eee ee 
ae 
1903] STUDIES IN SPINDLE FORMATION gt 
As the apices of the cones approach near to the cell wall, the 
nuclear membrane breaks down, and the space once occupied by 
the nuclear sap becomes filled with the ingrowing fibers of the 
base of the cones (fig. 20). By the time the nuclear wall has 
disappeared, some of the cones unite at their apices. This union 
continues in the same fashion as that described in Gladiolus, 
Iris, and Disporum. /%g. 27 shows a condition in which the 
nuclear wall has entirely disappeared, and five cones are seen in 
section. Some of these cones are evidently the result of the 
union of several primary cones. By the time this stage is reached 
the fibrils have become long independent threads, converging to 
the apices of the cones, and many of them have become attached 
to the chromosomes. 
The fusion of the cones is probably a very rapid process, as 
the multipolar stages were only obtained from material fixed in 
the field, immediately after being dissected from the plant. They 
are never found as frequently as the bipolar stage. 
fig. 22 shows the cones uniting into two groups, pointing in 
Opposite directions. Before fusing, the cones point outward in 
all directions, so that in a section few of them show (fig. 27), 
but as they collect in groups a median section shows many more 
in the same plane. When the bipolar condition is finally reached, 
the usual three sets of fibrils are sharply differentiated. The 
connective fibrils are clearly defined and appear to be thicker in 
the region of the chromosomes. The continuous fibrils extend 
uninterruptedly from pole to pole and are much finer than the 
connective fibrils. Extending laterally from the poles, numerous 
mantle fibrils are to be seen, with their free ends projecting into 
the cytoplasm (fig. 23). 
As shown by the series of figures, the process of spindle 
development i in the first division of Hesperaloe is essentially the 
Same as that described for Gladiolus, Iris, and Disporum. On 
account of the scarcity of the material, the development of the 
spindle for the second divisions was not observed. 
THE POLLEN MOTHER-CELLS OF HEDERA HELIX. 
From the above description, it seems quite evident that there 
isa very ae resemblance in the method of spindle forma- 
