921 
Gates.—Pollen Formation in Oenothera gigas. 
somes of the sporophyte. But it is well known that the synaptic period 
lasts for some time, and it is now evident that many changes in the con¬ 
dition of the nuclear contents may take place during this period. 
The succeeding stages, up to the heterotypic mitosis, have not been 
studied in detail in O. gigas, and as they were investigated with special 
care in O. rubrinervis (Gates, 08 ), they will not be touched upon in the 
present paper. 
Heterotypic Mitosis. 
In O. gigas, as I have shown to be the case in the other Oenothera 
forms, the chromosomes are loosely scattered on the heterotypic spindle, 
so that it is frequently impossible to say which members constitute a pair. 
PL LXVIII, Figs. 20-22 represent pollen mother-cells in the metaphase 
of the heterotypic mitosis. In Fig. 22 the chromosomes show a rather 
irregular alinement in two series which have begun their journey to the 
poles. Usually the nucleolus disappears immediately upon the solution 
of the nuclear membrane after diakinesis, but occasionally it persists for 
a short time as a pale-staining body, as shown in Fig. 22. In connexion 
with a previous paper (Gates, ’ 09 #) these stages were figured, but the 
plates were lost and I have since been unable to find again the particular 
cells from which they were drawn. 
Fig. 23 represents an early anaphase in polar view, all the chromo¬ 
somes but one being brought into view by focusing for a short distance 
in a vertical plane. A considerable amount of variation in the apparent 
size of the chromosomes is due to differences in the depth of stain. 
Figs. 24 and 25 are anaphases in somewhat oblique view. In both cases 
fourteen chromosomes can be counted in the group passing to the lower 
pole, while some of the chromosomes are missing from the upper pole. 
Fig. 26 is a later anaphase in side view. 
As the chromosomes pass towards the poles of the heterotypic spindle 
there is much variation in their shape. Very often they appear square, 
oblong, triangular, or globular, with no indication of a bipartite structure. 
In other cases they may be somewhat elongated and with their arms bent 
in the form of a V. All these forms will be seen in Figs. 24 and 25. Many 
of the chromosomes in Fig. 26 and one of those in Fig. 24 show a sharp 
median constriction. I11 the chromosomes in Fig. 26 this appears as 
a definite lighter area through the median plane of the chromosome, and 
leads to the conclusion that this must be a transverse segmentation of the 
chromosome, if one can judge from their constant manner of orientation on 
the spindle. Other chromosomes in Fig. 26 show the ‘ tetrad ’ appearance 
so characteristic of the chromosomes in early interkinesis. This is due to 
a split (apparently longitudinal) of the more elongated chromosomes in the 
anaphase. There is thus wide variation in the time at which the split 
