156 GEORGE W. TANNREUTHER 
persists until late development, it serves as an unmistakable 
point of orientation. 
Figures 8 and 9 show the four-cell stage from the left and 
right sides. The dorsal ventral axis of A, B and C is about the 
same as that of D, but immediately after the formation of the 
first generation of ectomeres, the cells A, B and C shorten and 
become more rounded (fg. 16). In later stages of development 
these cells often become very much flattened and cause the 
developing embryo to appear unusually large, when viewed 
from the upper or lower poles. 
Third cleavage (eight-cell stage): In the formation of the first 
generation of ectomeres (d', c', b! and a'), each of the four macro- 
meres divide obliquely. The ectomere end of the cleavage spin- 
dle is uppermost. The macromere D divides first; d! is budded 
off from D towards the upper pole, in the direction of the hands 
of a watch (dexiotropic), (figs. 10-11). We have thus a five- 
cell stage. Each of the macromeres C, B and A next bud off 
a small cell towards the upper pole. These are not formed 
simultaneously, but in the invariable order ¢', b! and a'. Thus 
there occurs successively, a six, a seven and an eight-cell stage 
(figs. 11-14). In figure 13, an upper pole view, D and C have 
divided and A and B are preparing to divide. In both A and B 
the asters of the ectomere end of the spindles are visible. The 
position of the opposite end of the spindles are indicated by 
circles. This figure shows the oblique nature of the cleavage 
spindles. The spindle in A points to the space between A and 
B. The spindle in B points to the space between B and C. In 
figure 14, an eight-cell stage, the exact relation of the ectomeres 
and macromeres are shown as they normally appear from the 
apical pole. The position of the first generation of ectomeres 
is obvious. They suggest a possible rotation, after their forma- 
tion, through an angle of about forty-five degrees in the direc- 
tion of the hands of a watch. If actual rotation did occur there 
would be no difficulty in explaining their final position. But 
the fact that the cleavage spindles are oblique and the position 
of the completely divided nucleus can be definitely determined, 
before there is any indication of the cytoplasmic division of the 
