[9] 



THE EVOLUTION OP THE FINS OF FISHES. 



989 



tieally unimpaired, thus afifording the necessary proof of the serial ho- 

 mology of the eutire series of median fin-rays and tbeir intermediary 

 supports. Previous authors failing; to attack this i)art of the problem 

 by the light of the ontogeny of a diphycercal, eel-like type, have missed 

 the solution of one of the most important minor parts of a rational 

 theory of the median fins, since it is otherwise impossible to prove the 

 existence of such a homology iu forms with atrophied intervals between 

 the vertical fins. The existence of the protopterygian stage also tends 

 to prove this view to be correct. The skeletogeuous tract, from which 

 the whole of the median fin-rays and their supjiorts are developed, is 

 continuous in the median line of the urosome, above, below, and over 

 the end of the chordain fish embryos; such continuity affords an expla- 

 nation of why the median fin-rays form an uninterrupted series in the 

 case of perfectly diphycercal forms, or where the archaic has not yet 

 been replaced by a specialized mode of development. Such an archaic 

 condition is actually retained with but slight impairment by the em- 

 bryo of the salmon, which has a nearly continuous series of embryonic 

 fin-rays. 



(4.) Heterocercy. — Heterocercy affects only the end of the chorda! axis 

 which is bent upwards (Fig. 5), as a result of which it and the later 

 formed terminal vertebral segments are consoli- 

 dated into a urostyle (in manj^ Teleostei), above and 

 below which epaxial and hypaxial elements are 

 formed, of which the former are, however, often 

 aborted and the latter widened as supports for the 

 caudal system of rays. This condition appears 

 to result from two causes, (1) great activity of growth in the terminal 

 hypaxial i)art of the x^i'iuutive fin-fold, in consequence of which the 

 chorda is shoved upwards, and (2) by the actions of the animal in 

 using the resulting expanded hypaxial caudal ray-bearing fold in swim- 

 ming; the strokes of the fin in action, owing to the resistance ottered 

 hy the water tend to throw up the somatic axis, just as an oar tends to 

 be thrown upwards when used in sculling. 



Since the hypaxial fold may be developed at some distance from the 

 end of the tail, in the more specialized forms (Lejyidostens, Gasiero-stcus), 

 a more or less free portion of the lophoccrcal caudal axis is left to 

 project during the growth of the true caudal, as shown in Figs. 6 and 



Fu^. 



F^.7 



7. This part of the larval axis, which may be called the opisfhure, 

 subsequently degenerates, or it may persist as a i)rolougation of the 



