[19] THE EVOLUTION OF THE FINS OF FISHES. 999 



dicbotomous rays in Anguilla, Fig. 4, PI. 4, aud it would seem that the 

 apodal coudition and the elongate body had reacted so as to prevent 

 the development of heterocercy, oi it may be that this form has descended 

 from a type which began to be heterocercal, but which lost that condi- 

 tion to some extent in the course of becoming ai^odal, when the tendency 

 to preserve the tail in its primitive form began to assert itself and pre. 

 vented further caudal degeneration, and that loss of terminal segments 

 which we have seen has occurred to some extent in the course of the 

 evolution of the fishes. 



Kolliker,* as a result of his investigations, proposes the following pro- 

 visional scheme of classification for the structural conditions found to 

 obtain at the hinder end of the axial columns of fishes : 



A. The end of the vertebral cohimn incompletely or entirely unossified. 



I. The end of the vertebral column without a spinal canal, but consists — 



1. Of the chonla alone, Esox. 



2. Of the chorda principally, which is, however, covered by a short more or 



less complete cartilaginous sheath, Salmo, Alosa, Elops. 



3. Of a complete cartilaginous tube which incloses the end of the chorda, 



Cyprimis. 



II. The end of the vertebral column consists of a cartilaginous sheath, which 



incloses both the ends of the chorda and the spinal cord, Polypterus, Lejnd- 

 ostetis, Amia. 



B. The end of the vertebral column completely osssified. 



I. The end of the column is not segmented, but consists of a longer or shorter 



bent bone {urosiyle, Huxley), which is to be regarded as an ossification en- 

 veloping the chorda, and which resembles, more or less, anteriorly, a verte- 

 bral body. All (?) Acanthopteri, Malacopferi in part, 



II. The vertebral column ends with a simple vertebral body, Flagiostomi with a 



fully ossified vertebral axis. 



These histological and morphological results of Kolliker's researches 

 are of great interest, but they throw but little light upon the questions 

 which we are attempting to elucidate on the basis of embryological and 

 mechanical hypotheses, until they are viewed through the medium of 

 the latter. The facts given in the table above merely show, as all of 

 our studies have shown, that all the types of caudal differentiation found 

 in fishes are stages tending toward the concentration of the caudal skel- 

 eton towards a point in advance of the hinder end of the chorda. The 

 differences observed in the details of structure at the posterior end of 

 the chorda are such as merely indicate phases of specialization. This 

 is quite clearly demonstrated by the fact that a series of conditions 

 might be jjicked out from this table which would roughly correspond 

 with a series of the stages of development of a young fish. 



IV. - The serial homology of the hypaxial and epaxial ele- 

 ments OF THE CAUDAL FIN. 



Huxley has held that the hypural pieces of Gasferosteus were in real- 

 ity composed of haemal and interhsemal pieces, as indicated by develop- 



* Ueber das Ende der Wirbelsaule der Ganoideu und einiger Telostier. 4to. Leip- 

 zig, 1860. 



