PROPERTIES CONFERRED BY CONSTITUENTS 259 



gressively more dilute. If unchecked by any balancing or opposing 

 influence, this distillation will continue until the solution becomes 

 infinitely dilute, i. e., practically equivalent to distilled water. In 

 this case, therefore, as in the case of thin solid membranes, we can 

 only measure the attraction of the solution for water by measuring 

 the change in the condition of the solution requisite for its neutraliza- 

 tion. This we may accomplish by heating the solution and thus 

 increasing the mobility of the molecules which it contains, and so 

 increasing the number of collisions per second of water-molecules 

 with the supernatant layer of air. The temperature to which we 

 must raise the solution in order to equalize the rates of distillation 

 to and from the water and the solution is proportionate to the increase 

 in the collisions per second which is requisite to produce this equaliza- 

 tion, and this in turn must obviously be proportionate to the con- 

 centration of the dissolved substance. Thus if the dissolved sub- 

 stance constitutes one-tenth of the total molecules in the solution, 

 we must raise the temperature of the solution sufficiently to increase 

 the total collisions per second by one-tenth, in order to render the 

 rate of distillation equal to that of pure water at the lower tempera- 

 ture. If this rate of distillation is sufficient to cause ebullition, i. e., 

 to render the pressure of water-vapor equal to that of the atmosphere, 

 it is evident that the temperature required to attain it will be higher 

 in the case of the solution than in the case of pure water. Hence, 

 the Boiling-point of water is raised by dissolved substances, and that 

 in proportion to their molecular concentration. 



There is yet another way in which we may equalize the rates of 

 penetration of water from opposite sides of the membrane, and that 

 is by cooling the water, and thus reducing the mobility of its molecules 

 relatively to those of the solution. Now when a solution freezes, it 

 is not the dissolved substance that freezes, but the solvent, in this 

 case water. The dissolved substance, in fact, with certain intelligible 

 exceptions, crystallizes out and becomes mechanically separated from 

 the solvent when the latter freezes. In such a case the membrane 

 is furnished by the surface separating the crystals of ice from the 

 remainder of the solution. If, now, reverting to the diagram in 

 Fig. 9, the water in the left-hand compartment be sufficiently cooled, 

 relatively to the solution, water will pass over from the warm solution 

 into the cool chamber of pure water, because of the greater mobility 

 of the molecules in the warm solution. 1 Hence, in order to accom- 

 plish the withdrawal of water from the dissolved substance which 

 occurs at the freezing-point the water must be cooled to a temperature 



1 Ultimately, however, the greater mobility of the molecules in the solution will 

 fail to compensate for the progressively decreasing proportion of water molecules present 

 in the solution, and the water compartment would have to be further cooled in order 

 to continue withdrawal of water from the solution. This is the phenomenon of "under- 

 cooling" which freezing salt solutions display. The correct freezing-point is that at 

 which the first crystal of ice separates, and which is marked by a sudden slight rise of 

 temperature due to the disengagement of the latent heat of fusion ol the ice. 



