RICHARDS. CALOMEL ELECTRODE. 5 



the case of the hydrochloric acid, to warrant further investigation. But 

 the exigencies of a research of this sort, the necessity of a hirge number of 

 pure materials, and the long duration of each experiment, postponed the 

 completion of the work for a much longer time than had been anticipated. 



First, a word about the materials employed in the work which follows. 

 The mercury was carefully purified chemically, and was also distilled in a 

 vacuum. Four grades of mercurous chloride were used ; in the first crude 

 experiments " pure " calomel of commerce sufficed ; subsequently a new 

 portion of the salt was precipitated from an exceedingly dilute solution of 

 mercurous nitrate, which was free from every impurity except mercuric 

 nitrate, by the addition of hydrochloric acid ; the third sample was made 

 in this same way from very carefully prepared mercurous nitrate which 

 had long been digested with mercury ; and this third sample upon being 

 twice successively sublimed yielded the fourth. When sufficiently washed 

 with successive portions of a decinormal solution of potassic chloride, each 

 of these preparations yielded " decinormal electrodes" of precisely the 

 same potential, showing that as far as this work was concerned they were 

 all alike. For most of the work the third sample, as being at the same 

 time convenient and trustworthy, was adopted. 



In order to free the hydrochloric acid which was used from possible 

 traces of bromine and iodine, it was somewhat diluted and twice treated 

 with a few crystals of potassic permanganate, each time boiling until all 

 the chlorine was expelled. The acid was then twice successively distilled 

 in a retort provided with a platinum condenser, rejecting the first and last 

 portions of the distillate. 



The soluble chlorides were also of a high grade of purity. Many had 

 been prepared in the course of recent atomic weight determinations in 

 this laboratory, and were far purer than is required in an investigation of 

 the present sort, while others were prepared for this research. Especial 

 pains were taken to exclude all anions except chlorine. The solutions 

 were alwa.ys made in such a way as to be as neutral as possible ; they 

 were in the first place prepared of the desired strength, and this strength 

 was subsequently verified by a determination of the chlorine according to 

 a modification of Volhard's method proposed independently by Sanger* 

 and myself.f Although parallel experiments with boiled and unboiled 

 solutions showed no difference in the temperature coefficient, the solutions 

 were usually boiled for a short time in Jena glass vessels to expel most of 

 the air. It soon appeared that the most serious cause of irregularity con- 



* These Proceedings, XXVL 34. t These Proceedings, XXIX. 67. 



