8 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



The next series of results had for its object the determination of the 

 effect of changing temperature. Heat evidently assisted the decomjDO- 

 sitiou of mercurous chloride, for the precipitate in the warmer tube had 

 always appeared the grayer of the two. Upon cooling to the same tem- 

 perature, however, the potential always came back to zero, unless evapo- 

 ration had been allowed to make one solution more concentrated than the 

 other. Experiments in which the potential did not return to zero were 

 rejected. 



SERIES IV. 

 Changk of Temperature Coefficient with Change of Temperature. 



In the course of these experiments it had been noticed that an elec- 

 trode upon shaking always yielded a higher potential than after it had 

 remained standing. This was especially the case with concentrated solu- 

 tions and high temperatures. This fact has been observed by Cogges- 

 hall,* who recommends a layer of sand or small pebbles as a means of 

 preventing the agitation. The irregularity in question appeared to show 

 a lack of homogeneity in the solution, and it seemed important to explain 

 the matter before proceeding further. 



With this end in view, several experiments were made with bromides 

 instead of chlorides. Here the irregularity was found to be vastly 

 greater, and the precipitated bromide became gray at once even in cold 

 solutions. Normal solutions gave results which varied over the wide 

 limits given below, and even these cannot be considered the possible 

 limits. Decinormal solutions were much less uncertain, but these seemed 

 to be as much affected as normal solutions of chlorides. 



* Z. Phys. Chem., XVII. 62. 



