Six species that cross the isthmus live on the humid Gulf lowlands and on the humid lowlands of Chiapas and Guatemala, but not on the semi-arid Plains of Tehuantepec; these include _Bolitoglossa occidentalis_, _Eleutherodactylus rhodopis_, _Microbatrachylus pygmaeus_, _Phrynohyas modesta_, _Phrynohyas spilomma_, and _Rana palmipes_. Of these, _Microbatrachylus pygmaeus_ also occurs in scattered humid environments to the west of the isthmus on the Pacific lowlands.

Two species are endemic to the isthmian region. _Bolitoglossa veracrucis_ is known only from the humid northern slopes of the isthmus. _Hylella sumichrasti_ occurs on the Pacific slopes of the isthmus and extends to the east into western Chiapas.

In a.n.a.lyzing the distribution of the amphibians with respect to those that are restricted to either the Pacific or Gulf lowlands or those that cross the continental divide in the isthmus, we find that 25 per cent of the species are restricted to the Gulf lowlands, 17 per cent are restricted to the Pacific lowlands, and 53 per cent cross the isthmus. In a.n.a.lyzing the distribution patterns with respect to those that extend across the isthmus of Tehuantepec from east to west, we find that 14 per cent of the species do not extend east of the isthmus into Central America and that 19 per cent do not range west of the isthmus into Mexico proper; 61 per cent of the species range to the east and to the west of the isthmus. Of the 36 species of amphibians inhabiting the isthmus only nine species (25 per cent) range across the isthmus, that is, occur on the Gulf and Pacific lowlands, and also range to the east and to the west of the isthmus. To these wide-ranging species the diversified environments of the isthmus do not present a barrier to distribution. The other 27 species (75 per cent) either do not cross the isthmus from east to west or from north to south; thus, probably in one way or another the isthmus presents a barrier to their distribution.

THE AMPHIBIAN FAUNA OF THE FOOTHILLS AND ADJACENT HIGHLANDS

To amphibians inhabiting the foothills and mountains of southern Mexico and northern Central America, the isthmus presents a great barrier to dispersal. For example, salamanders of the genus _Thorius_, the _mexica.n.u.s_ and _augusti_ groups of the genus _Eleutherodactylus_, the _bistincta_ group of the genus _Hyla_, and the genus _Tomodactylus_ occur on the Mexican Plateau and southward into the mountains of Oaxaca. Nevertheless, no members of these groups are present in the Guatemalan-Chiapan Highlands. The genera _Chiropterotriton_, _Magnadigita_, _Pseudoeurycea_, and _Ptychohyla_, as well as the _eximia_ group of _Hyla_ are represented by different species in the Guatemalan-Chiapan Highlands than in the mountains of Mexico on the other side of the isthmus. Several species of _Plectrohyla_ occur in the Guatemalan-Chiapan Highlands, but none is known from the Mexican Highlands, although one species occurs in the Tuxtlas.

Living in the humid forests of the foothills are salamanders of the genus _Lineatriton_, frogs of the _spatulatus_ group of _Eleutherodactylus_, _Anotheca coronata_, _Hyla miotympanum_, and _Phyllomedusa moreleti_. All of these occur in the foothills of the Sierra Madre Oriental in eastern Mexico and in Los Tuxtlas.

_Lineatriton_, _Hyla miotympanum_, and the _spatulatus_ group of _Eleutherodactylus_ do not occur in the foothills of the Guatemalan-Chiapan Highlands; those amphibians reach the end of their ranges at the isthmus. _Phyllomedusa moreleti_ and _Anotheca coronata_ are found in the northern foothills of the Guatemalan-Chiapan Highlands, and _Phyllomedusa moreleti_ is found in the foothills on the Pacific slopes of the Chiapan Highlands.

Although the above a.n.a.lysis is not so detailed as that of the lowland inhabitants, it does show that all of the genera and species of amphibians known to inhabit the foothills and highlands adjacent to the isthmus, only two species of amphibians cross the isthmus from one highland ma.s.s to the other. Thus, it is evident that the Isthmus of Tehuantepec presents a great barrier to dispersal of these groups of amphibians.

ESTABLISHMENT OF PRESENT PATTERNS OF DISTRIBUTION

From the foregoing a.n.a.lysis of geographical and ecological distribution in the Isthmus of Tehuantepec we may strive for an interpretation of the events that led to the establishment of patterns of distribution displayed not only by the amphibians, but other terrestrial vertebrates as well. The thesis that I am proposing below is based on the premise that in southern Mexico and northern Central America climatic fluctuation during the Pleistocene was of sufficient magnitude to cause vegetational shifts, both vertically and lat.i.tudinally, resulting in the establishment of alternating continuous and discontinuous lowland and highland environments, although this climatic fluctuation was not so great as to eliminate tropical lowland environments from the region.

I feel that the present patterns of distribution of the amphibians in the Isthmus of Tehuantepec may be explained on this premise.

Many authors dealing with the herpetofauna of Middle America have followed Schuchert's (1935) suggestion of a seaway in the isthmus during the Cenozoic. Thus, Burt (1931), Duellman (1956, 1958a), Gloyd (1940), Oliver (1948), Smith and Laufe (1946), and Stuart (1941) employed the presence of a seaway to explain distribution and speciation in various genera. Durham, Arellano, and Peck (1952), Olson and McGrew (1941), and Stirton (1954) have provided geological evidence that there probably was no Cenozoic seaway in the Isthmus of Tehuantepec. Even if there were a seaway in the Pliocene or Miocene (the dating of this possible seaway is open to question), its presence is not necessary to explain the present patterns of distribution in the isthmus.

In recent years the study of natural biotic environments, palynology, and Pleistocene chronology in Middle America has produced a wealth of data, which although still fragmentary begins to form a picture of past climatic events in that part of the world. Sedimentary studies by Hutchinson, Patrick, and Deevey (1956) and Sears, Foreman, and Clisby (1955) have provided evidence of drastic climatic shifts in Mexico during the Pleistocene. Further evidence of bioclimatic fluctuation is provided by Martin and Harrell (1957) and Martin (1958); the latter has suggested that there was a displacement of the tropical zones in southern Mexico and northern Central America by as much as 3000 feet during the glacial maximum. Much of the evidence of such drastic vertical shifts in environments is based on the presence of Pleistocene montane glaciers on Mexican volcanoes (White, 1956) and Chirripo in Costa Rica (Weyl, 1955). Dorf (1959) supports this idea of drastic climatic change.

In his studies of the avifauna of Mexico and Guatemala Griscom (1932 and 1950) made an important issue of the continuity of the bird fauna in what he called the Subtropical Life-zone, which essentially consists of cloud forest, a widespread, but discontinuous, habitat on the Gulf (windward) slopes of the Mexican and Central American highlands at elevations between 1000 and 2000 meters. To account for this apparent uniformity in the avifauna Griscom hypothesized a continuity of cloud forest environment in the Pleistocene; this would result in the depression of cloud forests to the coastal lowlands and the displacement of tropical lowland environments far to the south in Central America. Stuart (1951) objected to this displacement of lowland tropical rainforest; he stated that a descent to sea level of a subtropical zone would have brought about either widespread extermination of the tropical fauna or acclimatization of that fauna to subtropical conditions.

Although palynological studies and some faunal studies of subtropical and temperate animals suggest a drastic climatic fluctuation that might have eliminated tropical environments in southern Mexico and northern Central America, there is much biological evidence indicating the existence of tropical environments in this region even during the glacial maximum. Especially significant is the diversity of species inhabiting the present tropical environments; many of these have differentiated from related taxa to the south.

In the Pleistocene, climate fluctuated and vegetation shifted correspondingly in southern Mexico and northern Central America. Most of the palynological studies and many studies of Pleistocene chronology deal with montane regions, either the Mexican Plateau or the mountains rising from the plateau. No such studies have been made in lowland tropical environments. During glacial advances the tropical lowland environments in Mexico probably were not eliminated, for the great diversity of animals in these environments supports the hypothesis that they have been in existence for some time, although periodically they may have been discontinuous.

In order to understand the nature of bioclimatological events in the Pleistocene in lowland tropical environments of southern Mexico, certain factors that are of little importance in the interpretation of Pleistocene chronology in the highlands must be considered. These factors are: 1) climatic moderation by oceans, 2) fluctuation in sea level, and 3) fluctuation in level of the water table as affected by sea level.

It is well-known that large bodies of water moderate the temperature on adjacent land. Furthermore, it is known that faunas of marine invertebrates shifted lat.i.tudinally in the Pleistocene; Trask, Phleger, and Stetson (1947) recorded cold-water Foraminifera then as far south as the Sigsbee Deep in the middle of the Gulf of Mexico. Large bodies of warm water, such as the Gulf of Mexico, Caribbean Sea, and Pacific Ocean of today, probably were not sufficiently cooled at the time of glacial advance to affect greatly the temperature of the winds blowing across them. Even if these bodies of water were somewhat cooler than now, the prevailing winds blowing from them onto the lowlands of Mexico and northern Central America would have aided in maintaining relatively high temperatures there. These warm winds probably counteracted the cooling effect of glaciation in the lowlands and thereby maintained tropical conditions near the seas.

Although no adequate studies of Pleistocene beach lines have been made in southern Mexico, such information is available for peninsular Florida on the other side of the Gulf of Mexico (Cooke, 1945).

Fluctuation in sea level in the Pleistocene has been used by Hubbell (1954), Goin (1958), and Duellman and Schwartz (1958) to explain present patterns of distribution of animals in Florida. If Cooke's interpretations can be applied to the western side of the Gulf of Mexico, even generally, it would be supposed that sea level varied from about 300 feet lower than at present during the Illinoian Glacial Period to about 275 feet higher than at present during the Aftonian Interglacial Period. Lowering of sea level would expand the lowlands in the isthmus; rising sea level would restrict them, leaving only the central ridges and many islands in the isthmus, but never forming a seaway between the Gulf of Mexico and the Pacific Ocean.

Probably the level of the water table in the coastal lowlands and the gradients of the streams in the lowlands and foothills was closely correlated with fluctuation in sea level. If sea level fluctuated as much as 575 feet in the Pleistocene, changes in the level of the water table must have been of considerable magnitude.

During times of glacial advances the lowlands of the isthmus probably were more extensive and had more semi-arid tropical environments than at present, with patches of rainforest existing in sheltered valleys along the major streams. In the course of bio-climatic fluctuation the semi-arid environments (scrub forest and/or savanna) were continuous at times from the Pacific lowlands across the isthmus to the Gulf lowlands. At those times such typical inhabitants of the semi-arid environments as _Rhinophrynus dorsalis_, _Engystomops pustulosus_, and _Hyla staufferi_ could have made their way across the isthmus. At times of most extensive glaciation, such as the Illinoian, temperatures in the isthmus probably were low enough to permit the growth of pine-oak forest and cloud forest continuously across the central ridges from the Mexican to the Chiapan-Guatemalan highlands. At those times such highland members of the fauna as _Chiropterotriton_, _Pseudoeurycea_, _Magnadigita_, and the _eximia_ group of _Hyla_ could have crossed the isthmus. During Wisconsin time, climate probably fluctuated less than during previous glaciations; probably no montane environments, except cloud forest, were represented in the isthmus during the Wisconsin.

Even at this relatively late date such animals as _Lineatriton lineola_, _Anotheca coronata_, and _Phyllomedusa moreleti_ could have crossed the isthmus.

During the interglacial periods, which in the isthmian region were characterized by warmer temperatures, higher sea level and consequently more restricted areas of lowlands, and possibly more rainfall than in the glacial periods, the continuity of pine-oak forest and cloud forest from east to west across the isthmus was interrupted. Probably, too, the semi-arid environments were restricted, and the rainforests were more widespread. At those times animals now inhabiting the rainforests of the Gulf lowlands and those inhabiting the Pacific lowlands of Chiapas and Guatemala could have crossed the isthmus. In this group are species such as _Bolitoglossa occidentalis_, _Eleutherodactylus rhodopis_, _Microbatrachylus pygmaeus_, and _Rana palmipes_.

The amount of differentiation in isolated populations of amphibians in southern Mexico and northern Central America gives some idea of relative lengths of time of isolation from related populations. Those populations inhabiting high mountain environments on either side of the isthmus are specifically distinct. Some populations inhabiting cloud forests lower on the mountains are specifically distinct from related populations on the other side of the isthmus; between others there is no recognizable differentiation. Even though many populations are isolated from other populations of the same species in the lowlands of the isthmus, there is no apparent speciation. This indicates that the lowland environments and their inhabitants have been isolated from one another for a shorter time than have the highland environments and their inhabitants.

ACCOUNTS OF SPECIES

For each species of amphibian known to occur in the lowlands of the Isthmus of Tehuantepec, localities where one or more specimens were collected are listed, and variation, ecology, and life histories are discussed. A total of 2833 specimens has been examined for the purposes of this study. Individual specimens cited in the text are listed with catalogue numbers and abbreviations of the name of the museum, as follows:

AMNH American Museum of Natural History KU University of Kansas Museum of Natural History MCZ Museum of Comparative Zoology, Harvard College UIMNH University of Illinois Museum of Natural History UMMZ University of Michigan Museum of Zoology USNM United States National Museum

=Gymnopis mexica.n.u.s mexica.n.u.s= Dumeril and Bibron

_Oaxaca_: El Barrio (3); Matias Romero; Tehuantepec (2).

_Veracruz_: Cosamaloapan; Cuatotolapam (2).

The two specimens from Cuatotolapam were collected by Ruthven in an area of mixed savanna and forest. The three specimens (USNM 30535-7) listed above from El Barrio were collected by Sumichrast; possibly they came from another locality. The city of Tehuantepec is divided into seven districts called "barrios." The two specimens listed from Tehuantepec (MCZ 1604) merely bear the data "Tehuantepec, Mexico." They may have come from the town, the district, or from anywhere in the isthmus. The specimen from Matias Romero has 109 primary and 67 secondary annuli, a length of 400 mm., and a diameter of 19 mm.; the one from Cosamaloapan has 106 primary and 58 secondary annuli, a length of 397 mm., and a diameter of 19 mm. Data on the other specimens were recorded by Dunn (1942:475).

=Bolitoglossa occidentalis= Taylor

_Oaxaca_: Rio Sarabia (2); Ubero. _Veracruz_: La Oaxaquena; 14 km. E of Suchil.

The specimens from Oaxaca are only tentatively a.s.signed to _occidentalis_. All are immature and lack maxillary teeth. Taylor (1941:147) stated that the maxillary teeth are absent in young _occidentalis_. One from Rio Sarabia is a male with a body-length of 29 mm. and a tail-length of 22 mm. The dorsum is reddish brown streaked with dark gray; the venter is dark gray. Two small individuals (one from Sarabia and one from Ubero) have body-lengths of 19 and 21 mm. and tail-lengths of 10.5 and 11 mm. In life they were pale yellowish tan above with a brown triangular mark on the occiput, but with no middorsal stripe. Both were found in the axils of elephant ear plants (_Xanthosoma_).

This species has been noted by Goodnight and Goodnight (1956:146) on the Atlantic lowlands at Palenque, Chiapas, and by Shannon and Werler (1955:362) at several localities in Los Tuxtlas, Veracruz. I have collected it at Vista Hermosa on the eastern slopes of the Sierra Madre Oriental above Tuxtepec in northern Oaxaca. Both _B. occidentalis_ and _B. rufescens_ have been reported from Palenque, Chiapas (Taylor and Smith, 1945:547). Reexamination of specimens from northern Chiapas and Tabasco is needed to verify the sympatric occurrence of these two similar species.

=Bolitoglossa platydactyla= Tschudi

_Oaxaca_: La Oaxaquena; Tolosita (2). _Veracruz_: Acayucan; Cuatotolapam; 25 km. ESE of Jesus Carranza; 14 km. E of Suchil; 2.7 km. N of Tula.

Known only from the Gulf lowlands in the isthmian region, this species has been taken in a variety of habitats within the humid forest area: under outer leaves of banana plants, under a rock along a stream, under a log in a plowed field, and on a reed in a pond at night. Three adult males have an average snout-vent length of 44 mm. and a tail-length of 41 mm. In life the color of the dorsum varied from orange-yellow to orange-tan, usually being more orange on the tail. The iris was a reddish orange.

=Bolitoglossa veracrucis= Taylor

_Veracruz_: 35 km. SE of Jesus Carranza (21).

This species is known only from the type series collected at night on a limestone cliff by Walter W. Dalquest. If this salamander is restricted to this type of habitat, it should be found in the region of extensive limestone outcroppings in northern Chiapas and southern Tabasco.

=Rhinophrynus dorsalis= Dumeril and Bibron

_Oaxaca_: Ixtepec; Limon; Salina Cruz (18); Tehuantepec (57); Tuxtepec (3). _Veracruz_: Amat.i.tlan (3); Cosamaloapan (5); Novillero (2); San Lorenzo.