424 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
pelagic eggs. But Wilson’s figure (1) does not show conclusively that the protoplasmic 
cap is not the result of reagents. The oil-globule figured by Wilson is not of the shape 
of the oil-globules seen in living eggs and the space occupied by the protoplasmic cap 
may be due to contraction. I have seen similar spaces in other eggs filled with stainable 
material. Its position remaius fixed at the entodermic pole of the egg till the closing of 
tlie blastopore, and is therefore an excellent landmark for orientation. 
Segmentation . — The first segmentation plane is meridional and divides the germ 
iuto two equal blastomeres. The surface views of this stage are represented in 
figs. 1 aud 3. The groove separating the sphere varies in depth according to the 
stage of division. The constriction between the yolk and germ also differs in 
different eggs. The nuclear figures can not be distinctly made out, and it was but 
once that I was able to definitely distinguish the resting nuclei (fig. 3). Sections show 
that the dividing plane reaches entirely through the germ to the yolk before the 
second division begins and that the line dividing the germ from the yolk, though 
irregular, is well marked (figs. 4 and 5). The division is further indicated by the fact 
that yolk and germ readily part. The yolk shows no signs of segmentation, and if the 
yolk ever does segment in teleosts it ought certainly to sliow some signs of it where so 
little resistance is to be overcome as in Cymatogaster. The conditions described by M. 
Kowalewski, 1883, for the goldfish are probably not due to the segmentation of the 
yolk, but to the retarded segregation of yolk from germ and the consequent indistinct 
boundary between the two. During this first segmentation there is still considerable 
protoplasm scattered through the yolk, and the lower margin of the blastoderm, 
though quite distinct, is not yet as well defined as in later stages. The germ, on kill 
ing with strong O. 0. A. aud staining with Grenacher’s hsematoxylin, is seen to be 
finely granular, with a few larger granules, many of them arranged along the division 
planes. A granular network of protoplasm extends also through the yolk. A few 
yolk-spheres are, on the other hand, found in the germ (4), some of them being arranged 
along the division plane. The yolk nucleus at this stage is stained darker than the 
germ, has an irregular outline, and is apparently composed of protoplasmic granules 
similar to those of the germ, but more compact, and of a few yolk-spheres (fig. 4, yk.pr.). 
The second segmentation plane is also meridional aud at right angles to the first. 
This is seeu both in segmentation spindles of mounted 2-eelled eggs and in the 
division plane itself (fig. 2). 
Authors in general have been agreed that the third segmentation planes are 
parallel to the first in fishes. List, 1887, on the other baud, claims that the second 
segmentation appears almost simultaneously with the first in Crenilabrus tinea and is 
equatorial, and that the third is again meridional and at right angles to the first; 
while Brook, 1886, maintains that in the herring the third is the equatorial segmen- 
tation. Hoffmann, 1888, claims to have found the fourth segmentation to be hori- 
zontal in the salmon, but since Brook found the third to be horizontal or equatorial, 
he concludes that this is the rule, considering Amphioxus, Cyclostoms, and Amphib- 
ians. The fourth cleavage of Rauber, Kowalewski, Ryder, Agassiz, and Whitman, 
that is, the cleavage which according to them divides the germ into 16 cells, is not the 
fourth, but the fifth, etc. (Hoffman, 1888, p. 532). Wilson considers the third cleavage, 
the one parallel to the first, the homologue of the fourth in the frog. 
