LINNEAN SOCIETY OF LO]S'UO:S. 29 



large iiiterptervgoid vacuities only appearing at llie beginning of 

 the Permian period. 



As in so many cases of linlis among extinct animals, however, 

 the known series is far from being complete. The Crossopterygii 

 hitherto found have the upper jaw suspended to the cr.-inium, 

 while even the earliest Amphibia exhibit it directly fused witli the 

 cranium. As the Dipnoi among fishes show the same fusion of 

 the upper jaw, they were for some years placed nearest to the 

 Amphibia ; but recent studies suggest that the fused (or auto- 

 stylic) condition has arisen more than once as a mechanical 

 adaptation to the working of powerful teeth, and it may be pre- 

 dicted that definite links between Crossopterygian and Amph.ibian 

 jaws will sooner or later be discovered. 



More ditlicult to understand are the differences which have 

 lately been recognised in the basicranial axis. In the earliest 

 linown Amphibia, the Stegocephala, just as in those now existing, 

 the basicranial axis, with its large parasphenoid bone, extends 

 backwards as far as the occiput. In the Devonian and later 

 Palaeozoic Crossopterygii, and in the Coelacanth family which sur- 

 vived until the Cretaceous period, this axis, underlain by the 

 parasphenoid, extends backwards only so far as a point beneath 

 the hinder nuii-gin of the frontal bones where there Js always a 

 transverse line of weakness in the cranial roof. At this point it 

 ends abruptly, and its termination, round in transverse section, is 

 impressed behind with a conical hollow for the notochord, as if it 

 Were the basioccipital itself. There is no doubt, indeed, that the 

 fully developed notochord, surrounded only by non-os>if1ed 

 cartilage, extended as far forwards as tliis point, which appears to 

 have been close to the pituitary region. The condition of the 

 basal part of the skull in the adult early Crossopterygii thus 

 corresponds closely with the temporary embryonic condition of 

 the sftme part in a mod^-rn fish-skull. The arrangement has been 

 clearly seen in several specimens of the Upper Devonian Holo- 

 ptycliias and the Carboniferous Megalichthys, but it is especially 

 well-known in EuMlienopteron (fig. 2) from the Upper Devonian 

 of Canada, in which the ossified otic-ocdpital region has also 

 been observed*. It is equally clear in the Devonian and the 

 Cretaceous genera of the family Coelacanthidce f. 



Between some of the Crossopterygii and some of the Stegoce- 

 phala there is not much difference in the vertebral axis. The 

 reduction and concentration of the median fins towards the end 

 of the tail in the early Crossopterygii (fig. 1) may be regarded as 

 marking the beginning of the disappearance of these structures. 

 The links between the paired fins of the Crossopterygii and the 

 four- or live-toed limbs of the 8tegocephala, hox^ever, nve still 

 wanting. It can only be stated thai the cartilages in some of the 

 short-lobed fins, such as those of Eusthenopteron, approach more 



* W. L. Bryant, Bull. Buffalo Soc. Nat. Sei. vol. xiii. (1919), pp. 6-19, 

 pis. ii.-x. 



+ E. A. Stensio, Palaoiit. Zeitsclir. vol. ir. (1922), pp. 167-210, pis. iii.-v. ; 

 D. M. S. Watson, Ann. & Mag. Nat. Hist. [9] vol. viii. (1921), pp. 320-337. 



