332 BOTANICAL GAZETTE [November 



{^fig. J7). The endosperm at this time is closely in contact 

 with the tips of the cotyledons and even forces its way In 

 between them somewhat {figs. 35, 36). At the outer or upper 

 part of the endosperm it is somewhat thickened and forms thus 

 a collar which tightly grasps the tips of the cotyledons and fills 

 completely the rent in the seed-coat {figs. 37, 38, esp). This 

 collar evidently serves not only to hold the seed on the tips of 

 the cotyledons, but also to prevent the exit and loss of any dis- 

 solved food material on its passage between the perisperm and 

 the embryo. This collar must also prevent pretty effectually 

 the entrance of water from without to dilute this food material, 

 or of bacteria and fungi to destroy it. The walls of the endo- 

 sperm cells are comparatively thin throughout, and no indication 

 of thinner spots in their walls or those of the cells of the coty- 

 ledons was seen. 



The endosperm cap, pushed on by the swelling cotyledons, 

 finally crushes in the previously emptied perisperm cells, until 

 it fills about half the bulk of the seed. After this the absorp- 

 tion may go on until practically all the starch is removed from 

 the remaining perisperm cells, while the walls of these cells 

 remain distended just as when full of food material {figs. 36, 

 38). When the seedling is about 15^"°" high, the starch of the 

 perisperm being then nearly exhausted, the wasted remains or 

 the seed and fruit * slip off from the tips of the cotyledons. 

 Then the curve in the hypocotyl straightens out, and the cotyle- 

 dons open out widely to allow the still very small plumule to 

 develop and expand {fig. jp). During the transference of the 

 starch from the perisperm to the embryo, cells of the former can 

 be seen in which the end near the endosperm is nearly devoid or 

 contents, except for a faintly staining, apparently slimy substance, 

 while the opposite end is still closely packed with starch grains. 

 The chemical changes taking place during this process and the 

 exact mode of transmission have not been made out satisfacto- 

 rily as yet, but it is hoped that these may be followed out m 

 detail as soon as a supply of fresh seeds can be obtained. 



The ripe seed of Heckeria unibellata differs from that of Pipe- 

 romia pelliicida somewhat In shape and size {figs. 2^,32), but 



