132 Eastman — Dipnoan Affinities of Arthrodires. 



Ceratodus als der primitivere erkannt ist, muss audi die 

 Annahme Dollo's, dass die Heterocerkie bei Dipterus etwas 

 Primitives bedeute, fallen. Der Schwanz von Ceratodus 

 zeigt zwar gewisse Riickbildungen ; fur eine ehemahlige 

 Heterocerkie bei ihm ist aber kein Beweis erbraeht. Ich kann 

 nach alledem, falls iiberhaupt eine Verwandtschaft zwischen 

 Dipterus unci den recenten Dipnoern bestelit, diese im Gegen- 

 satz zu Dollo nur im Sinne von Woodward unci Bridge auffassen." 



Elsewhere the same author expresses himself unreservedly 

 as to the relative specialization of the two types : "So ware 

 eines doch sicher, dass Ctenodus unci namentlich Diptterus 

 weit holier differenzirt sincl als Ceratodus." Smith Wood- 

 ward's opinion is that Dipterus and its allies are more special- 

 ized than any existing Dipnoan, and that the more generalized 

 types have alone survived to represent the group at the present 

 clay.* Though he does not distinctly say so, his remarks 

 would imply that the archetype of Ceratodus was ancestral to 

 Ctenodipterines, the latter becoming finally extinct. 



So much for a statement of the problem. Remains to 

 inquire whether Palaeontology can point the way toward a solu- 

 tion, even if indirectly. It has been said that no generalized 

 form standing in the relation of ancestor to both Dip>terus and 

 Ceratodonts — the latter alone having persisted, — is known 

 from the Devonian. Were such a form to be brought to light, 

 or should its existence become a necessary postulate from 

 other facts, it is evident that the question as to descent of mod- 

 ern Dipnoans would be stripped of much perplexity. In that 

 event, also, there would be no necessity for an appeal to the 

 imperfection of the palasontological record, such as has hereto- 

 fore existed. 



]STow we confidently believe that the relations between 

 recent and fossil Dipnoans will appeal- in much clearer light 

 through comparison of JYeoceratodus with Arthrodires, taking 

 Jlulostoma and Dinichthys as typical examples of the latter. 

 There is no novelty in the idea that Arthrodires are related to 

 Dijmeusti, dewberry's original suggestion to that effect hav- 

 ing found several warm supporters, notably Cope and Smith 

 Woodward. It must be said, 'however, that the question has 

 remained eminently controversial, first one and then another of 

 rival interpretations gaining headway. That a singular lack of 

 unanimity prevails as to the sytematic position of Arthrodires 

 must be apparent to anyone having a casual acquaintance with 

 the literature. The reason for such diversity of opinion evi- 

 dently lies in non-recognition of homologies between the struc- 

 tural type of Arthrodires on the one hand, and those of fossil and 



* Woodward, A. S., Catalogue Fossil Fishes British Museum, Part II 

 (1891), introduc. p. xx. 



