388 



second fingers are equal and the third toe is shorter than the 

 fifth, in the other the first finger is the shorter and the third 

 toe is equal to the fifth. The shagreening shows better in these 

 two which are neither hard, like the type, nor soft like the 

 Barro Colorado Island specimen. 



Remarks. — A specimen of lineatus from Bolivia, kindly loaned 

 me by Mrs. Gaige, shows transverse vomerine teeth; a narrower 

 head, and a thicker body; flash markings in groin and on thigh; 

 a protruding or overhanging snout; snout not so pointed; canthus 

 not so marked; third toe longer than fifth. 



The specimens had all been identified by me with lineatus, until 

 Mrs. Gaige pointed out to me the differences between the Barro 

 Colorado Island specimen and South American ones. These dif- 

 ferences are more exaggerated in the Barro Colorado Island speci- 

 men than in the three others, but enough remains to show a dis- 

 tinguishable form, with which I take the liberty of associating her 

 name. 



As Ruthven remarked to Noble (Noble 1917, Bull. Amer. Mas. 

 Nat. Hist. 37, p. 794) that a specimen he secured in British 

 Guiana resembled a Dendrobates in life, so also the two which I 

 took in Talamanca were so like Phyllobates that for some years 

 they were identified as the very similarly colored P. lugubris 

 with which they were associated in life. 



I use Lithodytes advisedly. The shoulder girdle of both Fort 

 Randolph and the Barro Colorado Island specimens is Lepto- 

 dactylus-like, although the ossification of the sternum is not 

 complete in the Barro Colorado Island specimen. The terminal 

 phalanges are distinctly T-shaped in both. Noble (loc. cit.) on 

 a series of five Guiana specimens ranging in length from 22.5 

 to 45.5 mm., has maintained that while young specimens have 

 a T-shaped terminal phalanx, adults have a simple pointed 

 phalanx and has referred the species to Leptodactylus. His 

 figures are not convincing to me, and I have assured myself 

 that M. C. Z. no. 6033 from Trinidad, 48 mm. long, has the 

 terminal phalanx of toe IV extremely T-shaped. Neither 

 lineatus nor gaigei is reminiscent in habit of any Eleutherodactylus 

 or Leptodactylus known to me, and under the circumstances I 

 prefer to recognize the genus Lithodytes Fitzinger, type lineatus, 

 only other known species gaigei. 



