190 CALIFORNIA ACADEMV OF SCIENCES. 



ment of the antheridium of Marsilia and that of the or- 

 dinary Polypodiace£e, is in the imperfect development of 

 the dome-shaped cell by which in that group the central 

 cell of the antheridium, from which the sperm-cells later 

 arise, is at once completely separated from the outer wall 

 of the antheridium mother-cell. 



In Marsilia the contents of the central cells show a dis- 

 tinct separation into a granular inner and a nearly trans- 

 parent outer part, and the boundary between the two is 

 often sharply marked before any actual division has taken 

 place (pi. Ill, fig. 12). 



From the two cells into which the central part of the 

 antheridium is now divided, a varying number of sterile 

 cells are cut off, which are transparent, and more or less 

 completely surround the two central cells which are at 

 once distinguished by their densely granular contents. 

 Not infrequently (pi. in, figs. 17 and 18), a sterile cell is 

 formed between these, completely separating them, and 

 extending entirely across the antheridium. Sometimes, 

 as in the cases figured, this is followed by two walls 

 formed simultaneous!}^ that run parallel with the outer 

 wall of the antheridium. and reach to this central sterile 

 cell. In such cases, a cross section of the two cells 

 which are to give rise to the sperm-cells is very sym- 

 metrical and nearly semicircular in outline. This re- 

 gularity, however, while ver}' frequent, does not always 

 occur, and the peripheral cells may be cut off apparently 

 without any regularity (pi. in, fig. 19). After the sperm- 

 mother-cells are differentiated, however, the divisions in 

 them show great regularity. Each one divides into two 

 nearly equal cells by a vertical wall (pi. in, fig. 18), and 

 this is then followed by a horizontal wall. The next 

 division is vertical, and each of the resulting eight cells 

 then divides once more, making sixteen in each group of 



