50 REPRODUCTION AND DEVELOPMENT OF TELEOSTBAN FISHES. " 



before "hatching out of a solid mass of mesoblast cells situated below 

 tlie pbarynx, and continuous on either side with the lateral meso- 

 blastic plates which lie on either side of the notochord. I have not 

 studied the histological process by which the heart itself is formed. 

 Suffice it to say that it is produced by the formation of the central 

 raesoblastic cells into a tube, which, as soon as it has a lumen, 

 communicates with the space between the ventral epiblastic body 

 wall and the periblast. The cavity in which the tubular heart is 

 contained is due to a splitting of the mesoblast, and is continuous 

 superiorly and posteriorly with the cavity formed on each side of the 

 embryo by the splitting of the lateral mesoblastic plates. The 

 cavity containing the heart is in reality the first part of the ventral 

 region of the body-cavity to be formed. A section farther back, 

 through the centre of the yolk-sac at the stage immediately after 

 hatching, shows a cavity on each side of the embryo in the meso- 

 blastic plates ; the splanchnic mesoblast forms the roof of the cavity 

 between the periblast and the wall of the yolk-sac. The mesoblast 

 only extends a little way from the embryo, so that laterally and 

 ventrally the wall of the yolk-sac is formed solely of epiblast. The 

 lateral body-cavities communicate anteriorly with the ventral body- 

 cavity which contains the heart. 



What then does the blood space between periblast and wall of 

 yolk-sac correspond to in those Teleostean ova which possess a 

 vitelline circulation ? Obviously the vitelline veins and their tribu- 

 taries communicate with the posterior end of the heart, just as does 

 the single extensive blood space in pelagic ova. This blood space 

 corresponds, and in a sense is homologous with the cavities of the 

 vitelline veins. Suppose the vitelline veins and their capillaries to 

 open out sufiiciently so as to coalesce, and we have a single space 

 extending over the surface of the yolk and communicating with the 

 posterior end of the heart. This supposed continuous space would 

 then correspond with the venous space in the pelagic ovum in a 

 general way j but would it correspond exactly ? The vitelline veins, 

 where they exist, are canals running through a layer of splanchnic 

 mesoblast covering the periblast. Now, in a pelagic ovum the meso- 

 blast is, as far as we can judge from our present knowledge, limited 

 to the embryo and the embryonic ring, and at a later stage to a narrow 

 region at the sides of the embryo, which region is derived from the 

 embryonic ring after the latter has disappeared as such. The space, 

 therefore, over the rest of the yolk between the periblast and the 

 epiblast, i. e. the segmentation cavity, can only close at these early 

 stages by contact occurring between the epiblast and periblast. 

 Now, such contact seems to occur at the stage immediately after the 

 enclosing of the yolk has taken place, but it is certain that at a 



