positions at the opposite micropykir and antipodal - ends 

 of the sac. Each now passes through two simultaneous 

 divisions (Figs. 19 and 20), resulting first in two nuclei 

 being formed at each pole, then two groups of four. The 

 upper or micropylar series differentiates into an egg- 

 apparatus consisting of two pear-shaped synergids and a 

 rounded, more centrally placed egg, the fourth nucleus 

 remaining free. Three cells of the antipodal group take 

 walls and become arranged in inverted pyramid form, the 

 fourth nucleus here also remaining free. The two free 

 cxtr;i-pol;n- nuclei by slow mo\'ements gradually come 

 together and fuse. At the time of fertilisation they arc 

 probably joined by one of the generative nuclei from the 

 pollen tube, the complex of three forming the endosperm 

 nucleus. 



ANTIPODAL CELLS. 



Up to the eight-celled si age the course of development 

 in the sac is normal. Afterwards an unusual number of 

 antipodal cells iipj^ear, probably, as Koernicke-' seems to 

 think, by amitosis or direct division of their nuclei ; at 

 least in the many hundreds of antipodal nuclei examined 

 he did not observe a single spindle stage. On the other 

 hand, he failed to see a " typical " direct-division nucleus. 

 Hofmeister found 6-12 antipodals, Koernicke 36 or more, 

 in a mature embryo-sac, and Cannon reports like numbers 

 for A^'cna fatua. The antipodal cells enlarge enormousl}- 

 and form an arched plate of tissue at the chalazal end of 

 the sac. They " serve to transform the foodstuffs brought 

 to the embryo-sac." The chromatin of their nuclei is very 

 irregular, and the peculiar plate-like bodies (Fig. 21) ha\e 

 been figured and described as nucleoli ^. They stain light 

 brown with Flemming's triple-combination. This vigorous 

 develophient of antipodal tissue is said to be common in 



