VALUE OF THE BRITISH ASSOCIATION UNIT OF RESISTANCE. 265 
current the change that takes place is over in a time which is small compared with 
the time of swing % 
Now our reversals were made with an ordinary rocking commutator dipping into 
mercury cups, and only occupied a small fraction of a second. We therefore made a 
series of experiments to see if we could find the effect produced by holding the 
primary circuit open for definite small periods. 
Onweversing in the usual rapid manner we obtained a throw which varied between 
119*5 and 119*3 millims. 
When our primary circuit was held open for one second the throws were 118*9, 
119*3, 119*3, and 119*1, while if the contact was broken for two seconds the effect on 
the throw was marked. Thus if by chance in any experiment the primary circuit was 
open for as much as a second, so that the battery current took something over a second 
to get steady, it would only produce an effect of about 1 in 1000 in the result; we feel 
quite certain that in no case the period of break occupied more than a small fraction, 
from a quarter to one-sixth, of a second. 
Again, it might happen that the somewhat powerful induced current passing through 
the galvanometer coils might alter—temporarily or permanently—the magnetic moment 
of”our needle. 
A permanent alteration would of course be indicated by variations both in the throw 
produced by the induction current and in the time of swing. In the actual experi¬ 
ments after the induction current had been passed in one direction through the coils 
of the galvanometer, and the throw observed, it was passed in the other direction 
to stop the vibrations of the needle, and it was possible that each current might have 
produced real permanent changes in the magnetic moment, but of exactly equal 
amount, so that we had observed no appreciable alterations in the time of swing 
which we could assign to this cause. 
We therefore took a series of measurements of throws in which the current was 
only allowed to pass in one direction through the coils, the vibrations of the needle 
being stopped by means of an external damper ; the times of swing also were observed 
at intervals. 
The table below gives the result. 
ime of swing. 
Throw. 
115*8 
23*383 
116*0 
116*0 
115*7 
116*0 
115*9 
23*389 
Mncoci.xxxrn. 2 m 
