Cell Structuxe, Growtli and Division in the Antheridia of Polytrichum etc. 147 
with the notion that the material of the fiber is flowing from its ends 
in ward to the developing cell plate. Even after the plate extends entirely 
across the cell, however (Fig. 65), a few fibers are still present; and the 
final disappearance of these more persistent fibers is apparently to be 
accounted for by their resorption into the undifferentiated cytoplasm. 
The cell plate, during its construction, stains exactly like the spindle 
fibers; after the completion of cell division, the former position of the plate 
is occupied by what its staining reactions show to be a new cell wall. 
The Splitting of the cell plate and the deposition of a wall between its 
separated halves liave not been observed, doubtless because of the small- 
ness of the cells; but from the similarity of all the processes concerned 
in cell division, so far as they can be followed, to those observed in higher 
plants, there can be no doubt that the final steps also are the same here 
as in the better known and more fnlly studied cases. 
The partition wall between the daughter cells is at first very thin 
(PI. VI, Figs. 1, 11); but it grows later to about the same thickness as 
that of most of the previously formed walls (Figs. 19, 22, et al.). The 
very oldest walls within the antheridium, however — those separating the 
original segments of the apical cell — are in general perceptibly thicker 
than those which result from any of the later divisions; in consequence, 
the androgones are divided by these thicker walls into groups, each group, 
together with the adjoining wall cells, being derived by repeated division 
from a single segment. For some reason — possibly because of the ex- 
istence of intercellular protoplasmic connections in the one case and their 
absence in the other — plasmolysis is more likely to occur away from the 
older,* thicker walls than from the later-formed ones; consequently, it is 
not uncommon in fixed material to see the groups of closely related andro- 
gones separated from one another by clear spaces of some width, as well 
as by relatively thick walls. It is often true, too, that the cells of a certain 
group are all in about the same stage of development or division, while 
those of an adjacent group are in a very different stage. 
The walls which separate the wall cells from the androgones — mark- 
ing the plane of the first division of each segment — are also especially 
notable for their thickness (e. g., the wall at the upper side of Fig. 31, 
PI. VII, and that at the right side of Fig. 61, PI. VIII). 
The Androcyte Mother Cells. 
The Cytoplasmic Structures and Spindle Formation. 
The development of the androcyte mother cells previous to their 
division seems to involve a greater proportional growth than has occurred 
10 * 
