IAL GRADIENTS IX HYDROZOA. 31 



of reduction indicated by rate of appearance of the brown color 

 and as gradients in amount of reduction indicated by depth of 



r when the reaction proceeds to completion in excess of the 

 agent and the preparations are dehydrated and cleared. The 

 possibility that a differential permeability along the axis, either of 

 the plasma membrane of naked stages, or of the perisarc in stages 

 possessing it, is a factor in the gradient in rate of reduction need 

 not In- di-nied. It is highly improbable, however, that such dif- 

 < -ability are fundamental factors in the gradients 

 of naki-d for KMnO, in the concentrations used penetrates 



ra]>idl\ and < . < -n alters the visible aggregation of the protoplasm 



>re the i-ednetiiin gradient becomes distinct. Any differences 

 in penne ability which may have been present must be destroyed 

 early in its ad ion. Moreover, the possibility that differential 

 iii-abiliiy determines the gradient in amount of reduction as 

 indicated by depth of color due to oxides precipitated throughout 

 the cell or cells, seems to be eliminated. And finally, the fact 

 that tl u-iits in rate and in amount of reduction are the 



.same i"t a particular stage and are also identical with the 

 gradients in susceptibility indicates clearly enough that the dif- 



it methods of demonstrating the gradients are merely dif- 

 ferent ways of making evident one aspect or another of the gen- 

 eral quantitative gradation in condition of the living protoplasmic 



in. 



That the gradients indicated by reduction of KMnO 4 are closely 

 assi .dated with the processes of living in the protoplasm is further 

 .shown b\ tin fact that individuals of various stages killed in vari- 

 ous ways and then placed in permanganate reduce much less of 

 tin- a-ent than living individuals and show, either no traces of 

 axial gradients, or in some cases slight traces for a short time 

 after the action of the killing agent. For example, when develop- 

 mental are placed in alcohol, HgCL or various other agents 



': for hist .'logical purposes, then washed and brought into 

 K.Mn< >, after various periods in the killing agent, it is found that 

 a few M-cmuls in most agents is sufficient to obliterate in large 

 measure the differences on which the reduction gradient depends, 

 though slight traces of a gradient may persist in some cases, even 

 for several hours. In material carried up to 80 per cent, alcohol, 



