a Si a tN ih a a ts 
1897] THE LIFE HISTORY OF SALIX 155 
factory. One-of these (fig. 23) indicates that the longitudinal 
arrangement of spindles in the antipodal end and the transverse 
arrangement in the micropylar is continued in this stage. A 
portion of the contents, probably a micropylar spindle, has been 
washed out from this section, for the clear space between the 
mitotic figures is not a vacuole. It is common enough to find 
the eight-celled stage just as the two polar nuclei are fusing to 
form the primary endosperm nucleus. It would seem that after 
the second division the development proceeds rapidly, otherwise 
in examining a large number of sections from material repre- 
senting stages from the first division of the primary nucleus to 
the eight-celled stage one should find the eight-celled stage as 
frequently as any other. As a matter of fact, the uninucleate 
condition is found most frequently ; macrospores with two nuclei 
are not so frequent; those with four nuclei are comparatively 
rare; and those with eight nuclei are very exceptional. These 
observations show that the macrospore remains for some 
time in the uninucleate condition, a fact further indicated by 
the differences in the degree of maturity of ovules containing 
such macrospores. The increasing infrequency of the succeed- 
ing stages indicates that after germination has begun, develop- 
ment proceeds with increasing rapidity until the gametophyte 
has reached its fertilization period. 
The macrospore may reach the eight-celled stage without 
increasing very much in size, or during the divisions which result 
in the four-celled and eight-celled stages there may be consider- 
able enlargement at the expense of surrounding cells (compare 
Jig. 16 with fig. 2g). After the eight-celled stage is reached, 
the macrospore increases greatly in size, as may be seen by com- 
paring the figures of p/. X/// with those of g/. X/V, all of which 
are drawn to the same scale. This macrospore was a puzzle 
to me for more than a year. I had not yet’ found the third 
division resulting in the eight-celled stage, but the egg appara- 
tus and primary endosperm nucleus seemed to demand such a 
stage. I could find no antipodals, and Treub’s Casuarina (22) 
ht antipodals added to the perplexity. An effort to con- 
. 
