Campbell: The embryo-sac of Pandanus 215 



mother- cell in all cases where there is no previous division is to be 

 regarded not as a single megaspore, but as an aggregate of four. 

 He thinks that the nature of the reduction division which precedes 

 the formation of the embryo-sac necessarily involves the produc- 

 tion of four megaspores, or at least of four nuclei which represent 

 them. It undoubtedly is true that so far as we know at present 

 the reduction division cannot be omitted where fertilization is to 

 occur, but as Brown has pointed out, it does not follow that the 

 presence of the heterotypic division in a cell in the nucellus is 

 alone sufficient ground for holding that such a cell is necessarily 

 a megaspore mother-cell, inasmuch as the reduction division may 

 occur at various points in the life history of a plant. He further 

 points out the well-known fact that all stages of reduction in the 

 sporogenous tissue may be traced from the ferns, where a single 

 archesporeal cell gives rise to a mass of tapetal cells and spo- 



rogenous tissue, to such angiosperms as LiYtum, where the spo- 

 rogenous cell at once forms a single megaspore. He says, *' It 

 does not seem reasonable to suppose that the division of the 

 mother-cell into four megaspores may not also be left out and the 

 mother-cell function directly as a megaspore. In this case the 

 heterotypic division might be pushed forward and take place in 

 the embryo-sac." The fact that the embryo-sac is actually, if not 

 theoretically, part of the sporophytic structure, and the extensive 

 reduction of the sporogenous tissue in the ceils of the angiosperms 

 in general make it somewhat rash to assume that the limits between 

 sporophyte and gametophyte are as sharply drawn as they are in 



the pteridophytes. 



Brown bases his opinion that the four nuclei in the young em- 

 bryo-sac of Peperomia represent separate megaspores, upon the 

 fact that in the two first nuclear divisions in P. Sinteiiisii transverse 

 cell walls are formed, and in P. pelhicida cell plates accompany the 

 two first nuclear divisions. In the third division which results in 

 the ei»ht-nucleate stage, cell plates are wanting. But, as in the 



last division, by which the sixteen-nucleate stage arises, cell plates 

 are formed, and, in the case of certain of the nuclei, cell walls also, 

 this seems rather inadequate grounds for assuming that the embryo- 

 sac represents four spores instead of a single one. 



While the single megaspore in Peperomia and other similar 



