244 J. Olsen — Investigation into the Existence of 



If the deflections in any direction are large at the begin- 

 ning of the observations, the direction cannot be reversed bv 

 this charge, and this method and the effect is only a differential 

 one, but if tlie cell is in a more stable condition the dii'ections 

 can be reversed, and this was done scores of times. 



The induction produced when the jars are discharged is of 

 course in the opposite direction from what it is when charg- 

 ing, but it is not large and does not appear in the third reading 

 of 2-minute intervals. Diffusion acts against the heaping up 

 of positively or negatively charged ions in any part of the 

 liquid, and this often appears in the last observations of a 20- 

 minute series. 



A solution of potassium hydroxide and a solution of potas- 

 sium sulphate were also tested by this method, using platinum 

 electrodes as before, with the same results as with dilute sul- 

 phuric acid. 



2 



IHt 



Incidentally it may be mentioned that the induction pro- 

 duced, in the galvanometer circuit, by bringing up a charged 

 body to the cell was of about the same order of magnitude 

 whether there was any liquid in the cell or not. 



One acid, one salt, and one base were used in these experi- 

 ments. The only reason for choosing these particular ones 

 was, that it was necessary to employ some substance which 

 on being decomposed and set free would not affect the pla- 

 tinum electrodes, and it was desirable also to have substances 

 whose ions were known to have a considerable velocity. 



Distilled water was always used. 



It was now thought that it would be interesting to study the 

 effect of an electromotive force, less than the decomposition 

 value, upon an electrolyte. Accordingly the following experi- 

 ments were performed : 



]S^o electrostatic charge was here employed. 



A solution of different percentages of sulphuric acid was 

 put into a cell having four platinum electrodes. The cover of 

 the cell was so made that the inner or middle electrodes could 

 be rotated independently of the outer. 



The two inner electrodes a and a' were connected to the 

 galvanometer circuit, and the two outer electrodes A and A' 

 were to be connected to a Daniell. a and A were about 8*0 ""^ 

 apart, and a and «' about 50-0 """' apart. The circuit to the 



