236 THE CYCLE OF CELL-DIVISION 



granules; but that, since it is not necessary for it 

 actually to colour these granules, as shown by the last 

 experiment, it seems probable that the stain induces 

 divisions by virtue of some substance contained in it 

 which does not colour granules. It is not the stain 

 itself which induces divisions; it is some constituent 

 of it, and the action of that constituent is greatly 

 augmented by atropine and extract. 



The next point is that when mitosis is induced on a 

 microscope slide with stain, death is premature. Even 

 if there is not sufficient stain to colour the nucleolus- 

 centrosomes, death rapidly follows. We believe that 

 this dye contains at least two constituents which can 

 be utilised differently by the cell's protoplasm a 

 substance \vhich, by combining with the cell-gran- 

 ules, causes the cell to reproduce itself, and a poison 

 which kills it. Both diffuse into the cell together; 

 mitosis is induced and then the cell dies prematurely. 

 If the stain is sufficiently concentrated, the chromatin 

 after it is dead will combine with it, and the chromatin 

 then turns bright scarlet. From prolonged observation 

 of these induced divisions we think that the scarlet 

 coloration of the chromatin is a post-mortem effect. 

 The stain as it diffuses into the cell induces division as 

 it combines with the granules, which die and become 

 coloured one by one. All the time the stain is passing 

 farther into the cell, and later and later stages of mitosis 

 are being induced. Ultimately the nucleolus-centro- 

 some is reached and the cell dies; and thus it is seen 

 dead in the act of mitosis with its chromosomes and 

 centrosomes stained bright scarlet. If, on the other 



