3S8 BOTANICAL GAZETTE [November 



often being detected in the space between the wall of the ovary and the 

 micropyle; and in most cases traces could be seen in the micropyle 

 and continuing into the embryo sac {figs. 45, 49). 



In the pollen grain the division into tube end generative cells 

 takes place before shedding. In this division a wall forms in all 

 cases examined, but it soon disappears. In one anther over one 

 hundred late anaphases and early telophases were counted in which 

 the wall was formed, and only about a dozen late telophases were 

 without a wall. As these showed no traces of the spindle, the proba- 

 bility is that the wall was formed here as in the others, but had 

 disappeared. The generative nucleus has a mass of cytoplasm about 

 it that differs markedly in its staining reaction from that of the tube 

 nucleus. The gametophyte number of chromosomes (eleven) was 

 easily counted here. Figs. 36-41 represent a series in this division. 

 Fig. SQ shows the wall forming in the telophase, while fig. 40 shows 

 it completed. The latter figure is also interesting in showing a 

 chromosome left out of one of the nuclei at its organization. 



The division of the generative nucleus was not seen; but as pollen 

 from the bud to the time of shedding was studied, it must take place 

 after poUination. When the pollen tube is passing from the ovary 

 wall to the micropyle, the division has already taken place {fig. 45). 



Only one attempt at pollination was made. Two flowers with 

 stems about 25"=°^ long were brought into the laboratory, placed in 

 water, and the whole put under a bell jar with some ventilation to 

 keep them in as nearly normal condition as possible. Each was 

 pollinated from the other. Four days later one of them was fixed, 

 being apparently in good condition; but the other was dead six 

 days after poUination. When the first was sectioned and stained, the 

 ovules Avere found to have gone to pieces. The pollination was made 

 in the hope of getting some idea of the development of the o\^le in 

 relation to the time of poUination. The stigma was sectioned and 

 showed that the pollen grains had practically aU developed tubes, 

 many of them being traced into the conductive tissue, in some cases 

 for some distance, and the grains were almost aU empty. If the 

 development of the tube was delayed, a condition like that in fig. 41 

 was found, in which the tube is just forming and the generative 

 nucleus is beginning to show signs of acti\dty. Soon a complete 



