GREGORY: FISH SKULLS 105 



is a wide difference between the systems of Watson (1925, Fig. 2, p. 197) and of Goodrich 

 (1930, p. 304, Fig. 311) in the identification of the parietals, frontals and other bones of the 

 dipnoan skull-roof. More concretely, the paired bones in front of the intertemporals of 

 Dipterus are called frontals by Watson, parietals by Goodrich, while the large median 

 element above the orbit is equated with the paired tetrapod nasals by Watson, but with 

 the paired frontals by Goodrich. Watson's proposed solution of the problem makes it 

 easier to visualize or conceive the homologies of most of the Dipterus skull-roof bones 

 with those of both tetrapods and crossopterygians; also Goodrich's "frontals" of Dipterus 

 are anterior to the interfrontal while the frontals of Eryops are behind the interfrontal. 



In Dipterus the skull-roof is broken up into a mosaic of many small bones (Pander, 

 1858); in the higher dipnoans certain of the paired bones became dominant, this finally 

 culminating in the modern Protopterus and Lepidosiren in two greatly .elongate, rod-like 

 bones that run parallel with the mid-line on the top of the skull. In these two genera also 

 the fan-shaped dental crests acquire sharp edges and the fish has apparently become carniv- 

 orous. This line of specialization w'as begun in the Carboniferous genus Uronemus. 



In the light of all the evidence to the contrary, it is rather surprising to find Dr. Xaef 

 (1926) defending the theory that the dipnoans have given rise to the amphibians, and figur- 

 ing Uronemus as a structural ancestor to a purely hypothetical form that gave rise to a 

 suspiciously dipnoan-looking tetrapod! It is quite true that Uronemus and Conchopoma 

 apparently lack the typical dipnoan fan-shaped dentition on the roof of the mouth and 

 inner sides of the mandible, but Watson and Gill (1923, p. 214) give reasons for the belief 

 that these genera are true dipnoans which have lost the typical dipnoan plates and have 

 acquired a secondary dentition of patches of small tubercles on the parasphenoid and inner 

 sides of the mandible. But the skull-roof and paired fins of Conchopoma are described by 

 Watson and Gill as essentially similar to those of the existing dipnoan Ceratodus, while the 

 cranial roof bones of Uronemus are arranged "very nearly as in Ctenodus." Thus these 

 Carboniferous genera are definitely excluded from close relationships with the contemporary 

 Amphibia, which have a very different type of skulls, dentitions and paired appendages. 



In conclusion, the dipnoans, as a whole, have specialized toward an aberrant pattern of 

 the skull-roof and fan-shaped dental clusters on the roof of the mouth and inner sides of the 

 mandible, while the earliest tetrapods retained practically the complete carnivorous denti- 

 tion and every bone of the skull-roof of the contemporary crossopterygians except the oper- 

 cular series. 



Crossopterygii 



The skull patterns of the palaeozoic crossopterygians are of great morphological 

 importance because it has been shown, chiefly by Watson, that on the one hand they stand 

 not far from the ancestral line to those of the land-living vertebrates, while on the other 

 hand they are related to those of the oldest actinopterygian fishes. Hence they furnish 

 the necessary bridge (Fig. 8) across which the nomenclature of the tetrapod skull has been 

 transferred to the skull of the primitive fish. The skulls have been described in many 

 monographs and papers by Pander, Traquair, Watson, Goodrich, Bryant, Stensio and 

 others. 



Some of the salient characteristics of the more typical palaeozoic crossopterygian skulls 

 are as follows: they are highly predaceous types, typically with large jaws and strong teeth 

 with labyrinthine bases; the teeth comprise rows of smaller caniniform teeth on the lateral 



