Chemical Affinity in terms of Electromotive Force. 195 



A number of direct experiments were also made on this 

 point, using twin cells constructed as follows : — Three beakers 

 were arranged containing solutions of copper, cadmium, and 

 zinc sulphates of the same molecular strengths, and plates 

 of electro-copper, bright (or electro-) cadmium, and amalga- 

 mated zinc respectively. The copper and cadmium beakers 

 were connected by a siphon tube (with ends covered with 

 bladder) filled with the cadmium sulphate solution : and the 

 cadmium and zinc beakers were similarly connected by a 

 siphon tube containing the zinc sulphate solution. The cop- 

 per, cadmium, and zinc plates were then connected with cups 

 Nos. 1, 2, and 3 respectively of a switch-board like that repre- 

 sented in fig. 3, Part V. (§ 106); so that by connecting cups 

 1 and 2 with the electrometer the E.M.F. of the cadmium- 

 copper cell was determined, whilst when cups 2 and 3 were 

 connected the E.M.F. of the cadmium-zinc combination was 

 determined. These readings having been made several times, 

 the zinc and copper plates were transferred to another pair of 

 beakers, containing the same zinc and copper sulphate solu- 

 tions united by a zinc-sulphate siphon, so as to constitute a 

 normal Daniell cell after Raoulf s pattern, and the E.M.F. of 

 this combination determined. Several pairs of zinc and cop- 

 per plates were thus used — each pair being read first in the 

 zinc-cadmium-copper combination, then in the normal Daniell 

 cell, and then again in the ternary combination. In each case 

 the difference between the sum of the average electromotive 

 forces of the zinc-cadmium and cadmium-copper couples dif- 

 fered from that of the zinc-copper combination by quantities 

 no greater than the errors of observation of the electrometer- 

 scale (about +0'1 per cent, when a sufficient number of read- 

 ings were taken) ; whilst the average of the small differences 

 observed with different pairs was actually 0, the small + and 

 — differences due to errors of observation completely balancing 

 one another. This final result (that no discernible difference 

 was to be found between the sum of zinc-cadmium and 

 cadmium-copper couples, and zinc-copper couples containing 

 the same plates) was obtained in each of several sets of expe- 

 riments made respectivelv with solutions of molecular strength 

 MS0 4 47H 2 0, MS0 4 100 H 2 O, and MSO 4 1250H 2 O. 



Rate of Fall in E.M.F. through so-called Polarization occurring 

 in Zinc-Cadmium and Cadmium-Copper Cells for definite 

 amounts of Increase in the Fates of Current-flow. 

 123. The experiments made with normal Daniell cells 

 described in Part V. (§§ 103-105) were repeated with zinc- 

 cadmium and with cadmium-copper plates (exposing surfaces 

 02 



