184 PROCEEDINGS OF THE AMERICAN ACADEMY. 



phenol phthalein, thus showing that the carbonate had })een satis- 

 factorily eliminated. No barium could be found by careful qiuilitative 

 testing in a solution, but a minute trace of sulphate was found there. 

 This impurity Avas quite too small in amount to have any effect on the 

 specific heat. 



The Concentrations of the Purest Sodium Hydroxide Solutions 

 No. Date NaOH(%) Mols.H,0 



1 April, 1909 2.1710 100.06 



2 " " 2.1703 100.10 



3 May " 2.1709 100.07 



100.08 

 4 October, 1909 2.1685 100.18 



A solution containing enough carbonic acid to combine with 1.3 

 percent of the alkali present was also made and standardized in the 

 same way, the amount of carbonate present being calculated from the 

 difference between the end points found with the two indicators in 

 the usual way. The specific heat of this solution also was taken in a 

 few experiments. They did not, however, deviate more from those 

 with the pure solution than the possible error of the experimentation, 

 and need not therefore be recorded. Their evidence at any rate was 

 enough to verify the prediction made above that the presence of a 

 small amount of carbonate would have had an almost if not quite 

 negligible effect. 



Potassium Hydroxide. 



In the case of potassium hydroxide a similar method was used. 

 Two samples were prepared, one made up directly from the purest 

 obtainable commercial product and containing 1 .8 percent of its alkali 

 combined with carbonate, and the other freed from this carbonate in 

 precisely the manner described in the case of sodium. The analyses 

 of these solutions are recorded below. They were essentially identical 

 in total alkalinity and no difference whatever could be detected in 

 their specific heat. This again supports the conclusion reached in 

 the case of contamination with sodium carbonate. The object of 

 preparing the two solutions in the case of potassium as well as that of 



