332 BULLETIN 120, UNITED STATI<:S NATIONAL MUSEUM. 



nent, and a Tlyla bearing ZeUeriella is in the Bahamas. We have 

 therefore here five parallel lines of zoogeographic evidence of connec- 

 tion between the West Indies and the Bahamas on the one hand and 

 the American continent on the other hand. The species of ZeUeriella 

 are all so much alike that subgeneric groups of species are not to be 

 recognized. In consequence we can not determine whether the 

 Zelleriellas of the Antillean and Bahaman toads, Leptodactylids and 

 Hylas, are more nearly related to the continental species or to other 

 West Indian species. The contour of the floor of the Caribbean Sea 

 and of the Gulf of Mexico indicates that the connection of the West 

 Indian lands was with Central America b}^ probably two ridges (fig. 

 237), one from Cuba to Yucatan and another from Jamaica to Hon- 

 duras (Vaughan, 1919). Arldt places this connection in the Miocene, 

 but Vaughan's more recent study places it in the Pliocene. The 

 southern tip of the Florida peninsula was probably connected with 

 West Indian lands for a brief period, but the distribution of Anura 

 and Opalinids seems not significant in relation to this connection. 



It is noteworthy that the Leptodactylidae did not pass northward 

 beyond Central America, except for two species that have penetrated 

 into Texas. They are found in portions of South America seem- 

 ingly no better suited to them than are portions of North America. 

 It looks as if the desert region of northern Mexico and southwestern 

 United States held back these moist-skinned forms, for they, like the 

 Eanas, are not resistant to drouth. If so, the desert character of the 

 area mentioned must be of long standing, having existed at least 

 since the Pliocene, as noted in connection with discussion of the 

 distribution of the Ranas (p. 374) . Observe also that the Pelobatidae 

 in the north do not spread southward past the Mexican desert 

 (fig. 224). Some other groups, approaching from the south, find 

 in northern Mexico a barrier to their further migration: Gastro- 

 phrynidae (fig. 229), only one genus (3 species) being known north 

 of Mexico; Dendrohatinae (fig. 231); the archaic Urodele family 

 Coeciliidae (fig. 256, p. 382) ; and two equatorial families of fresh- 

 water fishes, Cichlidae (fig. 257) and Characiindae (fig. 258). We 

 thus see the probable reason why ZeUeriella is not abundant in North 

 America. 



The evidence from ZeUeriella, then, indicates that the Lepto- 

 dactylids are an Antarctic family and that Bufo had a more north- 

 ern origin (p. 362). ZeUeriella itself is an Antarctic genus. ZeUe- 

 riella is older than the time when northern and southern South 

 America became joined, and probably more recent than the time 

 when Australia separated from Papuasia, and of course then is 

 more modern than the time of separation of Australasia from 

 Malaysia. This would place the origin of ZeUeriella apparently in 

 the Miocene, or but little earlier. 



