VIVIPAROUS FISHES OF THE PACIFIC COAST. 
425 
I have seen nothing of a horizontal third cleavage in the various species of eggs 
I have examined, and, in the face of so much evidence to the contrary, am doubtful 
whether the third cleavage is equatorial in the sense used by Hoffmann, i. e., horizontal. 
If the first equatorial segmentation is the third cleavage and horizontal, then, indeed, 
the large yolk accumulated at one pole is an intimate part of the lower cells with 
which it is represented to be connected and does not divide for lack of sufficient energy 
in the comparatively small cytoplasm overlying it. But in Cymatogaster , in which the 
yolk is much smaller than the germ, it is just as distinct from the germ as it is in large- 
yolked species, and much more so than is represented for some large-yolked species 
by other authors (Kowalewski). It shows, however, no signs of segmentation or 
other interest in the changes of the cytoplasm; on the contrary, it is during the early 
stages of segmentation as impassive as so much dough might be. While I am not 
prepared to believe that it is not an intimate part of the ovum, I do believe that, 
in those large-yolked species in which the yolk and germ are well separated before 
segmentation begins (pelagic eggs, for instance), the nuclear spindles guided by the 
forced shape of the germ may be displaced from their primitive direction. In other 
words, owing to the presence of a large yolk the four blastomeres have a greater lateral 
than vertical dimension. The effect on cleavage is what it might be expected to be. 
The third cleavage spindles, instead of being vertical, have been forced by the depressed 
condition of the germ to lie in a horizontal or slightly inclined position, and the third 
cleavageof theteleostean egg instead of being normal (equatorial) has become parallel 
to the first cleavage. A similar result has been obtained by compressing the segment- 
ing eggs of the frog between two glass plates. By the compression the horizontal 
cleavage was changed into vertical cleavage. 
An examination of the figure of the external features of an 8-celled egg, and the 
cross-section constructed from a wax model and made after a series of 13 oblique sec- 
tions of an 8-celled egg, will lend force to the supposition that the third segmentation 
was originally equatorial. For now that the yolk has been greatly reduced, and the 
germ is no longer a comparatively flat particle, but has approached the spherical form 
again, the third segmentation is no longer vertical, but is oblique (figs. 6 and 8); that 
is, it has approached the original horizontal or equatorial segmentation. This third 
segmentation is probably not similar to the horizontal segmentation which Hoffmann 
claims for the salmon, and which is said to divide the germ into an upper and a lower 
layer. It may, however, be similar to that described by Brook in the herring, since 
the proportion of yolk to germ in Clupea approaches more to the proportions found 
in Cymatogaster. In the 8-celled stage, all of the cells, with perhaps one or two 
exceptions, touch the yolk, and all without exception form part of the external surface 
of the germ; they do not form distinct layers of four upper and four lower cells. 
Of the 16-cell stage I found but one egg. It was killed, as usual, with Flemming’s 
strong osmic chromic acetic, but stained with Gren. alcoholic borax carmine. It was 
examined in toto and a series of optical sections made, the planes passing through the 
resting nuclei. The cells are still but one layer deep, unless possibly one occupies the 
center of the group of cells (figs. 11-14). 
After the 16-cell stage the blastoderm becomes two cells deep. Such striking 
bilateral symmetry as is seen in pelagic eggs I have not observed, and after the 16-cell 
stage is passed the determination of the first and second cleavage planes becomes 
largely conjectural. The stages between 16 and 70 cells are represented in vertical 
and horizontal sections in figures 16 to 23. 
