304 
AMERICAN JOURNAL OF BOTANY 
[Vol. 9. 
view is shown in the upper part of the cell of figure 4, where two spindles 
are crossed by continuous plates. That this is not a central section of the 
spindle but a superficial view is obvious from the fact that the upper nu- 
cleus does not appear in this section. But it is equally apparent that the 
appearance of the equator of the two upper spindles would be exactly the 
same in their central portion if the section had been cut so as to show a chip 
off the upper nucleus and similar chips off the two lower nuclei. Such 
a section would give evidence of a continuous plate across the central spindle, 
until a careful examination of adjacent sections showed this interpretation 
to be incorrect. It is doubtless this mistake which was made by Lubimenko 
and Maige and by other observers in reporting at least some of the instances 
of quadripartition by cell plates. A central section and careful focusing 
invariably, in all instances which I have seen, reveal the fact that this plate 
does not extend across the center of the spindle and that it is connected 
with the plasma membrane at its periphery. 
From the time the furrow begins on the very boundary of the cell, there 
is an increase in the number of fibers of the central spindle. They seem to 
appear in centripetal order rather than centrifugally as in typical cell-plate 
formation. The furrow is very slender indeed, which adds to its resem- 
blance to a cell plate as compared with the furrows of the other plants which 
I have studied. It is so thin, in fact, that it appears to consist during its 
development simply of the plasma membrane itself, and that cell-wall 
material is not found between the daughter cells until later. However, 
observation of cells which are somewhat plasmolyzed indicates that this 
interpretation is incorrect, and that a projection of the cell wall does extend 
into the furrow in some cases at least. Even at this stage of quadripartition 
the cell walls are usually not thickened to a marked degree on all sides. 
Figure 4 shows an enormous thickening of the wall above and a consider- 
able thickening below, but the sides of the cell wall are rather thin. The 
middle lamella of the old mother cell is quite evident in some instances 
(fig. 4), while in other cases it seems to be lacking (figs. 2,3). 
A superficial view of the furrow indicates that it is perforated, as if 
it were made up of thickenings on spindle fibers as in typical cell-plate for- 
mation. Careful observation, however, reveals the fact that this perforated 
appearance obtains only on the inner margin of the centripetally forming 
furrow. Mrs. Farr (9) in her study of Cobaea found that the furrow as it 
advances inward presents not a simple cutting edge, but that its edge is 
wavy, the projections extending in between the fibers of the central spindle. 
Such a condition would give the appearance in section which is described 
above. 
During quadripartition the furrows are apparently all simultaneously 
initiated within a given mother cell and proceed in their development at 
about the same rate, so that it is the portion of the central spindles between 
the exact central point and the exact center of the entire tetranucleate cell 
