GOODWIN AND HASKELL. — CONDUCTIVITY OF SOLUTIONS. 401 



being rejected. The acid solutions were made by adding to definite por- 

 tions of this water weighed portions of a standard acid of strengtli about 

 0.01- normal. The standard hydrochloric acid solution was obtained by 

 absorbing in water hydrogen chloride gas made from twice-precipitated 

 sodium cliloride and pure sulphuric acid. Its strength was determined 

 graviraetrically by precipitating with silver nitrate and weighing the 

 silver chloride in Gooch crucibles. The several determinations made 

 showed an average deviation from the mean value of less than 0.05 per 

 cent. A standard nitric acid solution approximately 0.01 normal was 

 made by diluting chemically pure acid through which air had been 

 drawn to remove any lower oxides of nitrogen dissolved in it. This 

 was standardized against barium hydroxide solution, whose strength 

 had been determined by two independent methods. Five 0.01 normal 

 solutions of potassium chloride were used to obtain the cell-constant; 

 these were made from a sample of the salt which had been twice precipi- 

 tated with pure hydrochloric acid and then crystallized from water, 

 dried, and ignited to incipient fusion. A 0.01 normal sodium chloride 

 solution made from salt prepared in the same way was used as a check. 



In all cases, for the sake of greater convenience and accuracy, the 

 content of the solution was determined by weighing, and not by vol- 

 ume measurement; but the values thus obtained of the number of milli- 

 equivalents of solute in 1000 grams of water were invariably (both in 

 the calculation of the cell-constant and of the specific and equivalent 

 conductivity of the acid solutions) considered to be numerically identical 

 with the concentration in milli-equivalents per liter of solution. That 

 no appreciable error (not even one as great as 0.05 per cent) is intro- 

 duced by this assumption will be readily seen, since none of the solu- 

 tions were stronger than 0.01 normal. 



III. The Conductivity Apparatus and Method of 

 Mea.sukement. 



Tlie conductance of the solutions was measured by the Wheatstone 

 slide-wire bridge arrangement in the form recommended by Kohlrausch, 

 using an alternating current produced by an induction coil and a tele- 

 phone as an indicator of zero potential. Tlie bridge consisted of a mar- 

 ble drum on which was Avound five meters of manganine wire. Five or 

 six settings Avere made for each solution, so as to eliminate as far as 

 possible any personal error in determining the exact point corresponding 

 to tlie minimum sound in the telephone. The average deviation of the 



VOL. XL. — 2G 



