98 BOTANICAL GAZETTE [august 



the number of sperms per antheridium varies from 128 to 1024 

 (approximately). Correlated with this is a great difference in the 

 size of the antheridia. The sperms themselves also show consider- 

 able variation in dimension, as a comparison of figs. 28 and 29 will 

 show. The nucleus of the penultimate cell at the time of separa- 

 tion is in the resting condition. The cytoplasm has a very fine 

 and uniform structure, and in most cases is entirely free from 

 plastids or other inclusions. Vacuoles are present only very 

 rarely. 



While the penultimate cells are rounding off from one another 

 there appears in the cytoplasm near the nucleus a very minute 

 granule which stains intensely with iron-hematoxylin (fig. 2). Its 

 diameter lies between 0.25 and 0.3/*. Very faint cytoplasmic 

 radiations extend out from it in all directions, forming a very weakly 

 developed aster. In other cells of the same antheridium the granule 

 is seen to be dumb-bell shaped, and in still other cells distinctly 

 double, showing that it divides to two (figs. 3,4). These paired 

 bodies are the blepharoplasts. Immediately after division their 

 diameter increases to 0.5/4. Their radiations become more pro- 

 nounced and the nucleus often becomes flattened or slightly 

 indented at the point where they lie, as in fig. 8. 



The origin of the single granule cannot be stated with certainty. 

 When first made out it holds a position near the nuclear membrane, 

 a fact which would suggest its nuclear origin, but no other evidence 

 in favor of this interpretation was obtained. The nuclear mem- 

 brane shows no indication of recent disturbance. Moreover, it is 

 highly improbable that such a granule could be distinguished within 

 the nucleus because of its small size, its similarity in staining reac- 

 tion to the chromatin network, and the density of the latter. Some 

 light may be shed upon the question by exceptional cells like that 

 shown in fig. 5. Here are scattered through the cytoplasm many 

 very small intensely staining bodies, a few of which occur in pairs. 

 When first seen these granules lie in all positions with respect to the 

 nucleus and the plasma membrane. Some of the paired granules 

 are distinctly larger than the single ones; the pair nearest the nu- 

 clear membrane is always the largest, has the most evident radi- 

 ations, and is without doubt the same structure shown in fig. 4- 



