Gates.—Pollen Formation in Oenothera gigas. 911 
The cytological methods employed need no description. For fixing 
solutions Bouin’s fluid 1 was found satisfactory, in addition to various chrorri- 
acetic solutions. Heidenhain’s iron-alum haematoxylin was the stain most 
frequently employed, and a counter-stain with orange G was often useful 
in bringing out such structures as the cytoplasmic connexions between 
pollen mother-cells. The fact that the cells and nuclei are considerably 
larger in O. gigas than in the other forms has been a distinct advantage in 
the study of several of the stages of nuclear development. Several of these 
preparations were made in the laboratory of Professor Gregoire, to whom 
I am much indebted for laboratory facilities during my stay in Louvain. 
Description. 
Archesporium and Synapsis. 
In a longitudinal section of a very young anther a single row of 
archesporial cells appears. The nuclear reticulum of these cells is at first 
extremely delicate, so much so that only traces of such a structure can be 
seen. It is always composed of finer threads than the cytoplasmic reticulum 
at this time. The tapetum is uninucleate and remains so until about the 
time of synapsis, when a simultaneous mitotic division occurs in all the 
tapetal nuclei. Frequently, though not always, the archesporial cells divide 
to form the cells which become the pollen mother-cells. In some cases, 
however, these cells, by subsequent growth and without division, form the 
pollen mother-cells directly. Simultaneous growth of the cells, nuclei, and 
nucleoli of the archesporial cells occurs, until they are enormously larger 
than their original size. Such a pollen mother-cell is shown in PI. LXVII, 
P"ig. 1. This cell is just ready for the beginning of the nuclear condition 
known as synapsis. Fig. 2 shows the beginning of this process, Fig. 3 
a somewhat different condition occurring at this time, and Fig. 4 a slightly 
later stage as it sometimes occurs. As will be shown later, there is much 
variation in the relations between nucleus and cytoplasm at this period, but 
in all cases the result is a rather sudden increase in the size of the nucleus. 
Fig. 5 shows the typical synaptic condition after the nuclear expansion has 
been completed. 
In another paper (Gates, ’ 08 ) I have called attention to the fact that 
in O. rubrinervis there is a large amount of growth of the nucleus at the 
time of synapsis, without corresponding growth of the cytoplasm. With 
further measurements it has been shown in a recent paper (Gates, Tl) that 
the appearance- of contraction of the nuclear contents during the period of 
synapsis is largely, though not wholly, the result of this sudden nuclear 
expansion without a corresponding growth or change in position of the 
chromatin content of the nucleus. 
1 A note on my use of this fixing agent is to be found in Science, N. S., xxxi. 234, 1910. 
3 o 2 
