﻿Molecular Thermodynamics. 633 



on the " second " solute. We thus obtain a simple first- 

 approximation law for the solubility lowering of a slightly 

 soluble perfect solute produced by the addition of other 

 solutes. 



If the " second " solute be an electrolyte G M will not 

 vanish, but since at the very moderate concentrations to 

 which (97) would apply, G will probably be practically 

 entirely due to the electrostatic forces between the ions, 

 G, may be expected to vanish if (s) is a non-electrolyte. 

 Thus (97) may be expected to apply to the solubility lowering 

 of slightly soluble non-electrolytes — the more inert gases, 

 for instance — produced by the addition of small quantities 

 of salts, providing the complete ionization theory is accepted. 

 (97) therefore constitutes a prediction, compatible with 

 perfect behaviour on the part of the non-electrolyte of the 

 empirical law of Setche?iow, which latter should, however, be 

 re-written in terms of solvent-weight (instead of solution- 

 volume) concentrations. 



For binary electrolytes we see that 1000C (roughly) 

 equals the equivalent or twice the equivalent concentration 

 according as the ions are bivalent or univalent respectively, 

 iind that 



2t 1 f =(t c -tt a + 2t s ), (98) 



where t c is peculiar to the cation, t a to the anion, and t s to 

 the slightly soluble non-electrolyte. 



The data tabulated by Rothmund seem on investigation 

 to be of somewhat doubtful accuracy, and the concentrations 

 examined too high. Moreover, they are expressed (as per- 

 centages) in terms of Setchenow's law, that is, in terms of 

 solution-volume concentrations, and the recalculation is thus 

 necessary. 



They serve at once, however, to show that our prediction is 

 quantitatively quite of the right order of magnitude to suffice. 

 In particular, the smallness of the few negative values should 

 be observed. 



Among the more reliable figures there is also some evidence 

 of the parallelisms which the additive form of (98) would 

 predict for the recalculated values, while the greater degree 

 of solvation clearly assigned to the ions of higher valency, 

 (Mg ++ , S0 4 ~~, 00 3 ""), is quite what we should expect. 

 Closer examination and more accurate measurement will be 

 necessary to determine how far and for what range of con- 

 centration G s can be neglected here. 



Finally, the single case (as yet) of (probably) accurately 

 known "true" general terms may be introduced. 



