96 MORPHOLOGY OF ANGIOSPERMS 



Xyctaginaceae and Conyza by Guignard, 12 in Alchemilla by 

 Murbeck, 94 in Sium, in which case they are relatively small and 

 remain near one another in a parietal position until the em- 

 bryo-sac has become much enlarged, in Nicotiana by Guig- 

 nard, 107 and in Juglans nigra by Karsten, 115 in which there 

 may be no fusion. In this connection the recent experiments 

 of Shibata 122 upon Monotropa uniflora are of interest. He 

 found that the polar nuclei may fuse in the absence of pollina- 

 tion, but that fusion is hastened by pollination. For example, 

 when pollination occurs the polar nuclei fuse in about five days, 

 but when pollination is prevented the fusion does not occur for 

 ten days or more. 



It seems to be generally true that the polar nuclei either 

 fuse in contact with the egg, as observed by Guignard 12 in 

 Eriobotrya, Cuphea, Nicotiana, 1 " 1 and other forms, or the 

 fusion-nucleus migrates to that position just before fertiliza- 

 tion, as in Tricyrtis (Ikeda 10 °), or after fertilization, as re- 

 ported by BalickaTwanowska 68 for the Scrophulariaceae and 

 allied families. The last observer suggests that this position of 

 the primary endosperm nucleus has to do with the nutrition of 

 the fertilized egg; but the case of Tricyrtis suggests a function 

 during fertilization. It is certainly true that in most cases 

 this nucleus is finally either in contact with the egg or very 

 near to it. In Sagittaria ( Schaffner 47 ) and Potamogetoii 

 (Holferty 97 ) the polar nuclei fuse in the antipodal end of the 

 sac, but at the first division of the fusion-nucleus one daughter- 

 nucleus moves toward the egg-apparatus. The evidence seems 

 to show that the polar nuclei and the fusion-nucleus have 

 freedom to " wander " through the sac, and that there is at some 

 time a relation in position to the egg. For example, in Tri- 

 cyrtis Ikeda 1 "' ! has described the fusion-nucleus as passing first 

 to the antipodals, and then passing to the egg just before fer- 

 tilization. 



The antipodal cells are either naked or invested by walls, 

 and are exceedingly variable as to their arrangement, number, 

 and persistence. The ordinary statement that the number of 

 cells is limited to the three primary ones, and that they are more 

 or less ephemeral, taking no part in the activities of the embryo- 

 sac, has proved to he far from true in the majority of cases 

 investigated. It is impossible to classify them as ephemeral 



