178 PROF. G. B. HOWES AND W. RTDEWOOD ON [Mar. 6, 



7. That the naviculare is a centrale and not the basal segment of 

 the pre-hallux. 



8. That the pre-hallux never consists of more than four pieces, 

 and that it conforms to the structural requirements of a sixth digit. 



9. That the outer free border of the pre-hallux may undergo a 

 process of fragmentation, giving rise to insignificant nodules of 

 cartilage indistinguishable from those for which, in certain other 

 Vertebvata, the value of lost rajs has been claimed ; and that the 

 grounds upon which this claim is based are unsatisfactory. 



10. That there is a tendency towards loss of independence of the 

 hallux-tarsal in the Biscoglossidce, Pelohatidcd, and some Ranida;, 

 that element in them remaining cartilaginous and small, or fusing 

 with one of the adjacent elements. 



11. That Pipa is alone exceptional among living forms, in that 

 the third digit exceeds the fourth one in length. 



Perusal of the body of this paper will show that in all the higher 

 families of Anura there is a general tendency towards confluence of 

 three or more of the carpal elements, but consideration of the fact 

 that such modifications by fusion may not be even geuerically 

 constant, shakes our faith in them as guides to affinity. The fact 

 which stands out most conspicuously is that the least modified 

 conditions of both fore and bind feet are most nearly combined in 

 the DiscofflassidcE. Hochstetter has recently shown (20) that in 

 BomUiiator the posterior cardinal veins are retained for life, in a 

 slightly modified form, and his observation has been supplemented 

 by one" of us and extended (22) to Alytes and Discoglossus. Adding 

 these facts to those so well known concerning the vertebral and 

 other characters of this family, there can no longer be any doubt 

 that its members are, by far, the least modified of all living Anura. 



The digital formula of the Anura is ': — 



Fore. Hind. 



Ph. 2, 2, 3, 3. Ph. 2, 2, 3, 4, 3. 



The only other Amphibia of which we have any knowledge whose 

 phalanges approach this in order of arrangement are the Stegocephalia 

 of the Permian. We unfortunately know nothing of their carpus 

 and tarsus. Baur has lately tabulated (1, p. G4) the digital formulae 

 of all known Urodela, and perusal of his tables will show how 

 completely all the members of that order are, in this respect, modified 

 as compared with the Anura. Thus it is seen that while the limb- 

 skeleton of the Frogs and Toads is specialized in the extreme for 

 physiological purposes, there is retained in it a leading morphological 

 feature which carries us back to some of the oldest known represen- 

 tatives of the class Amphibia ; and we must look to the Stegocephalia 

 themselves or to some closely ahied forms fur the ancestors of these 

 familiar creatures. 



1 During the passage of these pages thro'jgh the press, Mr. Boiilenger has 

 called attention (see below pp. 204-206) to the fact that in certain Banidce a 

 supernumerary phalanx is intercalated between the penultimate and terminal 

 ones of each digit in both fore and hind feet. Ihe iormula of these animals is 

 thus: 3,3,4,4. 3,3,4,6,4. 



