4o6 BOTANICAL GAZETTE [December 



micropylar polar and in another the antipodal. To say that the 

 upper polar has the greater attraction because it is sister to the 

 eg-g is not sufficient, for in A. Cormtti the polar most distantly 

 related to the ^%^ has the greater attraction. Nor will prox- 

 imity account for it, for the micropylar polar is nearer the end. 

 of the tube than the antipodal. It might be supposed that inci- 

 dentally this lends weight to Strasburger's view that the male 

 cell is carried by the streaming of the protoplasm, but in this 

 case we should have to suppose the streaming to be towards the 

 antipodal polar. If the polars are in any way carried by the 

 stream, we should expect that the streaming would more often 

 be from the end of the tube towards the farthest polar, but that 

 the polars are carried to any considerable extent is doubtful. 

 Another objection is the proximity of the micropylar polar in 

 some cases to the end of the tube. It seems as though the male 

 cell would unite with the micropylar polar if the latter had the 

 same affinity for it as the antipodal polar. On the whole, I am 

 inclined to believe that the antipodal polar has for some 

 unknown reason a stronger attraction for the male cell than has 

 the micropylar. 



The sac after fertilization is greatly increased in size by the 

 development of the endosperm. The first division of the endo- 

 sperm nucleus occurs soon after fertilization, but the actual 

 division was' not observed. The second and third divisions 

 .rapidly follow, forming eight endosperm nuclei [fig- 50). Up 

 to this time the nuclei divide simultaneously and no walls are 

 formed. The number of chromosomes in the endosperm could 

 not be determined, but there were more than in the megaspores. 

 In many cases there was a tendency towards the massing of the 

 endosperm about the egg. The cytoplasm of the sac seems to 

 increase in volume as the number of free nuclei increases. In 

 the eight-nucleate stage of the endosperm the formation of the 

 walls by indentation begins [fig. §6). The division into sixteen 

 nuclei seems to be nearly simultaneous, since few cases could 

 be found where there were between eight and sixteen nuclei. 

 By this time the cells have become somewhat walled off {Jig- 3 

 and the divisions from this time on are not simultaneous. The 



