720 CALKINS. [Vol. XV. 



progressive process of division, and from a study of the various 

 modes of aggregation of the chromatin fragments in different 

 nuclei, it appears that they divide first into two nearly equal 

 parts, these into four, the four into eight, etc. In most cases 

 the parts thus formed soon become scattered so that it is 

 almost impossible to follow them ; but in some favorable cases 

 the parts remain in groups in the places occupied by the origi- 

 nal karyosomes until as late as the eight-part stage (Fig. 4). 

 I have never been able to follow the fragmentation beyond this 

 stage. The chromatin-granules thus formed leave their origi- 

 nal positions and become concentrated at the side of the 

 nucleus which lies away from the sphere (Fig. 5). But even 

 here it can be made out that the larger pieces continue their 

 fragmentation into smaller and smaller portions. The final 

 result of this disintegration is the formation of a great num- 

 ber of minute chromatin granules similar to those of the 

 nucleus of Actinosphaerium (Brauer). 



As they break up, the chromatin masses often, but appar- 

 ently not always, give rise to beaded fibers which have no 

 definite arrangement in the nucleus (Figs. 6, 7, 12). Ishikawa 

 ('94) in his second paper describes these fibers as chromosomes 

 and pictures them in his Figs. 33, 38, 40, 44, and 47. He 

 thought they were formed from the parts (" microsome-discs ") 

 of the disintegrated nucleoli (karyosomes) which were put 

 together "one after the other like chains of mammalian blood 

 corpuscles." I am convinced, however, that although these 

 fibers appear in the manner described by Ishikawa, they are 

 not the actual chromosomes. They do not pass directly into 

 the nuclear plate, nor do they split longitudinally. This stage 

 is a prophase of division, as shown by the condition of the 

 sphere (Fig. 12). It is antecedent to the stage of the nuclear 

 plate, and later than that of the resting nucleus. It must, 

 therefore, correspond to the spireme-stage of mitosis, though 

 it cannot be interpreted as a monospireme, for in total prepa- 

 rations the ends of the figures are in full view and are in no 

 way connected with each other (Figs. 6, 7). 



After the above changes the granules which compose the 

 fibers of the spireme break down into still more minute 



