358 Mr. B. D. Steele. The Measurement of Ionic Velocities 



It is found that the agreement is, on the whole, tolerably good, 

 whence it is inferred that in torsion, the effect of local action dies out 

 more rapidly than in tension or compression. The only case of 



obvious divergence is with regard to the shear rcf>. This shear persists 

 inside, even at sections where no stress of this kind is applied to the 

 outside of the cylinder, but it continually diminishes as we recede from 

 the ends. 



In the exact solution, the cross-sections do not remain undistorted, 

 • the transverse displacement increasing more rapidly than the radius. 

 The distortion is small at sections where there is no external applied 

 stress, but is very obvious near the ends. 



Further, when the applied transverse shear varies discontinuously, 

 as in this case, the other stress becomes infinite at the points of dis- 

 continuity. This suggests why it is that abrupt changes in the section 

 of such a cylinder are dangerous. The projecting parts acting upon 

 the inner core will introduce a sudden change in the transverse shear. 

 It has been noticed that propeller shafts usually break at such points. 



" The Measurement of Ionic Velocities in Aqueous Solution, and 

 the Existence of Complex Ions." By B. D. Steele, B.Sc, 1851 

 Exhibition Scholar (Melbourne). Communicated by Pro- 

 fessor Eamsay, F.K.S. Eeceived May 10, — Eead June 6, 

 1901. 



(Abstract.) 



The method of measuring ionic volocities described by Masson has 

 been extended in such a manner that, by the present method, the use 

 of gelatin solution and of coloured indicators is not necessary. 



An aqueous solution of the salt to be measured is enclosed between 

 two partitions of gelatin which contain the indicator ions in solution, 

 the apparatus being always so arranged that the heavier solution lies 

 underneath the lighter. On the passage of the current the ions of the 

 measured solution move away from the jelly, followed at either end 

 by the indicator ions ; the boundary is quite visible in consequence of 

 the difference in refractive index of the two solutions. The velocity 

 of movement of the margins is measured by means of a cathetometer, 

 and the ratio of the margin velocities gives at once the ratio of the 

 ionic velocities. 



It is found that, for the production and maintenance of a good 

 refractive margin, a certain definite range of potential fall is required 

 for any given pair of solutions, and this range differs very much for 

 different boundaries — for example, the margin potassium acetate 



