48 Alexander G. Ruthven 



have only been taken at Don Diego, Aracataca, Fundacion, Valencia, Las 

 Pavas, and Arroyo de Arenas. 



As has been pointed out, the desert and dry forest is replaced by low- 

 land forest in the valleys of the Aracataca, Fundacion, and Ariguani, and 

 by a dryer but still heavy forest in the valleys at Valle de Upar, San Juan 

 de Cesar, and Fonseca, while at Loma Larga, Arroyo de Arenas, and Don 

 Diego the forest is heavy. The occurrence of the four species named above 

 at least indicates a wet forest element in the general fauna at the places 

 mentioned, and since two of the cloud forest forms (Phyllobates subpunc- 

 tatus and Phrynonax poecilonotus) have been taken at Don Diego, and 

 since others (Lepidoblepharis intermedins and Cryptobatrachus fuhrmanni) 

 have been taken in the Choco, it may be confidently expected that at least 

 a part of the cloud forest fauna reaches the lowlands in the large valleys 

 at the west end of the range and has been derived along this route from 

 the lowland forest fauna of the Magdalena basin. 



The large number of forms apparently peculiar to the cloud forest 

 suggests that this almost completely isolated habitat possesses an endemic 

 fauna. This can be determined only when much more is known about the 

 reptile and amphibian fauna of Colombia. 



The Distribution of the Desert and Dry Forest Fauna: It is well 

 known to herpetologists that the faunas of Venezuela and western Colombia 

 are different in many respects. The dift'erences suggest a break in the distri- 

 bution of a number of forms in eastern Colombia or western Venezuela. 

 Too little is known of the distribution of most of the species to permit of 

 the locating of the place or places where the change takes place for many 

 species, but the distribution of the lowland forms in the Santa Marta Moun- 

 tains region apparently throws some light on the problem. 



The chart (p. 47) shows that on the south side of the range many of 

 the lizards do not range beyond certain localities in a particular direction. It 

 is evident that since the principal range of the species is not indicated and 

 there are many chances that the distributions are not accurately known, 

 this summary cannot be relied upon to reveal the eastern and western limits 

 of all of the forms included. There are, however, several forms which 

 here approach the range of near relatives or whose range is definitely enough 

 Icnown to indicate the nature of the relationships of the faunas of the two 

 regions. Many forms are common to Venezuela and Colombia. Some 

 eastern forms, e.g. Hyla venulosa (not included in the chart), occur as far 

 west as Fundacion, while one, Tropidodactylus onca, stops at Riohacha. 

 Ameiva blfrontata, a Venezuelan form, is replaced between Fonseca and 

 Valle de Upar by Ameiva bifrontata divisus, a Colombian form. One 

 Colombian species, Phyllobates subpunctatns (not shown in the chart), does 

 not get farther east than the Don Diego region, and is replaced by a nearly 

 related form in \'enezuela, while Ameiva maciilata and Cnemidophorus 

 lemniscatus gaigei enter western Venezuela at least before being replaced 

 by other forms. These facts indicate that there is no sharp break between 

 the faunas of Venezuela and Colombia, and that the eastern and western 

 forms which enter the region are limited in range at different places. 



