240 U. S. NATIONAL MUSEUM BULLETIN 206 



inside are not homogeneous and that a foreshadowing of the condition 

 found in Crossodactylus may here occur. 



Cycloramphus ohausi (WandoUeck) , from Petropolis, while a larger 

 frog and probably a true Cycloramphus because of its inguinal gland, 

 seems somewhat to suggest Crossodactylodes pintoi in its broad, de- 

 pressed head and particularly in the presence of eight or nine spines 

 on the first finger, a character not found in any other Cycloramphus. 

 A thorough anatomical study is needed of all these forms to ascertain 

 just how close is the relationship. 



Specimens examined 

 BRAZIL: 



Rio de Janeiro: Maca6, USNM 102606 (type), USNM 102607-11 (paratypes), 

 and MP 104 (paratype), Museu Paulista. 



Genus Crossodactylus Dumeril and Bibrou 



1841. Crossodactylus Dumeril and Bibron, p. 635. (Genotype, Crossodactylus 

 gaudichaudii Dumeril and Bibron.) 



Generic diagnosis. — Tongue medium, oval, adherent everywhere, 

 covered with irregular confluent folds. No palatine teeth. Tym- 

 panum distinct; Eustachian tubes very small. Fingers slender, weak, 

 slightly depressed, completely free, dilated into disks which are convex 

 below, flat above. Toes slightly flattened, enlarged in the same man- 

 ner as the fingers and fringed on each side; outer tarsal border fringed. 

 Transverse apophyses of sacral vertebra not dilated. 



Crossodactylus gaudichaudii, the type species of the genus, is com- 

 mon near the city of Rio de Janeiro from whence the type presumably 

 came. Most of the specimens have a white throat and chest, with 

 at most a small gray longitudinal mark on the chest between the fore- 

 arms. A more or less prominent dark longitudinal stripe begins at 

 the nostrils, continues along the side of the head and in a few instances 

 down the dorsolateral region halfway to the groin. The dorsum is a 

 uniform gray or olive, except for a dull interorbital chevron-shaped 

 darker marking, sometimes followed behind the shoulders by a 

 )(-shaped pair of marks. The number of spines at the base of each 

 first finger varies with age, reaching a maximum of six. The nostril 

 lies between one-half and two-thirds the distance from eye to tip of 

 snout. 



A very closely aUied form is C. aeneus from Teres6pohs. In the 

 series at hand, some of the characters supposed by Miiller to distinguish 

 it from gaudichaudii partly merge, since the snouts in both forms vary 

 from roimded to angular-edged, and the width of the interorbital 

 space varies. The nostril, said to lie closer to the tip of the snout in 



