142 Eastman — Dvpnoam Affinities of Arthrodires. 



exists in favor of this supposition ; on the contrary, authorities 

 like Traquair and Smith Woodward agree that it may as well 

 have been diphycercal. Now it still remains to be proved 

 that the diphycercy of JYeoceratodus is not a primitive fea- 

 ture, that does not faithfully reproduce ancestral characteristics. 

 One is, indeed, at perfect liberty to believe that the tail of 

 Ceratodonts and Arthrodires never advanced beyond the diphy- 

 cercal condition, Dipterus and its allies alone becoming hetero- 

 cercal. Inasmuch as the median fins of Coccosteus are well 

 separated, their continuity in Neoceratodus would seem to be 

 a secondarily acquired character; but the burden of proof 

 surely rests on those who hold that original heterocercy has 

 become suppressed through abortion of the extreme end of the 

 axis, and coalescence of the dorsal and anal fins. 



]STo enlightenment could be more welcome than that which 

 would acquaint us with the structure of the paired fins in 

 Arthrodires, as to whether they were biserial or uniseriaJ, 

 Crossopterygian-like or Pleuracanthus-like. Of the pectoral 

 pair no trace whatever has been preserved, nor do we even 

 know that a girdle was present. Failure of the latter to be 

 preserved might be attributable to a cartilaginous condition 

 resembling that of Neoceratodus, but we should expect to find 

 at least some traces of fin rajs, were these structures developed. 

 Complete atrophy of the pectoral pair would indicate, of 

 course, high specialization. Obscure traces of a pelvic pair 

 have been detected in some specimens of Coccosteus, but 

 nothing is known of their configuration or structure. There 

 seems to be no doubt *as to the occurrence of a pelvic arch, 

 and all that different writers have affirmed of it is consistent 

 with the view that it was constructed essentially as in modern 

 Dipnoans. In the latter it has remained cartilaginous; in Coccos- 

 teus and JDinichthus it was ossified. 



Conclusions. — The aggregate of facts brought together 

 through comparison of Arthrodires with modern Dipnoans 

 seems to uphold the following general propositions : 



1. Neoceratodus bears intimate resemblances to Arthrodires 

 on the one hand, and to Ctenodipterines on the other, but 

 represents a more primitive structural type than either. 



2. It is impossible to regard Neoceratodus && the degenerate 

 descendant of ooth the earlier, more specialized groups, nor of 

 either of them to the exclusion of the other ; since, however, 

 it partakes of- the characters of both, community of origin is 

 necessarily presupposed for all three orders, Sirenoids, Cteno- 

 dipterines and Arthrodires. 



3. Arthrodires and Ctenodipterines may be regarded as 

 specialized offshoots which diverged in different directions 

 from primitive Dipnoan ancestors ; and the more generalized 



