.4 0 5 0.4 0 0 0.0 1 22, 33.0 1 14 6.2 1 11 2.2 1 5 0.0 2 19 28.8 2 18 13.4 2 14 6.1 2, 2 0.0 3 14 16.5 3 14 19.3 3 1 11.5 3 4 0.0 4 7 7.1 4 17 20.9 4 9 16.1 4 4 0.2 5 8 2.4 5 5 18.0 5 15 18.1 5 a 0.6 6 2 0.7 6 6 13.0 6 3 16.9 6 6 1.4 i 1 0.2 7 9 8.0 iL 9 13.5 1 Uf 2.7 —————__________—. 8 2 4.3 8 12 9.5 8 4 4.6 X? = 37.3 x?.005(3) =12.8 9 9 2.1 9 if 5.9 9 5 6.5 C.D. = 1.74 10 2 1.0 10 3 3.3 10 2 8.6 11 3 0.4 11 6 Iai 11 3 10.3 12 0 0.1 12 3 0.8 12 6 11.3 13 2 0.0 13 0 0.3 13 7 11.5 ee 0 0.1 14 5 10.8 2=61.3 x2.005(5)=16.8 15 2 0.1 15 4 9.5 C.D. = 2.30 16 1 0.0 16 3 7.8 17 1 0.0 17 3 6.1 ———_________———- 18 2 4.5 X2=81.7 x2 .005(6) = 18.6 19 3 3.1 C.D. = 2.64 20 5 2.0 21 3 1.3 22 1 0.8 23 2 0.4 24 3 0.2 25 1 0.1 26 2 0.1 PAE 4 0.0 28 0 0.0 29 0 0.0 30 2 0.0 31 0 0.0 32 0 0.0 33 1 0.0 34 0 0.0 35 1 0.0 36 1 0.0 emerge as a major area of faunal homo- geneity for snakes. However, some indi- cation of the distinctness of this area is apparent in the union of areas 1, 2 and 3 in Figure 5. What, if anything, is represented by the minor clusters in Fig- ures 3, 4 and 5 is not clear, but I have confidence in the reality of the major areas, and in the existence of rather sharp faunal breaks between them, for One can intuit the areas and their ap- proximate boundaries from inspection of Figure 6 in which for each major taxon the limits of distribution are superim- X2 = 88.2 x?.005(7) =20.3 C.D. = 5.30 posed on a single map. For frogs and lizards the area southwest from northern Quintana Roo to the vicinity of Laguna de Términos marks an area of faunal transition, and the same may be said for a few species of snakes. Similarly, the region of central Belize and northern El Petén is an area where limits of distri- bution of many species approximately coincide. Species density.—With only two ex- ceptions, the limits of distribution of frogs indicated in Figure 6 are the north- ern limits of species. From south to 14 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 1 2 3 4 5 6 7 8 9 Fic. 5.—Cluster analysis of 108 grid squares on the basis of presence or absence of snake species. Squares clustered at the 0.90 level of similarity or higher are united, assigned a number, and cir- cumscribed by a dotted line. On the map a solid line encloses major areas of faunal homogeneity. The cophenetic correlation coefficient is 0.85. Lizards Snakes Fic. 6.—Limits of distribution of amphibians and reptiles in the Yucatan Peninsula. Each line represents the inferred limits for a single species. northwest species drop out and are not replaced. The result is the dramatic faunal attenuation illustrated in Figure 7. The number of frog species dimin- ishes from a maximum of 22 in southern El Petén, to a minimum of nine at the northwest corner of the peninsula. A less dramatic decrease in species density oc- curs from east to west in the northern third of the peninsula. Lizards and snakes manifest a different species den- sity pattern. For both groups species YUCATAN HERPETOFAUNA 15 density is greatest at the base of the peninsula, diminishes toward the center, and then increases toward the north end (Fig. 7). Endemism.—The number of endemic species of amphibians and reptiles per grid square is indicated in Figure 8. En- demism is unquestionably greatest at the north end of the peninsula, where as many as 20 of the 26 peninsular endemics occur in a single grid square. In contrast, por- tions of El Petén have no endemics. Amphibians are underrepresented among the endemics: they account for 21.3% of the entire herpetofauna, but constitute only 11.5% of the total number of en- demics. Lizards and snakes are over- represented. Respectively they constitute 26.2% and 44.5% of the herpetofauna, but comprise 30.8% and 53.8% of the endem- ics. The single endemic turtle constitutes 3.8% of the endemic fauna. DISCUSSION A number of biologists have sug- gested that the Yucatan Peninsula could be partitioned on the basis of biological criteria. Smith (1940) utilized the dis- tributions of lizards of the genus Scelo- porus to define two provinces in the Yucatan Peninsula; Stuart (1943) used distributions of salamanders to recognize biotic areas in Guatemala, including El Petén; and Savage (1966) subdivided the herpetofauna of the peninsula into two geographical assemblages. More comprehensive treatments, which com- bine information for many groups of or- ganisms, include that of Goldman and Moore (1945) who recognized but a single province in the peninsula. Smith (1949) recognized provinces similar to those in his 1940 paper, but added an additional province. Stuart (1964) dis- tinguished the northwest corner of the peninsula from the remainder of the pen- insula. One generalization that emerges from these studies is that the north and northwest portion of the peninsula repre- sents an area biotically distinct from the central and southern portions, though there exists no consensus as to where the boundary between these areas lies. My results support the view that the north- ern and southern portions of the penin- sula are dissimilar biotically, and they further indicate that the peninsula could be more finely divided. Yet I have cho- sen not to formalize the areas of herpe- tofaunal homogeneity by naming them. Identification of these areas is not an end in itself, but rather serves as a point of departure, for the existence of such areas raises interesting questions concerning the historical development of these areas and their faunas and the nature of their geographical limits. These historical and ecological questions are discussed in the sections which follow. Lizards L — Fic. 7.—Species density patterns of amphibians and reptiles in the Yucatdn Peninsula. The figures in each square represent the number of species known or presumed to occur within that square. 16 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Reduction in species density at the ends of peninsulas—the so-called penin- sula effect—has been documented in Florida for amphibians and _ reptiles (Keister, 1971), and in Florida, Yucatan, and Baja California for birds (Mac- Arthur and Wilson, 1967), and mammals (Simpson, 1964). The phenomenon is thus a general one and is attributable to the isolating effects of peninsulas (Rick- lefs, 1973). It is of interest therefore, that among amphibians and reptiles in the Yucatan Peninsula, only frogs exhibit this expected reduction in numbers of species. And even here something other than a peninsular effect is operating, for the reduction in numbers of species is decidedly asymmetrical (Fig. 7), a pat- tern not explainable solely on the basis of isolation. With fewest species in the middle of the peninsula, snakes and liz- ards depart even further from the ex- pected pattern. Stuart (1958) in discus- sing the herpetofauna of the Tikal- Uaxactin area of northern El Petén, attributed the depauperization there in part to the fact that the area is transi- tional between the dry thorn forests of the outer end of the peninsula, and the wet forests of southern E] Petén. Appar- ently the same situation obtains for much of northern E] Petén, and southern Campeche and Quintana Roo. The con- cept of ecotone might lead one to expect more rather than fewer species in such a transitional area, but apparently this is an area which lies beyond the northern- most limits of many southern species, and beyond the southernmost limits of many northern species, especially the en- demics (Fig. 8). The factors that set these limits are discussed in the next section. Endemics Fic. 8.—Numbers of endemic species of am- phibians and reptiles in the Yucatan Peninsula. The figure in each square represents the num- ber of endemic species known or presumed to occur within that square. In summary, the limits of distribution of amphibians and reptiles in the Yuca- tan Peninsula are contagious, indicating the existence both of areas of faunal transition and areas of faunal homoge- neity. This is true for the entire herpe- tofauna and for all major taxonomic subdivisions. Anuran species density di- minishes dramatically from south to northwest. For snakes and lizards, spe- cies density is highest at the base of the peninsula, lowest at the center, and in- termediate at the northern end. The number of endemic species is greatest at the northern end and diminishes rapidly to the south. SECTION II: ECOLOGICAL CORRELATES OF SPECIES DENSITY The patterns of distribution, species density, and endemism identified in the preceding section are the end products of a complex interplay of factors oper- ating through ecological and evolution- ary time. To understand these patterns and to evaluate them in the light of cur- rent ecological and biogeographic theory, YUCATAN HERPETOFAUNA 17 it is convenient to consider separately two aspects of the problem, namely the historical development of the patterns, and the ecological factors that are im- portant in maintaining the patterns. This distinction between history and ecology is partly artificial and is not always easily maintained, yet it is useful because two rather different sets of questions are in- volved. For example, several species otherwise restricted to the dry north end of the peninsula occur as disjuncts on the savannas of El Petén and Belize (see discussion below). How these disjunc- tions came about is an historical ques- tion; what restricts the species to savanna regions and to the north end of the pen- insula is an ecological question. Such ecological considerations form the sub- stance of the present section. Here I seek to ascertain whether the patterns identified previously can be related to features in the environment. Specifically my purpose is to seek correlates of her- petofaunal species density, and to use the results of this analysis to weigh the merits of various hypotheses that have been invoked to explain species density gradients. Tunkas @ Fic. 9.—Map of the Yucatan Peninsula showing location of study sites. METHODS Two sets of environmental variables change conspicuously through the Yuca- tan Peninsula and seem likely to be im- portant in controlling numbers of co- occurring species of amphibians and reptiles. These are the amount and sea- sonality of precipitation, and the struc- tural heterogeneity of vegetation. Unfor- tunately, rainfall data for the peninsula leave much to be desired; for many areas no data are available, or, where avail- able, they are often incomplete, or span only a few years. I have relied upon the published data of Page (1933, 1938), the rainfall map of México compiled by Garcia (1965), and unpublished data for the Xpujil, Campeche area taken by Robert Wade of the University of Wis- consin from the files of the Division Hidrométricas, Peninsula de Yucatan, Mérida, Yucatan. Quantitative data on vegetation structure in the Yucatan Pen- insula are few. Consequently, I estab- lished study sites in each of seven dis- tinct vegetation types along a rough north-south transect through the penin- sula. Figure 9 gives the names and loca- tions of the sites, each of which is de- scribed below. Each site was situated in relatively undisturbed vegetation. Using the point-quarter technique (Cottam and Curtis, 1956), I sampled woody vegeta- tion at random points along each of ten parallel 100 m transects situated 10 m apart, except as noted below. For each plant 1 m or greater in height, I meas- ured (or estimated) total height, height to first foliage, and plant diameter, using a clinometer when necessary. Occasion- ally I could identify plants to species. More often I designated apparent “spe- cies” of woody plants, and assigned sam- ple plants to these “species” on the basis of canopy shape, growth form, leaf mor- phology, bark color and texture, color of wood and sap, odor of crushed foliage, and the appearance of fruits and flowers. My “species” are phena which probably correspond to taxa at or near the bio- logical species level of differentiation. I - 18 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY made no floristic comparisons between sites. At the Poptin site I staked out an area of 1 ha, within which I placed at random ten quadrats each 10 m square. Within each quadrat I identified and measured all woody plants | m or greater in height in the manner described above. Site descriptions._The Poptun site is located approximately 6.4 km north of the town of Poptun (16° 21’ N, 89° 26’ W), Department of E] Petén, Guatemala, at an elevation of about 550 m. The site is in pine savanna, with Pinus caribaea the dominant plant, but broadleaf forest penetrates the savannas along stream channels and ravines. Grasses are the predominant herbs, and scattered shrubs and low trees dot the landscape (Fig. 10). Wadell (1938) described and illus- trated the pine savanna region of Pop- tim. The site lies within the subtropical humid forest formation of Holdridge (1967), and has a tropical rainy climate (Afw of the Koeppen classification; Viv6é Escoto, 1964). I worked the Poptun site from 4 to 14 July 1974. Fic. 10.—Typical pine savanna at the Poptin study site. The EI Ceibal site is located near the south bank of the Rio de la Pasion, approximately 1.2 km west of the archae- ological site of El Ceibal (16° 34’ N, 90° 03’ W), Department of El Petén, Guate- mala, at an elevation of approximately 150 m. The area supports a luxurious mesophytic forest dominated by the co- rozo palm (Orbignya cohune). Trees oc- casionally reach a height of 50 m; many are 30 to 40 m high with interlocking crowns which produce a closed canopy through which little light penetrates to the forest floor. Lianas and bromeliads abound, and small palms and members of the genus Piper are common under- story plants (Fig. 11). Lundell (1937) described this vegetation type, which, owing to the dominance of the corozo palm, is termed a corozal. The site lies within the tropical humid forest forma- tion of Holdridge (1967), and has a trop- ical rainy climate (Afw of the Koeppen classification; Vivé Escoto, 1964). I worked the E] Ceibal site from 19 to 28 July 1974. The La Libertad site is located ap- proximately 4.9 km southwest of the Fic. 11.—Interior view of forest at the El Ceibal study site. YUCATAN HERPETOFAUNA 19 town of La Libertad (16° 47’ N, 90° 07’ W), Department of E] Petén, Guatemala, at an approximate elevation of 210 m. The site is situated on a savanna char- acterized by open expanses of grass through which are scattered small shrubby flat-topped trees, chiefly the nanze (Byrsonima crassifolia) (Fig. 12). Islands of typical forest edge trees such as Bursera simaruba, and Cecropia spp. dot the flat landscape. Lundell (1937) reported in detail on the botany of the central Petén savannas, and Stuart (1935) in his discussion of the herpetofauna of these savannas described and illustrated the vegetation. The La Libertad site lies within the humid tropical forest forma- tion of Holdridge (1967), and has a tropical rainy climate (AfW of the Koep- pen classification; Vivé Escoto, 1964). I worked at the La Libertad site from 15 to 20 October 1976. The Tikal site is located approxi- mately 4.8 km south-southwest of the famous archaeological site of Tikal (17° 20’ N, 89° 39’ W), Department of El Petén, Guatemala, at an approximate elevation of 283 m. The site is situated in a medium high forest, the canopy of which averages 25 to 35 m in height and is sufficiently open to permit penetration of considerable light. Common tree spe- cies include Brosimum alicastrum, and Achras zapota. Occasional Swietenia macrophylla are encountered. The thorny escoba palm (Crysophila argentea) and Fic. 12.—Typical savanna in the vicinity of the La Libertad study site. various species of Piper are common in the understory (Fig. 13). Bartlett (1935) gave a detailed account of the forest in the Tikal area. The site lies within the dry tropical forest formation of Hold- ridge (1967), and has a tropical rainy climate (Amw of the Koeppen classifi- cation; Vivé Escoto, 1964). I worked the Tikal site from 9 to 27 August 1974, and from 21 to 24 October 1976. The Xpujil study site is located ap- proximately 10.2 km west of the village of Xpujil (18° 30’ N, 89° 24’ W), Cam- peche, México, at an elevation of ap- proximately 250 m. Vegetation in this area, which has been characterized by Duellman (1965a) as quasi-rainforest, is a lower forest than at Tikal, but many of the same species occur, including Achras zapota, Cedrela mexicana, and Bursera simaruba. Palms are uncommon. The canopy is party closed and the under- story is a dense tangle of small vines, shrubs and saplings. The Xpujil site lies within the dry tropical forest formation of Holdridge (1967), and has a tropical rainy climate (Amw of the Koeppen clas- sification; Vivé Escoto, 1964). I worked the Xpujil site from 1 to 12 October 1974. The Santa Rosa site is located ap- proximately 12 km east-southeast of the town of Santa Rosa (19° 58’ N, 88° 53’ W), near the west edge of Laguna Chi- chancanab, Yucatan, México, at an ap- proximate elevation of 31 m. Here the forest is comparable in height to that at Xpujil, but more open. The dominant tree is Bursera simaruba. Small palms, shrubs, and saplings comprise the under- story, and grasses and other herbs cover the forest floor, especially where the can- opy is sufficiently open to allow pene- tration of considerable light. The Santa Rosa site lies within the very dry tropical forest formation of Holdridge (1967), and has a tropical wet-and-dry climate (Aw of the Koeppen classification; Vivé Escoto, 1964). I worked at the Santa Rosa site from 12 to 20 November 1974. The Tunkas site is located approxi- mately 12.3 km west of the town of 20 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Tunkas (20° 54 N, 88° 45’ W), Yuca- tan, México, at an approximate elevation of 33.5 m. The site is situated in a low, scrubby thorn forest, 3 to 7 m high, dominated by various species of decid- uous legumes. Palms are absent. This vegetation type has been well described by Bequaert (1933), and illustrated by Paynter (1955). The Tunkas site lies within the very dry tropical forest for- mation of Holdridge (1967), and has a tropical wet-and-dry climate (Aw of the Koeppen classification; Vivd Escoto, 1964). I worked the Tunkas site from 25 to 31 October 1974. To characterize the precipitation re- gime at each site, I inferred the amount and seasonality of rainfall from records for nearby stations. For the Tunkas site I used data from Page (1933) for Izamal, located approximately 18 km to the west. For the Santa Rosa site I used data from Page (1933) for Peto, located approxi- mately 29 km to the northwest. For the Xpujil site I used unpublished data col- lected by Robert Wade for Zoh Laguna, approximately 15 km north of the village of Xpujil. For the Tikal site I used data from Page (1938) for El Paso de los Caballos, situated 67 km to the west. For both the La Libertad and E] Ceibal sites I used data from Page (1938) for Paso Real, approximately 27 km south- southwest, and 14 km west of the two sites, respectively. For the Poptun site I used the data of Stuart (pers. comm.) for the village of Poptin, approximately 6.4 km to the south. These data repre- sent records for as few as a single year (Poptin), five years (Paso Real), nine years (Izamal), ten years (Paso de los Caballos), 12 years (Peto), and 17 years (Zoh Laguna). The figures for annual precipitation are mean values. I used the percent of mean annual precipitation falling from May through October as a measure of seasonality of rainfall. For each site I compiled a list of species of amphibians and reptiles en- countered by me or presumed to occur there. Because many species are rare and/or cryptophilic, I inferred the pres- Fic. 13.—Interior view of forest at the Tikal study site. ence of some species on the basis of collections made by me or by others in similar vegetation at nearby areas. Lists of species known or presumed to occur at each site are given in Lee (1977). Data analysis.—Following Pianka (1971), I estimated the areal cover of each plant using the formula for the area of a circle (A = 0.7854 d?, where d is the maximum diameter of the plant crown). I estimated foliage volume for each plant using the formulae for oblate 4 and prolate spheroids (V = 3 za*b and 4 V = 3 wab2, where a and b are the major and minor semi-axes, respectively). As a measure of heterogeneity I esti- mated diversity in the vegetation param- eters using the information theory sta- tistic of Shannon (Shannon and Weaver, 1949; H = - ¥ p; log pi, where p; is the proportion of plants in the sample be- longing to the ith category). This index is a composite, sensitive both to numbers of categories (richness) and to equita- bility of numbers of individuals among categories (evenness). YUCATAN HERPETOFAUNA 21 I estimated plant species diversity where in the above formula the p,’s are the proportions of all individuals in the ith species; species cover diversity where the p;’s are the proportions of total plant cover attributed to the ith species; and species volume diversity where the pj,'s are the proportions of total plant volume attributed to the ith species. I used the same formula to calculate plant height diversity, plant cover diversity, and plant volume diversity where the p;’s are the proportion of all individuals within each of 20 height categories, 40 cover cate- gories, and 100 volume categories, re- gardless of species. To characterize each study site on the basis of vegetation heterogeneity, I extracted principal com- ponents of variation from a matrix of correlation coefficients between the di- versity indices of each site. For this I used the Biomedical Computer Program BMDP4M (Dixon, 1975). Principal com- ponent analysis constructs new orthog- onal (independent) axes which are linear combinations of the original variables. The axes are oriented so as to explain maximally the dispersion in the multi- variate data cloud. Thus, a large propor- tion of variation in the original data set can be parsimoniously explained by only a few components. OTU’s can then be projected onto the component axes and their relationships assessed. See Cooley and Lohnes (1971) for further discussion of this technique. To assess the relative contributions of variables, both singly and in combination toward explaining variation in herpeto- faunal species density, I performed cor- relation and multiple regression analyses. For the latter I used the stepwise re- gression program BMDP2R of the Bio- medical Computer Programs (Dixon, 1975). This program seeks that linear combination of variables that maximally explains variation in the dependent vari- able, in this case species density. It en- ters variables one at a time, in descend- ing order of their unique contribution toward explaining variation, while simul- taneously accounting for correlation be- tween the independent variables. The program calculates a coefficient of multi- ple determination (R?*) which represents the proportion of variation in the de- pendent variable explained by the com- bined effects of the independent vari- ables. I accepted as best that regression model which accounted for the greatest proportion of variation in the dependent variable (highest R?). This is normally an unreliable criterion because R? can never diminish with the addition of more variables. In this instance virtually all variation is explained by only a few vari- ables and over specification of the model does not seem to be an issue. Cooley and Lohnes (1971) give further details of this technique. For each site the following variables were included: eae A) Latitude 2. (LONG) Longitude 3. (AMPH) Number of species of amphibians known or presumed to occur at each site Number of species of snakes known or pre- sumed to occur at each site Number of species of lizards known or pre- sumed to occur at each site 6. (TOTAL) Number of species of amphibians and reptiles known or presumed to occur at each site 7. (ANRN) Mean annual rainfall 8. (PCTRN) Percent of mean an- nual rainfall occurring from May through Oc- tober 4. (SNK) 5. (LZD) 5 (ESI) Plant species diversity 10. (SCD) Species cover diversity HeaCSVD) Species volume diver- sity 12. (PHD) Plant height diversity 13s PED) Plant cover diversity 14. (PVD) Plant volume diversity 22, MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Site scores on the first principal component Site scores on the sec- ond principal compo- nent 15. (PCI) 16. (PCII) RESULTS Species-area curves for woody plants at each site are presented in Figure 14. With the exception of El Ceibal, the curves approach the horizontal asymp- tote, indicating that all, or nearly all spe- cies within the sampling area are repre- sented. The assumption of the Shannon diversity statistic that the total number of species be known (Krebs, 1972) is thus met, or only weakly violated. Fig- ure 14 also illustrates the marked floristic impoverishment of the two savanna sites, Poptin and La Libertad. The diversity scores and scores on the first two principal components for each site are presented in Table 3. In O El Ceibal poo oo Cumulative Species + + + + + + + 24 48 72 96 120 144 168 192 216 Cumulative Points Fic. 14.—Species-area curves for woody plants at seven study sites in the Yucatan Peninsula. For Poptun the curve represents the cumulative numbers of species in ten 100 m2 quadrats. For other sites the curves represent cumulative species against cumulative numbers of points. general the savanna sites are the least heterogeneous both floristically and structurally. This agrees with the quali- tative assessment that savannas, in con- trast to the other sites, represent rela- tively simple environments in terms of woody vegetation. Results of the principal component analysis are summarized in Tables 4 and 5 and in Figure 15. The first and second components subsume 59% and 38% of the variation respectively. All six diversity indices load positively on the first prin- cipal component, which is therefore in- terpretable as a general heterogeneity factor. On the second component, PSD, SCD, and SVD load negatively, whereas PHD, PCD, and PVD load positively. Component two thus represents a con- trast between those variables which have as their richness component of diversity the number of species at each site (spe- cies-dependent indices), and those which have as their richness component of di- versity the number of height, cover, and volume categories (species-independent indices). The remaining components are dificult to interpret, but are relatively unimportant, accounting for less than 3% of the variation. They are not considered further. Ordination of the study sites on the first two principal components is il- lustrated in Figure 15. With low scores on the first component, La Libertad and Popttn again emerge as the least hetero- geneous of the seven sites; E] Ceibal is most heterogeneous, followed by Santa Rosa, Xpujil, Tikal, and Tunkas. El Cei- bal scores high on component two, indi- cating that the species-independent in- dices are the most important contributors to heterogeneity at that site. In contrast, TaBLE 3.—Summary of vegetation statistics for seven sites in the Yucatan Peninsula. Variables Site PSD SCD SVD PHD PCD PVD PCI PCII La Libertad 1.63 141 1.19 0.29 0.08 0.02 1.44 0.28 Poptuin 1.62 1.53 1.45 0.09 0.05 0.01 —1.42 0.07 Tikal 2.66 2.53 2.28 1.39 0.34 0.51 0.43 0.33 Tunkas 3.30 2.64 2.53 0.87 0.18 0.03 0.24 —0.68 EI! Ceibal 2.88 1.73 2.22, 1.63 1.06 0.97 0.91 1.88 Xpujil 3.16 3.18 3.01 1.04 0.06 0.06 0.54 —1.05 Santa Rosa 3.33 3.27 3.00 1.03 0.19 0.22 0.75 -—0.83 YUCATAN HERPETOFAUNA 23 TasBLE 4.—Factor loadings on the principal components extracted from a correlation matrix of six indices of structural and floristic diversity of woody vegetation at seven study sites. Component Variable I I Ill IV V VI PSD 0.917 —0.329 0.203 0.087 —0.034 —0.008 SCD 0.695 —0.706 —0.126 —0.044 —0.038 0.018 SVD 0.870 —0.476 0.030 -0.113 0.050 —0.010 PHD 0.922 0.328 -0.157 0.125 0.036 0.003 PCD 0.525 0.831 0.161 —0.049 0.009 0.019 PVD 0.581 0.802 —0,.112 —0.070 —0.042 -0.014 TaBLE 5.—Tabulation of eigenvalues, percent of trace, and accumulated percent of trace for each component of principal component analysis performed on a correlation matrix. The original data matrix consisted of six indices of structural and floristic diversity of woody vegetation at seven study sites. Accumulated Principal Component Eigenvalue Percent of Trace Percent of Trace 3.545 59.1 59.1 Il 2.279 38.0 97.1 Wil 0.121 2.0 oom IV 0.045 0.7 99.8 Vv 0.008 0.2 100.0 vI 0.001 0.0 100.0 and LONG were entered as independent variables. Of those variables considered, the single best predictor of AMPH is ANRN (Fig. 16), which accounts for 89% for Xpujil, Santa Rosa, and Tunkas, species-dependent indices are most im- portant. For La Libertad, Poptin, and Tikal, species-dependent and _species- independent indices contribute about equally to heterogeneity. Product-moment correlation coeffi- cients for all pairwise comparisons of ecological variables and herpetofaunal species densities are presented in Table 6. LIZ shows a significant positive corre- lation with SNK, PHD, and PVD, and a highly significant positive correlation with TOTAL. AMPH shows a highly sig- nificant negative correlation with LAT, a significant positive correlation with LONG and PCII, a highly significant positive correlation with ANRN, and a significant negative correlation with PCTRN. SNK shows a significant posi- tive correlation with LIZ, TOTAL, and PHD. Because many of the environmen- tal variables covary, it is of interest to regress simultaneously species density on these variables. Table 7 summarizes the results of the stepwise multiple re- gression analyses for which AMPH, LIZ, and SNK were treated as dependent variables and a separate analysis run for each. All other variables except LAT of the variation, followed by PCTRN and SVD. In combination these variables explain 99% of the variation in AMPH. For SNK, PHD is the best predictor (Fig. 17), followed by PCTRN, SVD, @E! Ceibal @La Libertad @ Tikal @Poptun PC Il @ Tunkas @ Santa Rosa @Xpuyil Fic. 15.—Ordination of seven study sites on the first and second principal components extracted from a matrix of correlation coefficients between indices of structural and floristic diversity. 24 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY and PVD; these four combine to explain 99% of the variation in SNK. PVD is the best predictor of LIZ (Fig. 18), followed by PCD, PCTRN, and ANRN; together these four explain 99% of the variation in LIZ. DISCUSSION Gradients in species density on local, regional, and global scales have long been of interest to biologists, and many hypotheses have been advanced to ex- plain them. It is generally agreed that the processes specified in these hypoth- eses need not act to the exclusion of one another, but instead may operate in con- cert, the exact combination varying with the situation. Nonetheless it is conven- ient to examine each hypothesis sepa- rately before inquiring as to how they might work in combination. The various hypotheses have been summarized so often (Pianka, 1966a, 1967, 1974; Rick- lefs, 1973; Krebs, 1972; Uetz, 1974) that a thorough summary is not necessary here. Instead I will examine only those hypotheses that are relevant to the pres- ent study. The Time Hypothesis.—According to this hypothesis, biotas diversify through time, hence older communities should be more diverse (and contain more species) than younger ones. It is useful to dis- tinguish between ecological time, which refers to the time available for dispersal and colonization, and evolutionary time, the time available for speciation. Areas that have only recently become available for colonization may be depauperate be- cause insufficient time has elapsed for an equilibrium number of species to be- come established. One might argue that the northern third of the Yucatan Penin- sula, which was submerged until some- time in the Pleistocene, represents such a non-equilibrium situation. Species den- sities of snakes, lizards, and amphibians are indeed lower at the north end than at the base of the peninsula, which has apparently remained land positive at least since the Miocene. But for snakes TasLE 6.—Correlation coefficients (r) between ecological variables and numbers of species of amphibians and reptiles in the Yucatan Peninsula. SVD PHD PCD PVD PCI SCD LONG ANRN PCTRN- PSD LAT SNK TOTAL LZD AMPH * Significant at the 0.05 level ** Significant at the 0.01 level YUCATAN HERPETOFAUNA 25 TaBLE 7.—Summary of results of multiple regression analysis of ecological variables and numbers of species of amphibians and reptiles in the Yucatan Peninsula. Amphibians Snakes Lizards Variable Variable Variable Step entered 2 Step entered R2 Step entered 2 1 ANRN 0.89 1 PHD 0.66 1 PVD 0.73 O4 PCTRN 0.95 2 PCTRN 0.76 2 PCD 0.85 3 SVD 0.99 3 SVD 0.95 3 PCTRN 0.92 4 PVD 0.99 4 ANRN 0.99 El Ceibal @ Poptun @ @ Tikal @ La Libertad Xpujil @ AMPHIBIAN SPECIES DENSITY Santa Rosa e @ Tunkas 84 108 132 156 180 MEAN ANNUAL RAINFALL (cm) Fic. 16.—Regression of amphibian species den- sity on mean annual rainfall for seven study sites in the Yucatan Peninsula. The regression equation is Y = 5.25 + 0.24X; r= .944; p<.0l. @ Tikal 36 @ Tunkas e El Ceibal 32 @ Santa Rosa 28 @ Xpujil SNAKE SPECIES DENSITY 26 @ La Libertad 24+ @ Poptun Of 045 @8 © Io 2 ie ls {fe PLANT HEIGHT DIVERSITY Fic. 17.—Regression of snake species density on plant height diversity for seven study sites in the Yucatan Peninsula. The regression equa- tion is Y = 23.97 + 8.25X; r = .814; p<.05. ®@ Tikal e EI Ceibal LIZARD SPECIES DENSITY 0.0 0.2 0.4 06 08 1.0 PLANT VOLUME DIVERSITY Fic. 18.—Regression of lizard species density on plant volume diversity for seven study sites in the Yucatan Peninsula. The regression equa- tion is Y = 10.41 + 16.63X; r = .864; p<.05. 26 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY and lizards, the lowest numbers of spe- cies occur not in the youngest area, but rather in the somewhat older, central portion of the peninsula. And for anu- rans, the reduction in numbers at the north end is decidedly asymmetrical. Furthermore, so many species are pan- peninsular that it is difficult to accept the idea that more species could occur in the north, but have simply not yet made the journey. There remains the possibility that more species could co- exist at the north end, but that there has been insufficient time for the evolution of forms sufficiently specialized to par- tition the environment finely. The pre- sumed recency of some of the Yucatecan endemics lends some credence to this view. Concerning Dipsas brevifaces and Sibon sanniola, and the great variation in lepidosis which obtains in those spe- cies, Peters (1960) wrote: “Since the Yucatan Peninsula was flooded for the most part during the Pliocene and Pleis- tocene, it is likely that both of these species are of fairly recent origin, and are quite possibly still in a state of evo- lutionary flux.” However, the same ob- jections may be advanced as with ecolog- ical time. The fewest species either do not occur in the youngest areas (snakes and lizards), or the pattern is asym- metrical (anurans). The existence of a substantial number of endemic species at the north end further argues against this view. Finally, the discovery of a fossil Lepidophyma of Pleistocene age at the northwest corner of the Peninsula (Hatt et al., 1953), far to the north of the present range of this mesophilic genus (see Appendix, Plate 15), suggests that the reduction in numbers of species at the north end may have resulted not from failure to differentiate or disperse into the area, but from failure to persist there. I conclude that the time hypoth- esis by itself is not adequate to explain the observed patterns of species density in the Yucatan Peninsula. The Spatial Heterogeneity Hypothe- sis.—Environments that are physically complex are expected to have more spe- cies than relatively simple environments. Here it is useful to distinguish between macro- and microspatial heterogeneity. The former refers to topographic relief on a geographic scale; the latter to habi- tat complexity on a local scale, such as vegetation structure, texture of substrate, etc. A number of studies (Simpson, 1964; Cook, 1969; Keister, 1971) have examined species density patterns in North America and have concluded that topographically diverse areas (moun- tains) support more species of mammals, birds, and amphibians than do non- montane areas at comparable latitudes. Reasons for this seem clear: topographic complexity can lead to isolation of popu- lations that promotes speciation; such areas are also likely to contain more hab- itats and consequently to support more species. In the Yucatan Peninsula, major topographic relief is wanting; this aspect of spatial heterogeneity is thus not an issue and will not be considered further. Following the successes of Mac- Arthur and MacArthur (1961) and Mac- Arthur (1964) who showed that bird species diversity was correlated with foliage height diversity, a number of workers have sought to quantify micro- habitat heterogeneity and to relate it to species densities or diversities of various groups of organisms. Recher (1969) found that the regression equation derived by MacArthur for North American birds accurately predicted bird species diver- sity in Australia, thereby suggesting gen- erality of the relationship. But Tomoff (1974) concluded that a model that com- bined aspects of foliage height diversity and physiognomic cover diversity was a better predictor of bird species density in desert scrub. Using multiple regres- sion analysis, Pianka and Huey (1971) found that plant height diversity was the best single predictor of bird species den- sity in the Kalahari desert, followed by mean annual precipitation, numbers of species of perennial plants, mean percent cover by perennials, and plant species diversity. Rosenzweig and Winakur (1966) devised a model which incorpo- YUCATAN HERPETOFAUNA 27 rated qualities of soil surface, vegetation height, and vegetation density to account for species diversity in desert rodent communities. Pianka (1966b, 1967, 1971) explored the relationships between lizard species density and environmental vari- ables in a variety of lizard communities on three continents. He found that in the deserts of western North America plant volume diversity was a good pre- dictor of lizard species density, but that in the Kalahari desert, mean percent plant cover, and plant species diversity were the better predictors. Microspatial heterogeneity in its various forms thus has been shown to be an important cor- relate of species density or species diver- sity, but the aspects of heterogeneity that are important vary between and within vertebrate groups. In the Yucatan Peninsula lizard spe- cies density is significantly correlated with plant volume diversity, duplicating the findings of Pianka (1966b, 1967) for North American desert lizards. However the two studies are not strictly compara- ble because I used many more categories in calculating plant volume diversity. Plant height diversity is also significantly correlated with lizard species density in the peninsula, but plant height diversity and plant volume diversity are them- selves correlated and the multiple re- gression analysis indicates that plant height diversity has little or no unique explanatory power. Two aspects of veg- etation structure—plant volume diversity and plant cover diversity—thus appear to be especially important, while the amount and seasonality of rainfall, which together account for only 14% of the variation, are relatively unimportant in explaining variation in lizard species density. Of the parameters of vegetation structure considered here, only scores on the second principal component correlate significantly with amphibian species den- sity. Apparently those sites that have high species-dependent indices of plant diversity also have large numbers of am- phibian species, but the biological sig- nificance of this relationship, if any, is not clear. Both the correlation and mul- tiple regression analyses indicate that the amount and seasonality of rainfall are of paramount importance in accounting for variation in amphibian species density between sites, and that species volume diversity makes a small (4%) contribution. Scores on the second principal component possess little or no unique explanatory power, and are not entered into the mul- tiple regression equation. I conclude that amphibian species density is relatively independent of habitat structure, at least as I have been able to quantify it, and that the spatial heterogeneity hypothesis need not be invoked to explain the ob- served pattern of amphibian species density in the Yucatan Peninsula. Only a single variable, species height diversity, correlates with snake species density in a way suggesting possible causation. The multiple regression anal- ysis shows that in addition, seasonality of rainfall makes a substantial contri- bution to explaining variation in snake species density, as does species volume diversity of vegetation. Taken as a whole, snake and lizard species density is related primarily to aspects of vege- tation structure rather than to the taxo- nomic composition of the vegetation; amount and seasonality of rainfall ap- pear to be of secondary importance, and I therefore conclude that the spatial het- erogeneity hypothesis is in some way applicable to snakes and lizards in the Yucatan Peninsula. But what is the bio- logical meaning of these relationships? Several explanations can be offered. First, it seems reasonable to assume that a more complex environment can be more finely partitioned by specialists, and MacArthur and MacArthur (1961) and MacArthur (1964) have developed this argument for birds. According to these authors, organisms can exploit a complex environment either by special- izing on one or a few resources, and foraging widely for these; or by utilizing a wide range of resources, and foraging over a restricted area. Only where the 28 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY resources are highly concentrated are the disadvantages of specialization out- weighed by the advantages. A similar argument might apply to lizards and snakes in the Yucatan Peninsula. A num- ber of studies, mostly involving lizards of the genus Anolis, have demonstrated that some lizards partition the habitat vertically, both within (Andrews, 1971; Schoener, 1967, 1969) and between (Schoener and Schoener, 197la, 1971b; Jenssen, 1973) species. In general these studies have been restricted to Antillean species which characteristically exist in much higher densities than do their mainland congeners, and among which competition for food is thought to be more intense (Andrews, 1976). My sub- jective impression is that in the Yucatan Peninsula population densities of snakes and lizards are low, perhaps too low for competition to promote fine habitat par- titioning. Henderson and Fitch (1975) found no evidence of vertical partition- ing of the habitat between Anolis seri- ceus, a pan-peninsular species, and its sympatric congeners, even where A. sericeus occurred in unusually high num- bers. Furthermore, in the present study, the correlation between snake and lizard species densities are actually weakened if only arboreal and semiarboreal species are considered. Finally, although Pianka (1967) found that plant volume diversity correlated well with the number of liz- ard species in North American deserts, only three of his 15 lizard species are to any extent arboreal, and only one is highly specialized for such an existence. So although this explanation has intuitive appeal, and although the structurally complex forests of El] Ceibal and Tikal do support many arboreal species, it re- mains to be demonstrated that mainland species of snakes and lizards partition the vertical component of the environ- mental mosaic. A second possible explanation blends aspects of the spatial heterogeneity hy- pothesis with an hypothesis involving predation. On the basis of manipulations of intertidal invertebrates, Paine (1966) concluded that predators can exert a reg- ulatory force over their prey such that species are held below carrying capacity, thereby reducing competition and pro- moting the coexistence of more species. The best empirical evidence for this hy- pothesis involves structurally heterogene- ous environments such as intertidal zones (Paine, 1966, 1969) and coral reefs (Por- ter, 1972). Such structurally complex environments should provide numerous safe sites in which individuals of prey species can avoid elimination by their predators. Experimentation (Huffaker, 1958; Huffaker et al., 1963) has shown that for some simple predator-prey sys- tems, stable oscillations in numbers of predators and prey can be obtained only under conditions of considerable spatial heterogeneity; in structurally simple situ- ations the systems become self-anihilat- ing. Thus, some minimal level of envir- onmental heterogeneity seems necessary for predation to be effective in promot- ing the coexistence of species; this effec- tiveness might vary with the degree of heterogeneity to produce the species density patterns observed for snakes and lizards in the Yucatan Peninsula. Evalu- ation of this suggestion requires informa- tion on the intensity of predation, infor- mation not presently at hand. The Productivity Hypothesis.—All things being equal, areas of greater pro- ductivity can support more individuals than areas of lesser productivity. The re- sultant large population sizes can result in greater genetic variation, which in turn could promote speciation (Connell and Orias, 1964). Furthermore, because each species need use less of the total range of resources, the same array of re- sources can support more species in a productive environment (Pianka, 1974). I have no direct measure of productivity for my study sites in the Yucatan Pen- insula. However, because productivity is known to be correlated with annual rain- fall (Odum, 1959; Whittaker, 1970), rainfall data provide a crude index of productivity. The base of the peninsula, which receives the greatest annual rain- YUCATAN HERPETOFAUNA 29 fall and is thus presumably the most productive, does support the greatest numbers of species of amphibians and reptiles. But for snakes and lizards, the dry—and therefore least productive— north end of the peninsula supports more species than does the wetter and presumably more productive central por- tion. The species density patterns of amphibians are most consistent with the productivity hypothesis. However, the correlation between amphibian species density and amount of rainfall is open to other interpretations. Does lack of rain limit the numbers of amphibian species indirectly through control of productiv- ity, or does it exert a more direct effect by imposing physiological demands re- lated to problems of water balance? Al- though these interpretations do not ex- clude one another, I favor the latter for several reasons. Nearly all of those anu- rans occurring at the xeric northwest corner of the peninsula possess charac- teristics that can be interpreted as adap- tations to minimize evaporative water loss. For instance, with few exceptions such species tend to be large (e.g., Rana pipiens, Bufo marinus, Bufo valliceps, Phrynohyas venulosa, Smilisca baudinii, Triprion petasatus). Their surface to volume ratio would convey a relative advantage in terms of cutaneous evapo- rative water loss, in contrast to those small species which drop out along the rainfall gradient (e.g., Hyla picta, H. microcephala, H. staufferi, H. ebraccata, and Syrrhopus leprus). Triprion peta- satus possess a coossified skull (Trueb, 1970), across which evaporative water loss is probably reduced, as has been shown for two other species of frogs with cranial coosification (Seibert et al., 1974); this is of obvious advantage dur- ing phragmosis (plugging holes with parts of the body) for which Triprion is known to use its head (Stuart, 1935). Finally, two species, Leptodactylus labi- alis and L. melanonotus, construct foam nests in which the eggs hatch and the larvae undergo partial development. In L. melanonotus the foam nest floats on the surface of the water, whereas in L. labialis the nest is constructed in bur- rows at the water’s edge. Heyer (1969) discussed the adaptive trend toward ter- restriality demonstrated by members of the genus Leptodactylus. He considered foam nests to be adaptations that convey a degree of independence from the aquatic environment, thereby reducing exposure to aquatic predators and the risk of desiccation of a temporary water source. All but one of the anuran species occurring at the northwest corner of the peninsula are widely distributed through- out México and Central America. Their adaptations to xeric conditions cannot be viewed as a response to the specific con- ditions of aridity in Yucatan, but rather represent characteristics which preadapt them to that situation. These consider- ations lead me to conclude that problems of evaporative water loss and water bal- ance have been important in setting dis- tribution limits of amphibians in the Yucatan Peninsula and that the produc- tivity hypothesis, although consistent with the distribution data, does not offer a compelling explanation for the ob- served patterns of amphibian species density. Other Hypotheses.—Additional hy- potheses and combinations of hypotheses have been advanced to explain species density gradients. They are relevant to the present discussion, but are difficult to evaluate with the data at hand. Com- petition, for example, is generally thought by ecologists to be a potent force in shaping community structure and it may play an important role in controlling numbers of coexisting species. How and to what extent this is so in the Yucatan Peninsula is not clear. If in fact lizards and snakes partition the structural habi- tat (see above), such partitioning is pre- sumably an adjustment made in response to past competitive interactions. The com- plementary distribution of certain species pairs suggests competitive exclusion; e.g., Kinosternon acutum and K. creaseri, (Ap- pendix, Plates 6, 7), Laemanctus long- 30 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY ipes and L. serratus (Appendix, Plate 12), Sceloporus chrysostictus and S. tea- pensis (Appendix, Plates 12, 13), Both- rops nasutus and B. yucatanicus, (Ap- pendix, Plates 26, 27), Bothrops asper and Crotalus durissus (Appendix, Plates 26, 27). Such ecological replacement would not produce species density gra- dients, but more complex and diffuse competitive interactions might contrib- ute to the observed patterns. Closely coupled with hypotheses concerning competition are ideas about climatic sta- bility and predictability. Stable and/or predictable environments may allow finer adaptations and greater specializa- tion because less energy is expended or held in reserve for maintenance. Evalu- ation of this suggestion as it applies to the Yucatan Peninsula awaits acquisition of information comparing competitive ability, reproductive performance, and energy allocation within and between species. In conclusion, it appears that the two basic species density patterns in the Yu- catan Peninsula—one manifested by am- phibians, the other by snakes and lizards —have rather different underlying causes. Amphibians seem to be responsive to, and apparently are controlled by, essen- tially abiotic factors, of which amount and seasonality of rainfall are especially important. These presumably act to set distribution limits through the imposition of conditions beyond the physiological tolerances of certain species. In contrast, snakes and lizards seem to be controlled, perhaps indirectly, by biotic factors, par- ticularly features of environmental struc- ture such as plant height, cover, and volume diversity. These conclusions can be generalized to include other tropical amphibian and reptile communities. Bar- bault (1976) studied herpetofaunal spe- cies diversity on savannas in the vicinity of Bouake, Ivory Coast. He found that lizard species diversity (estimated using the Shannon index) was positively re- lated to habitat structure diversity, whereas amphibian species diversity in- creased as a function of both the length of the rainy season and the number of breeding sites. He felt that snake species diversity was controlled indirectly by both habitat diversity and weather act- ing through changes in prey community structure. The similarity in the findings of these two studies, conducted on different con- tinents and involving phylogenetically unrelated communities, strongly suggest that fundamental differences exist in the relative importance of biotic and abiotic factors in controlling species densities of tropical amphibians and reptiles. SECTION III: EVOLUTION OF A NEOTROPICAL PENINSULAR HERPETOFAUNA In the preceding sections I found it necessary to treat the taxonomic compo- sition and patterns of distribution of the peninsular herpetofauna as static. In reality these two attributes of the herpe- tofauna are in perpetual flux. New spe- cies evolve or are added to the fauna by immigration; other species become extinct locally or regionally. The éffects of speciation, immigration, extinction, and emigration, all of which proceed against the background of a changing environ- ment, insure that the herpetofauna of today is not what it was in the past, nor what it will be in the future. Thus there remains the question of the development of the taxonomic composition and pat- terns of distribution through time, a sub- ject that has long been of concern to biologists interested in the Yucatan Pen- insula (e.g., Gadow, 1905) because the area has served as a notable center of vertebrate differentiation and, possibly, dispersal (Miller, 1973). The following questions are addressed in this section: (1) To what extent is the peninsular herpetofauna autochtho- nous? (2) Where did the allochthonous YUCATAN HERPETOFAUNA 31 elements come from, and by what route? (3) Of several potential source areas, which have been most important in sup- plying faunal elements to the peninsula? (4) To what extent can the present-day composition and patterns of distribution be interpreted in terms of past vegeta- tion and climatic changes? (5) How have the patterns of distribution been modified by millennia of settlement and intensive agriculture by the Maya? METHODS Leon Croizat’s biogeography (Nelson, 1973) questions the assumption that the objective of historical biogeography is to find centers of origin and patterns of dis- persal. Croizat et al. (1974) argue that to seek centers of origin and dispersal routes is to search for that which often does not exist. In those instances where they do exist, they are attributes of indi- vidual taxa having little general explan- atory power. A more profitable approach, they contend, is the “vicariance” or pan- biogeographic method advocated by Croizat (1958, 1962). The method in- volves the compilation of the distribu- tions of many species in order to ascer- tain general patterns. The distribution of a species or group of related species is circumscribed or connected by a line, producing an area termed a track. When this is done for many taxa, areas of con- cordant tracks may become apparent; these are known as generalized tracks. Tracks represent the geographic relation- ship between the members of the group and the generalized track represents an estimate of the geographical distribution of an ancestral biota which has been fragmented (vicariated) to produce the observed pattern. The method then leads to inferences concerning those historical events responsible for effecting the vi- cariance. These permit formulation of testable hypotheses of considerable gen- erality. Although apparently not explic- itly stated by Croizat (Ball, 1975), his followers have been quick to emphasize that the appropriate units for track anal- ysis are monophyletic groups. Although I believe that use of the terms track, generalized track, and vicariance as ap- plied to biogeography represents an un- necessary proliferation of jargon, I also believe that this method of analysis has two principal strengths: (1) It makes no a priori assumptions about centers of origin and the role of dispersal; thus the facts of distribution are allowed to speak for themselves. (2) Because the method involves the evaluation of large numbers of individual distributions, it leads to formulation of general hypotheses. The methods of “vicariance” biogeography are further discussed by Croizat et al. (1974), and by Rosen (1974). Recent examples of application of the method are those of Rosen (1975) and Wiley (1976). Where possible I have utilized the method of Croizat, outlined above, to formulate hypotheses about the historical development of the peninsular herpeto- fauna, especially the origins and evolu- tion of the endemic elements. Because the method requires information about the cladistic relationships of the taxa un- der consideration, and because often these relationships are very imperfectly known, this analysis must be considered a first approximation. However, the hy- potheses that stem from this analysis are amenable to test. Additional distribu- tional data and/or improved understand- ing of cladistic relationships may corrob- orate (but cannot verify) these hypoth- eses. More importantly, as they become available, data from paleoecology and paleoclimatology can be marshalled as potential falsifiers. RESULTS Of the 164 species of amphibians and reptiles in the Yucatan Peninsula, 112 (68.3%) are widespread in the Gulf and/or Caribbean lowlands of southern México and Central America. Of the 112, 54 (48.2%) also are widespread on the Pacific versant. The presence of these wide-ranging species in the penin- 32 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY sula is certainly no surprise; indeed, it is the absence of some such species (e.g. Coniophanes picevittis) that is note- worthy. The origin of this portion of the peninsular herpetofauna thus involves the larger question of the origins of the Middle American herpetofauna, a sub- ject treated at length by Savage (1966). Primarily on the basis of modern distri- bution patterns, Savage (1966) charac- terized the genera of Middle American amphibians and reptiles as belonging to four historical assemblages: Old North- ern, Middle American, South American, and Young Northern. He concluded that the herpetofauna of Middle America is not transitional between that of the Neo- tropics and the Nearctic, but rather is sufficiently distinct to stand alone as a separate major herpetofauna. If we ac- cept Savage’s interpretation, 15.1% of the peninsular genera belong to the Old Northern assemblage, 48.4% to the Mid- dle American assemblage, 15.1% to the South American assemblage and 4.3% to the Young Northern assemblage. The re- maining genera cannot be easily referred to a particular assemblage. At the spe- cific level, 13.4% of the peninsular species show Old Northern affinities, 49.4% show Middle American affinities, 12.8% show South American affinities, and 4.9% show Young Northern affinities. Thus, at the generic and specific levels, the peninsular herpetofauna as a whole shows its great- est affinities with the Middle American assemblage, a conclusion wholly ex- pected on the basis of geography alone. Few genera and species appear to be Nearctic or Neotropical derivatives. Fic. 19.—South peninsular disjuncts. A. Sceloporus chrysostictus. B. Cnemidophorus angusticeps. C. Conophis lineatus. D. Masticophis mentovarius. E. Stenorrhina freminvillei. YUCATAN HERPETOFAUNA 33 INTRA-PENINSULAR PATTERNS Southern disjuncts.—Several species occur more or less continuously through the northern portion of the peninsula, and are represented by disjunct popula- tions to the south (Fig. 19). Several of these southern disjuncts are sufficiently differentiated to have been accorded subspecific status. Five species of rep- tiles that manifest this pattern are inhab- itants of subhumid to xeric situations; all tend to avoid heavy forest. Those popu- lations isolated in the south are generally associated with savannas or areas of sec- ond growth. Three additional species— Rhinophrynus dorsalis, Triprion peta- satus, and Crotalus durissus—are thought by some authors to conform to this pat- tern. However, accumulation of addi- tional locality records suggests that they are continuously distributed throughout the region (Appendix, Plates 1, 5, 27). Northern disjuncts——The opposite pattern obtains for several other species. These are widely distributed through the base of the peninsula, with isolated pop- ulations to the north, and especially to the northeast (Fig. 20). In general these are mesophilic forest-dwelling species, and the disjunct populations tend to oc- cur in an area of unusually high rainfall at the northeast corner of the peninsula (see discussion of climate in Section I above). In summary, within the peninsula there are two complementary intraspe- cific patterns of distribution. Species in- habiting xeric to subhumid situations are widespread in the north and occur as disjuncts on savannas and in disturbed areas to the south; mesophilic species, wide-ranging through the base of the peninsula, occur as disjuncts to the north- east, especially in an area of high rainfall. EXTRA-PENINSULAR PATTERNS The Yucatan-West México Pattern.— No fewer than five species of amphibians and reptiles endemic to the peninsula have their apparent closest living rela- tives distributed on the Pacific versant of México (Fig. 21). The genus Triprion, with only two species, is represented by T. petasatus in the peninsula and by T. spatulatus from Jalisco to Guerrero and on the Pacific versant at the Isthmus of Tehuantepec. Symphymus, likewise con- taining only two species, is represented in Yucatan by S. mayae and on the Pa- cific versant at the Isthmus of Tehuan- tepec by S. leucostomus. Of the four species of Enyaliosaurus, Duellman (1965b) considered the peninsular en- demic E. defensor most closely related to E. clarki of the Tepalcatepec Valley of Michoacan. This pattern is recapitu- lated by Eumeces schwartzei, a penin- sular species considered by Taylor (1935) to be most closely related to E. alta- mirani of Michoacan. Finally, Cnemi- dophorus anquisticeps of Yucatan was considered by Beargie and McCoy (1964) to be closest to C. costatus, a wide-ranging species in west Mexico. Additional peninsular endemics of un- certain affinities have relatives which, if sister taxa, would further corroborate this pattern, as for example, Dipsas brevifaces of Yucatan and D. gaigeae of Colima and Jalisco. Bothrops yucatani- cus may also reflect this pattern, for it is a member of a closely related group of four species of hog-nosed vipers, all of which are restricted to subhumid habi- tats, and two of which—B. hesperis and B. dunni—are restricted to western Méx- ico (Campbell, 1976). The Yucatan-East México Pattern.— Three species of reptiles occur as appar- ent isolates at the north end of the peninsula, and are represented by popu- lations on the Atlantic versant of México (Fig. 22). In the case of Sceloporus ser- rifer the apparent disjunction could be a collecting artifact. Such is not the case for Terrapene mexicana and Agkistrodon bilineatus, both of which occur in Ta- maulipas, far to the northwest of their Yucatecan relatives. Agkistrodon_bili- neatus is also widespread on the Pacific versant of México and Central America. The relationships between the three pop- ulations of A. bilineatus are unknown. If 34 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fic. 20.—North peninsular disjuncts. Extra-peninsular distributions are rough approximations. A. Hyla ebraccata. B. Corytophanes hernandezi. C. Eumeces sumichrasti. D. Sphenomorphus cher- riei. E. Dendrophidion vinitor. the affinities of the peninsular form lie with the population on the Pacific ver- sant, then the species represents, in mod- ified form, an example of the Yucatan- West Mexico pattern. The Maya Mountain-Nuclear Central America’ Pattern.The Maya Moun- tains of Belize support essentially a low- land fauna. However, two species of frogs known from the vicinity of the Maya Mountains, Rana maculata and Agalychnis moreletii, typically occur in montane situations. They are apparently isolated from the geographically nearest populations of their species in the high- lands of Guatemala and Honduras by unsuitable lowland habitat in the De- partments of El] Petén and _ Izabal, Guatemala. F. Scaphiodontophis annulatus. Miscellaneous Patterns._Two en- demic species of amphibians, Bolito- glossa yucatana and Eleutherodactylus yucatanensis, are nearly restricted to the north end of the peninsula where they occur in mesic situations such as caves and cenotes. In naming E. yucatanensis, Lynch (1964) acknowledged its close relationship with E. alfredi to the south- west. B. yucatana is one of the three members of the dofleini species group (Wake and Lynch, 1976); the closest liv- ing relative of B. yucatana is perhaps B. schmidti to the southwest (Wake, pers. comm.). The fossil Lepidophyma arizeloglyphus is known only from a Pleistocene cave deposit in the northwest corner of the peninsula (Hatt et al., 1953), far to the north of the modern YUCATAN HERPETOFAUNA 35 —-- Triprion spatulatus -» Enyaliosaurus clarki --—-Eumeces altamirani —— Cnemidophorus costatus --- Triprion petasatus ~~ Enyaliosaurus defensor Eumeces schwartzei —— Cnemidophorus angusticeps = Symphymus mayae --- Triprion spatulatus = Symphymus leucostomus Fic. 21.—The Yucatan-West México pattern of distribution. Extra-peninsular distributions are rough approximations. limits of this mesophilic genus. Thus, two mesic-adapted species are confined to the north end of the peninsula, but have their closest relatives in wetter areas to the south. An additional spe- cies, presumably mesophilic, existed at the north end of the peninsula until sometime in the Pleistocene. The foregoing distribution patterns involving peninsular endemics or species with isolated populations in the penin- sula can be summarized as follows: (1) Five species pairs exhibit a Yucatan- West Mexico pattern. Two additional pairs may also exhibit this pattern. (2) Two, and perhaps three, species have a Yucatan-East México pattern of distri- bution. (3) Two species centered on the north end of the peninsula have their closest relatives at the base of the penin- sula or immediately adjacent. (4) Within the peninsula five species are widespread at the north end and occur as isolates to the south; six species are widespread through the base of the peninsula and occur as isolates to the north. (5) Two species are widespread through the high- lands of Central America and occur in apparent isolation in the Maya Moun- tains of Belize. DISCUSSION The intra- and interspecific disjunc- tions identified above involve pairs of species, or sets of populations which, al- though separated geographically, occur in similar environments. These patterns represent the remnants of once continu- ous distributions that have become frag- mented. In isolation, the fragmented populations have differentiated to vari- ous degrees to produce species pairs (e.g., Triprion petasatus and T. spatula- 36 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY © Terrapene mexicana Sceloporus serrifer ---- Agkistrodon bilineatus Fic. 22.—The Yucatan-East México pattern of distribution. Extra-peninsular distributions are rough approximations. tus), subspecies pairs (e.g., Terrapene mexicana yucatana and T. m. mexicana), and populations showing little or no dif- ferentiation (e.g., Hyla ebraccata). Par- simony requires the assumption that the common ancestor of each pair inhabited an environment similar to that presently occupied by its descendants. Therefore, we may infer that disjunct populations inhabiting similar habitats betoken a more widespread and continuous distri- bution of that habitat at some time in the past. We need to know what histor- ical events were responsible for effecting the breakup of these habitats. I believe that two sets of events—Pleistocene changes in climate and vegetation, and pre-Colombian human influences—have been of overriding importance in shap- ing these patterns of distribution. PLEISTOCENE CHANGES IN CLIMATE AND VEGETATION The conventional view of the tropics in general, and the Neotropics in partic- ular, as ancient and stable environments must be abandoned in the light of the paleobotanical and _paleoclimatological evidence that has been accumulating steadily, especially during the past 15 years. Griscom (1942) was one of the first to advance this view when he sug- gested that Pleistocene climate changes resulting in the expansion and contrac- tion of vegetation zones were important in shaping patterns of bird distribution in Central America. His idea that the montane forests of Central America were sufficiently lowered to completely pinch out the lowland rainforest is probably incorrect. However, it now seems certair. YUCATAN HERPETOFAUNA 37 that the Neotropics have not been ex: empt from the Pleistocene climatic changes that so profoundly affected the northern hemisphere. Ideas about Neotropical climatic and vegetation change are central to theories concerning the evolution of species den- sities of Amazonian birds (Haffer, 1969, 1974) and frogs (Crump, 1974), and dif- ferentiation of Amazonian lizards ( Van- zolini and Williams, 1970). Such ideas have also played a central role in the biogeographic analysis of Neotropic dis- persal centers (Miiller, 1973). Parenthet- ically, it is interesting to note that the two opposing views of the tropics—one that they are unchanging, the other that they have been subject to much change —are both invoked to explain the same phenomenon, namely the extraordinary numbers of species of plants and animals in the tropics. Little evidence is at hand concerning the nature of these changes as they ap- ply to northern Central America and southern México. However, considerable palynological data are available for trop- ical South America (see Van Der Ham- men, 1974, for summary), and southern Central America (Bartlett and _ Bar- ghoorn, 1973). Also available are paleo- temperature curves calculated from deep-sea sediments in the Caribbean (Emiliani and Rona, 1969; Lynts and Judd, 1971), and documentation of long- term fluctuations in water levels of La- guna Chichancanab, Yucatan (Covich and Stuiver, 1974). Taken together these data allow a qualitative assessment of changes in the climate and vegetation in the Yucatan Peninsula, especially during Pleistocene and Holocene times. Before pursuing this question it is appropriate to inquire as to the nature of the climatic and vegetation changes suggested by the facts of amphibian and reptile distribution. Assuming that the extent to which isolated populations have differentiated is at least roughly propor- tional to the length of time since they became separated (admittedly a ques- tionable assumption, for it requires equal rates of evolution), it is possible to sug- gest the nature and sequence of the en- vironmental fluctuations which affected the separations. Thus, full species pairs presumably reflect an earlier divergence than do subspecies pairs, which in turn are older than those fragmented popula- tions showing little or no differentiation. Those species pairs exhibiting the Yucatan-West México pattern show a decided preference for subhumid to xeric situations. Trueb (1970) interpreted the disjunct distribution of Triprion in terms of a period of Pleistocene aridity when continuous subhumid to xeric habitat may have extended from the Pacific side of México across the Isthmus of Tehaun- tepec to the Gulf coast, and thence into the Yucatan Peninsula. Rossman and Schaefer (1974) noted the similarity be- tween the distributions of Triprion, Enyaliosaurus, and Symphymus. The ad- dition of the Cnemidophorus angusti- ceps-C. costatus and Eumeces schwartzei- E. altamirani species pairs to this pattern strengthens the argument that continu- ous subhumid habitat existed on both coasts of central and southern México. The subsequent onset of wetter condi- tions and the expansion of mesophytic vegetation, especially in the vicinity of the southern Gulf Coast, served to iso- late the subhumid environment of the Yucatan Peninsula from that of west Mexico, thereby promoting the differen- tiation of at least five pairs of amphib- ians and reptiles. The presence of three mesic-adapted species isolated at the outer end of the peninsula, far to the north of their rela- tives, suggests that the peninsula was once a more mesic environment than it is today. We may suppose that under wetter conditions mesophytic forests ex- tended northward in the peninsula, and that the progenitors of those species presently restricted to the base of the peninsula were more widely distributed. With the onset of drier conditions and the retreat of the wet forests, many spe- cies disappeared from the north end; others became restricted to small pockets 38 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY of mesic habitat associated with cenotes and caves (e.g., the ancestors of Eleu- therodactylus yucatanensis and Bolito- glossa yucatana) where they underwent differentiation in isolation. This same sequence of events might also account for the presence of the northern disjuncts discussed above. Alternatively, they could be the result of a more recent pe- riod of humid conditions, for the isolated populations of the six species exhibiting this pattern have undergone little or no differentiation. The Yucatan-East México pattern also involves species generally restricted to subhumid habitats, but these are differ- entiated only at the subspecies level. We may again hypothesize a continuous sub- humid habitat around the Gulf of Méx- ico uniting the Yucatan Peninsula with northeastern Mexico. Martin (1958) was first to call attention to the similarity between the faunas of Yucatan and Ta- maulipas, and to suggest the existence of a dry lowland connection between the two areas. Several species thought by Martin to exhibit the Tamaulipas-Yuéa- tan disjunction are now known to be more widely distributed through the in- tervening area than he supposed (e.g. Hypopachus variolosus and Laemanctus serratus). His most impressive example of a Yucatecan endemic with northern af- finities was “Opheodrys” mayae, subse- quently shown by Rossman and Schaefer (1974) to be a member of the Middle American genus Symphymus rather than of the genus Opheodrys. Nonetheless, the presence of Terrapene mexicana and Agkistrodon bilineatus in Tamaulipas and Yucatan argues for the existence of continuous dry forest between the two areas. Possibly this connection was coe- val with the Yucatan-West México con- nection. Alternatively it could represent a more recent connection, for the Yuca- tan and Tamaulipas populations are only subspecifically distinct. The Maya Mountain-Nuclear Central America pattern is difficult to account for in terms of Pleistocene climatic and vegetation change. Moderate depression of montane forest might connect the Belizian population of Agalychnis more- letii with those in the highlands of Gua- temala and Honduras, but such lowering of vegetation zones would hardly provide suitable habitat for Rana maculata, a species which characteristically breeds in lotic situations. The Belizian speci- mens referred by Lee (1976) to Rana maculata are peculiar in several respects, and the possibility exists that they are not conspecific with populations of Rana maculata to the south. If the above interpretations are even approximately correct, two conclusions follow. First, during the late Pleistocene much of Middle America was subject to alternating periods of aridity and wet- ness. Second, the Yucatan Peninsula has been both drier and wetter than it is today. We now need to know to what extent these conclusions are consistent with the known facts of paleoclimatology and paleobotany. Palynological studies in northern South America (Van Der Hammen, 1974) have documented a pe- riod of aridity from about 21,000 to 13,000 B.P. when effective precipitation was less than during the Holocene (ca. the last 10,000 years), and an earlier period from about 90,000 to 21,000 B.P. when precipitation was greater than during the Holocene. In Panama pollen from about 7,300 to 4,200 B.P. suggest a drier climate than at present (Bartlett and Barghoorn, 1973). The generalized Caribbean paleotemperature curve of Emiliani and Rona (1969) is approxi- mately consistent with these findings if periods of low temperature are assumed to coincide with periods of aridity. Covich and Stuiver (1974) documented fluctuations of water levels in Laguna Chichancanab from about 22,000 to 8,000 B.P., culminating in a phase of re- duced lake volume or perhaps complete desiccation. Thus, different lines of evi- dence from paleoclimatology, palynol- ogy, limnology, and zoogeography all are consistent with the idea that major changes have occurred in and adjacent to the Yucatan Peninsula with respect to YUCATAN HERPETOFAUNA 39 climate during the Pleistocene. Although the timing, magnitude, and sequence of these changes are imperfectly known, one may confidently assert that the alter- nating wet-dry periods suggested by the facts of reptile and amphibian distribu- tions were real. PrE-COLOMBIAN HUMAN INFLUENCES Recent archaeological excavations in northern Belize have demonstrated the existence of an Early Formative Maya civilization at about 2500 B.C. (Hammon et al., 1976). This pushes back the be- ginnings of the Maya Early Formative period nearly 1500 years and establishes the Maya culture as one of the oldest in Middle America. Other studies in the vicinity of Edzna and Xpujil, in the state of Campeche, have shown that the Clas- sic (300-900 A.D.) Maya were far more sophisticated agriculturalists than previ- ously believed. Terraced fields (Turner, 1974) and ingenious irrigation systems (Matheny, 1976) allowed Mayan farmers to bring large areas under cultivation. Estimates of the number of people that a single Mayan farmer could support are being revised upward, and as a result ideas about population sizes and densi- ties are being reevaluated. Whatever the impact of the Mayan civilization on the biota of the Yucatan Peninsula, the effect was of greater duration and intensity than previously thought. Amphibians and reptiles featured prominantly in Mayan thought, to judge by their representation in carvings, paintings, and masonry. Some species were evidently of mythical significance: a monsterous rattlesnake with a human emerging from its jaws is a common motif at Puuc style sites in Yucatan, es- pecially those showing Toltec influence. The Maya undoubtedly transported liv- ing amphibians and reptiles from one locality to another for ceremonial pur- poses or as food items. Stuart (1958) concluded that the plastron and carapace of Dermatemys mawii found in a burial urn at Uaxacttin were probably carried into the area from some other locale. Such relocations were probably common- place, but of a local nature; their effect upon general patterns of animal distri- bution probably was insignificant. Of far greater importance was the extensive habitat modification occasioned by Mayan agricultural practices. Present- day Mayan farmers practice the slash- burn shifting agriculture of their ances- tors. As a result, the countryside is a patchwork of active and abandoned farm plots in various stages of succession, and the vegetation of nearly all of the north- western corner of the peninsula—today the area of most intensive cultivation—is held in a subclimax stage. During the Classic period, virtually all of the Yuca- tan Peninsula may have been under cul- tivation. Lundell (1934) believed that primeval forest was either rare or non- existent in the peninsula. It has been suggested that El] Petén, which today is an area of continuous tropical forest, was, at the height of the Classic period, an area of intensive cultivation similar to present-day Ohio (Turner, quoted by La Fay, 1975). The possible anthropogenic origins of the savannas of central El Petén have been mentioned previously (see discussion of vegetation, Section I). More subtle environmental modifica- tions have been attributed to the exten- sive deforestation by ancient Mayan ag- riculturists. Covich (1976) documented changes in abundances of freshwater gastropods in Lago Petén Itza which he attributed to major fluctuations in nutri- ent inflow and sedimentation rates caused by destruction of the surrounding for- ests. As discussed by Lundell (1937), the dominance of certain tree species in the vicinity of Mayan ruins is attributable to ancient Mayan horticulture. Some trees were of religious significance (e.g., Ceiba pentandra). Others, such as Achras za- pota and Brosimum alicastrum were en- couraged, if not actually cultivated, for their edible fruit. The effects of such widespread en- vironmental modification on the distri- butions of amphibians and reptiles are difficult to assess, but must have been 40 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY considerable. Those species which re- quire open situations (e.g., Sceloporus chrysostictus and Cnemidophorus an- gusticeps) and which today are largely restricted to the disturbed subclimax sit- uations at the north end of the peninsula very likely were much more widely dis- tributed in the past. Today such species also occur in isolation on savannas and in disturbed situations to the south. Fragmentation of their once continuous distributions could have resulted from expansion of the forests following col- lapse of the Mayan civilization and the near-abandonment of the Petén centers at about 950 A.D. Beargie and McCoy (1964) interpreted the distribution of Cnemidophorus angusticeps as the result of Pleistocene aridity. In support of this view one might argue that the subspe- cific differentiation exhibited by several of the southern disjuncts, including C. angusticeps, could not have evolved in the relatively short time since the decline of the Maya. However, a millennium seems sufficient for such differentiation, especially considering the rates of evolu- tion which could obtain in small, isolated populations under intense selection. Be- cause these southern disjuncts are nearly always associated today with areas of human disturbance, I favor an anthro- pogenic explanation for this pattern of distribution. The foregoing considerations suggest the following sequence of events has been important in shaping patterns of distribution of amphibians and reptiles in the Yucatan Peninsula. (1) Aridity was widespread and west México and Yucatan were connected by subhumid habitat. (2) With the onset of mesic conditions, west México and Yucatan be- came separated; mesophilic species be- came more widespread in the peninsula. (3) With the return of arid conditions, northeast México and Yucatan were con- nected by subhumid habitat; some meso- philic species were isolated at the north end of the peninsula. (4) With the es- tablishment of somewhat wetter, essen- tially modern conditions, northeastern Mexico and Yucatan became separated by mesophytic forest through much of the Gulf lowlands. (5) Mayan agricul- turists deforested much of the peninsula; non-forest species of amphibians and reptiles expanded their ranges. (6) With the collapse of Maya civilization, the for- est regenerated. Non-forest species re- ceded and became restricted to the north end of the peninsula or persisted in the south as relicts on savannas and dis- turbed areas. In conclusion, the herpetofauna of the Yucatan Peninsula taken as a whole shows overwhelming affinities with the herpetofauna of Middle America. How- ever, peninsular endemics and those spe- cies represented by disjuncts at the north end show affinities with xeric-adapted forms of western and northeastern Méx- ico. The bulk of the peninsular endemics appear to have evolved in situ when iso- lated in the peninsula by changing en- vironmental conditions during the late Pleistocene. Disjunct distributions within the peninsula are partly attributable to wet-dry alternations in climate and partly to deforestation by the Maya and subsequent reforestation following the decline of the Mayan civilization. SUMMARY AND CONCLUSIONS Owing to its peninsular configuration and lack of topographic relief, the penin- sula of Yucatan offers an excellent op- portunity to study patterns of animal distribution and to assess the relative contributions of several factors thought to be important in setting distribution limits and controlling the numbers of co- occurring species. The primary objectives of this study were to ascertain the taxonomic compo- sition of the herpetofauna of the Yucatan Peninsula; to identify patterns of distri- bution, species density, and endemism; YUCATAN HERPETOFAUNA 4] and to account for these patterns in light of ecological and historical factors. The known herpetofauna of the Yu- catan Peninsula numbers 164 species representing 25 families and 93 genera. I collated locality records for each species and summarized the records as spot maps, from which I inferred the limit of distribution of each species. Statistical analyses of these data show that the lim- its of distribution are contagious, indi- cating the existence both of areas of rapid faunal change and areas of faunal homogeneity. This is true for the entire herpetofauna and for all major taxo- nomic subdivisions. Using cluster analy- sis I identified and delineated four areas of faunal homogeneity for frogs, five for lizards, and three for snakes. For frogs and lizards these areas are largely con- gruent; the pattern for snakes differs from that of frogs and lizards. Amphib- ian species density diminishes dramati- cally from south to north, and especially to the northwest. For snakes and lizards species density is highest at the base of the peninsula, lowest at the center, and intermediate at the north end. Ende- mism is greatest at the north end of the peninsula. Disproportionately few spe- cies of amphibians are endemic, whereas snakes and lizards are overrepresented among the endemics. Measurement of vegetation structure at seven sites, each located in a distinct vegetation type, indicates that various parameters of vegetation heterogeneity, estimated using an information theory statistic, are important correlates of snake and lizard species density. For amphib- ians the amount and seasonality of pre- cipitation are most important. Thus, am- phibians appear to be responsive to, and limited by, abiotic factors. Snakes and lizards seem sensitive to biotic factors, especially the heterogeneity of the struc- tural habitat. In contrast with the results of other studies on peninsular distributions, there is no evidence that a “peninsular effect” involving isolation and distance from a source area is important in shaping spe- cies density patterns of amphibians and reptiles in the Yucatan Peninsula. Nor is it necessary to invoke ecological or evo- lutionary time hypotheses to explain the observed patterns of herpetofaunal spe- cies density. The majority of species of amphib- ians and reptiles in the Yucatan Penin- sula are forms widely distributed throughout the mesic Gulf and Carib- bean lowlands. At both the generic and specific levels, the peninsular herpeto- fauna shows its greatest affinities with the herpetofauna of Middle America. The xeric-adapted fauna of the north end of the peninsula exhibits affinities with the faunas of western and north- eastern Mexico. The bulk of the penin- sular endemics appear to have evolved in situ when isolated in the peninsula by changing environmental conditions dur- ing the Pleistocene. Disjunct distribu- tions within the peninsula are partly attributable to wet-dry alterations in cli- mate. Those disjunctions involving non- forest species are interpretable in terms of anthropogenic influences. RESUMEN Debido a la configuracién peninsular y a la falta de relieve topografico, la peninsula de Yucatan ofrece una ex- celente oportunidad para el estudio de las normas de la distribucién animal y para evaluar las contribuciones relativas de varios factores que se consideran im- portantes en la asignacion de limites de distribucién y en el control de nimeros de especies coexistentes. Los principales objectivos de este estudio son varios: cerciorarse de la composicién taxondmica de los anfibios y reptiles de la peninsula de Yucatan; identificar las normas de _ distribucidn, densidad de especies y endemismo; y ex- plicar estas normas desde el punto de vista de factores ecoldgicos e historicos. En la peninsula de Yucatan hay 164 especies conocidas de anfibios y reptiles que estan representadas por 25 familias y 93 genera. He cotejado informes de A2, MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY localidad por cada especie y resumido la informacién en mapas acotados de los cuales he deducido el limite de distri- bucién de cada especie. Analises esta- disticos de estas referencias muestran que los limites de distribucién estan agrupados; esto indica la existencia de areas de rapido cambio de fauna y areas de homogeneidad de fauna. Esto ocurre en todos los anfibios y reptiles y en todas las subdivisiones taxondmicas. Mediante el analisis de agrupacion he identificado y delineado quatro areas de homogeneidad de fauna en las ranas, cinco en las lagartijas y tres en las cu- lebras. Con respecto a las ranas y las lagartijas estas areas son mayormente congruentes pero la norma entre las culebras difiere de la de las ranas y las lagartijas. La densidad de las especies anfibias diminuye dramaticamente de sur a norte, y en particular al noroeste. Entre las especies de culebras y de lagartijas la densidad es mas alta en la base de la peninsula, mas baja en el cen- tro e intermedia en el norte. El ende- mismo es mayor en la parte norte de la peninsula. Desproporcionadamente po- cas especies de anfibios son endéemicas mientras que culebras y lagartijas estan sobrerrepresentadas entre las endemicas. Las medidas de estructura de vege- tacién en siete sitios diferentes, cada uno ubicado en un tipo de vegetacion dis- tinto, indica que varios parametros de heterogeneidad vegetal se correlacionan con la densidad de las especies de cu- lebras y lagartijas de una manera impor- tante. Esto se basa en lo obtenido medi- ante la estadistica de diversidad de Shannon. Entre los anfibios la cantidad y la periodicidad estacional de precip- itacidn es muy importante. Por lo tanto, parece que los anfibios reaccionan a y estan limitados por factores abidticos. Las culebras y las lagartijas parecen sus- ceptibles a factores bidticos, especial- mente a la heterogeneidad del habitat estructural. En contraste con los resultados de otros estudios sobre la distribucién pen- insular, no existe prueba de que un “efecto peninsular’ que abarca_aisla- miento y distancia desde un punto de origen es importante en moldear las nor- mas de la densidad de especies de an- fibios y reptiles en la peninsula de Yuca- tan. Tampoco es necesario apelar a hi- potesis de periodos evolucionarios 0 eco- Idgicos para explicar las normas obser- vadas en la densidad de dichas especies. La mayoria de las especies de an- fibios y reptiles de la peninsula de Yuca- tan son formas ampliamente distribuidas a lo largo de las zonas Iluviosas del Golfo y las tierras bajas del Caribe. En ambos niveles, genérico y especifico, los anfibios y reptiles de la peninsula in- dican su mas grandes afinidades con los anfibios y reptiles de México y la Amer- ica Central. La fauna xerdfita de la parte norte de la peninsula revela afini- dades con las faunas del oeste y noreste de México. La gran parte de los endé- micos peninsulares parece haber evolu- cionado en el lugar de origen cuando fueron aislados durante el pleistoceno debido a cambios ambientales. Distri- buciones esporadicas dentro de la penin- sula se atribuyen en parte a las altera- ciones de humedad y sequedad en el clima. Las distribuciones disyuntivas que abarcan especies en las zonas des- provistas de bosques se interpretan en términos de influencias antropogenas. LITERATURE CITED ANDREWS, R. M. 1971. Structural habitat and time budget of a tropical Anolis lizard. Ecology, 52: 262-270. ANDREWS, R. M. 1976. Growth rate in island and mainland anoline lizards. Copeia, 1976:477- 482. Bat, I. R. 1975. Nature and formulation of biogeo- graphical hypotheses. Syst. Zool., 24: 407-430. BARBAULT, R. 1976. Notes sur la composition et la di- versité spécifiques d’une herpéto- cenose tropicale (Bouaké, Cote YUCATAN HERPETOFAUNA 43 d'Ivoire) Bulletin de LF.A.N. 38: 446-456. BARONI-URBANI, C. and M. W. BuseEr. 1976. Similarity of binary data. Syst. Zool., 25:251-259. Barrett, A. S. and E. S. BARGHOORN. 1973. Phytogeographic history of the Isth- mus of Panama during the past 12,000 years (A history of vegeta- ation, climate, and sea level change). In Graham A. Vegetation and vege- tational history of northern Latin America. Elsevier Scientific Publish- ing Company. Bartlett, H. H. 1935. A method of procedure for field work in tropical American phytogeography based upon a_ botanical reconnais- sance in parts of British Honduras and the Petén forest of Guatemala. Camegie Instit., Wash., Publ. 408. BEARD, J. S. 1955. The classification of tropical Amer- ican vegetation types. Ecology, 36: 89-100. Bearcieg, K. and C. J. McCoy. 1964. Variation and relationships of the teiid lizard (Cnemidophorus angusti- ceps). Copeia, 1964:561-570. BEQUAERT, J. C. 1933. Botanical notes from Yucatan. In Shattuck, G. C. The peninsula of Yucatan. Carnegie Instit. Washing- ton, Publ. 431. CAMPBELL, J. A. 1976. A new terrestrial pit viper of the genus Bothrops (Reptilia, Serpentes, Crotalidae) from western México. J. Herp., 10:151-160. Cote, L. J. 1910. The caverns and people of northern Yucatan. Bull. Amer. Geogr. Soc. 42.(5):321-336. ConnELL, J. H. and E. Ortas. 1964. The ecological regulation of species diversity. Amer. Natur., 98:399-414. Cook, R. E. 1969. Variation in species density of North American birds. Syst. Zool., 18:63-84. Cootey, W. W. and P. R. Lounges. 1971. Multivariate data analysis. John Wiley and Sons, Inc., New York. Cope, E. D. 1884. Notes on Central American and Méx- ican Batrachia and Reptilia. Proc. Amer. Phil. Soc., 22(1884):171-180. Cottam, G. and J. T. Curtis. 1956. The use of distances measures in phytosociological sampling. Ecology, 37:451-460. Covicu, A. 1976. Recent changes in molluscan species diversity of a large tropical lake (Lago de Petén, Guatemala). Limnol. Oceanogr., 21:51-59. Covicn, A. and M. STurvEr. 1974. Changes in oxygen 18 as a measure of long-term fluctuations in tropical lake levels and molluscan populations. Limnol. Oceanogr., 19:682-691. Croizat, L. 1958. Panbiogeography. Published by the author. Caracas. CroizatT, L. 1962. Space, time, and form: the biolog- ical synthesis. Published by the author. Caracas. CroizaT, L., G. J. NELson, and D. E. RoseEn. Centers of origin and related con- cepts. Syst. Zool., 23:265-284. Crump, M. L. 1974. Reproductive strategies in a tropical anuran community. Misc. Publ. Univ. Kansas Mus. Nat. Hist., 61:1-68. Dixon, W. J. (ed). 1975. Biomedical computer programs. Univ. California Press, Berkeley. DuELLMAN, W. E. 1965a. Amphibians and reptiles from the Yu- catan Peninsula, México. Univ. Kan- sas Mus. Nat. Hist., 15:577-614. DUELLMAN, W. E. 1965b. A biogeographic account of the her- petofauna of Michoacan, México. Univ. Kansas Publ. Mus. Nat. Hist., 15:627-709. DuUELLMAN, W. E. 1978. The biology of an equatorial herpeto- fauna in Amazonian Equador. Misc. Publ. Univ. Kansas Mus. Nat. Hist., 65:1-352. Epwarps, C. 1957. Quintana Roo: Mexico’s empty quar- ter. Report of field work carried out under ONR contract 222(11) NR 388 067. Univ. of California, Dept. Geogr. EMILiAnt, C. and E. Rona. 1969. Caribbean cores P6304-8 and P6304- 9: new analysis of absolute chronol- ogy. A reply. Science, 166:1551- 1552. ENGERRAND, J. and F. UrsIna. 1910. Informe acerca de una excusién geo- légica preliminar afectuada en el es- tado de Yucatan. Bol. Inst. Geol. Mex., Mexico City. Gapow, H. 1905. The distribution of Mexican amphib- ians and reptiles. Proc. Zool. Soc. London, June 6, 1905:191-244. Garcia, E. 1965. Distribucién de la preccipitacién en la Republica Méxicana. Publ. Instit. Geogr. Univ. Nac. Aut. Mex., 1:173- 191. 44 MISCELLANEOUS PUBLICATION GoLpMaAN, E. A. and R. T. Moore. 1945. The biotic provinces of México. J. Mammalogy, 26:347-360. Grisco, L. 1942. Origin and relationships of the fau- nal areas of Central America. Proc. Eighth Amer. Sci. Cong. 3 (Biol. Sci. ):425-430. HaFFER, J. 1969. Speciation in Amazonian forest birds. Science, 165:131-137. HaAFFER, J. 1974. Avian speciation in tropical South America. Nuttall Omith. Club, 14: 1-390. Hacmerer, E. M. 1966. A numerical analysis of the distribu- tional patterns of North American mammals. II. Reevaluation of the provinces. Syst. Zool., 15:279-299. HacMetrer, E. M. and C. D. Stutts. 1964. A numerical analysis of the distribu- tional patterns of North American mammals. Syst. Zool., 13:125-155. Hammon, N., Princ, D., BERGER, R., Swirsur, W. R., and A. P. Warp. 1976. Radiocarbon chronology for early Maya occupation at Cuello, Belize. Nature, 260:579-581. Hartt, R. T., FisHer, H. I, LANGEBARTEL, D. A., and G. W. BRAINERD. 1953. Faunal and archeological researches in Yucatan caves. Cranbrook Instit. Sci., Bull. 33. Herpryy, A. 1891. Geological researches in Yucatan. Proc. Acad. Nat. Sci. Philadelphia, 42:136-158. HENDERSON, R. W. and H. S. Frrecu. 1975. A comparative study of the structural and climatic habitats of Anolis seri- ceus (Reptilia: Iguanidae) and its syntopic congeners at four localities in southern México. Herpetologica, 31:459-471. Heyer, W. R. 1969. The adaptive ecology of the species groups of the genus Leptodactulus (Amphibia: Leptodactylidae ). Evolu- tion, 23:421-428. Howtprince, L. R. 1967. Life zone ecology. Tropical Science Center, San Jose, Costa Rica. HuFFakeEr, C. B. 1958. Experimental studies on predation: dispersion factors and predator-prey oscillations. Hilgardia, 27:343-383. Hurraker, C. B., SHea, K. P., and S. G. HARMAN. 1963. Experimental studies on predation: complex dispersion and levels of food in a acarine predator-prey interac- tion. Hilgardia, 34:305-330. MUSEUM OF NATURAL HISTORY JENSSEN, T. A. 1973. Shift in the structural habitat of Anolis opalinus due to congeneric competition. Ecology, 54:863-869. Koester, A. R. 1971. Species density of North American amphibians and reptiles. Syst. Zool., 20:127-137. Kress, C. J. 1972. Ecology: the experimental analysis of distribution and abundance. Harper and Row, New York. La Fay, H. 1975. The Maya, children of time. Natl. Geogr. 148:729-767. Ler jaG- 1976. Rana maculata Brocchi, an addition to the herpetofauna of Belize. Her- petologica, 32:211-214. liveise 1EMGE 1977. An ecogeographic analysis of the her- petofauna of the Yucatan Peninsula. Ph.D. dissertation. University of Kan- sas. 213 pp. LEOPOLD, A. S. 1950. Vegetation zones of México. Ecology, 31:507-518 LUNDELL, C. L. 1934. Preliminary sketch of the phytogeog- raphy of the Yucatan Peninsula. Car- negie Instit. Washington, Publ. 436. LUNDELL, C. L. 1937. The vegetation of Petén. Carnegie Instit. Washington, Publ. 478. Lyncu, J. D. 1964. Two new species of Eleutherodac- tylus from México (Amphibia: Lep- todactylidae). Herpetologica, 20: 246-252. Lynts, G. W. and J. B. Jupp. 1971. Late Pleistocene paleotemperatures at Tongue of the Ocean, Bahamas. Sci- ence, 171:1143-1144. MacArtuur, R. H. 1964. Environmental factors affecting bird species diversity. Amer. Natur., 98: 387-398. MacArtuur, R. H. and J. W. MacArtuur. 1961. On bird species diversity. Ecology, 42:594-598. MacArtuur, R. H. and E. O. Witson. 1967. The theory of island biogeography. Monog. Pop. Biol. 1:1-203. Martin, P. S. 1958. A biogeography of reptiles and am- phibians in the Gomez Farias region, Tamaulipas, México. Misc. Publ. Mus. Zool., Univ. Michigan, 1-102. MaTHEny, R. T. 1976. Maya lowland hydraulic systems. Sci- ence, 193:639-646. Mercer, H. C. 1896. The hill-caves of Yucatan; a search YUCATAN HERPETOFAUNA 45 for evidence of man’s antiquity in the caverns of Central America, being an account of the Corwith Expedition of the Department of Archaeology and Palaeontology of the University of Pennsylvania, J. B. Lippincott, Phila- delphia. MULLER, P. 1973. The dispersal centers of terrestrial vertebrates in the Neotropical realm. Biogeographica II. W. Junk, Publish- ers, The Hague. NELSON, G. J. 1973. Comments of Leon Croizat’s bioge- ography. Syst. Zool., 22:312-320. Opum, E. P. 1959. Fundamentals of ecology (2nd ed.). Saunders, Philadelphia. Owen, L. H. 1929. The geology of British Honduras. The Clarion, Ltd., Belize, British Honduras. PaIneE, R. T. 1966. Food web complexity and species di- versity. Amer. Natur., 100:65-75. Paine, R. T. 1969. The Pisaster-Tequla interaction: prey patches, predator food preference, and intertidal community structure. Ecology, 50:950-961. RAGE: Je L: 1933. The climate of Yucatan. In Shattuck, G. D. The peninsula of Yucatan. Carnegie Instit. Washington, Publ. 431. PacE, J. L.. 1938. The climate of Petén. In Morley, S. G. The inscriptions of Petén. Car- negie Instit. Washington, Publ. 437, vol. 4. PAYNTER, Jr. R. A. 1955. The ornithogeography of the Yucatan Peninsula. Peabody Mus. Nat. Hist., Bull., 9: 1-347. Pearse, A. S. 1938. Fauna of the caves of Yucatan. Car- negie Instit. Washington, Publ. 491. PETERS, J. A. 1960. The snakes of the subfamily Dipsa- dinae. Misc. Publ. Mus. Zool., Univ. of Michigan, 114: 1-224. PETERS, J. A. 1971. A new approach in the analysis of biogeographical data. Smithsonian Contr. Zool., 107:1-28. PIANKA, E. R. 1966a. Latitudinal gradients in species diver- sity: a review of concepts. Amer. Natur., 100:33-46. Pranka, E. R. 1966b. Convexity, desert lizards and spatial heterogeneity. Ecology, 42: 1055-1059. PIANKA, E. R. 1967. Lizard species diversity. Ecology, 48: 333-351. PIANKA, E. R. 1971. Lizard species density in the Kalahari desert. Ecology, 52:1024-1029. PIANKA, E. R. 1974. Evolutionary ecology. Harper and Row, New York. PranKA, E. R. and R. B. Huey. 1971. Bird species density in the Kalahari and the Australian deserts. Koedoe, 14:123-130. Porter, J. W. 1972. Predation by Acanthaster and its ef- fect on coral species diversity. Amer. Natur., 106:487-492. REcHER, H. F. 1969. Bird species diversity and habitat di- versity in Australia and North Amer- ica. Amer. Natur., 103:75-80. RIcK.LeEFs, R. E. 1973. Ecology. Oregon. RoHLF, F. J. 1970. Adaptive hierarchical _ clustering schemes. Syst. Zool., 19:58-82. Rosen, D. E. 1974. Review of Space, Time, and Form: the biological synthesis. Syst. Zool., 23:288-290. Rosen, D. E. 1975. A vicariance model of Caribbean bio- geography. Syst. Zool., 24:431-464. RosENZWEIG, M. L. and J. WINAkwR. 1966. Population ecology of desert rodent communities: habitats and environ- mental complexity. Ecology, 50:558- Die RossMAN, D. A. and G. C. SCHAEFER. 1974. Generic status of Opheodrys mayae, a colubrid snake endemic to the Yu- catin Peninsula. Occas. Pap. Mus. Zool., Louisiana St. Univ., 45:1-12. ROTRAMEL, G. L. 1973. The development and application of the area concept in biogeography. Syst. Zool., 22:227-232. SApPER, K. 1937. Mittelameria. In Handb. Reg. Geol., 8(4):1-160. SAVAGE, J. 1966. The origins and history of the Cen- tral America herpetofauna. Copeia, 1966:719-766. SCHOENER, T. W. 1967. The ecological significance of sexual dimorphism in size in the lizard Anolis conspersus. Science, 155:474- ATT. SCHOENER, T. W. 1969. Size patterns in West Indian Anolis Chiron Press, Portland, 46 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY lizards: I. Size and species diversity. Syst. Zool., 18:386-401. SCHOENER, T. J. and A. SCHOENER. 197la. Structural habitats of West Indian Anolis lizards I. Lowland Jamaica. Breviora, 368: 1-53. SCHOENER, T. W. and A. SCHOENER. 1971b. Structural habitats of West Indian Anolis lizards II. Puerto Rican up- lands. Breviora, 375:1-39. SCHUCHERT, C. 1935. Historical geology of the Antillean- Caribbean region. John Wiley and Sons, New York. SEIBERT, E. A., H. B. Littywuire, and R. WASSERSUG. 1974. Cranial coossification in frogs: rela- tionship to rate of evaporative water loss. Physiol. Zool., 47:261-265. SHANNON, C. E. and W. WEAVER. 1948. The mathematical theory of commu- nication. Univ. Illinois Press, Urbana. SIMPSON, G. G. 1964. Species density of North American Recent mammals. Syst. Zool., 13:57- WSs SmirH, H. M. 1940. An analysis of the biotic provinces of México, as indicated by the distribu- tion of the lizards of the genus Scelo- porus. Los Anales de la Escuela Na- cional de Ciencias Bioldégicas, 2(1): 95-102. SmirH, H. M. 1949. Herpetogeny in México and Guate- mala. Ann. Assoc. Amer. Geog., 39(3):219-238. SneaTH, P. H. A. and R. R. SOKAL. 1973. Numerical taxonomy. W. H. Free- man and Co., San Francisco. Stuart, L. C. 1934. A contribution to a knowledge of the herpetological fauna of El] Petén, Guatemala. Occ. Pap. Mus. Zool., Univ. Michigan, 292:1-18. Stuart, L. C. 1935. A contribution to a knowledge of the herpetology of a portion of the sa- vanna region of central Petén, Guate- mala. Misc. Publ. Mus. Zool., Univ. of Michigan, 29:1-56. Stuart, L. C. 1943. Taxonomic and geographic comments on Guatemalan salamanders of the ge- nus Oedipus. Misc. Publ. Mus. Zool., Univ. of Michigan, 56:1-38. Stuart, L. C. 1958. A study of the herpetofauna of the Uaxactin-Tikal area of northern El Petén, Guatemala. Contr. Lab. Vert. Biol., Univ. of Michigan, 75:1-30. Shoieae, IL, (CG. 1963. A check-list of the herpetofauna of Guatemala. Misc. Publ. Mus. Zool., Univ. of Michigan, 122:1-150. Stuart, L. C. 1964. Fauna of Middle America. In Wau- chope, R. and R. C. West (eds.). Handbook of Middle American In- dians. Univ. Texas Press. Stuart, L. C. 1966. The environment of the Central American cold-blooded vertebrate fauna. Copeia, 1966:684-699. Tamayo, J. L. 1964. The hydrography of Middle America. In Wauchope, R. and R. C. West (eds. ). Handbook of Middle Amer- ican Indians. Univ. of Texas Press. Taytor, E. H. 1935. Lizards of the genus Eumeces. Kan- sas Univ. Sci. Bull., 23: 1-643. THomMpson, E. H. 1897. Cave Loltun, Yucatan. Peabody Mus. Mem., 1:51-70. TomoFfr, C. S. 1974. Avian species diversity in desert scrub. Ecology, 55:396-403. TRUEB, L. 1970. Evolutionary relationships of casque- headed tree frogs with co-ossified skulls (family Hylidae). Univ. of Kansas Publ., Mus. Nat. Hist., 18: 547-716. TuRNER, B. L. 1974. Prehistoric intensive agriculture in the Mayan lowlands. Science, 185: 118-124. UETz, G. 1974. Species diversity: a review. The Biologist, 56:111-129. VAN DER HAMMEN, T. 1974. The Pleistocene changes of vegetation and climate in tropical South Amer- ica. J. Biogeog. 1:3-26. VANZOLINI, P. E. and E. E. WiLiiaMs. 1970. South American anoles: the geo- graphical differentiation and evolu- tion of the Anolis chrysolepis species group (Sauria: Iguanidae). Arqui- vos de Zoologia 19( 1-4): 1-298. Vinson, G. L. and J. H. BrinEMAN. 1963. Nuclear Central America, hub of An- tillean transverse belt. In Childs, O. E. and B. W. Beebe (eds.). Back- bone of the Americas. Assoc. Petr. Geo., Tulsa, Oklahoma. Vivo Escoro, J. A. 1964. Weather and climate of México and Central America. In Wauchope, R. and R. C. West (eds.). Handbook of Middle American Indians. Univ. of Texas Press. WaDELL, H. 1938. Physical-geographical features of the Petén, Guatemala. In Morley, S. G. YUCATAN HERPETOFAUNA 47 The inscriptions of Petén. Carnegie Instit. Washington, Publ. 437, vol. 4. Wacne_enR, P. L. 1964. Natural vegetation of Middle Amer- ica. In Wauchope, R. and West, R. C. (eds.). Handbook of Middle American Indians. Univ of Texas Press. Wake, D. B. and J. F. LyNncu. Yucatan Peninsula. 1976. The distribution, ecology, and evolu- tionary history of plethodontid sala- manders in tropical America. Nat. Hist. Mus. Los Angeles Co. Sci. Bull. 25:1-65. WEsT, R. C. 1964. Surface configuration and associated geology of Middle America. In Wau- chope, R. and R. C. West (eds.). Handbook of Middle American In- dians. Univ. of Texas Press. WuitTAker, R. H. 1970. Communities and ecosystems. Mac- millan, New York. WIitey, E. O. 1976. APPENDIX Plates 1-27 summarize the distribu- tions of amphibians and reptiles in the The index below lists the taxa in alphabetical order by genus. The star on the map for Lepi- dophyma flavimaculatum (Plate 15) in- dicates the fossil record of Lepidophyma arizeloglyphus. Question marks indicate doubtful records. INDEX TO GENERA AND SPECIES Taxon Plate No. Adelphicos quadrivirgatus 16 Agalychnis callidryas 3 A. moreleti 4 Agkistrodon bilineatus 26 Ameiva chaitzami 14 A. festiva 14 A. undulata 14 Amistridium veliferum 16 Anolis biporcatus 9 A. capito 10 A. lemurinus 10 A. pentaprion 10 A. rodriquezi 10 A. segrei 10 A. sericeus 10 A. tropidonotus 11 A. uniformis 11 Aristelliger georgeensis 9 Basiliscus vittatus Wa Boa constrictor 15 Bothrops asper 26 B. nasutus 26 B. nummifer 26 B. schlegeli 26 B.. yucantanicus PAT Bolitoglossa dofleini 1 B. mexicana 1 B. rufescens i B. yucatana 1 Lepisosteidae). Misc. Publ. Univ. Kansas, 64:1-111. Taxon Plate No. Bufo marinus 3 B. valliceps 3 Celestus rozellae 15 Centrolenella fleischmanni 3 Chelydra serpentina 6 Chrysemys scripta 7 Claudius angustatus 6 Clelia clelia 16 C. scytalina 16 Cnemidophorus angusticeps 14 C. cozumela 14 C. deppei 15 C. rodecki 15 Coleonyx elegans 8 Coluber constrictor 16 Coniophanes bipuncatus 16 C. fissidens Lif C. imperialis 17 C. meridanus li C. quinquevittatus V7 C. schmidti 17 Conophis lineatus U7 Corytophanes cristatus 11 C. hernandezi us Crocodylus acutus C. moreleti Crotalus durissus Ctenosaura similis Dendrophidion vintor Dermatemys mawii Dipsas brevifaces Drymarchon corais Drymobius margaritiferus Elaphe flavirufa E. triaspis Eleutherodactylus alfredi laticeps E by ts bs hy loki . rostralis . rugulosus . yucantanensis nyaliosaurus defensor Eumeces schwartzei The phylogeny and biogeography of fossil and recent gars (Actinopterygii: 48 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Taxon E. sumichrasti Ficimia publia Gastrophryne elegans Geophis carinosus Hemidactylus turcicus Hyla ebraccata H. loquax H. microcephala H. picta H. staufferi Hypopachus variolosus Iguana iguana Imantodes cenchoa I. gemmistratus I. tenuissimus Kinosternon acutum K. creaseri K. leucostomum K. scorpioides Laemanctus longipes L. serratus Lampropeltis triangulum Lepidophyma flavimaculatum L. mayae Leptodactylus labialis L. melanonotus Leptodeira frenata L. septentrionalis Leptophis ahaetulla L. mexicanus Leptotyphlops phenops Mabuya brachypoda Masticophis mentovarius Mastigodryas melanolomus Micrurus diastema M. hippocrepis M. nigrocinctus Natrix rhombifera Ninia diademata N. sebae Oedipina elongata Oxybelis aeneus O. fulgidus Oxyrhopus petola Physalaemus pustulosus Plate No. 13 18 5 19 NW e ell sell aed ell coal [pall eel alll eee SCOONNOUSMONNAYAAADODOONUAAH AK PO Taxon Phrynohyas venulosa Phyllodactylus tuberculosus Pliocercus andrewsi P. elapoides Pseustes poecilonotus Rhinophrynus dorsalis Rana maculata R. palmipes R. pipiens Rhadinaea decorata Rhinoclemys areolata Scaphiodontophis annulatus Sceloporus chrysostictus S. cozumelae S. lundelli S. serrifer S. teapensis Sibon dimidiata S. nebulata S. sanniola Smilisca baudinii S. cyanosticta Sphaerodactylus glaucus S. lineolatus Sphenomorphus cherriei Spilotes pullatus Staurotypus triporcatus Stenorrhina degenhardti S. freminvillei Symphimus mayae Syrrhophus leprus Tantilla canula T. cuniculator T. moesta T. schistosa Tantillita lintoni Terrapene mexicana Thamnophis marcianus T. proximus Thecadactylus rapicauda Tretanorhinus nigroluteus Triprion petasatus Tropidodipsus fasciata T. sartorii Typhlops microstomus Xenodon rabdocephalus Plate No. a Solifoglossa yucatana e Bolitoglessa dofleint_ 4 i) 100 ——=—=_ Kilometers 600 moter contour YUCATAN HERPETOFAUNA 92 —— Bolitoglossa rufescens Kilometers 600 meter contour Fhinophrynus dorsalis Kilometers 600 meter contour Oedpina elongata Kilometers 600 meter contour Kilometers 600 meter contour PLATE 1 50 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Kilometers 600 meter contour 21 to) 100 200 —— Kilometers 600 meter contour 90 T T Leptodactylus labralis 21;— ° 100 200 Kilometers 600 meter contour LS PLATE 2 Kilometers 600 meter contour 92 90 see ee ( lon Ps o [ications rugulosus Kilometers 600 meter contour Kilometers 600 meter contour LO 2 okey CEA YUCATAN HERPETOFAUNA Kilometers 600 moter contour she 1 A Kilometers 600 meter contour 92 90 calle ls T Centro/enella fleischmann! Kilometers 600 meter contour Kilometers 600 meter contour Kilometers 600 meter contour PLATE 3 51 52 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Agalychnis moreleti Kilometers ¢ Kilometers 600 meter contour 600 meter contour Hyla /oquax Kilometers Kilometers 600 meter contour 600 meter contour Kilometers 4 Kilometers 600 meter contour 600 meter contour PLATE 4 YUCATAN HERPETOFAUNA Smilisca cyanosticta Kilometers C Kilometers 600 meter contour 600 meter contour 4 S Ta ee 92 90 88 T T 21 Kilometers Kilometers i t 600 meter contour 600 meter contour 92 90 88 PLATE 5 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Frana maculata a = to) 100 200 Kilometers 600 meter contour Frana piplens Kilometers 600 meter contour Frana pa/mipes Kilometers 600 meter contour Kilometers 600 meter contour Claudius angustatus t) 100 200 ——— Kilometers 600 meter contour PLATE 6 74) = 100 200 Kilometers 600 meter contour YUCATAN HERPETOFAUNA me 4 kinosternon /eucostomum __ Kilometers 600 moter contour Kilometers Kilometers meter contour ete) 600 meter contour Staurotypus triporcatus Kilometers Kilometers ~) 600 meter contour z 600 meter contour PLATE 7 55 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 90 — Fhinoclemys areolata Kilometers 600 meter contour Kilometers 600 meter contour Kilometers 600 meter contour 92 ——= Terraqpene mexicana Kilometers 600 meter contour Kilometers 600 meter contour Kilometers 600 meter contour PLATE 8 YUCATAN HERPETOFAUNA Sphaerodactylus lineolatus 9/8 aa ae Kilometers 600 moter contour Kilometers 600 meter contour Thecadactylus rapicauda Kilometers ~) 600 meter contour e *% Phyllodactylus tuberculosus_ "= Kilometers 600 meter contour 92 PLATE 9 —— Anolis byporcatus Kilometers 600 meter contour \ 57 58 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Kilometers 600 meter contour Kilometers 600 meter contour Kilometers 600 meter contour Tr Kilometers 600 meter contour Kilometers 600 meter contour PLATE 10 Kilometers 600 meter contour YUCATAN HERPETOFAUNA 92 T Anolis tropidonotus A Kilomotors 600 moter contour Basiliscus vittatus Kilometers Kilometers 600 meter contour 600 meter contour Kilometers Kilometers 600 meter contour ano mate exten PLATE 11 60 21 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Kilometers 600 meter contour 21 Kilometers 600 meter contour poe ay \ 92 Kilometers 600 meter contour 21 Laemanctus serratus Se e v @ O t) e @ @ @ (:) 100 200 e ( 600 meter contour t) 88 Kilometers 600 meter contour PLATE 12 Sceloporus cozumelae Kilometers 600 meter contour Sceloporus lundelli Kilometers 600 meter contour Sceloporus teapensis Kilometers 600 meter contour Eumeces sumichrasti Kilometers ~) 600 meter contour YUCATAN HERPETOFAUNA Eumeces schwartze/ Kilometers 600 meter contour Kilometers ~) 600 meter contour PLATE 13 61 62 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Amerva festiva Kilometers 600 meter contour PLATE Kilometers 600 meter contour 21}— 0 100 200 Kilometers 600 meter contour ae | Cnemidophorus cozumela Kilometers 600 meter contour 14 YUCATAN HERPETOFAUNA © Cnemidophorus deppel 4 Cnemidophorus rodeck! Ce Kilomoters 600 meter contour Lepidophyma mayae Kilometers 600 meter contour Boa constrictor Kilometers 600 meter contour PLATE 15 Celestus rozellae Kilometers 600 meter contour Kilometers 600 meter contour 63 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Typhlops microstomus Kilometers Kilometers 600 meter contour 600 meter contour Amistridum veliferum :j © Clelia clelia 4 Clelia scytalina Kilometers Kilometers 600 meter contour - 600 meter contour Coluber constrictor Kilometers Kilometers 600 meter contour - 600 meter contour PLATE 16 YUCATAN HERPETOFAUNA a Conophanes meridanus : : 3 ire qT 0 m “iin © Coniophanes fissidens_ a4 =< Coniophanes imperialism , we e Kilometers 600 meter contour 600 meter contour Y Comnophanes quinquevittatus Kilometers Kilometers 600 meter contour 600 meter contour Conoptis lineatus Kilometers Kilometers ~) 600 meter contour 600 meter contour PLATE 17 66 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Kilometers 600 meter contour Kilometers 600 meter contour Drymobius fi aes acon e ee Kilometers 600 meter contour 92 90 88 Drymarchon corals Kilometers 600 meter contour (> (Wr lon i >> Ae ee ae Kilometers 600 meter contour Kilometers 600 meter contour PLATE 18 YUCATAN HERPETOFAUNA Kilometors 600 meter contour /mantodes gemmistratus Kilometers 600 meter contour 19 a ma 4 : a t ee t e Kilometers 600 meter contour PLATE 19 /mantodes cenchoa /mantodes fenuissimus Kilometers 600 meter contour Leptodeira frenata Kilometers 600 meter contour 67 68 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Leptodeira septentrionalis Kilometers Kilometers 600 meter contour 600 meter contour Leptophis mexicanus Masticophis mentovarius Kilometers Kilometers 600 meter contour 600 meter contour Natrix rhombifera Kilometers Kilometers 600 meter contour 600 meter contour PLATE 20 Kilometers 600 moter contour Oxybelis aeneus Kilometers 600 meter contour Oxyrhopus petola Kilometers >, 600 meter contour YUCATAN HERPETOFAUNA Nina sebae Oxybelis fulgidus Kilometers 600 meter contour Pliocercus andrews! Kilometers 600 meter contour PLATE 21 69 70 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 90 Pliocercus elapoides CA = Kilometers 600 meter contour | etecomens Frhadinaea decorata 21 = Kilometers 600 meter contour Pseustes poecilonotus Kilometers 600 meter contour LP H as "36 Me \ lee ; Scaphiodontoplis annulatus Co) 100 200 — Kilometers 600 meter contour 21;— to) 100 200 Kilometers 600 meter contour 92 — Sibon nebulata Kilometers 600 meter contour PLATE 22 YUCATAN Sibon sanniola 100 200 ——— Kilometers 600 motor contour e e 1 ee Kilometers 600 meter contour Symphimus mayae Kilometers 600 meter contour HERPETOFAUNA 7 100 Kilomater 600 meter contour Stenorrhina treminville) — "=~ ah e ¢ e een ee Kilometers 600 meter contour Kilometers 600 meter contour PLATE 23 72 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Tantilla cuniculator Tantilla moesta Kilometers Kilometers 600 meter contour 600 meter contour Tantilla schistosa S Tantillita lintoni Kilometers Kilometers 600 meter contour 600 meter contour 2 , aul 9 2 : Thamnophis marcianus ere Thamnophis proximus 100 Kilometers Kilometers 600 meter contour 600 meter contour PLATE 24 YUCATAN HERPETOFAUNA Tretanorhinus nigroluteus Kilometers 600 meter contour Tropidodipsas sartoril Xenodon rabdocephalus Kilometers Kilometers 600 meter contour 600 meter contour Micrurus diastema Kilometers Kilometers —) 600 meter contour 600 meter contour PLATE 25 73 74 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 90 I Micrurus mgrocinctus Agkistrodon bilineatus Kilometer: jometers, Kilometers 600 meter contour 600 meter contour Kilometers Kilometers 600 meter contour 600 meter contour S se 0 CI a nee. SOR Orr c 92 Sle Bothrops nummifer 0 100 200 aS! Kilometers Kilometers $ 600 meter contour 600 meter contour 19 19 wf oN = Vong 5 ) om Hy BIDS 6 36 OF fi 4 < 5 } \ . lat q : Ary =X ° = 17 17 S& e as 92 90 PLATE 26 YUCATAN HERPETOFAUNA 75 92 —— Bothrops yucaranicus mo en Crotalus durissus _ Kilometers 600 moter contour Herp L655 -L43 od ey phic analysis of the e i UNA Vl 3 2044 0 Date Due RECENT MISCELLANEOUS PUBLICATIONS UNIVERSITY OF KANSAS MUSEUM OF NATURAL HISTORY*® 52. Reproductive cycles in lizards and snakes. By Henry S. Fitch. Pp. 1-247, 16 fig- ures in text. June 19, 1970. Paper bound, $5.00. 53. Evolutionary relationships, osteology, and zoogeography of leptodactyloid frogs. By John D. Lynch. Pp. 1-238, 181 figures in text. June 30, 1971. Paper bound, $7.00. 54. The dentition of glossophagine bats: development, morphological characteristics, variation, pathology, and evolution. By Carleton J. Phillips. Pp. 1-138, 49 figures in text. September 24, 1971. Paper bound, $5.00. 55. Middle American lizards of the genus Ameiva (Teiidae) with emphasis on geo- graphic variation. By Arthur C. Echternacht. Pp. 1-86, 28 figures in text. De- cember 14, 1971. Paper bound, $3.00. 56. Systematics of the chiropteran family Mormoopidae. By James Dale Smith. Pp. 1-132, 40 figures in text. March 10, 1972. Paper bound, $5.00. 57. A systematic review of the Teiid lizards, genus Bachia, with remarks on Heter- odactylus and Anotosaura. By James R. Dixon. Pp. 1-47, 15 figures in text. February 2, 1973. Paper bound, $1.50. 58. Systematics of three species of woodrats (genus Neotoma) in central North Amer- ica. By Elmer C. Birney. Pp. 1-173, 44 figures in text. April 18, 1973. Paper bound, $7.50. 59. Systematics and evolution of the Andean lizard genus Pholidobolus (Sauria: Teiidae). By Richard R. Montanucci. Pp. 1-52, 8 figures in text. May 14, 1973. Paper bound, $1.75. 60. Mammals of the Black Hills of South Dakota and Wyoming. By Ronald W. Turner. Pp. 1-178, 14 figures in text. April 3, 1974. Paper bound, $7.50. 61. Reproductive strategies in a tropical anuran community. By Martha L. Crump. Pp. 1-68, 13 figures in text. November 15, 1974. Paper bound, $4.50. 62. A demographic study of the ringneck snake (Diadophis punctatus) in Kansas. By Henry S. Fitch. Pp. 1-53, 19 figures in text. April 3, 1975. Paper bound, $3.00. 63. Morphology of the bony stapes (columella) in the Passeriformes and related groups: evolutionary implications. By Alan Feduccia. Pp. 1-34, 16 plates, 7 figures in text. May 30, 1975. Paper bound, $3.00. 64. The phylogency and biogeography of fossil and Recent gars (Actinopterygii: Lepisosteidae). By E. O. Wiley, Pp. 1-112, 72 figures in text. November 12, 1976. - Paper bound, $2.75. 65. The biology of an equatorial herpetofauna in Amazonian Ecuador. By William E. Duellman. Pp. 1-352, 198 figures in text. August 30, 1978. Paper bound, $15.00. * Postage and handling 5% of total order or a minimum of $0.50. Publications of this and other series are available from the Publications Secretary. oe