182 PROCEEDINGS OF THE AMERICAN ACADEMY. 



(see Davis, : 04, Am. Naturalist, vol. xxxviii, p. 435, and literature 

 there cited). Stevens (:03) has described the origin of the spindle 

 from the linin in Synchytrium, but his material was apparently less 

 favorable for the exact determination of the process. In Phyllactinia, 

 Harper (: 05) has shown that the spindle-fibres of the mitotic figure are 

 identical with the linin -fibres of the resting-nucleus, but the method 

 of spindle-formation differs from the present case, since the process is 

 controlled by centrosomes in Phyllactinia. From the method of origin 

 in Entomophthora, it is clear that the spindle-fibres are identical with 

 the linin- fibres and are attached to the chromosomes permanently. In 

 this process of spindle-formation there is no place for centrosomes, 

 and no structures resembling centrosomes were seen in any case. 



During this migration of the spindle-fibres the chromosomes are 

 grouped in a more or less definite equatorial plate (Figure 5). The typi- 

 cal metaphase with the e(iuatorial plate and the bipolar spindle, if it 

 occurs in Entomophthora, must be of very short duration, as it is the 

 only phase that has not been seen in examining a large number of 

 sections showing various stages of mitosis. The nearest approach to 

 the metaphase figure is the condition shown in Figure 5. The absence 

 of the typical metaphase may perhaps be explained if we assume that 

 the chromosomes split and begin to separate before the spindle becomes 

 definitely bipolar. The next older stage of mitosis seen in the prepa- 

 rations studied, is one where the chromosomes have divided and the 

 daughter-chromosomes are beginning to draw apart. The spindle, 

 occupying the entire nuclear-cavity, elongates at this time, and the 

 daughter-chromosomes are clearly half the size of those in the equa- 

 torial plate ; this stage (Figure 7) is unmistakable and was met with 

 repeatedly. As the daughter-chromosomes reach the poles the wall of 

 the nucleus becomes more rounded at these points. The chromosomes 

 remain very clear up to a late stage, and eight could often be counted. 



Two methods occur in the reconstruction of the daughter-nuclei. In 

 the first method, a vacuole appears between the two daughter-nuclei, 

 and the spindle-fibres in this region break down, leaving only a few 

 granules (Figure 9). The old wall of the mother-nucleus persists up 

 to this time, connecting the two daughter-nuclei. This old wall forms 

 the walls of the daughter-nuclei, except on the side toward the vacuole, 

 where new walls are laid down, probably by the plasma-membrane, 

 bordering the vacuole (cf. Lawson, : 03). The portion of the old wall 

 connecting the daughter-nuclei now disapjieai's and the latter draw 

 apart freely. The second method is much less common than the first, 

 but was met with several times. The portion of the wall of the 

 mother-nucleus and the spindle-fibres connecting the daughter-nuclei 



