156 
Charles E. Allen 
tions — a quite unusual occurrence at so late a stage; something similar 
appears at the lower pole in Figure 87, Plate VIII. 
After the nuclear membrane disappears (PI. IX, Figs. 93 — 96), the 
closely grouped chromosomes lie, as in the earlier divisions, at the center 
of a relatively clear space, which corresponds roughly with the region 
occupied shortly before by the nucleus. In transverse sections through 
nucleus and spindle during and after the nuclear contraction (PI. VIII, 
Fig. 89; PI. IX, Figs. 91, 92, 97), the spindle fibers appear as dark-stained, 
round or elongated bodies (the latter, of course, if the fiber runs diagonally 
to the plane of the section). The fibers lie in a zone of rather dense cyto- 
plasm; there is no connection visible, either in longitudinal or in cross 
sections, between spindle fibers and chromosomes. Unless, therefore, 
there are other fibers present that are not distinguishable with our present 
technique, the attachment of the mantle fibers to the chromosomes — 
plainly evident later — must be first effected in these cells after the dis- 
appearance of the nuclear membrane. 
The further history of the spindle is essentially the same as in previous 
mitoses. The amount of fibrous material is less than in the earlier andro- 
gonial divisions, but not noticeably less than in the divisions immediately 
preceding the present one. The convergence of the fibers toward the poles 
in the equatorial plate stage, a tendency toward which was noted at the 
corresponding stage in the earlier divisions, is now carried to the extent 
of forming again a decidedly sharp-poled spindle (PI. IX, Figs. 98, 99). 
When a central body is visible at either pole (Fig. 99), most of the fibers 
center approximately upon it, although the centering is by no means so 
exact as in the early prophases. In the development of a sharp-poled 
spindle (Figs. 98, 99) out of the more or less broad-poled condition 
which immediately preceded (Figs. 93 — 96), there has plainly been again 
some rearrangement of individual fibers. 
Karyokinesis. 
The structure of the androcyte mother nucleus and the behavior 
of its constituents during mitosis are in almost all respects similar to what 
has been described for the nuclei of the androgones. The nucleole-like 
mass in the resting nucleus is on the average smaller, at least than that 
found in the large androgones of the earlier generations (compare Figs. 66, 
67, 69, 70, 72—75, PI. VIII, with Figs. 1—10, PI. VI); but its nature 
seems to be essentially the same. 
When the migrating daughter centers reach their final positions 
(PI. VIII, Figs. 77, 78, 80), the transformation of the nuclear reticulum into 
