THE AXIAL GRADIENTS IN HVDROZOA. 93 



metric negativity correspond in all cases to the regions of highest 

 susceptibility and rate of reduction. 



REDUCTION AFTER KILLING. 



The gradients in rate and amount of reduction of perman- 

 ganate appear when the living animals are brought into per- 

 manganate, but when they are first killed by some other means 

 and then brought into permanganate the gradients are either 

 entirely absent, or only the merest traces of them remain. 

 Various means of killing have been used in these experiments, 

 among them HgCl 2 , HgCl 2 with 5 per cent, acetic acid, HNOs, 

 HC1, strong alcohol, formalin, hot sea water. If the animals 

 are brought into permanganate within a few moments after the 

 killing agent is applied, with only a brief washing in sea water, 

 if necessary to remove excess of the killing agent, slight gradients 

 may still appear. So far as my observations go, however, this 

 is not the case if the animals are subjected after killing to the 

 usual treatment of histological material. If, for example, they 

 are passed through the alcohols up to 70-80 per cent, and kept 

 in this percentage for a day or two, or if they are killed and kept 

 in formalin and then washed in water before being placed in 

 permanganate, the rate of reduction in different parts is uniform 

 and the only differences in depth of color, so far as could be 

 determined, are the apparent differences resulting from differ- 

 ences in thickness of the layers through which the light passes. 

 Moreover, in these killed animals the total amount of reduction 

 is much less than when the living animals are brought into 

 permanganate. Potassium permanganate, even in high dilu- 

 tions, is highly toxic and when the living animals are placed in 

 it they are of course killed by it, but it is evident from the facts 

 cited that the differences in physiological condition in different 

 regions play a part in determining the rate and amount of reduc- 

 tion. In the dead animals these differences are no longer present, 

 although some factor or factors concerned in them may persist 

 for a time after the killing agent is applied. Whatever the 

 nature of the differences in reaction to permanganate between 

 living and dead protoplasm, it is evident that the gradients in 



