84 MR S. W. J. SMITH ON THK NATURE OF ELECTROCAPILLARY Pill- . \O.MK\A. 



The mercury of the dropping electrode fell into the tube A, which had a vertical 

 fine adjustment attached, so that the position of the liquid in it could be varied with 

 respect to the point of the dropping electrode. The capillary electrode was immersed 

 in the liquid of the vessel B. The two electrodes were directly connected by a wire, 

 and A and B were connected by a siphon tube. 



The surface tension of the capillary electrode was measured under four different 

 conditions, viz., when 



1. A and B were filled with %n KC1. 



2. A and B were filled with %n KI. 



3. A was filled with n KC1 and B with \n KI. 



4. A was filled with \n KI and B with \n KC1. 



The capillary electrode formed part of the electrometer previously described, and 

 the curves which it gave for the \n KI and the \n KC1 respectively were first 

 determined. 



From the behaviour of the capillary meniscus it was easy to observe when the 

 dropping electrode jet broke in the surface of the solution in B. If the jet were 

 completely immersed in the liquid and if, then, the latter were gradually lowered, 

 the surface tension of the capillary electrode in general increased correspondingly to 

 a maximum value and then changed in the same way as it did when the dropping 

 electrode was absent showing that the communication between the electrodes had 

 then ceased. 



In cases 1 and 2 the mercury in the capillary electrode assumed (within the limits 

 of experimental error) its maximum surface tension when the dropping electrode 

 broke in the surface of the liquid in A. Thus, for %n KC1 the readings lay between 

 31'3 and 3T35 ; for $n KI the reading for several observations was 29'41. The 

 corresponding readings of maximum surface tension for the electro-capillary curves 

 were 31'4 and 29'43. 



In case 3 the capillary surface tension lay between 26'38 and 26'31, and from the 

 behaviour of the capillary meniscus when the drop electrode jet broke above the 

 liquid surface the surface tension gradually decreased it was evident that this 

 corresponded to a point on the ascending portion of the %n KI curve. The horizontal 

 distance between surface tensions of 29 '43 and 26'35 on the $n KI curve corresponds 

 to a difference in the applied E.M.F. of about '28 volt. So that, if the potential 

 difference between n KI and a KC1 can be neglected, the conclusion from the above 

 observations is that the potential difference between a Paschen electrode and the 

 solution (reckoned from the latter to the former) is about '28 volt, greater when the 

 solution is $n KI than when it is Jn KC1. 



The ol)servation in case 3 is corroborated by that in case 4. In this case, when the 

 electric contact between the dropping electrode and the solution was on the point of 

 breaking, the observed surface tension of the capillary lay between 30'48 and 30'5 ; 



