56 Al^EXANDER G. RUTHVEN 



phalanges," and "claw-shaped phalanges are to be expected in the lepto 

 dactylids." In the opinion of the writer, these conclusions are not convincing. 



Fry-- has shown of the Leptodactylidae that "the Australian members 

 of this family exhibit all stages between that of Heleioporus with consider- 

 ably dilated diapophyses and the [cylindrical] condition shown in South 

 American genera.'' It remains to be pointed out that not all of the neo- 

 tropical leptodactylids have cylindrical diapophyses,-^ and that among the 

 forms usually referred to the ?Iylidae there are considerable differences in 

 the form of these processes, cylindrical, slightly dilated and expanded dia- 

 pophyses being represented. It would thus seem to be clear that the form 

 of the sacral dipophyses is not to be depended upon as a character 

 distinguishing two families. 



If the form of the sacral diapophyses is not a reliable family character, 

 then the association of a particular shape with T-shaped or claw-shaped 

 phalanges does not indicate the diagnostic value of the form of the terminal 

 digital phalanx. The hypothesis that the claw-shape is derived from the 

 T-shape, and that the former is to be expected among the leptodactylids, is 

 interesting and will possibly be found to be true, but it must be admitted 

 that at the present time the claw-shape prevails among the hylids with 

 dilated diapophyses, and it is only found in the leptodactylids if such forms 

 of doubtful affinities as evand, goeldi, and fuhrinanni are placed in this 

 family on the basis of that unreliable character the form of the sacral 

 diapophyses. In view of the variations in the form of the sacral processes, 

 it would seem either that the two families should be combined or that they 

 should be separated on the basis of the form of the terminal digital phalanx, 

 until more data is secured. 



While it is uncertain whether one or more families are represented by 

 the species now referred to the Hylidae and the Leptodactylidae, some 

 progress may be made' in the classification by revising the genera and segre- 

 gating the species most alike in the totality of their characters. Bvansi, 

 goeldi, and fuhrinanni are alike in having claw-shaped digits, cylindrical 

 diapophyses with somewhat dilated cartilaginous head, bell-shaped gills in 

 the larval period.-^ and in that the female carries the eggs on the back and 

 the young are born in the adult stage. It is proposed that these forms be 

 united under the generic name of Cr3'ptobatrachus until the identity of 

 Peters' Hyloscirtus is established. The genus may be referred to the Hyli- 

 dae until the value of the form of the terminal digital phalanx is fully under- 

 stood. The writer is inclined to believe that the presence of cup-shaped 

 gills in the larval stage indicates a close relationship between Gastrotheca, 

 Cryptobatrachus, and Ceratohyla, whether or not it is believed that the 



-'^Proc. Roy. Soc. Queensland, XXVII, p. 73. 



2^ Noble has recently described a form with T-shaped phalanges and slightly 

 dilated diapophyses {Bui. Amer. Mus. Nat. Hist., XLII, pp. 441-443). 



2* Boulenger states that "no traces of gills are to be seen in the embryo of H. 

 goeldi," but in the description of Hyla evansi he states that the larva has "allantois- 

 like membranous respiratory organs. In this respect it agrees with Hyla go'cldi Blgr., 

 to which it is nearly related, and with Ceratohyla bubalus Esp." Proc. Zool. Soc. 

 London, X, p. 209. 



