Gates.—Pollen Formation in Oenothera gigas. 925 
proliferated mother-cell walls were a constant feature of certain anthers in 
this flower, while certain other anthers showed with equal constancy very 
narrow spindles and smooth regular mother-cell walls. The cause or 
meaning of this difference is not evident. In other flowers the most usual 
width of the homotypic spindle is between these two extremes, as shown 
by PI. LXVIII, Fig. 36, and PI. LXIX, Fig. 38. 
PI. LXVIII, Fig. 42, shows an early telophase of the homotypic mitosis. 
Many of the chromosomes have the median constriction which appears to 
be characteristic of them at this time, not only in the homotypic telophase 
but in somatic telophases as well, as I have shown for O. Lamarckiana in 
a previous paper (Gates, ’ 0 7 , pp. 18-20). Figs. 43 and 44 are early homo¬ 
typic telophases drawn with higher magnification, and nearly all the 
chromosomes show clearly a bivalent structure. Since the late anaphase 
stages of the heterotypic mitosis in O. gigas sometimes show a clear trans¬ 
verse constriction (Fig. 26), and since a similar constriction frequently 
appears in the early telophase of the homotypic mitosis, it appears probable 
that this may, in both cases, bear a relation to the succeeding karyokinetic 
division. It seems further probable that, in O. gigas at least, there is 
no significant difference between a longitudinal and a transverse fission 
of the chromosomes. 
Fig. 45 is a slightly later stage in the homotypic telophase. The 
nucleus has grown and a nucleolus has appeared, while the chromosomes 
are beginning the process of ‘ stringing out ’ to form an anastomosis of coarse 
threads. In PI. LXIX, Fig. 4 6, this process in the nuclei has gone somewhat 
further, the chromosomes having elongated considerably to form thick, 
somewhat ragged threads. This figure also shows the condition of the 
fibrillae in the cytoplasm just previous to the beginning of cell-plate forma¬ 
tion. In comparing PI. LXVIII, Figs. 42-45, and PI. LXIX, P"ig. 46, with 
those of the heterotypic telophase, it will be seen that while there has been 
no change in the size of the cell, the chromosomes, although the same in 
number, are only half the size of the bivalents in the heterotypic telophase, 
and the volume of the nuclei is correspondingly smaller. In this case the 
amount of chromatin evidently determines the size of the nucleus, for the 
number of bodies is the same in each case, and the size of the individual 
bodies is the differential factor. 
The Pollen-tetrads and Pollen-grains. 
Figs. 47 and 48 show the tetrad of young pollen-grains still within 
the mother-cell wall. After the tetrahedral walls come in, dividing the 
contents of the mother-cell into four pyramids with spherical base, whose 
apices meet at a central point, the walls begin to separate at the apex and 
the contents of each cell form a flattened disc, spherical in surface view, 
discoid in side view. This withdrawal of the cells leaves them arranged so 
3 P 
