VIVIPAROUS FISHES OF THE PACIFIC COAST. 
461 
-from the condition usually prevailing in fishes where it projects into the vitellus. He 
then states that this difference is of no importance (1881), for in a later stage the vesicle 
“peut faire saillie hors de I’embryon” and thus become as in other fishes. In other 
words, the process described by Agassiz and Whitman is here said to be inverted. In 
a stage with twenty-two protovertebrae it is figured as the enlarged posterior end of the 
alimentary canal. It would here, then, permanently form a part of the intestine, and 
in this respect agree with the middle vesicle in Cymatogaster. It elongates antero- 
posteriorly and becomes pyriform. It comes directly in contact with the nervous 
thickening above. It is only the first indication of the digestive tube. It is important 
on account of its relation with the nervous system and the notochord and the region 
corresponds to the region of the neurenteric canal of other vertebrates. He has never 
found a canal leading to the exterior. He considers the opinion of Kingsley and Conn, 
Agassiz and Whitman, Cunningham, Ziegler, that the vesicle lies between the periblast 
and the hypoblast, to be based on an error of observation. 
Frequently there exists below the vesicle or in its neighborhood a hemispherical 
depression in the surface of the vitellus. (It is possible that in this case the vesicle 
divides into two, a yolk vesicle and the intestinal vesicle seen in Cymatogaster.) He 
thinks the original opinion of Kupffer may still be defended when one considers that 
it is the first indication of the alimentary canal in the neurenteric region and that the 
allantois of higher vertebrates is but a diverticulum of the intestine appearing very 
early in front of the neurenteric canal. 
In Serranus, according to Wilson, the vesicle appears as an up-pushing of the 
hypoblast, which had previously become colummar, and a down-pushing of the 
periblast. It disappears by the proliferation of cells from its own walls. It is formed 
some distance behind the future anus and lies in the postanal gut. In this respect 
it greatly resembles the postanal vesicle of elasmobranchs, with which it is homologous. 
“Before Kupffer’s vesicle is folded off it represents the terminal dilatation of the 
archenteron itself, and in this phase is to be compared with the dilated posterior 
extremity of the archenteron in certain amphibian gastrulas.” 
McIntosh and Prince (1890) observed the vesicleiu a number of species. They find 
that it arises, as Kingsley and Conn have described, by the union of a number of 
granules or small vesicles. They found secondary vesicles quite frequently, some- 
times extending “all along the ventral line almost to the pectoral region.” Its 
contents are usually homogeneous and clear, though granules are occasionally present. 
They claim to have traced a neurenteric canal from the vesicle to the blastopore at the 
time the latter closes. 
Formation of liver and air bladder , mouth , thyroid gland , and hypophysis.— In a 
preceding chapter the intestine was described from its first appearance until, in 
larvae 1 mm. long, it forms a simple tube from the hyobranchial gill-slit to the anus. 
In the stages succeeding that the rudiments of the liver and of the air-bladder make 
their appearance. Parts of the tract are at the same time otherwise modified. 
In larvae 1-8 mm. long (figs. 150, 156, 157) the intestine is broad and depressed just 
behind the hyobranchial slit, and its walls are composed of ciliated cells 18 p high. 
In the roof the nuclei of the cells are slightly nearer the free ends of the cells and 
nearly all are on the same level. 
In the floor there are besides this layer of cells a few scattered ones among the 
bases of the others (figs. 156, 157). Towards the back part of the gullet the lumen 
