GOODWIN AND HASKELL. — CONDUCTIVITY OF SOLUTIONS. 407 



althougli some measurements and computations have also been made at 

 greater dilutions. The reasons for this selection are: first, that the 

 resistance measurements, owing to a sharper minimum, are considerably 

 more accurate (the average deviation of the separate readings from the 

 mean being less than 0.1 per cent) than in the more dilute solution; 

 and secondly, that the computed equivalent conductivities differ when 

 tlie initial concentration chosen is varied from less than 0.03 to 0.06 

 millinormal, but that they no longer do so to an important extent when 

 it is varied from 0.05 to 0.01 millinormal, showing that in the 0.0.3 

 millinormal solution the maximum influence of the water has not yet 

 been produced. 



It will be seen from the data given in the next section that the water 

 used in different series of experiments had a conductivity varying from 

 0.8 to 1.6 X 10"'', and consequently, the observed equivalent con- 

 ductivity at the higher dilutions also exhibit considerable variations. 

 In fact, the latter exist even where ditierent samples of water with the 

 same conductivity were used, showing that the character of the im- 

 purities varies and exerts a specific influence. Nevertheless, the cor- 

 rected values of the equivalent conductivity are substantially identical — 

 a fact which furnishes strong evidence of the correctness of the prin- 

 ciple underlying the computations. 



The above methods only serve to extend the series of reliable values of 

 the equivalent conductivity of acids from about 0.002 normal down to 

 0.0002 or 0.0001 normal. To what extent the values at the latter con- 

 centration differ from those at infinite dilution can only be estimated. 

 Probably the best basis for this estimate is the assumption of Kohl- 

 rausch which is most briefly expressed by the relation : 



(A^ — A400Q) for acids _ (A^ — A4000) for salts _ 0.9* 

 (A4000 — Aioo) for acids (A4000 — Aioo) for salts ~ 3.0 



This method of extrapolation has been applied to our results. Since 

 according to our measurements with hydrochloric acid described in 

 Sections V. and VI. A4000 — Aioo = 377.6 — 369.3 = 8.3, the so-deter- 

 mined value of (A^— A4000) ^^^ acids is only 2.5 units (or 0.7 per cent). 



V. The Conductivity Data. 



The experimental data and the values calculated therefrom will now 

 be presented. 



The temperature was in all cases 18.00° ± 0.02°. The standard 



* Kohlraiisch, Wied. Ann., 66, 794 (1898). 



