1903 | STUDIES IN SPINDLE FORMATION 83 
development of the spindle in the spore mother-cells of Osmunda. 
In this case the spindle originates from a zone of kinoplasm 
which surrounds the nucleus. The granules in the kinoplasm 
arrange themselves in rows, concentric with the nuclear wall, and 
are finally massed on opposite sides of the nucleus. From these 
masses two cones of fibrils are developed, which become the two 
cones of the spindle. As there are only two primary cones 
developed, the spindle is bipolar from the first. 
Osterhout (1902), in his recent work on Agave, records the 
following series of events leading to the formation of the spindle 
in the mother-cell. During the early stages there is developed 
a cytoplasmic membrane outside of the nuclear wall. This he 
calls the “‘spindle membrane,’ and he regarded it as a unique 
structure. There is no weft of fibrils formed, but the spindle- 
forming fibrils are radial from the beginning, and are attached to 
both the nuclear and spindle membranes. From these radial 
fibrils the cones are developed, and these, by fusing into two 
groups, bring about the bipolar condition. Probably the most 
remarkable observation that Osterhout has recorded is that the 
second spindle formation differs entirely from the first. Accord- 
ing to his observations, the spindle-forming fibrils of the second 
division are found in close contact with the nuclear wall, and 
their free ends radiate from it. They extend outward into the 
‘cytoplasm, and by the union of their ends form a series of cones. 
These fuse at their apices into two groups and form the bipolar 
‘Spindle in much the same fashion as in Equisetum. 
From the observations of the writers described above it is per- 
fectly obvious that there are considerable differences between 
the methods of spindle formation. It is also evident that there 
are certain important features which are common to nearly all of 
them. It would seem that there are several distinct types of 
‘Spindle formation, but the differences between them are too 
reat, and the number of forms worked out in detail are too few 
to allow of any generalizations. The number and character of 
these types can be determined only by additional observations. 
it is to this end that the following studies are recorded. 
