Bot.— Vol. I.] CAMPBELL— SPARGANIUM. 305 



developed. There seems little question, that as Westermaier 

 (1890) pointed out in the grasses, these enlarged antipodal 

 cells replace physiologically for the time being the endo- 

 sperm, and elaborate food materials for the growth of the 

 developing embryo-sac and embryo. 



The first divisions in both endosperm and antipodal cells 

 take place before there has been any marked increase in 

 the size of the embryo-sac; but later there is a rapid in- 

 crease in the size of the ovule, which probably coincides 

 with the first divisions in the embryo. 



Kornicke (1896) in his studies on the antipodal cells of 

 various Gramineas was unable to detect any karyokinetic 

 figures in these and was inclined to think that the divisions 

 of the nuclei might be direct. This view appeared to be 

 hardly likely, as the nuclei in the developing antipodal cell 

 of Sfarganium appeared entirely normal, and the cells 

 were actively growing and dividing, not a condition in which 

 one would look for direct nuclear division. Fortunately a 

 preparation was finally secured in which some of the antip- 

 odal nuclei were actually in process of division (fig. 29), 

 and it was seen that the division is the typical karyokinesis, 

 which it is safe to say is the normal type of nuclear division, 

 at least in the early stages. In the case under consideration 

 the endosperm-nuclei were also dividing. 



In most cases, possibly always, the first division of the 

 nucleus in the antipodal cells in S. simplex is not accom- 

 panied by the formation of a division-wall, and the enlarged 

 antipodal cells show plainly two conspicuous nuclei (fig. 

 29). Later, however, all the nuclei become separated by 

 walls and each of the very numerous antipodal cells has 

 evidently but a single nucleus (fig. 30). In some older 

 ones, however, two nuclei were sometimes seen, so that it 

 would seem that the last nuclear division, like the first one, 

 may be independent of cell division. There is thus a cer- 

 tain analogy with the behavior of the normal endosperm- 

 cells which would indicate that the antipodal cells are not 

 markedly different in nature from the endosperm-cells, and 

 may be really homologous with these. 



