94 BOTANICAL GAZETTE [AUGUST 
two masses of chromatin at opposite sides of the cell. While 
this is taking place (jig. 30) numerous long mantle fibrils extend 
from the sides of the masses of chromatin and at the same time 
the continuous fibrils increase in number. The mantle fibrils, 
however, are not confined to the lateral position on the chroma- 
tin mass, but radiate from all sides of it. These radiating fibrils 
persist for a considerable time, even after the chromatin has 
secreted a nuclear sap and surrounded itself with a membrane. 
Fig. 31 shows two mature daughter nuclei, with the chromatin in 
the spirem stage, each surrounded by a distinct membrane. The 
continuous fibrils between the two nuclei begin to disappear in 
the equatorial region of the cell, and each nucleus is completely 
surrounded by a system of radiating fibrils with their free ends 
projecting into the surrounding cytoplasm. When first observed, 
the writer mistook this condition for the radiating stage that 
Osterhout has figured in the formation of the second spindle in 
Agave. The two conditions are strikingly alike, but a careful 
study of the stages immediately preceding and following this 
showed conclusively that the radiating fibrils were the remnants 
of the first spindle and not the beginning of the second. As 
shown in the next stage (fig. 32), these radiating fibrils and the 
continuous fibrils disappear completely, and take no part what- 
ever in the formation of the second spindle. As fig. 32 illus- 
trates, the resting period of the daughter nuclei is a very short 
one. The chromatin breaks up into chromosomes before the 
last of the continuous fibrils have vanished. 
The first evidence of the beginning of the new spindle is the 
transformation of the cytoplasmic reticulum close to the daugh- 
ter nuclei into a weft of fibrils completely surrounding each 
nucleus (fig. 33). In every detail the series of events that leads 
to the formation of the second spindles is identical with that of 
the first. Almost every stage in the sequence was carefully 
examined, and the second was found to be a duplicate of the 
first series. It will therefore be only necessary to mention the 
critical stages. 
At first the kinoplasmic zone increases uniformly in thick- 
ness, and its fibrils run more or less parallel to the nuclear wall. 
