346 Mr. C. T. Regan on the 



still less valid, and might be applied with equal force to any 

 of the numerous Ganoid fishes with tuberculated bones, and 

 surely it is a retrograde step to suggest that structures which 

 in Cephalaspis have been shown to be posterior extensions of 

 the head-shield may after all be pectoral fins. 



In fact, the evidence that the Coccosteidge ai'e Teleostomi, 

 that the Asterolepidae are allied to the Coccosteida?, and that 

 the Cephalaspidse have been derived — through the Trematas- 

 pidas — from the Asterolepidae is so clear, that I am com- 

 pelled to regard the Ateleaspid structure as a modification of 

 that of the Cephalaspid. 



TELEOSTEI. 



The reasons for regarding the Teleostei and Chondrostei as 

 distinct orders and for including the Holostei with the former 

 are apparent in the diagnoses given above. The Holostei 

 may then be regarded as the first Teleostean suborder*, dis- 

 tinguished from the Malacopterygii by their well-developed 

 splenial and by one or more of the pectoral baseosts being 

 attached to the metapterygi.um. Whether certain features of 

 resemblance between Polypterus and the Holostei, of which 

 the articulation of the operculum to a posterior process of the 

 hyomandibular is the most important, are to be interpreted as 

 derived from a common ancestor or as due to convergence is 

 not yet clear. 



THE PALiEONTOLOGICAL EVIDENCE. 



It may be said that the conclusions as to the evolution of 

 the Teleostomi expressed above are not in accordance with 

 the palasontological evidence ; but to this I reply that they 

 are in accordance with the morphological evidence, which is 

 clear and sufficiently complete, whilst the geological record is, 

 and must be from the nature of the case, very incomplete. 

 The Teleostomi probably originated from Pleuropterygian 

 Elasmobranchii in the Lower Silurian, and the Crossopterygii, 

 with their specialized offshoots the Dipneusti and Placo- 

 dermi, must have rapidly evolved, since all are well represented 

 in the Lower Devonian, and the highly specialized Cepha- 

 laspidse are found in the Upper Silurian. In the same way 

 that generalized Reptilia gave rise to the host of forms which 



* Provisionally, for I am inclined to think that none of the characters 

 which have been used to distinguish between Holostei and Malacopterygii 

 will prove satisfactory. 



