294 DEDUCTION FROM THE GASEOUS THEORY OF SOLUTION. 



of the two iugredients is nu accident wliicli does not characterize all 

 cases. 



IsTow imagine a 50 jjer cent mixtare of aniline and water sealed up 

 in a tube, shaken, and gradually heated. Let us assume that the tube 

 is only large enough to contain the mixture and allow of expansion by 

 heat, so that evaporation may be neglected as too small to matei-ially 

 compli(;ate the result. The course of events will be exactly what I 

 have already described with reference to the hypothetical A layer and 

 B layer. There will be formed a saturated solution of water in aniline, 

 which we may call tlie aniline layer, and a saturated solution of aniline 

 in water — the water layer. Given the temperature, the percentage 

 strength of each layer may be read oft' from the curve. As the tem- 

 perature rises, the two layers will effect exchanges in such a way that 

 the aniline layer will become poorer aiul the water layer richer in 

 aniline, and at about 167° the two layers will have attained equal 

 strength and become merged into one. Were we to start Avith the 

 aniline and water in any other proportions by weight, there would 

 still be formed the two saturated solutions, but their relative amounts 

 would be different, and one or other would be used up and disappear 

 at a lower temperature than 167°. To attain the maximum tempera- 

 ture of complete solution, you must start with the exact proportions 

 which correspond to that temperature. 



But it is possible to learn even more from Alexeefif's work than he 

 himself has made evident. Let me call yi»ur attention to the curve 

 shown in Fig. 2*, the data for which I liuvr* calculated in the following 

 manner. 



Oc.c. 2-5 5 7-5 /O /2S /5 



Fig. 2.— Vokinie of saturated aqueoiis solution containing 1 gram of aniline. 



From Alexeefif's percentage figures was deduced the weight of water 

 capable of dissolving, or being dissolved by 1 gram of aniline at 



* In order to save space, only the npper portion of the curve is here represented, 

 as it shows all that is essential to the argument. Of the twelve experimental poiuts 

 one ajipears to be somewhat misplaced; hut this doe.s not affect that part of the 

 curve shown in the figure. 



