OF AQUEOUS SOLUTIONS. BARNES. 137 



In this table it will be noticed in the first series of observa- 

 tions, where the concentration of the constituent solution of 

 sulphuric acid is constant, while the constituent solution of hydro- 

 chloric acid has a variable concentration, that the calculated 

 values are all greater than the observed, and that the differences 

 gradually increase as the concentration of the hydrochloric acid 

 increases. This is also true for the second and third series, 

 except in the case of the weaker hydrochloric acid solutions 

 where the calculated is now less than the observed value. Prof. 

 MacGregor has shown in a note to my former ^aper, in which 

 I drew attention to a regular progression similar to the above 

 for series of solutions of potassium chloride and potassium sul- 

 phate, which were for the most part of moderate concentration, 

 that the regular progression observed may be due to two sources 

 of error. ' The second source, namely, the impossibility of draw- 

 ing with perfect accuracy the dilution-ionic-concentration 

 curves, has, I think, been considerably reduced, for in place of 

 drawing these curves, which for moderate concentration curve 

 quite rapidly, concentration-specific-conductivity curves were 

 employed, these curves having but slight curvature and being 

 thus easily interpolated. The other source of error, namely, 

 the using for the value of the ionization coefficient the quotient 

 of the specific molecular conductivity by the specific molecular 

 conductivity at infinite dilution, still remains. In the above 

 fourth series where the first two mixtures are of moderate con- 

 centration and the other four may be called dilute solutions, this 

 regular progression has disappeared and the differences are all 

 negative, except in the first mixture. The disappearance of 

 the progression is consistent with its being due to the above 

 sources of error ; for in dilute solutions they both produce a 

 very small effect upon the result. Two reasons may be assigned 

 for the signs being all negative in the fourth series: (1) The 

 use of the above values of the specific molecular conductivity at 

 infinite dilution ; for if either of the values used should not be 

 correct then it would clearly produce an error of the same sign 



1 Loc. cit. 



