102 REPORT— 1891. 



solution-volume is already not far from the specific volume of 

 pure aniline itself at the same temperature. 



To complete the comparison of the two curves, let me point 

 out that, just as we can from Fig. 3 calculate the distribution of 

 the alcohol between its liquid and vapour layers under given 

 conditions, so can we calculate from Fig. 2 the distribution of 

 the aniline between the aniline layer and the water layer under 

 given conditions. In the former case, if the total volume of a 

 tube containing n grammes of alcohol at, say, 230° be wx a*, 

 and if x be marked off between h and c on that line of tempera- 

 ture, then [x, h, and c standing for the volumes which can be 



x — c 

 read off on the horizontal base-line) n -jtz^ is the weight of the 



h — x 

 alcohol in the vapour layer, and n . j is its weight in the 



liquid layer, and the volumes of the two layers in cubic centi- 



x—c h—x 



metres are n .h . ~j — and n .c . j _ respectively, which are to- 

 gether equal to n . x. Just so also with the aniline and water 

 mixture (Fig. 2). li nxx be the total volume of the mixture 

 (both laj'ers together) containing n grammes of a,niline at, say, 

 140°, and if x be marked off as it was in the other case, then 



^—c . ... b—x 



n . T~, IS the weight of aniline in the water layer, and n . . _ 



is its weight in the aniline layer, and the total volumes of the 



1 7 ^~^ -, b—x . 



two layers are n . o . -, and n .c . j— — respectively, which are 



together equal to u . x. If the actual weights of aniline and 

 water in the mixture be given, the value of x can be calculated 

 with a fair approximation to accuracy in the manner used in 

 plotting the curve ; and so all the facts concerning the distri- 

 bution at all temperatures become known. 



Now, if it be remembered that this case of aniline and water 

 is not an isolated one, but typical of many cases experimented 

 on by Alexeeff, and if it be remembered also that there exists 

 no direct experimental evidence to show that the law which 

 governs these cases is not the general law regulating all simple 

 solutions, it must, I think, be granted that the facts do some- 

 what strongly support the hypothesis of a critical solution- 

 point, which I deduced in the first instance from the general 

 theory of solution. It may be summed up as follows : — 



(1.) In every system of solution which starts with a solid 

 and its simple solvent, there is a solution melting-point for the 

 solid which is lower than its true melting-point. Above this 

 temperature the system consists of two separate liquids, each 

 of which is a saturated solution. 



