140 Records of the Indian Musevm. [Vol. XV, 



Before entering into the question, I would again draw attention 

 to the presence of independent radials as supports of caudal fin-rays. 

 Dorsal caudal radials as I have called them are usually present, and are 

 the bones referred to by Huxley^ as epurals ; ventral caudal radials 

 are also frequently present though much less than their corresponding 

 dorsal homologues. In all Gadidae, Solea, Zeus, Gobius and others, 

 radials persist ventrally. 



Now since radials are the normal supports of the dermotrichia in 

 other median fins, we may assume that they once were in the caudal, 

 which is merely a part of the same system. Haemal arches, therefore, 

 cannot be regarded as the original supports of fin-rays, and if they are 

 considered to have taken over this function, they must be regarded as 

 having lengthened and expanded distally with a view effectively to fulfil 

 their new function. On the other hand, radials are the natural supports 

 of fin-rays. It is also interesting to notice which of the two elements, 

 radials and arches, are the more persistent when subjected to eliminating 

 influences. The flexion of the extremity of the chorda affected the 

 dorsal and ventral side of the fin differently ; epaxial elements had less 

 room assigned to them, while the hypaxial structures were afforded 

 scope for extended development. The response to this influence is very 

 marked ; on the ventral side fin-ray supports have expanded to fill 

 the widening cleft between them, but dorsally suppression has resulted. 

 But which structures dorsally have succumbed to this crowding-out 

 process ? Not the radials but the neural arches ; as long as dorsal fin- 

 rays remain to be supported, the radials retain their function, while 

 neural arches have disappeared or been reduced. We are forced there- 

 fore to regard radials as more persistent than arches. 



When the caudal fin became a definitely propulsive organ, more rigid 

 support was required for the fin-rays ; to accomplish this there were 

 clearly two ways open ; (1) by the transference of the supporting func- 

 tion from radials to haemal spines, and (2) by the mere fusion of the 

 radials with the haemal spines. The first alternative involves, one 

 might almost say, a preconceived purpose on the part of the s^Dines to 

 acquire a new function ; they must lengthen, expand and, in so doing, 

 eliminate structures already performing the work they are to take over. 

 Such a change of function is usually accompanied by a stimulus, in 

 response to which the change takes place ; but it is difficult to see what 

 stimulus could have been applied to the haemal arches to initiate a 

 change. The radials clearly had a stimulus, a mechanical one, and one 

 cannot avoid concluding that they responded thereto, thus obtaining 

 as it were a start on any other competitor. 



It has been shewn that radials are more persistent than arches dor- 

 sally ; is it likely therefore that in a region where extended scope for 

 development is afforded, as is the case ventrally, that structures already 

 well adapted for the support of fin-rays, and indeed actually fulfilling 

 that role, should abandon their function in favour of structures in no 

 way so fitted, especially when, under adverse conditions, as dorsally, 

 they tenaciously retain that function ? It is inconceivable, and one is 

 tempted to ask what would happen during the period of transference. 



I " On Some Parts of the Skeleton of Fishes/'. 9. J. M. S., 1859 



