398 bulletin: museum of comparative zoology. 



spotted with black. In that case the lateral yellowish stripe is very- 

 wide, and the dark cross-bars on the thighs are in sharp contrast to the 

 greyish ground-tones. Most of the adults have some dark spots on 

 the back, and most have the dark lateral band profusely reticulated 

 with white. 



Phyllobates infraguttatus Boulenger. 



Nearly a thousand specimens from several localities: — Palarabla 

 and Huancabamba (August), Perico and Bellavista (September), and 

 Querocotilla (October). 



We have not been able to find any character with which to dis- 

 tinguish the specimens from Palambla, Huancabamba and Queroco- 

 tilla from a specimen (M. C. Z. 3,214) of P. infracjiittatus taken at 

 Rio Chanchan, Ecuador. The majority of our specimens from the 

 mountains of northern Peru are dark grey below spotted with white, 

 similar to the Ecuadorian specimen of P. infraguttatus before us, but a 

 few of those from the movmtains and all of those from the lowlands 

 (Perico and Bellavista) have the ventral surfaces white with a pair 

 of large dark spots just anterior to the pectoral girdle and sometimes 

 a delicate marbling of the same tone along the sides of the belly. The 

 pair of dark spots is scarcely visible in the dark bellied specimens from 

 the mountains and no such spots are mentioned in Boulenger's descrip- 

 tion of P. infraguttaius. Nevertheless they are apparently invariably 

 present in some degree and form the most important distinguishing 

 character of the species. 



Peracca (Boll. Mus. Torino, 1904, 19, no. 465, p. 17) has shown that 

 Phyllodromus must be referred to Prostherapis, a genus indistinguish- 

 able from Phyllobates as defined by Boulenger (Proc. Zool. soc. London, 

 1888, p. 206). The notch on the posterior margin of the tongue is 

 extremely variable in all the species of Phyllobates and Prostherapis 

 which we have examined. In Phyllobates trinitatus the tongue is often 

 entire, while in our huge series of P. infraguttatus the tongue is some- 

 times emarginate and sometimes entire. In the four specimens of 

 P. latinasus before us the notch shows various degrees of de\elopment. 

 We have examined specimens of Prostherapis inguinalis and P. 

 boulengeri having an emarginate tongue. It is evident then that the 

 emarginate tongue does not distinguish Phyllobates from Prostherapis. 



In internal structure Prostherapis agrees essentially with Phyllo- 



