THE FEMALE GAMETOPHYTE 91 



quite possible if one of the fusing gametophytes is at a younger stage 

 of development than the other. 



In a Musa variety known as "I.R. 53," Dodds (1945) has re- 

 cently described one compound embryo sac with three egg appa- 

 ratuses and two pairs of polar nuclei; and another with two egg 

 apparatuses, one pair of polar nuclei, and an additional group of 

 7 large "polar-like" nuclei at the chalazal end. Juliano (1934) has 

 figured a peculiar embryo sac in a fallen flower of Sandoricum 

 koetjape with a normal egg apparatus, two polar nuclei, and a large 

 cytoplasmic vesicle extending from the chalazal end of the sac to 

 its middle and containing more than a dozen nuclei (Fig. 57C). 

 Since the antipodals are very ephemeral in this species, it is con- 

 sidered probable that the embryo sac proper was formed from the 

 third megaspore and that the multinucleate vesicle arose as a result 

 of some free nuclear divisions in the fourth megaspore. 4 



In some plants there is a migration of the nucellar nuclei into the 

 embryo sac. This migration is due to the fact that during the 

 growth and enlargement of the latter, the adjacent cells of the 

 nucellus become flattened and crushed. Their walls, which are 

 very thin and delicate, get ruptured, and the contents — both cyto- 

 plasm and nuclei, or only the latter — may "wander" into the em- 

 bryo sac and become incorporated in it. 5 Two instances of this 

 nature deserve special mention. In Hedychium gardnerianum 

 (Madge, 1934) the nuclei of the nucellar cells lying just below the 

 hypostase migrate "from cell to cell" through a small hole in the 

 walls until they reach the hypostase. Here their progress is stopped 

 for a time and groups of 20 or 30 nuclei collect together, surrounded 

 by the ragged cell walls of the ruptured cells. Some of the nuclei 

 now make their way around the hypostase into the cavity of the 

 embryo sac, where they are believed to serve a nutritive function. 

 In Pandanus (Fagerlind, 1940), which has no thick-walled hypos- 

 tase, the nucellar cells lying directly beneath and on the sides of 

 the young embryo sac show a marked tendency to enlarge. Their 

 nuclei become swollen and the plasma assumes an appearance simi- 



4 Another possibility, not mentioned by Juliano, is that the embryo sac proper 

 arose normally from the chalazal megaspore and the vesicle was of aposporic 

 origin. 



5 This is comparable to the condition in many gymnosperms in which the nuclei 

 of the jacket cells often make their way inside the egg. 



