92 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
Numerous cases, like that represented in plate 13, figure 32, have 
led me to suspect that in these cases, the large vacuole containing a 
mitosome derived from the intermediate fibers of the preceding 
division and the homologue of the nebenkern, is due to a failure on 
the part of the nucleus to expand and occupy the space which it nor¬ 
mally would occupy. I believe that such an expansion would nor¬ 
mally take place, and the mitosome consequently be reabsorbed by 
the nucleus (pi. 13, fig. 32, b). Even if in this case the mitosome 
has a centrosome in its interior as is suggested by the central, con¬ 
densed portion (pi. 13, fig. 32, a, b), the normal expansion of the 
nucleus would cause it to fill the vacuole in which the mitosome lies 
and crowd the centrosome into the narrow space between the nucleus 
and cytoplasm. This narrow space would have the form of a crescent 
if the kinoplasm surrounding the centrosome were sufficient in amount 
to spread out as the nucleus expanded. 
Considering the mechanism of mitosis, the relative position of the 
centrosome, chromosomes, and intermediate fibers in the anaphase 
of mitosis, we should hardly expect a priori to find the centrosome 
associated with the nebenkern derived from the intermediate fibers 
of the spindle, and, indeed, I find they are often separate bodies (pi. 
13, fig. 39). They may lie at opposite poles of the nucleus. 
In those spermatids which retain a fixed relation to the cyst wall, 
already referred to, there is a definite orientation, also, of the nucleus 
and nebenkern. The nucleus is always nearest the periphery of the 
cyst, or at the fixed pole of the cell, while the nebenkern is always 
nearest the free end of the cell. The nebenkern consequently occupies 
the position which the nucleus has in spermatocytes of the first order, 
the nucleus having now changed its position from the inner free narrow 
end to the outer fixed pole of the cell (pi. 13, fig. 39). This, of course, 
cannot be seen in those cells that lie free in the lumen of the cyst. 
But all these spermatids later acquire definite relation to the cyst 
wall, and then they all show this relative position of nucleus and 
nebenkern (pi. 14, figs. 40, 46). 
Relation of centrosow,e to nucleus and nebenkern .— Immediately 
after formation of the spermatid, the chromatin segments form a 
compact mass often slightly crescent-shaped lying between the aster 
and the intermediate fibers which give rise to the nebenkern (pi. 16, 
figs. 119, 121). As the nucleus develops, it expands more toward 
one side than the other, and consequently seems to move out from 
