586 



SCIENCE. 



IN. S. Vol. XX. No. 514. 



in several eases have been shown to accord 

 fairly well with these two hypotheses. Yet 

 their simultaneous validity is quite incon- 

 sistent with the principle in regard to the 

 ionization in mixtures. In fact, Avhen con- 

 sidered in the light of this principle, the 

 existing data lead to the conclusion that 

 the former hypothesis is not even approxi- 

 mately true, and that the latter one, at any 

 rate in cases where the ionization is far 

 from complete, is affected by a considerable 

 error. One example may be cited : when 

 thallium chloride and bromate, each of 

 which alone has a solubility of about 1/40 

 normal in water at 40°, are simultaneously 

 present as solid phases, the solubility of 

 each is reduced by the other to an extent 

 which shows that the concentration of the 

 un-ionized molecules is diminished bj^ about 

 15 per cent, and that the product of the 

 ion-concentrations is increased by about 5 

 per cent. This case is a typical one; but 

 what the quantitative law of the influence 

 in question is, can only be determined by a 

 further study of the phenomenon. In the 

 case of tri-ionic salts, the ion-concentration 

 product is even approximately constant, 

 only when the square— not when the first 

 power— of the concentration of the univa- 

 lent ion is employed. This has been shown 

 by experiments with lead iodide in the 

 presence of potassium iodide, with lead 

 chloride in that of other chlorides and 

 with calcium hydroxide in that of ammo- 

 nium chloride. 



I will close by calling your attention to 

 a remarkable principle in regard to the 

 properties of salt solutions, of a character 

 quite distinct from those thus far consid- 

 ered. That many properties of dilute salt 

 solutions can be expressed as the sum of 

 values assigned once for all to the con- 

 stituent radicals or ions was long ago 

 recognized, and has often been cited as a 

 corollary from the ionic theory. That this 

 additivity of properties persists up to fair- 



ly high concentrations is a fact, however, 

 that has received scant consideration, owing 

 to its apparent lack of relationship to that 

 theory. This fact is shown strikingly in 

 the case of certain highly specific optical 

 properties which are ordinarily found to 

 be dependent in a high degree on molecular 

 structure. Thus, the experimental data 

 fully warrant the statement of the prin- 

 ciple that the optical activity and the 

 color of salts in solution, when referred 

 to equivalent quantities, are independent 

 of the concentration and therefore of the 

 degree of ionization of the salts and are 

 additive with respect to the properties of 

 the constituent ions even up to concentra- 

 tions where a large proportion of the salt 

 is in the un-ionized state. Abundant data 

 might be cited in support of this principle, 

 especially with reference to optical activity. 

 But I can only illustrate the character of 

 the evidence by presenting a few of the 

 results obtained by Walden with the salts 

 of a-brom camphor sulphonic acid. In 

 1/30 normal solution he found the follow- 

 ing values of the molal rotary power: 



Lithium salt 275 



Sodium salt 272 



Potassium salt 273 



Thallium salt 273 



Acid itself 273 



Beryllium salt 274 



Zinc salt 272 



Barium salt 2,72 



The values are seen to be substantially 

 identical, although the conductivity shows 

 the acid to have an un-ionized fraction of 

 7 per cent., the salts of the univalent 

 metals one of 16 per cent., and those of the 

 bivalent metals one of 30 per cent., and 

 although the un-ionized molecules present 

 contain in some cases the elements hydro- 

 gen, lithium and beryllium of very small 

 atomic weights, and in two others the ele- 

 ments thallium and barium of large atomic 

 weights. 



