Oenothera Lamarckiana and O . gigas. 957 
An interesting type of irregularity, present in the heterotypic mitosis, 
is shown in Fig. 6 9. In this case two pairs of chromosomes had lagged 
behind the main group in passing to the poles of the spindle, and were 
associated in a small secondary spindle within the main structure. That 
such stray chromosomes, as well as the main group with a chromosome 
content smaller than the normal, may form independent nuclei is illustrated 
in Fig. 70. It is, however, improbable that fertile pollen-grains can be 
developed under such abnormal conditions. The fact that the chromosomes 
in these nuclei have lost their usual form and by anastomosing appear to 
be developing a chromatic reticulum, indicates that the nuclei do not 
proceed further towards the homotypic mitosis, but probably pass into 
a resting condition. 
The nuclei during the interkinesis between the heterotypic and homo¬ 
typic mitoses show many interesting features in agreement with the 
conditions in Lamarckiana and other Oenotheras. This is the period when 
the count of the chromosomes can be most easily made, as illustrated in 
Figs. 71 and 72, which show fourteen pairs of chromosomes derived from 
the fourteen split chromosomes previously described. It is clear that the 
halves of the split chromosomes swing so that their ends separate and 
thus become a pair united only in the middle region. Although the 
chromosomes sometimes proliferate and show a tendency to form an 
imperfect network, the pairs generally remain clearly defined throughout 
the interkinesis. The writer has not observed cases in which the chromo¬ 
some boundaries were no longer distinguishable, as reported by Gates (’ 09 , 
p. 538) for certain ‘ semi-resting ’ conditions, and it may be questioned 
whether the latter are normal, i. e. whether such nuclei enter the homotypic 
mitosis and give rise to fertile pollen-grains. 
The history of the heterotypic mitosis thus agrees in all essentials 
with that of Lamarckiana , some of the features being even more perfectly 
shown on account of the larger size of the spindles, the only complications 
resting with the double number of chromosomes, which are naturally more 
difficult to follow. 
The Homotypic Mitosis. With the breaking-down of the nuclear 
membrane following the interkinesis (Fig. 73), the fourteen pairs of 
chromosomes, previously scattered through the nuclear cavity, are brought 
by the development of the fibrillae to the centre of the spindle^ which is at 
first multipolar. At the equatorial plate of the final bipolar spindle (Fig. 74) 
these pairs of chromosomes become grouped very regularly, so that the 
members of the pairs are distributed in two sets of fourteen chromosomes 
each. The chromosomes at the metaphase of the homotypic mitosis have 
the form of short rods, having returned by the condensation of their 
material to about the same size as when they entered the period of inter¬ 
kinesis (compare Fig. 74 with Fig. 68). 
3 R 
