266 
MESSRS. R. T. GLAZEBROOK AND E. B. SARGANT ON THE 
Second series.—Throws in opposite direction. 
Time of swing. 
Throw. 
115-5 
23-389 
115-6 
116-2 
115-8 
23-392 
The throws throughout are very nearly equal; the time of swing, it is true, has 
increased by -009", but this change cannot be due to an alteration in the magnetic 
moment of the needle, for if the current passed in the one direction had decreased the 
magnetic moment and so increased the time of swing, that passed in the other direction 
in the second series would have increased the magnetic moment and so decreased the 
time of swing. The alteration in T then we must rather ascribe to a variation in the 
value of the horizontal intensity. This has been allowed for in the experiments by 
the repeated observations which were made of T. As regards the temporary effect of 
the induced magnetism produced by the current while it lasts, we notice that its 
direction would at each instant be perpendicular to the plane of the coils, so that 
the force between the magnet thus formed by the current and the current forming it 
would be along the axis of the temporary magnet and have no tendency to cause it 
to move. In fact, the field of force produced by the coils being uniform, the actual 
force on the induced magnetism will be zero. 
A number of observations were also made to see if the time taken by the battery 
current to become steady after a reversal was appreciable. The galvanometer circuit 
being open the battery circuit was reversed, and then the galvanometer circuit closed 
at an interval of from ’25" to -3" after. 
If after this interval the battery current had still been varying an induction current 
would have shown itself in the galvanometer, but though the reversals were made 
in three separate experiments, 13, 7, and 10 times respectively, m time with the 
period of the galvanometer, no effect was produced. 
We are sure, therefore, that the battery current has attained its steady value in 
less than -3” after reversal, while our former experiments have shown us that if the 
period of change were as great as 1 " the error produced in the throw would be less 
than 1 in 1000. 
The direct effects of the connecting wires and commutators on the galvanometer 
during the experiments were carefully tested for but no result could be found. 
Copper wire insulated with gutta-percha was used for the connexions, and the wires 
in each circuit were carefully twisted up together. 
Most of the apparatus employed belongs to the Cavendish Laboratory, where the 
