Fhlk : COXTRIBUTIOX TO THE LIFE-IIISTORY OF RUMEX. 143 



terior three cells free while the polars are approaching each other 

 \n Rumcx salicifoliics [Y'lg. 15). When the polars have met, 

 or are fusing in this plant, the cellulose walls seem to begin to 

 form, and a suggestion of such a structure may occasionally be 

 seen (Figs. 17 and 20) about the nuclei of the synergidaj. In 

 Riimcx vcrticillatiis, even shortly after the definitive nucleus is 

 formed (Fig. 18), I could only distinguish a suggestion of a 

 beginning of formation of a membrane about the lower end of 

 the egg while the sister nuclei which form the synergidas were 

 yet free. In Riunex salicifolins the synergidtc form earlier 

 than in Rimiex verticillatns and are more regular in form in the 

 former plant (compare Fig. 24 with Figs. 19 and 23 ). So far 

 as investigation has proceeded then, cell walls appear about the 

 three anterior nuclei (excluding the anterior polar) before the 

 meeting of the polar gametes in Polygonum; while in Rumex 

 the walls appear after the meeting of the gametes, or even after 

 their fusion to form the zygote constituting the definitive nucleus. 

 The matter is one of some taxonomic interest, which can only 

 be solved by a laborious study of all the species of the two genera 

 with reference to this particular problem in developmental history. 

 Regarding conditions presented in Figs. 15, 17 and 20 some 

 further statements are necessary. In these stages I was able to 

 note no difference in size of the three anterior nuclei, while at 

 later periods the nucleus of the ^gg had increased in size fo 

 that it was larger than those of the synergidje (Figs. 18 and 19^ 

 So far as I could ascertain, the nuclei of the three antipodals 

 were smaller than the three just discussed at all periods. Fig. 

 17 represents the three antipodals as best seen in Rumex salici- 

 f alius, and here their smaller size can be seen distinctly. Tiie 

 antipodals are difficult of observation at all times because of 

 their lying in the small caecum-like prolongation of the embryo 

 sac so that, except in very thin sections, other cellular struc- 

 tures of the surrounding macrosporangium almost always par- 

 tially or completely obscure them. In the sac of the same plant 

 showing a slightly earlier stage of development (Fig. 15), only 

 the deeply-stained nucleoli could be seen through the overlying 

 tissues of the macrosporangium, neither the nuclear membrane 

 nor the cellular wall about each of the three nuclei appearing. 

 In this plant the three-celled antipodal area was found persisting 

 in the latest phases studied after the establishment of the em- 

 bryonic sporophyte (Fig. 33) and seems to be quite constant, 



