378 



Dr. W. W. J. Nicol on the Mutual 

 Table III. 



Salts. 



Present 

 determination. 



Former 



result. 



Biidorff. 



KClandNaCl 



KOlandKNOg 



44-70 



51-87 



78-89 



134-21 



4560 



52-67 



76-93 



135-29 



45-6 18°-8 



54-30 



79-89 



NaCl andNaN0 3 



NaN0 3 andKNOg ... 



In parts of salt per 100 parts of water. 



I shall have to return to this question later ; so any discus- 

 sion of the general accuracy of the above results will he best 

 deferred till then. With regard to the composition of the 

 total quantity of mixed salts in the saturated solution little 

 need be said here. At first attempts were made to determine 

 the ratio of the two metals in the case of the mixed chlorides and 

 mixed nitrates by estimating the total chlorine and conversion 

 into sulphate respectively ; but the atomic weights of the two 

 metals sodium and potassium differ by so small an amount, 

 that all such indirect methods of analysis yield, of necessity, 

 but approximate results on which no great reliance can be 

 placed. In the case of the other pairs of salts the determina- 

 tion of the chlorine gives a direct method by which the ratios 

 of the two salts can be ascertained, and one which can be 

 trusted ; but there is a still better method applicable to all 

 four pairs of salts, which will be described further on. 



The remainder of the research was conducted as follows : — 

 Solutions of the second salt in each of the eight pairs given 

 on page 376 were prepared by weighing definite amounts of 

 water (about 50 cub. cent.) to the nearest half milligramme, 

 and adding to these the requisite quantity of the salt to form 

 a solution of a definite molecular composition (n molecules 

 of salt to 100 molecules of water). The salts used were 

 carefully purified, the potassium and sodium chlorides by pre- 

 cipitation with hydrochloric-acid gas from solutions of the 

 commercially pure salt, which had been previously treated 

 with ferric chloride, followed by precipitation of the iron and 

 other heavy metals that might be present by the pure car- 

 bonate of the metal. The salt thus precipitated with hydro- 

 chloric acid was thoroughly drained, washed repeatedly with 

 pure hydrochloric acid, and finally dried by heating to 200° C. 

 for several days. In some cases the salt was fused in a 

 platinum dish at the lowest possible temperature, and in no 

 <jase did the salt possess the slightest acid reaction. The 



