220 SCIENCE PROGRESS 



it is encumbered — a view that receives support from the hygro- 

 scopic character of the chloride as compared with those of the 

 other alkali metals. 



These general views have received an important extension 

 by the introduction of quantitative considerations. Bousfield, in 

 two recent papers (Proc. Roy. Soc, February 10, 1905, compare 

 Zcit. phys. Chem. 1905, 53, 257; Phil. Trans. 1906, 206, 101), has 

 revived the theory (which had generally been discarded on 

 account of the anomalies quoted above) that the velocity with 

 which an ion moves through a solution is governed by Stokes 

 Law in reference to the motion of a sphere in a viscous medium. 

 According to this view the mobility of an ion is inversely pro- 

 portional to its radius, and by assuming a suitable value for the 

 density of the combined water it is possible to calculate the 

 number of molecules associated with each equivalent of the ion. 



As a further result of the application of Stokes' Law it was 

 found possible to obtain evidence of a shrinkage of the ionic 

 radius as the concentration of the solution was increased. In 

 the most dilute solutions the ionic mobility is probably, as 

 Arrhenius assumed, a fixed quantity which does not vary with 

 the concentration, and the coefficients of ionisation can there- 

 fore be calculated accurately from the molecular conductivities. 

 In the case of a few weak acids and bases the simple mass-law 



a 2 k 



equation of Ostwald =— can be used to express the 



n 1 — a v 



course of the dissociation, but in most cases it is necessary to 



a," k 



use a variable index = — as proposed by Rudolphi. It 



is remarkable that in a large range of compounds the index n is 

 very near to the exact figure n = § proposed by van't Hoff, and 

 Bousfield has shown {Phil. Trans. 206, pp. 155-157) that in the 

 case of sodium and potassium chlorides the index approximates 

 more and more closely to this figure as the dilution becomes 

 more extreme, and the risks of variations of ionic mobility are 

 eliminated. The assumption appears, therefore, to be justified 

 that this equation with index f represents correctly the course 

 of the dissociation throughout the lower ranges of concentration 

 and that the deviations observed at moderate dilutions are due 

 to changes of ionic mobility as a result of alterations in the 

 viscosity of the medium, or in the dimensions of the ion. The 

 former quantity can be measured without difficulty, and when 



