On the Motion of a Sphere in a Viscous Fluid. 519 



lines, runs down to the right of the 3-2 (1'85) line, bending 

 towards the line. LoormVs lies between the 3-2 and 6—4 

 lines. A mean curve would be just to the right of the 3-2 

 line, and might readily run out at the 3-2 (1*85) intersection. 

 This would indicate single molecules in dilute solutions disso- 

 ciating into two ions, an early occurrence of doubling of 

 molecules, and steady increase in the extent of association as 

 dilution diminished, the double molecules formed dissociating 

 into 4, 3, or 2 ions, but not into more. Although the co- 

 efficients with which the curves are plotted are doubtful, the 

 curves are so nearly parallel to the axis of coefficients, that 

 even a considerable error in their determination would not 

 affect the above result. 



General Conclusions. 

 Although the observations on which the above discussion 

 is based are defective, and the particular conclusions drawn 

 are consequently tentative, I think it may be held with some 

 confidence: (1) that the curves of equivalent depression 

 against ionization-coefficient have positions, forms, and slopes 

 such as they might be expected to have on reasonable 

 assumptions as to mode of ionization and constitution in 

 solution, according to the Van 't Hoff-Arrhenius theory o£ 

 the depression of the freezing-point in solutions of electro- 

 lytes ; (2) that they are consistent with the depression-con- 

 stant having a common value of about 1*85 for all the 

 electrolytes examined, and that in the case of the electrolyte 

 for which we have the best data, its curve is not consistent 

 with a greater limit of error in this value than about O'Ol, 

 unless improbable assumptions are made with respect to the 

 constitution of the electrolyte in solution; and (3) that the 

 diagram enables us to reach in some cases conclusions of 

 considerable probability with respect to the constitution of 

 the electrolyte in solution and its mode of ionization. 



L. The Motion of a Sphere in a Viscous Fluid. 



By H. S. Allen, M.A., B.Sc. 



[Continued from p. 338.J 



[Plates I. & II.] 



III. 



8. Photographic Method of Determining Velocities. 



IT was only possible to apply the method already described 

 to cases in which the terminal velocity was small. In 

 order to extend the range of observation it was decided to 

 have recourse to a photographic method. Several ways of 



