87 



place. This is affi'ctt'd in such a way that the iiu'iiibers of the bivalent 

 are separated and one lueniher passes entire to each daughter nucleus, 

 thus bringing al)out a (jualitative division. The retreating chromosomes 

 undergo longitudinal tission as tliey pass to the poles. Having arrived at 

 the poles, they soon break up into smaller pieces, lose entirely their iden- 

 tity and form a complete resting nucleus. This condition lasts but a short 

 time when a spirem is again formed which segments into rod shaped 

 univalent segments. These segments have been quite generally assumed to 

 be identical with the ones which appeared at the poles of the spindle at 

 the close of the first mitosis, but since these segments were seen to lose 

 their identity entirely there is no basis in fact for such a supposition. 

 If these segments are not identical with the grand daughter segments of 

 the first mitosis then this division is not equational. The first division 

 in both genera is (uialitative and rcdnctional ; the second may be conceived 

 either as e(piational or qualitative. 



The investigations on which this preliminary report is bas(>d will ap- 

 pear later in a more complete form, together with figures illustrating the 

 entire lu'ocess of mitosis. 



