No, I.] THE GENUS OF FOSSIL FISHES. 



49 



nous processes. It might be said, possibly, that the basal 

 pieces are the proper arches, while the pieces which are borne 

 on them are the spinous processes. I hold that there are two 

 objections to this view. The first is, that what are sometimes 

 called free spinous processes are always unpaired pieces. The 

 second is, that when the lateral halves of the arches remain 

 distinct from each other and are prolonged into spines, as they 

 are in various fishes, Amia and Salmo, for instance, the spinous 

 portion is never, so far as we know, developed in the embryo 

 as pieces separate from the base of the arcuale. This is true 

 in the case of Amia, which I have investigated. We must, 

 therefore, seek some other explanation. The key to the under- 

 standing of the problem is, it seems to me, to be found in the 

 vertebral column of that primitive fish, Amia. We may call 

 this fish to our assistance since the Isospondyli are believed to 

 have had ancestors not far removed from Amia. 



In the middle region of the tail of Amia there are for each 

 muscular segment two vertebral rings, the one bearing the 

 arches, upper and lower, the other archless. If a transverse 

 section be taken through the middle of the arch-bearing ring, 

 there will be found an X of cartilage, the upper arms of which 

 are continuous with the cartilage of the neural arch. In like 

 manner the lower arms will be seen to be continuous with the 

 cartilage of the haemal arch. If a section is made similarly 

 through the archless disc, a similar X of cartilage is found ; but 

 the arms project beyond the outer surface of the disc but a 

 short distance. These archless discs are developed in Amia 

 from ossifications arising in the intercalated cartilages, upper 

 and lower, and the arms of the X are the unossified portions of 

 these cartilages. There appears to be no reason why these 

 intercalated cartilages should not sometimes take on a hyper- 

 trophied growth. In the sharks they often become consider- 

 ably larger than the true neural arches themselves. 



In case these intercalated cartilages should become thus 

 enlarged and arch-like, each might develop a bony investment 

 that would simulate the bony neural half-arch, and thus would 

 rest on the top of its proper epicentum.^ 



1 For figures illustrating the architecture of the vertebral column of Amia, see 

 the May number of the Amei-ican A\itnralist of the present year. 



