104 TRANSACTIONS OF THE AMERICAN PHILOSOPHICAL SOCIETY 



considerably diminished. Nevertheless, the many other profound differences between these 

 groups seem to afford ample justification for Tate Regan's conclusion. More recently 

 Smith Woodward (1920) has pointed out that in the Palaeozoic petalodont sharks of the 

 genus Climaxodus the teeth exhibit "a restricted area of highly vascular dentine much 

 resembling a tritor in the dental plate. of an ordinary Chimaeroid," also that "a relationship 

 between some of the Palaezoic Cochliodonts and the early Mesozoic Chimaeroids has already 

 been remarked upon by P. de M. Grey Egerton and O. Jaekel"; finally, "that the presence 

 of an apparently Chimaeroid character in Elasmobranch teeth which are noteworthy 

 for their slow and scanty succession may therefore have some special significance." In any 

 event, it may be regarded as established that the chimaeroids are modified Palaeozoic 

 sharks of some sort and in no sense intermediate between sharks and bony fishes. 



Dipnoi 



The oldest known dipnoan (Dipterus), of the Middle Devonian, was a contemporary 

 of the oldest true ganoid {Cheirolepis) and crossopterygian (Osteolepis) and even at that 

 remote epoch already exhibited the chief characteristics of dipnoan fishes, especially in the 

 skull and dentition. The latter consisted chiefiy of masses of conical tubercles arranged 

 in radiating rows, like the sticks of a fan, one pair of such clusters being on the roof of the 

 mouth; the opposing fan-like clusters were located on the splenials or inner bones of the 

 mandible. There was also a pair of small patches on the vomerine region. The lower 

 border of the mouth formed a hard curved rim; the front upper teeth were inconspicuous or 

 wanting; there were no teeth on the dentary bone and (except in Scaumenacia) none on the 

 premaxillae or maxillae. Each individual tooth of the fan-like clusters was composed of 

 tissues that corresponded with the several layers of the ganoine-covered scales. 



The histology of the scales and skull-bones, the form and arrangement of the fins, 

 afford strong evidence of remote common ancestry with the contemporary crossopterygians, 

 but the dipnoans were much more highly specialized in that their dentition, instead of being 

 of the predaceous type, was adapted chiefiy for crushing hard substances, perhaps the shells 

 of molluscs or crustaceans. Hence arose the necessity for a firm foundation for the dental 

 plates on the roof of the mouth and for well braced points of articulation of the mandibles 

 with the palatoquadrate. These effects were secured by firmly attaching the palatoquad- 

 rate to the underside of the cranium. The hyomandibular, becoming free from its sus- 

 pensory function, eventually dwindled into a vestige, which in modern dipnoans has been 

 the subject of many investigations (see Edgeworth, Goodrich, De Beer). 



Meanwhile in the line that led to the modern Neoceratodus the endocranium became 

 massive but at the same time much less osseus and more cartilaginous, while the jaw mus- 

 cles extended dorsad, lifting the dermocranial roof away from the endocranium. At the 

 same time the roofing bones, like the scales, tended to sink beneath a membranous surface 

 layer, to become thin and, in Ceratodus, more or less horny. 



Even in the Devonian Dipterus the bones of the skull-roof are only in part and with 

 considerable difficulty homologizable with those of the contemporary crossopterygian. 

 Watson and Day (1916), Watson and Gill (1923) and Goodrich (1930, p. 304) have tried to 

 equate the Dipterus elements with those of the earliest tetrapods but here the gap is still 

 greater. For there are three rows of bones above the orbits to the midline in Dipterus 

 (Fig. 8C) but only two in tetrapods (Fig. %B). From this and other circumstances there 



