84 0\' THE STRUCTURE OF BOXE IX FISHES. 



Chondrostei, it would seem that the Amioidei and Lepidosteoidei 

 have been derived from a common ancestral form which developed 

 the lepidosteoid scale and bone, diverging in this and other 

 respects from the remainder of the Teleostomi. 



It becomes now a matter of great interest to ascertain from 

 which group the Teleostei may have been derived. On general 

 anatomical grounds they would certainly be associated with the 

 Amioidei (2). Bat so far I have been unable to discover the 

 lepidosteoid structure either in the scale or in the skeleton of 

 any living or extinct species of Teleost, even after the examina- 

 tion of representatives of a very lai-ge number of families. In 

 the lower Teleostei (including the Leptolepidae) the bone is of 

 the ordinary structure, similar to that of Osteolejns or Palceoniscus ; 

 but, as is well known, in the higher forms it becomes generally 

 very much modified, chiefly owing to the loss of the bone-cells. 

 Only in the Fistulariidae does the structure of the bone i^ecall 

 that of the Amioid. Even here, however, the resemblance is not 

 close, and I have not been able to convince myself that the fine 

 canals described by Stewart (6) are really homologous with 

 lepidosteoid tubules. 



In connection with tlie phylogeny of the Teleostei it is 

 interesting to note that Oligopleitms vectensis A. S. W. has no 

 lepidosteoid tubules in its endoskeleton. OUgojyleurus esocinus 

 Th. I have not had an opportunity of examining ; l:)ut Spathmrus 

 and (Enoscopus, the only other genera belonging to the family, 

 have the typical lepidosteoid structure in the scales, dermal bones 

 and endoskeleton. Associated with a skeleton of Oligopleurus 

 vectensis in the British Museum is a plate, either a scale or a 

 dermal scale-like bone, of typical lepidosteoid structure ; it 

 cannot, however, be made out for certain whether this bone 

 belongs to the skeleton or not. O. vectensis certainly differs in 

 bone-structure from other Amioids, and probably should be 

 placed not with the Oligopleuridee, but with the Teleostei *. 



We may suppose that the common ancestor of all the Holostei 

 (Amioidei, Lepidosteoidei, and Teleostei) was some primitive 

 Actinopterygian with lepidosteoid scales, but in which the 

 lepidosteoid structvire had not yet penetrated to the endoskeleton. 

 If this supposition is correct, " Oligoplenrios " vectensis might 

 perhaps be a representative of such an nnspecialised group. 

 Placed by some authors among the Amioids, and by A. Smith 

 Woodward among the primitive Teleosts (7), it certainly appears 

 to be a somewhat intermediate form. 



As a provisional hypothesis, the view may be adopted that the 

 Amioids and Lepidosteoids on the one hand, and the Teleosts on 

 the other, diverged from a primitive group possessing lepidosteoid 

 scales and ordinary bone ; and that in the former the lepidosteoid 

 sti-ucture spread inwards over the whole endoskeleton. The 

 extreme modification of the scales in the Teleost series would 



* Mr. Rpgaii informs me that O. vectensis is without fulcra, and resembles the 

 Leptolepidiv in the structure of the tail, and shoiild be included in that family. 



