210 Proceedings of the Royal Physical Society. 



affecting not merely the general features of development, 

 but extending to histological detail in a way that makes it 

 quite impossible to put it down to mere convergence. 



The really important morphological differences between 

 the two groups are very few. The most striking, perhaps, is 

 the pentadactyle character of the limb in the urodele. This 

 difference becomes of less importance on the view of evolution 

 of the limb which I hold — involving the independent 

 development of cheiropterygium and ichthyopterygium from 

 a stylopterygial form. In the Dipnoi the limb has evolved a 

 less distance from the stylopterygium than in any of the 

 other groups. In Ceratodus it is a typical archipterygium; 

 in Protopterus and in Lepidosiren it shows less or more 

 complete reversion to the stylopterygial condition. In other 

 words, the Dipnoans are nearer the form which possessed 

 the type of limb ancestral to both cheiropterygium and 

 icthyopterygium than is any other group. 



Another important difference, that of the character of the 

 intestine, is also shown by embryology to be less fundamental 

 than it seems, for it appears that in the short gut with spiral 

 valve we have a type produced by the fusion together of 

 the turns of an elongated spirally coiled gut. From the occur- 

 rence of the spiral valve in Elasmobranchs, Crossopterygians, 

 Ganoid Actinopterygians, and Dipnoans, it is fairly clear 

 that at one period it must have been characteristic of the 

 ancient Gnathostomata generally. In the Pentadactyle verte- 

 brates we see a reversion to the still earlier condition where 

 the turns of the gut had not yet become fused. 



In the earlier stages of my work upon Lepidosiren, I was 

 much interested by the paper which Dollo ^ had just pro- 

 duced upon the phylogeny of the group, and special search 

 was made for any features which might support his plausible 

 arguments. Dollo's main thesis is that the present-day type 

 of Dipnoan, with its diphycercal tail and continuous median 

 tin, is really a degenerate descendant of forms with a hetero- 

 cercal tail and discontinuous median fins. Such questions, 

 dealing with the possibly degenerate character of forms 

 which are apparently primitive, are, in my opinion, of much 



1 Bull. Soc. Belcj. Giol. PaUont. , T. ix. 



