302 
AMERICAN JOURNAL OF BOTANY 
[Vol. 9, 
spindle, however, and hence is to be considered as an incomplete cell plate. 
This incomplete cell plate may persist after the nuclei are completely re- 
organized, but it always disappears during early interkinesis at least. In 
no instance was a pollen mother cell in interkinesis or subsequent stages 
found in which complete bipartition had taken place. 
Interkinesis is marked by the disappearance of the transitory cell plate 
and the subsequent enlargement of the two nuclei. The spindle fibers be- 
come progressively fewer and fewer, until in some cases it is practically 
impossible to find a single fiber connecting the two nuclei. The fibers are 
furthermore obscured by the formation of numerous large bodies in the 
cytoplasm during this period. These bodies approach the size of plastids, 
though they do not seem to have as definite boundaries as do those struc- 
tures. They seem to be slightly flocculent in consistency, and it may be 
that further studies in chrondriosomes will reveal their nature. 
The incomplete cell plate of Nelumbo seems to resemble that of Mag- 
nolia (7) to a considerable extent, except that it is perhaps somewhat more 
fully formed. Timberlake (23) reported such a structure as occasionally 
being found in the larch. He explains the failure to accomplish partition 
in these cases as due to a lack of formation of peripheral spindle fibers, and 
the same might be said of the condition in Nelumbo. Tangl (22) and Juel 
(11) found incomplete cell plates in Hemerocallis, but here they persist for 
some time after the telophases. A closer approach to the condition in 
Nelumbo is probably that reported by Juel (12) in Carex as occurring after 
both the heterotypic and the homoeotypic divisions. 
The period of interkinesis is apparently quite long, and during this time 
the nuclei enlarge very much. They become very much larger than the 
nuclei which result from the second division, and can be easily distinguished 
from the latter by their size without consulting neighboring sections to 
determine the number of nuclei within the cell. These nuclei in interkine- 
sis remain at some distance from the plasma membrane. No evidence at 
all of an incipient furrow such as was found in Magnolia (7) was revealed. 
The homoeotypic mitosis is accomplished by the formation of long, 
narrow spindles which appear very much like those of the heterotypic 
mitosis except that they are smaller. In some cases the two spindles are 
almost exactly parallel, while in others they are almost exactly at right 
angles (PI. XV, fig. 2). It seems likely that the usual condition is that the 
spindles are somewhere between parallel and at right angles to each other. 
At any rate, the division results usually in a tetrahedral arrangement of 
microspores, though there are some departures from this disposition. 
Observations were made on many cells showing the homoeotypic telo- 
phase spindles at various stages, but in no case was a cell plate or orange 
zone seen. When the chromosomes first reached the poles, there was in 
one or two cases an apparent thickening of the spindle fibers throughout 
the middle portion of their length, just as was observed in the heterotypic 
