142 MINNESOTA BOTANICAL STUDIES. 



nuclei have changed position somewhat. A difference in size 

 has frequently been observed here also, in some other plants the 

 lower pair being larger. I may cite Guignard* again and also 

 Mottier.f However, the last named investigator only states that 

 the nuclei forming the lower pair are much larger just before 

 dividing, distinctly stating}: that at an earlier period the four are 

 all of equal size. I did not observe this phase of development 

 often enough to be certain that the size is the same at all times 

 in Rtimex; but it seems from the instances observed that, though 

 coming from the nuclei differing somewhat in size, the four are 

 so nearly of the same size at all times that any difference would 

 be difficult to detect. During the change to the four-nucleate con- 

 dition the sac increases somewhat in size, as may be seen by 

 comparing Figs. 13 arid 14 with Fig. 12. 



The development from the four-nucleate to the eight-nucleate 

 condition must be even more rapid than that from the two-nucle- 

 ate to the four-nucleate phase, for I was neither able to find the 

 two tetrads in position, nor the division of the four nuclei leading 

 to its establishment. The nearest approach to it was observed 

 in Rumex salicifolius (Fig. 15), when the polar nuclei were ap- 

 proaching. The condition represented in Fig. 15 is an espe- 

 cially interesting step in the life-history of the gametophyte of 

 Rumex because of departure from the usual conditions and espe- 

 cially from the nearly related Polygonum. On examining the 

 figures of Guignard, Vesque, Strasburger, Ward and others, I 

 find that in fully five-sixths of their drawings they show cell 

 walls about the three anterior cells before the polars have fused. 

 Strasburger figures for Polygonum divaricatum three nuclei 

 enclosed in cellular membranes and one free nucleus in each 

 tetrad even before the two polars begin to approach each other. 

 My own observations on Polygonum erectum L. (Fig. 16), 

 Bursa bursa-pastoris ( L.) Britton and Silcne antirrhina L. gave 

 the same results so far as the anterior end of the sac is con- 

 cerned, though the evanescent antipodals of Bursa were not 

 satisfactorily studied in this respect. Methods which brought 

 out these walls in the three plants named above should show 

 them, if present in Rumex, yet in this genus I find all the an- 



*Guignard, L., 1. c. 188. 

 tMottier, David M. , 1. c. 137. 

 JMottier, David M., 1. c. 136. 

 Strasburger, E. 1. c. Fig. 17. 



