63 KANDA— VERBENA 1020} 



Figs. 48 and 49 illustrate two parts of the same embryo sac; the 

 endosperm tissue is seen to be fully formed in the micropylar 

 chamber, while the antipodal chamber is still uninucleate. 



A large amount of starch is present in the embryo sac, as was also 

 observed by Giignard (7) (Cestrum), D'Hubert (5) (Cactaceae), 

 Webb (18) (Astilbc), and Lloyd (12) (Calicium). This is ob- 

 servable not only a little before fertilization, but more especially 

 after fertilization has occurred digs. 43, 44, 46). Fig. 46 shows 

 >tarch not only in the micropylar and antipodal chambers, but also 

 even in the egg cell. It is evident that the starch grains in the 

 micropylar chamber are always larger than those in the antipodal 

 chamber. These starch grains are naturally closely related to 

 those in the nutritive jacket. I have already mentioned that 

 jacket cells loaded with starch grains may protrude into the sac. 

 Sometimes one gains the impression that the starch grains have 

 entered the sac through the destruction of the thin walls of the 

 jacket cells. Such a direct transfer of starch, however, is hardly 

 to be credited, partly because there are many fewer grains in the 

 sac than in the jacket, but mainly because the walls of the jacket 

 cells seem to be composed of very resistant material, since they 

 persist for a long time apparently intact. In the V. hastata 

 material I found occasionally an entire absence of starch grains 

 in the jacket cells, and in such cases the development of the embryo 

 sac is always remarkably retarded, and the egg apparatus is absent 



(fig. So)- 



The further development of the endosperm is the same as in 



Sagittaria, described by Schaffner (17). While the micropylar 



chamber is becoming filled with walled endosperm tissue through 



free nuclear division, the enlarged nucleus of the antipodal chamber 



still remains undivided. Sometimes it divides once or twice 



(tig. 51), forming two or three free nuclei which enlarge enormously. 



Meantime the endosperm tissue continues to develop, finally 



extending from the micropylar chamber into the antipodal chamber, 



forcing the large cell which occupies the antipodal chamber up 



to the antipodal end. At about this time the antipodal cells 



disintegrate (fig. 52). The large cell at the antipodal end of the 



chamber gradually diminishes in size, and linally disappears. 



