TIME TAKEN BY ACT OF MITOSIS 229 



correspondingly slow. But as far as we can see, mitosis 

 cannot occur completely in less than about three 

 minutes. It can take a very long time in its accom- 

 plishment ; but it cannot be completed in less than three 

 minutes. Hence, if mitosis can take place slowly, 

 without the cell being killed by the stain, complete 

 mitosis can occur ; but if the nucleolus-centrosome 

 stains in less than a minute or so, death will occur 

 before the cell has had time to divide. This fact 

 governs the w r hole of this experimentation, for when 

 inducing cell-division with the aniline dye it must 

 be remembered that the mitosis has to occur after the 

 cell-granules have begun to stain, but before death is 

 occasioned by the staining of the nucleolus-centrosome. 

 We have the power of accelerating and delaying the 

 diffusion of the stain into the cells by adding or sub- 

 tracting alkali, or by increasing or decreasing the 

 concentration of the stain by rules which can be plotted 

 in an equation, and therefore by such an equation we can 

 ascertain the rate of cell-division as induced by the 

 chemical agent. But throughout it must be appre- 

 ciated that it is the stain which is inducing the 

 cell-division, and that if the stain is not sufficiently 

 concentrated no division will occur at all. On the 

 other hand, it must also be remembered that an excess 

 of stain will poison the cells too quickly. A cell must 

 absorb a certain amount of stain before it will divide, 

 and the absorption depends on the concentration of the 

 stain in the jelly and on the alkali. One may place 

 living blood-cells on a jelly w r hich contains the best 

 ingredients for inducing cell-division; but unless the 



