344 Olive. — Sexual Cell Fusions and Vegetative 
the two nuclei begin at once to divide, even before they come to lie close 
together in the fusion cell. This is proved, as I have already suggested, by 
the immediate formation of vacuoles in the nucleoles, and by numerous 
instances in which the nucleoles are cast out into the cytoplasm before the 
two gamete nuclei have approached each other (as, e. g., in Fig. 32). Such 
nuclei as are shown in Fig. 35 also illustrate this point. The dumbbell- 
shaped centrosomes in this instance are obviously in a state of division, 
although cell fusion has only recently taken place, and the two nuclei still 
lie in the bases of their respective gametes. 
Fig. 12 represents a condition just following cell fusion, although, 
in this instance, the origin is not shown. The long, budding growth con- 
tains the two conjugate nuclei, each of which has begun the prophases 
of division. Here again, as in Fig. 35, are seen two dumbbell-shaped 
centrosomes. Each centrosome appears to be contained within the nucleus ; 
but careful focussing inclines one rather to the belief that it is not an intra- 
nuclear structure, but that it lies instead on the surface of the nuclear 
membrane. Figs. 1 and 2 show this point more clearly, for there can be no 
doubt in these instances that the centres, as well as the central spindle 
which is formed between them, lie, at least for a time, on the nuclear mem- 
brane. Further, the connecting isthmus of each dumbbell-shaped structure 
in Fig. J2 corresponds, in my opinion, to the long, slender strand, or 
central spindle, extending between the centres in Fig. 2. 
The nuclear membrane soon breaks down, and the nucleole is thus cast 
out into the cytoplasm, as Figs. 13-15 show. Apparently this phenomenon 
is about to take place in one of the nuclei of each of the following figures, 
12, 29, 32, and 34. The central spindle early becomes a strongly-developed 
structure, apparently of a densely filamentous nature, which stretches 
between the two diverging centrosomes. The chromatin at this early stage 
forms an ill-defined, more or less thready mass at one side of the central 
spindle, much as Hermann and others figured for the nuclei of the Sala- 
mander and other animal cells (Figs. 13-15, 24). 
The central spindle is frequently at first somewhat curved, or bowed in, 
on the side on which the chromatin is massed, as is shown for certain spindles 
in Figs. 14 and 24. Further, during the divisions just following conjugation, 
the two spindles are often variously oriented with respect to each other. 
In Fig. 14, for example, the two spindles are placed almost at right angles 
to each other. In Fig. 13, while the mitotic figures are about parallel to each 
other, they are twisted about so as to lie across the long axis of the cell. 
In Figs. 15, 17, 18, and 20 the more common position is shown, the two 
spindles lying parallel to each other and in the long axis of the cells. 
An early prophase of conjugate division is shown in Fig. 13. The 
nucleoles are somewhat near the central spindles, which are at this stage 
comparatively short. The poles of the spindles are each terminated by 
