2 Records of the Indian Museum. [Vol. XX, 



The time has come to rid ourselves of the empirical methods 

 which have necessarily prevailed so long in zootaxy, and to en- 

 deavour to group species, as far as possible, according to their 

 probable phylogenetic relationships, and for this purpose it is desir- 

 able to fix one's ideas by establishing a theoretical prototype 

 from which phyletic lines ma}' be drawn up. This I have re- 

 cently attempted to do in a paper ' of which an abstract is here 

 given. 



The Firmisternia are certainly derived from the Arcifera, as 

 shown by their ontogeny, and among the latter the Cystignathidae 

 agree in so many points with the Ranidas that we may legiti- 

 mately assume a direct genetic relationship between them. 



Dealing first with external characters, I consider the com- 

 plete separation of the outer metatarsals by web as the most primi- 

 tive character, which is not correlated with the degree of develop- 

 ment of the web between the toes. In all the series of allied forms 

 in which the phyletic tendencies are clearly indicated, we can follow 

 the gradual reduction of the membrane ; the species with fully 

 webbed toes are, therefore, the least specialized, so far as this 

 character is concerned. 



The dilatation of the tips of the digits is a specialization 

 culminating in the large scansorial discs of the species adapted to 

 an arboreal life. 



The primitive frog must have had a well-developed tympanum 

 covered by a very thin skin. The reduction of the organ, which 

 may be completely concealed under the thick skin, leads to its 

 suppression [R. flcskei). 



The prominent glands on the back have often a tendency to 

 run together in longitudinal ridges, irregular at first and ulti- 

 mately localized, such as the dorso-lateral fold which characterizes 

 a great number of species. However, this fold may again disap- 

 pear, as some species of Hylorana seem to show ; it is therefore 

 necessary to distinguish, from the other characters takeir in con- 

 junction, whether the absence of the fold is primary or secondary. 



The skeleton affords important indications for judging of the 

 ph5dogenetic relationships. I have limited myself to the use of 

 characters which may be ascertained without difficulty on spirit 

 specimens. Such are those derived from the skull, which can be 

 examined by lifting up the skin, usually feebly adherent, aird the 

 pectoral arch, which can be exposed by making an incision along 

 the breast and removing the muscles. The shape of the terminal 

 phalanges is easily revealed by the dissection of the tip of one of 

 the fingers or toes. 



I regard as the most typical a skull such as that of R. hexa- 

 dactyla, in which the nasals are large, and in contact with each 

 other and with the frontoparietals, the upper surface of the eth- 

 moid being covered over. From such a type we may derive on the 

 one hand those forms in which the cranial roof is expanded, and 



Bull. Soc. Zool. France, 1918, p. Ill 



