1917] SAWYER—IRIS 163 
within the ovary. The pollen tubes occasionally branch. This 
/ was observed both in tubes grown on the stigmas and in those 
grown on nutrient solutions (fig. 8). In the cell sap sugar solution 
some grains produced 2 tubes (fig. 9), indicating 2 germination spots. 
Longitudinal sections of hand pollinated stigmas (fig. 11) show 
pollen grains (fig. 5) which have essentially the characteristics of 
the grains in the dehiscing anthers, and among them grains which 
are producing pollen tubes. A tube (fig. 10) caught just as the 
tube nucleus and the generative cell were passing the tube dis- 
tinctly shows the latter preceding, contrary to the order that is 
usually reported. In other cases, the tube nucleus led the way 
from the grain into the pollen tube (fig. 7). The number of cases 
thus far observed at just this stage is not sufficiently large to war- 
rant a statement as to which of these conditions is prevalent in 
Tris versicolor. Apparently it is usual for the tube nucleus to soon 
gain the leading position. One instance (fig. 14) of the tube nucleus 
in the act of passing the generative cell was observed. The peculiar 
elongated and pointed anterior end of the tube nucleus is very 
interesting and suggests a self-motile body. 
The generative cell certainly usually passes into the tube before 
its nucleus divides to form the male nuclei, but occasional grains 
suggest the possibility that it sometimes divides in the grain. The 
division of the generative cell has not been observed, but in tubes 
_ dissected from the stigma, and in those grown in the culture media, ce 
the generative cell has frequently been seen apparently containing — 
2 nuclei (fig. 15m). The contents, which microchemical tests show 
_ contain much starch, stain so heavily that it is difficult to differ- ee : 
 entiate the nuclei satisfactorily. The material seems to indicate a : : eS 
_ Somewhat wide 1 a in | the position of the generative oe in i oS 
