MR. J. S. TOWNSEND ON MAGNETIZATION OF LIQUIDS. 
543 
a — iooj J„ (*) — 1? an d Jj { x ) — g 
when x is small, so that the above ratio reduces to : 
1 
- • T> t e ^ 
32tt10- 3 1 ’ 
and letting t — which is about the time the commutator allowed the secondary 
to be connected in the positive direction, say, with the galvanometer after the process 
of changing the current in the primary from 0 to i = ^ ( 1 — e was completed, 
we get a value for this ratio far less than unity e~ 50/3 ; in fact, the commutator 
might have been going twice as fast without the currents induced in the liquid 
having any effect on the value of k in the eighth decimal place. 
Results of the Experiments. 
In carrying out an experiment for the determination of k, it is necessary in all 
cases to syphon the liquid out of he solenoid immediately after m has been found, 
and test the balance again to see whether it is restored by replacing m. When this 
condition is fulfilled, it is evident that the mutual induction of the system of coils 
has not been altered while the experiment has been going on. 
It does not always happen that an exact balance can be got by the method of 
changing the number of turns in the primary of c, for if a certain adjustment gives a 
deflection of two divisions positive, a reduction of one turn in the primary of c 3 gives 
a deflection of three divisions negative, so that in this case the correct reduction can 
be taken as two-fifths of 65, which is sufficiently accurate, since all that is necessary 
in order to determine k in the 7th decimal place is to be able to detect an induction 
of 20 C.G.S. units. 
In all the following experiments the magnetising force was varied from 1 to 9 
C.G.S. units, and in no case was there any change in k. 
Ferric Chloride. 
The values of k were found for four solutions of ferric chloride, the quantities of 
the salt per cub. cemtim. being proportional to I, 2, 3, and 4. 
The values of 10 7 & were 
69, 148, 220, and 298. 
These numbers are not proportional to 1, 2, 3, and 4 ; but when we add 7'7 (the 
value of — 10 7 A: for the solvent) to each, we get values for k which are proportional 
