804 REPORT— 1898. 



assumption that no double salt exists in solution, of the conductivity and sp. 

 gravity of mixtures of KL.SO, and GuSO, solutions (Archibald), and of the con- 

 ductivity of mixtures of K.SO,, and MfrSO, solutions (McKay) have been found 

 not to agree with observation, except at much lower concentrations than the 

 above, a fact which is probabl}' connected with the formation of double salts by 

 these substances. The limit of experimental error in the above observations was 

 about 0*25, 01, and 0-005 per cent, for conductivity, surface tension, and sp. 

 gravity respectively. l''or accounts of the experiments see current volumes of 

 ' Trans. Koy. Soc. Can. and Trans., N.S. Inst. Sci.' 



Scbrader, in his ' Elektrolyse von Gemischen' (Berlin, 1897), has tried to 

 determine the ionisation coefficients in mixtures of electrolytes with a common 

 ion by means of electrolytic measurements, using expressions for the coefficients 

 in terms of the conductivity of the mixture, the amounts of the distinctive ions 

 transferred by the current and the Ilittorl's transference numbers for these ions. 

 He requires, however, to make a necessarily doubtful assumption as to the relation 

 between tlie transference numbers in the mixtures and in simple solutions ; and 

 some of the quantities involved in his expression for the coefficients are incapable 

 of exact measurement. We cannot, of course, test his values by calculating the 

 conductivity of his mixtures. But if N,, N., are their concentrations with respect 

 to the two electrolytes, and V,, V., the dilutions of their isohydric constituents, 

 his coefficients (a,, Oo) should satisfy the equations: aJYi = aJY„, and NjVj + 

 N2V3 = 1 (see last year's paper). Inserting his values of the a's and N's, we may 

 then find the V's and the corresponding a/ V's. The a/V's corresponding to the 

 V's thus found may also be obtained with great accuracy from Kohlrausch's and 

 Archibald's observations of conductivity on the assumption that no double molecules 

 are formed in solution, and that both the II's of HoSOj act as cations. When 

 Schrader's results are thus tested, his a/V's are found to be in error in the case of 

 solutions containing KI and KCl bv from 1 to 10 per cent., and in the case of 

 solutions containing IT.SO^ and C'uSO, by from 3'6 to 60 per cent. In the former 

 case the discrepancy is not greater than might be expected from the defects of the 

 method. In the latter it is ; and tliis result is interesting, as it is difficult to 

 account for so great a discrepancy except by the assumption of the existence of 

 molecules of the acid sulphate in solution. 



•7. The Carhon-Consuming Cell of Jacques. By S. Skinner. 



The cell consists of an iron vessel containing fused caustic soda in which a 

 carbon rod is placed, the carbon forming the positive element when the cell is 

 sufficiently hot. At lower temperatures the ii'on is positive. The cell polarises 

 rapidly, and, following the method of Jacques, the polarisation may be removed by 

 blowing air into the fused soda. In the experiments here described, in the place 

 of forcing in air, sodium peroxide is added to the fused soda. The electrolyte then 

 consists of caustic soda, sodium ferrate, and sodium pero.xide. This form of cell 

 shows little polarisation. To show that the oxidising substance depolarises at the 

 surface of the iron, experiments were made in a crucible nearly divided into two 

 parts by an iron plate reaching almost to the bottom of the vessel. The carbon 

 was placed in the liquid caustic soda on one side, and the sodium peroxide added to 

 that on the other. It was found that there was an immediate increase in electro- 

 motive force on adding the peroxide, thus demonstrating that the depolarising 

 action of the oxidiser was at the iron surface. Another experiment was made in an 

 iron U-tube, the U being loosely plugged with iron wire. The carbon rod was 

 immersed in the soda on one side, and the pero.xide was added in the other limb. 

 The same effect was observed. 



Experiments were also made to find the rate of consumption of the carbon, and 

 hence to deduce the electro-chemical equivalent. These measurements were 

 complicated by the disintegration of the carbon in the molten liquid even when no 

 current was passing. It is hoped by using more dense and pure carbon to prevent 

 local action and disintegration, and therefore to obtain an experimental value for 

 the electro-chemical equivalent of carbon. 



