316 Sir J. Conroy. Refractive Indices and Densities of 



in the corresponding solutions of sodium chloride ; hence if the chlo- 

 rides of the two metals had, when present in solution in molecular pro- 

 portions, equal powers of retarding the velocity of light, the solution 

 of the potassium salt might be expected to have a lower refractive index 

 than that of the sodium salt, as the unit volume contains, in addition 

 to the salt, less water. 



At present we are not in a position to distinguish between the 

 retardation due to the substance in solution and that due to the solvent. 

 If, as a first attempt, we assume (which is, of course, improbable) that 

 the presence of the salt merely causes the water to occupy a greater 

 volume without altering any of its properties other than those which 

 depend on its density, we can calculate approximately the refractive 

 index of the water in the solution. 



When two gases are mixed and no mutual action is known to occur, 

 we regard each gas as unchanged except that its density is reduced by 

 the admixture. If a mixture of liquids or the solution of a salt, where 

 no mutual action is known to occur, be similarly regarded, we may 

 consider each of the liquids or the solvent as changed only in respect of 

 its density. If again we attribute the change of refractive index with 

 temperature solely to the change in the density of the liquid, we may 

 make the hypothesis that the effect on the refractive index of water of 

 a change of density is the same when it is expanded by admixture and 

 when it is expanded by rise of temperature. No doubt such a hypo- 

 thesis, resting upon two hypotheses each of which is improbable, has 

 itself a very small probability, but I have thought it worth while to 

 reckon what the refractive indices of the water in the various solutions 

 would be on this hypothesis. The difference between this value and 

 the observed value of the index of the solution furnishes more probable 

 value for the influence which the salt may be supposed to exert on the 

 velocity of light than that obtained by subtracting the index of water 

 from the index of the solution at the same temperature. 



From Landolt and Bornstein's tables the temperatures were ascer- 

 tained at which water has the same density as that contained in the 

 various solutions, and then from the same tables the refractive index of 

 water for these temperatures and densities. 



The values for the differences between the observed indices of the 

 various solutions and the index of water at 18 are given in the second 

 and third columns of Table IV, and the differences between the 

 observed indices and the values for the indices of water, calculated on 

 the above assumption, in the fourth and fifth columns. 



The table shows that the differences between the refractive indices 

 of the solutions, and that of water at 18 increase with the molecular- 

 weights of the salts in solution, except in the case of potassium 

 chloride, but if the differences are calculated on the assumption that 

 the refractive index of the water in the solutions is less than that of 



