27 2 Allen.— The Spermatogenesis of Poly trichum juniper inum. 
is especially likely to be true shortly after its first appearance (Figs. 9, 10, 
13) ; it is possible that in such cases there is no differentiation as yet 
between its outer and inner portions, although sometimes (e.g. Figs. 5 and 
12) this differentiation certainly appears at a very early stage. The sections 
from which Figs. 14, 16, 17, 18, and 20 were drawn were stained first in 
gentian violet and then in safranin, thus reversing the usual order. Under 
this treatment the limosphere becomes deep red throughout ; probably its 
apparent homegeneity in these cases is due merely to the unusually deep 
colour and consequent opacity of its outer shell. This notion is supported 
by the differentiation shown in the limosphere of Fig. 15, which is drawn 
from the same antheridium as are Figs. 14 and 16. 
While the blepharoplast is growing, the arrangement of nucleus, 
blepharoplast, and limosphere seems decidedly variable. Cells may be 
found (compare those represented in Figs. 9-13) which show these three 
bodies in almost every conceivable spatial relation to one another as well as 
in almost every possible position within the cell. It does not follow, of 
course, that there is no relation between the positions of the nucleus, 
blepharoplast, and limosphere because no constancy in this respect is to be 
detected. When the androcyte is first formed, it plainly has a polar 
organization. Later, as we shall see, the different cell organs take up very 
definite positions with reference to one another, and the cell again displays 
a marked polarity. It is not necessary to assume that the polarity mani- 
fested during the earlier and the later history of the cell is lost in the period 
now under consideration, simply because, in the changes and shiftings 
incident to the formation of the limosphere and the elongation of the 
blepharoplast, these bodies take various positions with reference to the 
nucleus and to one another. 
Fig. 13 shows a condition that is not infrequently found — namely, the 
presence at the distal end of the antheridium of a small group of androcytes 
(represented by the two uppermost ones in this figure) that are much larger 
than any others in the same antheridium and are provided with proportion- 
ally large nuclei and blepharoplasts. Similarly placed groups of large 
androcytes have been found at somewhat later stages than that here shown, 
but it was impossible to distinguish the details of their structure and they 
appeared to be in process of breaking down ; no unusually large antherozoids 
have been found ; and it is not unlikely that in all cases the giant androcytes 
cease their development comparatively early and become disorganized. 
Possibly comparable in a measure with these are the ‘ double spermatids ’ of 
Mnium observed by Wilson (1911), which are of about double the normal 
size, each containing two blepharoplasts, two limospheres, and probably two 
nuclei. Ikeno (1903) also describes an unusual growth ol occasional 
androcytes in Marchantia ; however, he thinks these giant cells divide, each 
thus forming two androcytes of normal size. 
