88 



pectlnata is prol)ably not constant and is difficult to determine 

 exactly, it is prol)able that as a rule there are not less than 

 12 of these instead of the four found by Rauwenhoit in G. 

 jlabelhUd and 6r. rapestrls. 



The subsequent divisions do not appear to be absolutely 

 constant, nor is the final number of si^ermatocytes always the 

 same. The number is very much greater in G. pectinata than 

 in any of the species studied by RAUWENnoir. The exact numl)er 

 could not be counted, but there must be several hundred alto- 

 gether. In a single section of a large antheridium more than fifty 

 are visible (PI. XI, Figs. 72—73). The spermatocytes are not very 

 laroe and the nuclei were not well fixed in most of the material, 

 so no attempt was made to study the development of the sperma- 

 tozoids, which probably does 4iot show any particular features, 

 since the fully developed spermatozoids (PI. XI, Fig. 75) closely 

 resemble those of other ferns. The opening of the antheridium 

 was not observed in living specimens, but from sections of the 

 recently opened antheridium it is clear that the deliiscence is 

 brought about by the opercular cell. The spermatozoids closely 

 resemble those of other leptosporangiate ferns, but are smaller 

 than is often the case, e. g. O.smunda, Struthiopteris. The fully 

 grown antheridium is much larger than that of any other 

 leptosporangiate fern yet described and may reach a diameter 

 of 100 ^. 



G. polypodioide.'i agrees more nearly with Rauwenhoff's account 

 of G. flahellata and G. rupe.stris. There are, however, important 

 differences in which it approaches G. pectinata. The very young 

 antheridium seen in cross section (PI. XI, Fig. 76) shows a nearly 

 complete ring-shaped cell surrounding the base of the antheri- 

 dium, l)ut usually it can be seen that there are two intersecting, 

 strongly curved walls, so that the base of the antheridium is 

 composed of two nearly circular cells. These may later undergo 

 further divisions, and in the dome-shaped upper cell, when 

 seen from above, curved walls arc formed (Fig. 77, 2) running 

 over the top of the antheridium. Finally an opercular cell is 

 cut off. The total number of cells in the wall is about ten, 



