a LE ET AEN I FR 
ee 
1900 | DEVELOPMENT OF THE POLLEN GRAIN gI 
the spirem show numerous distinct chromatin granules on a com- 
paratively short ribbon. The ribbon soon becomes much 
thickened, with many sudden bends and irregular twists, larger 
chromatin masses being evident at definite places. By its 
bendings the whole ribbon seems to be marked off into definite 
segments, as in fig. ro. At this stage often the connection of 
the ribbon with the nucleolus may be seen. Immediately pre- 
ceding actual segmentation, as well as afterward, each forming 
Segment has its chromatin massed toward the ends. A longi- 
tudinal splitting of the segment is then evident, thus somewhat 
imitating a characteristic of heterotypic division. On separation, 
or complete segmentation, the form of the chromosomes is that 
of a double half link. 
With approaching division, the nuclear cavity increases 
Somewhat in size, and the chromosomes are in the vicinity of the 
nuclear membrane. The nucleus is ina position near that wall, 
along which, apparently, the microspore was adherent in the 
mother cell. With the gradual opening of the network forming 
the nuclear membrane, kinoplasmic threads enter the nuclear 
hollow largely from the region of the greatest cytolasm; and 
nO previous radial arrangement of these threads is noticeable. 
In the forming spindle of the microspore nucleus, an inter- 
esting variation from the normal type of multipolar spindle 
Prevails. The fibers readily take up a position more or less 
Perpendicular to the wall near which the nucleus is located. 
These fibers arrange themselves in contact with this wall at many 
points, or ina subpolar manner. At the opposite or free end 
there are at first, also, false poles. There is then a multipolar 
Spindle in an axial plane. Finally, at the free end, the clusters 
are usually drawn together into a single pole, or what is approxi- 
mately a Single pole. The complete spindle is thus truncated 
at oné end, and more or less conical at the other (fig. zr). 
When arranged in the nuclear plate, the ends of the chromo- 
somes point outward, and on account of the loop form there 
a '0: Ge. more chromosomes i the periphery than are 
really present. As in the case of the pollen mother-cell divisions, 
