190 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
the cytoplasmic division completed before the next nuclear division 
started. ‘This delay in the cleavage of the egg is so extreme in some 
instances that cytoplasmic division does not begin till a number of 
nuclei are present. Similar conditions are common in other coelenter- 
ates, and have already been indicated by other observers. 
The result of the first and second cleavages may be respectively two 
and four equal blastomeres, as is foreshadowed in Figure 58. ‘The 
eggs which I have seen at the end of the first cleavage, in cases where 
the division of the egg followed immediately on that of the nucleus, 
showed approximately equal parts; but the second cleavage may 
result in the formation of very unequal parts, and from this time on 
inequality is the rule. I have no doubt that the first division may also 
sometimes result in unequal blastomeres. Figure 59 illustrates the 
inequality in the size of the blastomeres resulting from the second 
cleavage. 
b. Later Cleavage.— Usually the cleavage planes for a considerable 
time correspond to meridional divisions in eggs that segment more 
regularly. An equatorial division rarely occurred before the 12-cell 
stage, and usually it was much later in making its appearance. Sec- 
tions from eggs with 20-30 nuclei each are shown in Figures 60-62; 
the meridional furrows, the delay of the equatorial cleavage and the 
polarity of the egg, all being well shown. ‘The elongated shape of 
these eggs is probably due to external causes, the development of 
several eggs in one gonophore having led to a compression which has 
influenced the direction of growth, and may have helped to determine 
the position of the planes of cleavage. 
Traces of segmentation cavities are seen in Figures 60-62, and such 
cavities, I believe, usually occur. In this I must differ from Allen 
(:00), who did not in any instance find a cleavage cavity in Tubularia. 
It is clear from these sections that the cleavage cavity is not a typical 
one, such as we know in eggs of echinoderms, for instance. Several 
separate spaces occur between adjacent cells, beginning as early as the 
8-cell stage. It is doubtful whether these spaces unite to form a 
single cavity in all cases, perhaps they do not even in the majority of 
cases; there is no question, however, that a single cavity does arise in 
some instances. For example, Figure 63 (Plate 8), a section of an 
egg with 40-50 cells, shows a rather irregular, but single segmenta- 
tion cavity, which is for the most part surrounded by a single layer 
of cells. This stage, I believe, corresponds to the typical blastula of 
other animals, from which it differs chiefly in its irregularity of form. 
That this stage, which marks the end of segmentation, does not always 
4 
7 
