[21 J THE EVOLUTION OF THE FINS OF FISHES. 1001 



arrived at respectiug the serial homology of the epiiral and hypural 

 bones in the tails of Eolocephali, Ganoids, and Teleosts. 



An examination of the tail of the Eel has, as already stated, led to the 

 belief that a part of the hypiiral bones are the homologues of iuter- 

 s})inous bones; but we find that even in this case they are present as a 

 pair of successive appendages on the inner side of the ultimate and 

 penultimate vertebriB. The next step in our inquiry will therefore be 

 the following : What is the probable cause of such a duplication or mul- 

 ti})lication of median ai)])eudicular elements on the under side of the 

 posterior end of the vertebral axis? The antepenultimate vertebra^ 

 which has nothing to do with supporting the caudal lobe, has a single 

 neural arch and two haemal arches, and to these correspond two inferior 

 interspinous elements. This would seem, therefore, to indicate that the 

 double hypural elements of the penultimate and ultimate vertebra? were 

 in part, or at least distally, the homologues of interspinous elements. 



The neural arches of the penultimate and ultimate vertebroe are want- 

 ing, and there are no superior interspinous pieces which belong to the 

 two last vertebral segments. This implies a degeneration and complete 

 atrophy of the dorsal interspinous pieces pertaining to the caudal ver- 

 tebrae, while it is evident, from what has been said above, that the in- 

 ferior homologues of the suppressed upper pieces are present. 



The inferior elements, or true hypurals, of the Eel aie present as four 

 hyi)axial apophyses formed in cartilage and ensheathed in membrane 

 bone in the young animal. In the old animal these four elements are 

 reduced to two, because of the fact that the sheaths of membrane bone 

 investing the two successive two in the young become co-ossitied. 



Further possible conclusions are derivable from a study of the struct- 

 ure of the tail of the Eel. The penultimate vertebra bears two inferior 

 arches, which are hajmal, but it supports only one neural arch. This 

 may imply that at one time this vertebra was compound, or double, 

 and that it is now fused into one, having lost one of its neural arches. 

 Such a duplication of centra occurs in the caudal part of the axial col- 

 umn of Amiay but here the appendages of the alternate segments are 

 suppressed. 



The concrescence or crowding together and fusion of the proximal 

 ends of the epural and hypural elements, so frequently noticed, seems, 

 however, to be really due to the tendency of the heterocercal caudal 

 structure to incline toward the development of that of the gephyro- 

 cercal condition, as pointed out in another place. 



The diplospondylism of part of the caudal axis of Amia is not ap- 

 l)arently due to a process of concrescence, but to the elision of some 

 of the skeletal appendicular parts of alternate segments. At the 

 other extreme the excessive multiplication of epaxial arches to a single 

 segment in LophobrancMi is an instance of the acquirement of super- 

 numerary arches, which it is difficult to account for unless it be supposed 

 that the single myotomes of the adults of the existing forms represent 



