124 PEOCEEDINGS OF THE AMERICAN ACADEBIY 



If we examine the tails of the Devonian fishes, — as we know them 

 from the restorations of Agassiz, Hugh Miller, Pander, Meckel, Pictet, 

 Huxley, and others, — we cannot fail to be struck with the exact paral- 

 lelism of these ancient fishes, as far as the structure of the tail is con- 

 cerned, with the structure of the successive stages of the tail of the 

 young Flounder, figured on Plate I. of this paper. 



We find, among the Devonian fishes, genera with truly leptocardial 

 tails, like those of PI. I. fig. 1, such as the genera Glyptok-emus, Gy- 

 roptichius ; also, genera with slightly modified tails (with the least 

 possible tendency to heterocercality), as in figs. 2 and 3, — Holoptichius 

 and Osteolepis ; next, such genera as Glyptolepis, where the hetero- 

 cercal tail is somewhat more marked, approaching nearer the form of 

 PI. I. fig. 4. But it must be remembered, that in all these genera, 

 although the outline of the tail-fin is much as in the figures here given 

 (PI. I. figs. 2-4), yet, as in Polypterus and still more in Ceratodus, the 

 scales extended over the dorsal column into the tail, in a triangular 

 shape ; and it is only in such genera as Dipterus that the heterocercal 

 character becomes more prominent, as in PI. I. fig. 4. 



This docs not by any means conclude the parallelism, which is still 

 more striking when we come to such forms as Phaneropleuron and 

 Tristichopterus, where the tail is lobed, the dorsal column extending 

 into the dorsal lobe exactly as in the stage represented in PI. I. 

 figs. 5, 6. 



In the Old Red, such genera as Acanthodes, Diplacanthus, Cheiro- 

 lepis, and the like, represent stages corresponding to those of PL I. 

 fio-s. 5, 6, 7, where we find the first indication of the separation of a 

 true caudal and of an embryonic caudal. In the subsequent modifica- 

 tions of the tail of fossil fishes (approaching Lcpidosteus), the tendency 

 has been gradually to lessen the upper embryonic caudal lobe, and to 

 give greater prominence to what is to become the caudal proper ; 

 although there is not, of course, the diiference in structure of the fin- 

 rays of the two sections to separate them, as in bony fishes. It is only 

 when we compare these older forms with such genera as Platysomus, 

 Semionotus, Lepidotus, and finally Pachycormus, that we trace the 

 gradual approach to an externally horaocercal tail, much by the same 

 process which we readily follow in the embryo fish through the corre- 

 sponding changes from PI. I. fig. 8, to PI. I. fig. 11. The gradual 

 shortening of the extremity of the chorda dorsalis, until it only extends 

 slightly in advance of the base of the caudal rays, is strictly analogous 

 to the disappearance of the embryonic caudal and the gradual develop- 



