1050 EEPORT OF COMMISSIONER OF FISH AND FISHERIES. [70] 



On PI. I, Fig. 1, the archicercal filament of Chimcera monstrosa shows 

 that there is a tendency in one of the lowest types to abort the end of 

 the caudal axis and. develop the true caudal farther forward. Turning 

 now to a repiesentative of the Teleostei the same thing- is repeated in 

 Gasirostomus Bairdii, in which the tail is actually lophocercal at the tip 

 in a si)ecimeu one-third grown, and with no epaxial or hypaxial rays 

 developed for a considerable distance in advance of its termination. 

 This last type is, however, archaic in some respects. Turning now to 

 Polypterus (PI. V, Fig. 2), the chorda is exserted posteriorly, and does 

 not support neural or haitnal arches for some distance behind the point 

 where the development of vertebral bodies is discontinued. The tail is 

 roundly fan shaped; some proximal concrescence of the hypnrals is 

 evident. The young of Lepidosteus (Fig. 6, PI. V) shows that the termi- 

 nation of the chorda does not develop inferior and superior arches. 

 Yet the embryonic rays are developed in the fin-fold jjosterior to the 

 last arclii's. In the next stage the opisthure (Fig. 5, op.^ PI, V) is fully 

 developed, the evidence of degeneration which it affords being com- 

 plete. In Fig. 3, PI. Y, of the adult tail, the end of the unabsorbed 

 remnant of the still more degenerate urochord extends into the dorsal 

 l)ortion of the caudal fin. Concrescence or crowding together of the 

 jnoximal ends of the sujjports of the rays is evident. Jn Plati/somus 

 (Fig. 4, PI. Y) the end of the chorda is uiore massive, more persistent, 

 with a less marked proximal crowding of the bases of the rays, contrast- 

 ing strongly in this respect with the primitive caudal skeletons of Coc- 

 costeufi, Fig. 1, and Centrina, Fig. 8. 



Turning now to the contenjplation of the changes which go on during 

 development, as shown on Plate I, the evolution of the heterocercy of 

 the Flounder's tail displayed in Figs. 3 to 10, after Agassiz, shows 

 (hat the crowding from behind forward of the i)roximal ends of the 

 epaxial and hypaxial pieces keeps pace with the gradually increasing 

 heterocercy. On Plate III, illustrating the formation of the caudal 

 skeleton of the Salmon, the same thing is again shown, actual coucr^es- 

 ceuce or blending of two hypaxial elements being in progress in Fig. 4. 

 , In the caudal skeleton of adult Teleosts, that of Barbus (PI. YI, 

 Fig. 3) illustrates the proximal crowding referred to. In Salmo fario 

 (Fig. 1, same plate) and in S. salar, this is illustrated again, but the 

 urochord or urostyle is not so extensively exserted as in Fig. 3. JSalmo 

 salar also has the end of the chorda more sharply bent upward, and the 

 crowding proximally of the hypural pieces is more pronounced. The 

 hinder epural pieces, ep' ep', are slid backwards somewhat in consequence 

 of the markedlj" upbent urostyle, so that if the latter were now aborted 

 a structurally gephyrocercal tail would result, because the epaxial and 

 hypaxial elements would then form a confluent series. In the caudal 

 skeleton of Cottus (Fig. 1, PI. YII) a more si)ecialized type of heterocercy is 

 shown, in which, as in Gasterosteus, Anguilla^ and Fistularia, the number 

 of hyi)axial processes are reduced. The urostyle in these is almost sup- 



