1910.] CAUDAL FIN OF THE TELEOSTOMI. 595 



structure in pa.ssin<,' from before, backwards. Tlie caudal fin of 

 Fohjodon, tlien, is obviously heterocercal, but of a more specialized 

 type than Acipenser, on account of its nearer approacti to the 

 honioeercal type. 



Amia calva. (Holostei.) (Plate XLVII. fig. 4.) 



The caudal fin of Amia is externally symmetrical, or at any 

 late the discrepancy in symmetry is veiy slight indecid, and 

 negligible. But dissection i-eveals a very unsynimetrical internal 

 structure ; at the root of the fin the vertebral column takes a 

 sudilen turn upwards, at a considerable angle. The centra are 

 well formed and retained throughout the length of the axis, and 

 consequently no urostyle is present. Nearly the whole of the 

 dermoti'ichia are supported by hypurals, a feature which it shares 

 with the majority of Teleostean fishes ; these h3'purals are slightly 

 expanded or somewhat club-shaped at their distal ends, and each 

 bears a single fin-ray. The last three hypurals belong to the last 

 centrum, but each of the other centra possesses but one hypural. 

 The final dorsal element (fig. 4, n.a^.) is strikingly large, and con- 

 tinues beyond the vertebral axis for some distance between the 

 .fin-rays. Kulliker remarks that it is cartilaginous and contains 

 the spinal cord ; he regards it as an incorporation of vertebi'a? and 

 neural arches. In the specimens I have examined, the same 

 structure proved to be bony, and grooved along the dorsal edge to 

 conduct the spinal cord ; it appears to correspond to a bone, fre- 

 quently found among the Teleostei, which T have regarded as a 

 persistent neural arch, retained, even when its adjacent homologues 

 are suppressed, as a protection for the delicate termination of the 

 spinal cord. According to Kolliker, there are three dorsal caudal 

 radials (d.c.r.) remaining as supports for the insignificant fin-rays 

 which constitute the epaxial contribution to the caudal fin ; but 

 these dorsal radials are very poorly developed. Most authorities 

 agree that the caudal fin of Amia is scarcely correctly described 

 by being called heterocercal, and therefore it is generally i-eferred 

 to as hemi-heterocercal. So near, however, does the structure 

 approach honiocercy that hemi-homocei'cal would probably be a 

 more suggestive designation : however that may be, the fin may 

 certainly be considered the most specialize;! type of heterocercal 

 forms. 



General remarks on the Ganoid Catulal Fin. 



It has been thought advisable to omit any detailed references 

 to other Ganoids than the three considered, for the present, since 

 they liave not been cleared up to my entire satisfaction and 

 material is lacking. But the Ganoid caudal is characteristically 

 heterocercal; truly protocercal forms ai"e absent, and homocercy 

 is nowhere attained within the group. The three types considered 

 illustrate how the tendency of specialization is to approach the 

 homocercal condition ; in Acipenser, the typical heterocercal form, 

 so characteristic of the Elasmobranchs, is present ; Pol^/ndou, by 



