ARTHENOGENESIS IN THALICTRUM PURPURASCENS — 397 
ear division is so rapid as to produce in a remarkably 
E great number of nuclei, during which the sac enlarges 
y in all directions. It may be well to note here that 
division of the egg takes place when the free nuclei 
¢ parietally placed, lining the whole sac. Coulter found 
sof Ranunculus “occasional evidence of endosperm- 
before the fusion of gametes, and even before the 
of the pollen-tube into the cavity of the sac.” In 
examination of normal material of 7. purpurascens | have 
unable to find any stages that show segmentation of the 
before the definitive nucleus divides, in all cases free 
division having begun before fertilization. In the 
i that I know to be parthenogenetic the free nuclear 
gan without any stimulus from fertilization, and 
re the egg divides. Undoubtedly fertilization, when 
lace, may exert an influence upon the definitive nucleus, 
1own to do upon other adjacent structures, but it isnot ] 
ecessary to its division. The parietal placing of the 
tm nuclei is followed by the formation of a 
Osperm gradually fills the cavity of the sac. : 
bed above, the egg becomes elongated far below the Be 
i there seems to be a great lack of stainable _ 
nboth nucleus and cytoplasm. In every case observed 
| of the sac was very dense about the egg, excep ] 
ediately in contact with it. This layer appears a 
ent consistency and stains very little, much resem-_ 
ne of broken-down endosperm tissue SO frequently 
nding an embryo in the seed. It suggests that the a = 
8 off an enzyme that digests the adjacent cytoplasm. ee 
pt was made to trace in detail the development | we 
} as that was not the purpose of the work. Without 
on the first division occurs, and is transverse (fg. r2) ane 
The next division is also transverse, and a row of three — Boe 
‘Is formed, after which a longitudinal division takes a 
terminal cell, differentiating the embryo-roper ae 
nsor. The persistence of the yee * . 
