194 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 



of fibers which extend from pole to pole are not straight 

 as if stretched, but are more or less wavy (fig. 19). This 

 may be due to the drawing together of the poles. That 

 the poles do approach each other appears probable from 

 the fact that, although in the equatorial plate stage, the 

 poles extend almost or quite to the cell-wall; yet the 

 daughter-nuclei are always some distance from it. 



The difference in the character of the threads of the 

 inner and outer portions of the cytoplasmic reticulum, 

 which has been mentioned above, increases as spindle form- 

 ation progresses. The threads of the inner portion become 

 less granular and more sharply defined, while those of the 

 outer portion behave in exactly opposite fashion, becoming 

 more and more granular, and finally coming to resemble 

 accumulations of granules rather than threads. Sometimes 

 in the bipolar stage only a few meshes remain and these 

 lie close to the cell-wall (fig. 19). At other times in the 

 same stage such a reticulum still extends throughout the 

 cell (fig. 18). 



Frequently, when the chromosomes have reached the 

 poles, and also after the daughter-nuclei are formed, 

 branching fibers extend in all directions from the poles 

 almost to the cell-wall. At this time the granular zone 

 may still be discernable, although it may be more or less 

 obscured by the presence of granular matter both inside 

 and outside of the zone (fig. 21). Before long the spindle 

 fibers are separated from the daughter-nuclei by granular 

 matter and each daughter-nucleus is surrounded by a dense 

 granular mass of cytoplasm (fig. 22). 



Many recent observations on spindle-formation in higher 

 plants show that the developing spindle passes through a 

 multipolar stage which becomes bipolar by the fusion of the 

 poles, or by the withdrawal of some of them. Multipolar 

 stages are figured and described by Farmer (1893 and 

 1895), Belajeff (1894), Strasburger (1895), Osterhout 

 (1897), Mottier (1897 a and b), Juel (1897), Webber 

 (1897) and Debski (1897). But the origin of the multi- 

 polar spindle has been worked out in only two or three 



