104 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
taneously in line to their respective poles. Here they soon develop 
a new nucleus, and the cytoplasm constricts in the usual way. 
25. The mitosis of the secondary spermatogones is quite similar 
to that of the mother branch cell and the primary spermatogones. 
26. After the last division of the secondary spermatogones with 
its peculiar prophases, the cells, now spermatocytes, enter on a 
period of growth. 
27. The cytocyst containing growing spermatocytes differs from 
the gonocyst in being hollow, the spermatocytes being arranged in a 
single layer and definitely oriented with reference to the cyst wall. 
28. Besides the ordinary small spermatogones derived from the 
mother branch cell, there are giant stalked spermatogones in the testes 
of the mature butterfly, especially late in the season. These differ 
from the ordinary ones in that they are stalked and also in that the 
period of growth precedes the period of multiplication. 
29. After being loosened from its stem, the giant spermatogone 
gives rise to giant spermatocytes and these to giant spermatids. 
30. The end product of the two kinds seems to be similar, as I 
have found no trace of giant spermatozoa, unless it be those that I 
have described as forcing their heads into gonocysts and apparently 
causing their disintegration. 
31. -Like the synapsis stage of the spermatogones, the spermato¬ 
cytes have a definite orientation with reference to the cyst wall, the 
outer pole of the cell being larger than that facing the central cavity. 
But, unlike the spermatogones, the spermatocytes have the nucleus 
in the inner narrow end of the cell. 
32. In the first maturation division the spindle is transverse or 
parallel to the cyst wall. In the second maturation the spindle is 
vertical to the cyst wall, hence one cell resulting from the division 
lies freely in the cyst cavity. 
33. In the prophase of the first maturation division, the centro- 
some lying in an indentation of the nucleus divides, each of the two 
resulting centrosomes moving to an opposite pole of the nucleus where 
asters are formed. 
34. A general division or splitting of the cytomicrosomes and 
cytoreticulum simultaneously with the division and separation of the 
centrosomes seems not at all improbable. 
35. While the nuclear membrane is still intact, and while the 
asters are fully developed, the chromatin inside the nucleus under- 
