458 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
With the further growth of the vesicle the yolk is usually infringed upon (fig. 145). 
Whether this is accomplished by the absorption of a part of the yolk, or whether it 
is simply crowded down, I am not able to say. I am inclined to believe that it is 
pushed down by the contents of the vesicle; otherwise the rearrangement of the yolk 
particles would soon fill up the gap made by absorption. However, in later stages of 
yolk absorption (fig. 136) the yolk is frequently eaten into in very much the same way 
without any apparent rearrangement of its parts. 
After the vesicle has reached its largest size it acquires a cellular floor which 
corresponds with the original outline of the yolk (fig. 146). The vesicle is thus cut 
in two, a larger yolk vesicle and an upper vesicle which forms the enlarged portion of 
the hind end of the intestine. The cellular floor seems to be formed by an ingrowth 
from all sides. Both sagittal and transverse sections show that in earlier stages 
hypoblast cells extend in below the margin of the vesicle. These would simply have 
to extend still further centripetally to form a floor for the whole vesicle. The depres- 
sion in the yolk is at the same, time converted into a spherical vesicle, which may be 
termed the yolk vesicle. It is sometimes partially filled with a brightly staining sub- 
stance, and as it is quite evident in entire specimens (figs. 84 to 86) I supposed it to 
be the yolk nucleus before I examined sections and learned the fate of that structure. 
In one larva the yolk vesicle was found at the ventral side of the yolk (fig. 85). The 
yolk vesicle remains longer than Kupfifer’s vesicle, and in fact the intestine extends 
some distance beyond it before it disappears. There seems to be no regularity in its 
disappearance, and probably no importance attaches to it after it is separated from 
the main vesicle. 
The cellular floor of the vesicle rapidly thickens as the tail begins to grow out, 
while its communication with the alimentary canal is maintained (fig. 147). The 
dome-shaped part is probably constricted off at this time (12 protovertebrae), and the 
remainder of the vesicle is reduced to the caliber of the rest of the alimentary tract. 
My sections tell conflicting stories about the exact processes. Figure 78 shows an irreg- 
ular cavity at the end of the mesenteron, which I consider the remains of the main 
vesicle. This cavity is connected by an area of disconnected cells with a vesicle lying 
dorsad of it. Figure 148, on the other hand, in which the tail is not as long as in 
figure 78, shows a dilation of the intestine dorsad, at the end of which lies a triangular 
cavity. It would appear that a part of the body of the original vesicle had been con- 
stricted off with the dome. The same appears to be true of figure 149. In both these 
larvae the cells below this cavity are disarranged. The connection of this dorsal 
cavity, or vesicle, with the alimentary canal is now represented by the disarranged 
cells only. The two different structures, vesicle and disarranged cells, may, however, 
be harmonized. The dorsal vesicle disappears by the proliferation of cells from its 
wall into its cavity. These cells would at first be expected to be arranged somewhat 
differently from those in the older structure. In figure 78, which represents the older 
condition, the process has gone far enough to obliterate all of the body of the vesicle 
which had remained with the dome. 
Neurenteric canal . — Several of the larvae described above leave no doubt as to 
the meaning of the dome-shaped structure of Kupfifer’s vesicle. It is part of the 
neurenteric canal. In all late stages, as in figure 148, there are fewer cells above the 
neurenteric canal than in neighboring regions, and in several cases a tubular connec- 
tion undoubtedly exists between the remains of the dorsal vesicle and the neural region. 
