Cell Structure, Growth and Division in the Antheridia of Polytrichum etc. 153 
As Figures 77, 78, and later figures show, the long axis of the spindle 
as a rule is approximately parallel to one of the sides of the cell; if the 
cell is elongated in one direction, the spindle axis commonly coincides 
or nearly coincides with the long axis of the cell. The same is true when 
the cell plate is being formed (Plate IX, Figs. 117 — 122). Occasionally 
the spindle has a more or less diagonal position (PI. VIII, Fig. 90; PI. IX, 
Fig. 117) ; but such an occurrence is exceptional, and not noticeably more 
frequent in this than in previous divisions (compare Figs. 49, 52, PI. VII). 
The final division within the antheridium is not, therefore, distinguished 
from its predecessors, like the corresponding division in certain liver- 
worts, by the plane in which it cuts the mother cell. 
After the condition now described has been reached, the nucleus 
swells, as in previous mitoses, increasing especially in length, until its 
membrane comes into contact at the sides with the spindle fibers and at 
the ends with the central bodies (PI. VIII, Figs. 80 — 82). In a median 
cross section of the cell at the conclusion of the swelling process (Figs. 83, 
85), a narrow zone of dense cytoplasm lies just outside the nuclear mem- 
brane, and in this zone a cross section of a spindle fiber occasionally appears 
as a round dark granule. For the most part, however, the fibers are too 
closely appressed against the nuclear membrane to be plainly seen in 
any view; which fact accounts for the apparent scarcity of the fibers in 
Figures 81, 82, 84 and 86. 
At or about the time that the membrane comes into contact with 
the central bodies, the latter become less conspicuous than before; and 
henceforth their continued existence is by no means so obvious a fact 
as it has been during the period already described. In most cells a dark- 
stained granule can be found in about the position hitherto occupied by 
each central body; and if, as is probable, this granule really is the central 
body, it is often less deeply stained than previously, it frequently appears 
smaller (Figs. 82, 84), and sometimes it escapes Observation entirely 
(PI. IX, Fig. 93). In other cases (PI. VIII, Figs. 86 — 88), such a body is 
visible at one pole but not at the other. From this stage onward, the ap- 
parently undifferentiated cytoplasm commonly appears densest in the 
neighborhood of the spindle poles; and it is not unlikely that it is their 
lessened affinity for stains which at times renders the central bodies in- 
distinguishable from the surrounding dense cytoplasm. Recognition of 
the central bodies is made less easy, too, by the fact that the spindle fibers 
are no longer uniformly attached to them and that there are usually no 
polar radiations; so that if, as sometimes happens (Figs. 88, 90), two 
or more dark-stained granules lie in the neighborhood of a single spindle 
