254 BOTANICAL GAZETTE : [ APRIL 
ative nucleus appearing to develop more chromatin, and thus 
always taking a deeper stain. Although no special staining was 
employed, centrospheres were frequently seen beside the resting 
generative nucleus. The pollen grain now rapidly grows to its 
mature size, and the generative nucleus immediately divides into 
the two sperm nuclei. These are small and spherical at first, 
and always stain so deeply that little or no structure can be seen 
in them (fg.7). This division of the generative nucleus takes 
place long before the anther has reached its mature size and is 
ready to dehisce. After the pollen is shed the sperm nuclei no 
longer appear spherical, but are bean shaped or spindle shaped, 
and the tube nucleus shows a difference in its reaction, since it 
now stains almost as deeply as the sperm nuclei themselves, and 
shows little or none of its internal structure (figs. 8,9). Whether 
the sperm nuclei organize definite cells I could not determine. 
The spindle shaped appearance may have been produced by the 
accumulation of a small amount of cytoplasm at the two ends, 
but if this was really the case my staining produced no differen- 
tiation between nucleus and cytoplasm. The division of the 
generative nucleus before pollination seems to be quite common 
in monocotyledons, and it is probable that this condition will be 
found to be the rule rather than the exception in this group. 
DEVELOPMENT OF THE FEMALE GAMETOPHYTE. 
Because of a lack of suitable material the development of 
the macrospore could not be worked out. The earliest stage 
found was a four-celled embryo sac (fig. ro). The two nuclei 
at the micropylar end arise by longitudinal division, while the 
two lower ones are produced by a transverse division. After 
the next divisions, which produce the typical eight-celled embryo 
sac, the nuclei begin to travel to their proper positions, while at 
the same time large vacuoles appear in their rear. The nuclei 
of the synergids, the nucleus of the oosphere, and the lower polar 
nucleus are about the same size, while the upper polar nucleus is 
by far the largest nucleus in the sac (figs. zr, 12). The three 
antipodal nuclei are pererat smaller than the others, and 
