﻿MICH ELL — RICH A RDIA AFRICAN A 



Endosperm formation 



Endosperm formation proceeds from the base upward. Exactly 

 when cell walls first appear has not been ascertained, but from a 

 comparison of different stages in endosperm development it seems 

 probable that they appear soon after the first 2 or 3 nuclear divi- 

 sions. Fig. 14, which represents the chalazal end of the embryo 

 sac in which fertilization has been demonstrated, shows the 2 nuclei 

 resulting from the division of the primary endosperm nucleus. 

 Fig. 156 shows the division of the lower of these nuclei. Fig. 15 

 shows the antipodals which have by this time lost all semblance of 

 active cells. Perhaps it is hardly fair to quote this as an instance 

 of the complete disorganization of the antipodals after endosperm 

 formation, as in this embryo sac an abnormal embryo has been 

 formed, but the same thing is seen in fig. 11 (a, b, c) and in fig. 17, 

 which represents a slightly later stage in the development of an- 

 other embryo sac, and in every preparation of embryo sacs made 

 at this stage. It is clear that the antipodals disorganize rapidly, 

 and in the embryo sacs shown in figs. 18 and 19 they are no longer 

 recognizable. 



The basal endosperm cells usually have a striking appearance. 

 Three to five of these cells become much larger than any other endo- 

 sperm cells, and may be seen with the naked eye (figs. 18 and 19). 

 The protoplasm and nuclei of these cells also undergo some change. 

 The protoplasm becomes coarsely granular and has a honeycombed 

 appearance. All nuclear outline is lost, while the nucleolus hyper- 

 trophies, shows great vacuolization, and finally fragments, spread- 

 ing its substance over practically the whole cell. The nuclei of 

 the neighboring endosperm cells often imitate those of the giant 

 cells in behavior. One thing is certain, and that is that these 

 cells are true endosperm cells and have nothing to do with the 

 antipodals. 



The function of these cells is not clear. It is interesting to 

 compare them with the antipodal cells described in Tricyrtis hirta 

 by Ikeda (10). Ikeda believes that the cytological features 

 exhibited by the antipodals in Tricyrtis bear a relation to their 

 nutritive activity. This seems to be true for these endosperm 

 cells of Richardia. Everything points to the conclusion that these 



