BARBOUR AND NOBLE: AMPHIBIANS FROM PERU. 399 



bates. The type of the latter genus, P. hicolor} is known only from 

 the t^pe-specmien. This is at present in the Museum d'Histoire 

 Naturelle in Paris, and thanks to the kindness of Dr. Louis Roule 

 we are able to make some statement as to its internal structure. Its 

 pectoral girdle is similar to that of Plv/llobafcs trinitatus, P. pratti, 

 P. infragutiatus, Prosthcrapis inguinalis, and P. boulengeri. The girdle- 

 form of P. trinitatus Garman (Fig. 1) may be taken as characteristic 



Fig. 1.- — Phyllobates trinitatus Garman. Shoulder-girdle, x IS. 



of all these species. In the sketch of the girdle of P. bicolor sent us 

 by Dr. Roule, the anterior part of the omosternimi is not represented. 

 • It may have been lost during dissection. The fact that the form of 

 the coracoids and metasternum agrees in all the species examined is 

 excellent indication of the species close affinity. We can see no reason 

 for maintaining Prosthcrapis distinct from Phyllobates. 



There remains another genus which should be referred to the 

 sj^Tionymy of Phyllobates. Colostethus, known only from the t^^je- 

 specimen of C. latinasus, was described by Cope as lacking the omo- 

 sternum. The t^'pe-specimen cannot be found in either the U. S. 

 National INIuseimi or the Academy of Natural Sciences of Philadelphia. 

 Specimens very similar to C. latinasus, but possessing an omosternum, 

 have been found in Ecuador and Colombia. Boulenger has recently 

 advised the senior author by letter that he considers his Phyllobates 

 pratti referable to C. latinasus. Boulenger evidently considers that 

 Cope overlooked the omosternum, a very probable oversight, in view 



• The locality of this species has been discussed by the senior author (Mem. M. C. Z., 1919, 

 47, p. 113). 



