Chemical Affinity in terms of Electromotive Force. 293 



deposition of external film as the current flowed producing 

 electrolysis), and being only a fraction of the variation pro- 

 duced on exchanging plates. With bright zinc plates, how- 

 ever, the alteration during a few minutes after setting up was 

 often considerable and irregular in direction. The numbers 

 given in the subsequent section are in all cases deduced from 

 the average readings of the cells during the first 15 or 20 

 minutes after first setting up, or after exchanging plates, as the 

 case might be; the mean of the two averages thus obtained 

 (or of the four, if three exchanges were made after taking the 

 first set of readings) being finally adopted as the mean value 

 for the particular experiment. 



As regards the experiments with the diffusion-cells, these 

 were ordinarily constructed of two beakers united by an 

 inverted Y-shaped tube, the ends of which were covered with 

 bladder or other thin membrane, and filled with either of the 

 two solutions used, one in one beaker, the other in the other: 

 in short, they were precisely similar to the forms of Raoult 

 cell used in the preceding portions of these researches. It was 

 thought desirable to see if any influence could be traced as 

 exerted by the membranes as such, and accordingly a number 

 of gravity diffusion-ceils were set up by pouring into a long 

 U-tube the stronger of the two solutions used until rather 

 more than half full, and then very cautiously running slowly 

 into one limb the weaker solution, so as to have a well-marked 

 line of demarcation between the two fluids. Two plates were 

 then immersed in the two solutions and exchanged after 

 taking readings, and so on. Both with zinc sulphate and 

 copper sulphate no measurable differences could be detected 

 between the average results obtained with these cells and 

 with the Raoult form of cell, both sets of cells having respec- 

 tively the same fluids and kinds of plates; clearly proving 

 that the membranes as such exercised no measurable influence 

 on the results. 



As an illustration of the general character of the experiments 

 made, the following observations with two opposed Daniell 

 cells may be cited as an example, typical in all respects saving, 

 first, that the values obtained before and after exchanging 

 plates differ more than usual (for which reason this experiment 

 is selected) ; and, secondly, that the galvanometer was regu- 

 lated so as to be considerably less sensitive than was found 

 desirable for other experiments,on account of the comparatively 

 large potential-difference to be measured. 



Plates used : amalgamated zinc and electro-copper. 



Solutions used : per 100 H 2 0, 2'0 CuS0 4 throughout ; 

 5-584 ZnS0 4 in one cell and -1012 ZnS0 4 in the other. 



