[73] EMBRYOGRAPHY OF OSSEOUS FISHES. 527 



hypoblastic or nethermost embryonic layer, which is uotably thickened 

 at an early stage along the ventral side of the body of the embryo, but 

 is still quite thin or almost wanting underneath the head; in fact it ap- 

 pears to be almost entirely undeveloped below the fore and mid brain 

 at the time the blastoderm closes. Its condition shortly after the closure 

 of the blastoderm in the embryo cod is shown in Fig. 31, at i, where its 

 solid riuliment is visible as a band of cells underlying the notochord in 

 the living egg. While its anterior extremity is not traceable to below 

 the anterior end of the head, the posterior extremity is lost in the caudal 

 mass at ?',with which it is continuous. This relation of continuity of the 

 hinder end of the intestine with the caudal mass shows that we must 

 regard this condition as homologous with that observed in AmpJiioxus, 

 Elasmobranchs, and other forms where the continuity of the neural canal 

 at its posterior extremity with the intestine is efi'ected through the inter- 

 mediation of a short post-anal section of the latter or a neurenteric canal. 

 This primitive continuity of the neural tube with the intestine has been 

 so fully elucidated by Kowalewsky, Hatschek, and Kupffer that it is 

 only necessary to refer to their memoirs on the subject and especially 

 to the general tre^itise and the monographs of Professor Balfour. While 

 I have found it imjiossible to convince myself by means of sections that 

 there ever exists a neurenteric canal in embryos of osseous fishes, I feel 

 assured that the solid nature of the posterior end of the neurula obscures 

 this relation and prevents the development of it. This does not, how- 

 ever, permit us to deny the jiossibility of a primitive union of the enteric 

 and neural tracts at the tail, and thus to realize a gastrula stage of devel- 

 opment for the Teleostei. Kupffer has placed some observations upon 

 record in regard to the connection of the vesicle named after him with 

 the hinder part of the neurula. I have already remarked of Kujiffer's 

 vesicle that it is an evanescent structure, and of uncertain significance in 

 relation to any organs developed afterwards. It has been observed by 

 me in the ova of Gadus, Alosa, Cj/bium, Tylosurus, Coregonus, Ajjeltes, and 

 two undetermined forms, so that it seems to be pretty generally present. 

 In the cod its relation to the yelk blastopore hi is shown in Figs. 26, 

 28a, 29rt, 29&, 30«, 30b, 31, and 32, at Kv. In 30b it seems to bo joined by 

 a fine canal to the blastopore, and in 28a it appears to be provided with 

 a cellular wall. Its relations are, not, however, constant, as may be 

 inferred from an inspection of the different figures in which it is rejire- 

 sented as present. After the stage shown in Fig. 32 had been passed, 

 I was no longer able to identify it with auy succeeding structure which 

 it could be sujiposed was derived from it. I therefore reserve my de- 

 cision as to its true nature. In some forms it appears long before the 

 closure of the blastoderm, in others coincidently with that phenomenon. 

 The great generalization, first distinctly formulated by llaeckel, that 

 animals generally, pass, in the course of their development, through a 

 gastrula stage, applies to the osseous fishes, and, notwithstanding the 

 uncertain fate of Kupfi;er's vesicle, it is evident that the caudal plate, 



