Chemistry and Physics. 75 



Armstrong discusses the apparently anomalous result that water 

 in the liquid state is necessary to the reaction just described. 

 He regards the interaction in this case as only another instance 

 showing the general fact that interactions which are commonly 

 assumed to occur between two substances, are possible more fre- 

 quently than not, only in presence of a third substance (which he 

 calls a catalyst). The case in question is parallel to that of the 

 oxidation of sulphur dioxide investigated by Dixon. " In gas- 

 eous mixtures chemical change appears so take place only when a 

 comparatively high electro-motive force, or its equivalent, is 

 employed ; one sufficient to produce disruptive discharge being 

 usually required. Regarding the interaction as a case of electro- 

 lysis, a gaseous mixture of HC1, O, and OH 2 therefore might be 

 expected to prove insensitive to light. But an aqueous solution 

 of hydrogen chloride is one of the best of liquid conductors and 

 it is easy, therefore, to understand that a relatively small electro- 

 motive force should suffice to electrolyze a liquid system of the 

 same three elements." — J. Chern. Soc, li, 806-807. G. f. b. 



3. On the Concentration of Solutions by Gravity. — Experi- 

 ment teaches that a homogeneous solution left to itself at con- 

 stant temperature, preserves sensibly its homogeneity. Gout and 

 Chaperon have examined this question mathematically in order 

 to see how far this result, taking gravity into the account, is in ac- 

 cordance with thermodynamic principles. And they find that under 

 these conditions, the permanent state of a solution is not one of abso- 

 lute homogeneity but that the density of the liquid increases from 

 the surface downward according to a determinate law ; so that in 

 time the primitive homogeneity of a solution will be destroyed 

 by gravity, and a new state of equilibrium will be established 

 within it. To show that the principle of Carnot is in contradic- 

 tion with the hypothesis of absolute homogeneity in the case of a 

 heavy solution in the permanent state, the authors suppose a per- 

 fectly homogeneous solution placed in a vessel of height H. Let a 

 very small portion of this liquid of volume V, at its upper surface, be 

 supposed temporarily isolated from the remainder of the solution. 

 If the weight doo of the solvent pass by distillation from this 

 isolated portion to the rest of the liquid at constant temperature, 

 evidently no work will be expended. Suppose now that, since the 

 weight of the solvent doo has gone from this isolated portion to 

 the rest of the liquid, the density of this portion increases in con- 

 sequence by an amount dD, its volume will diminish by dV ; and 

 now if by reason of this increased density, this portion sinks to 

 the bottom of the vessel, it will do a positive amount of work 

 equal to H (V—dV) dD. If after this we suppose the homogen- 

 eity of the solution to be re-established by diffusion the cycle will 

 be closed, the total work done in the cycle will be zero and hence 

 dD must be zero. It is not possible therefore for a solution to 

 be perfectly homogeneous in the permanent state unless its den- 

 sity does not vary for an infinitely small variation of the concen- 

 tration. But if, on the contrary, the solution is not homogeneous, 



