460 On Magneto-Electric Induction. 



They gave the following numbers, which represent the 

 number of divisions on a scale of equal parts, at which the 

 reflexion of the light could be seen when the mirror was in its 

 position of mean deflexion. 



a 2-1, b 2-6, c 5*5, d 6*8, e 8-9. 

 These numbers, being obtained only from a provisional ap- 

 paratus, must not be considered to represent the accuracy of 

 the process, but merely as showing that larger and more per- 

 fect apparatus is likely to give good results. 



With regard to the validity of this method of determining 

 the conducting-power of liquids a few words may be said. 

 1st. What will be the action of para- or dia-magnetic liquids 

 apart from conductivity ? Iron gives a greater turning effect 

 than copper when placed over a revolving magnet, owing to 

 its very great magnetizability and to its retaining magnetism 

 a perceptible time ; but whether liquids can acquire sufficient 

 magnetism and retain it long enough to vitiate the results it 

 is impossible to say; experiment will show. 2nd. Does the 

 torsion of the wire really represent the twisting effect on the 

 liquid at rest ? It would seem at first that it does not; but a 

 little consideration will show that no error in any way appreci- 

 able can be committed on the supposition that it does. The 

 action is this : the rotating magnets cause the liquid in the 

 vessel to rotate in the same direction; the friction between the 

 liquid and the vessel tends to make the vessel rotate too ; but 

 it is restrained by the torsion of the wire. As the inertia of 

 the liquid has to be overcome by a comparatively small force, 

 its velocity will increase very slowly ; and were there no 

 friction against the sides of the vessel it would, if only left 

 long enough, equal that of the magnet. But there is friction, 

 which increases with the velocity; and therefore as the velocity 

 increases the acceleration will diminish, and in time the velo- 

 city will be constant — that is, when the twisting force and the 

 friction become equal — -just as a light body when dropped in 

 air soon attains a constant velocity. This velocity depends on 

 the viscosity of the liquid ; and it might be well said that two 

 liquids having the same conductivity, but one viscous and the 

 other mobile, would attain very different constant velocities, 

 and, as the twisting force is directly proportional to the relative 

 motion between the liquid and the magnet, that different re- 

 sults would be attained. This is quite true ; but the question 

 turns upon the velocity which the liquid attains. If it is itself 

 comparable to the velocity of the magnet, then any changes 

 in it will affect the result but if it is nothing compared with 

 the magnet's velocity, then no error at all can be produced. 

 Now on looking back it will be seen that the rotation of the 



