1907] ^ LAND— EPHEDRA TRIFURCA 279 



must not be confused with the female gametophyte of gymnospermsl 

 Many things lead to the belief that the chromatin of the second male 

 nucleus, together with that of the jacket nuclei, has something to do 

 with the production of at least a part of the mass of small cells which 

 appears in the egg and jacket region. It is at the edge of the dis- 

 integrating male nucleus that spindles first appear, with either a 

 few granules or masses of chromatin at their poles. The absolute 

 equality of the male nuclei, for they differ in no visible way except in 

 position in the pollen tube, seems to suggest that the nucleus which 

 does not fuse with the egg nucleus functions in some other way; for 

 cells which cease to function are either eliminated or show signs which 

 point to their future elimination. If it can be shown beyond doubt that 

 the chromatin of the second male nucleus together with some of the 

 jacket-cell nuclei or the chromatin from these nuclei arc jointly respon- 

 sible for some of the small cells in the egg and jacket region, it follows 

 that we have the foreshadowing if not the real beginning of endosperm. 

 At least it may for the present be designated as physiological endo- 

 sperm. The apparently universal occurrence of triple fusion in angio- 

 sperms seems to preclude the possibility that they may furnish a 

 solution for this problem. It is not impossible that the Gnetales, 

 which in so many ways have reached the angiosperm level, may 

 have some member which will give an answer to the problem of the 

 origin of the angiospermic endosperm. The ephemeral mass of nutri- 

 tive cells is certainly larger in E. trifurca than the endosperm mass 

 of many angiosperms. 



The functioning proembryonal cells round out as the adjacent 

 nutritive mass is absorbed {fig. 10). The nucleus of a proembryonal 

 cell divides, giving rise to two free nuclei, which at first are equal in 

 size, but one soon gains an ascendancy over the other, becoming 

 noticeably larger. A large vacuole develops in the upper part of the 



r 



cell and the free nuclei pass down and to one side of the cell if it is 

 one of the upper proembryos, but to the bottom if it is the lowest one 

 in the archegonium. The larger nucleus is invariably placed lowest 

 in the cell. Two papillae, seemingly initiated by the nuclei, arc put 

 out by the cell wall from the points nearest the nuclei {fig, 11). Soon 

 the papilla nearest the larger nucleus rapidly elongates and the other 

 papilla disappears {fig. ij). In one instance four free nuclei, one 



