54 U. S. NATIONAL MUSEUM BULLETIN 206 



microdiscus Anders, it differs by the loose skin of the cranium, colour and size. 

 From G. fulvo-rufa Anders, by the colour, short leg, and smooth skin, from Gas- 

 trotheca ernestoi Mir. Rib., which is nearest to it, by the shape of the tongue and 

 vomerine teeth and the still shorter leg colour and the simpler markings. Miranda 

 Ribeiro makes no mention of unequal disks. 



Affinities: In spirits, the colour of some species of Gastrothecas is dark. This 

 does not generally occur in light frogs Hke ours, which do not turn darker but fade. 

 Black borders and striking designs are seen in many species of Gastrotheca, but 

 differ in detail. 



Genus Hyla Laurenti 



1768. Hyla Laurenti, p. 32. (Genotype Hyla viridis Laurenti.) 



Generic diagnosis. — Teeth present in upper jaw, or in both upper 

 and lower jaws; shoulder-girdle arciferous; lateral sacral processes 

 dilated ; bone of last toe joint claw-shaped. Pupil horizontal. Tongue 

 entire or slightly nicked, adherent or more or less free behind. Fin- 

 gers free or webbed; toes webbed, the tips dilated into larger or smaller 

 disks. Outer metatarsals united or slightly separated. Omosternum 

 cartilaginous; sternum a cartilaginous plate. Diapophyses of sacral 

 vertebra more or less dilated. 



Remarks. — A careful study of several thousand specimens of Bra- 

 zilian hylas has demonstrated the impracticability of attempting a 

 dichotomous key to all the species, since the known individual varia- 

 tion in many of the characters normally used in making a key is so 

 great that a category broad enough to include all the variants of a 

 single species might include some parts of several closely allied species 

 as well. 



The most that can be attempted now, therefore, is the assembling 

 of species together into several natural groups, based on a real affinity 

 in structural characters, often with color similarities, and a statistical 

 comparison of the critical measurements when sufficient material is 

 available for this treatment. Subgeneric names have been proposed 

 for most of these groups, but the consistent application of such group 

 names must await the combined study of internal and external char- 

 acters of the hylas of all the Americas, a truly Herculean task. In 

 consequence of this, the keys for these groups must be used with some 

 caution. 



Some groups are relatively easy to discern after the first survey 

 of species has been made. The largest frogs are contained mostly 

 in Groups 1 (with heavily glandular skin) and 2 (with nearly smooth 

 or slightly tubercular skin) while Group 4 is composed entirely of 

 the smallest species with an immaculate red or yellow femur, and 

 usually with a characteristic pattern elsewhere. Group 3 is com- 

 posed mainly of species medium to fairly large in size, most of which 

 have fairly well developed dorsolateral folds, relatively long snouts, 



