428 BOTANICAL GAZETTE | JUNE 
jig. 16, but here the generative cell has effected its ordinary 
divisions. Over twenty such cases were observed. I have called 
the cell nuclei marked g (jigs. 76, 78) male nuclei, that is, daugh- 
ter nuclei from the generative nucleus, because their nucleoli 
are small, their chromatin network coarse, and in staining they 
are cyanophilous, that is, they show a preference for basic dyes. 
Besides, some of these nuclei are surrounded by definite areas 
of cytoplasm devoid of starch, indicating the organization of the 
male cell. The tube nucleus, or the several nuclei to which it 
may give rise, has larger nucleoli, finer chromatin network, and 
is uniformly erythrophilous. Hence it seems safe to conclude 
that the nuclei marked g in figs. 76, 78 are generative in origin, 
and not vegetative like the tube nucleus and its derivatives. 
The small cell marked gr (figs. 76, 78) is hard to interpret. Its 
nucleus is cyanophilous, and its cytoplasm is free from starch. 
In these respects it resembles the generative nucleus and its deriv- 
atives, and if the tube nucleus were the only other nucleus in the 
spore I should call the small cella much reduced generative cell. 
However, a study of all the cases discovered leads me to suggest 
that the small cell is a prothallial — ee with the 
single prothallial cell of het hytes. The small 
cell cut off from the microspore of Populus inonsdifird. figured but 
not described in my paper on Salix, adds probability to this 
hypothesis. If this interpretation is correct, it supports the view 
that the whole spore development, as it ordinarily appears, is an 
antheridium. In this case the tube nucleus and its cytoplasm 
is probably the homologue of the wall cells of such an antherid- 
ium as that of Isoetes; the pollen tube would become an out- 
growth from the antheridium wall; and the two male cells would | 
p homologize with the spermatozoid mother cells. At least it 
: seems out of the question to speak of the pollen tube as the 
oe male gametophyte. 
Another peculiar phenomenon was noted in L. 
on ore cell sh eoatat: but one mocleee the} 
= about twenty cases there was a distinct wall dividing the ‘micro- = 
spore chev ne net equal parts (figs. r9-20). Both cells con- 
and w. 
tc sitesi i itn i sola 
