THE SPECIFIC INDUCTIVE CAPACITY OE A DIELECTRIC. 
5 
whole arrangement is fastened down to an ebonite board. As the current passes 
intermittently through the coils L M of the electromagnet, the vibrating piece 
strikes alternately against the parts G and T ; when it strikes against G the opposite 
plates of the condenser are connected with the two poles of the battery ; when it 
strikes against T the condenser is discharged (see fig. 1).” 
To be able to allow for conduction or absorption in the dielectric by means of 
observations with forks of different rates, we must know the comparative times 
during which the vibrating piece P is in contact with G, in contact with T, and 
travelling from one to the other. I therefore fixed a stiff arm to the end of P and 
took tracings with it on smoked paper, moved across while the interrupter was 
vibrating. These tracings showed that when the interrupter is going steadily (this 
is easily known by the sound) the vibrator is in contact with G and T for equal times, 
and the time occupied in passing from one to the other is negligible compared to the 
time of contact. 
The battery used was six Leclanche cells, and the variable resistances were taken 
from a box by Elliott. Tuning forks were used, making 99, 64, and 49 complete 
vibrations per second. The forks were worked by a current from the storage cells 
in the Laboratory. 
The coefiEicients of expansion with temperature 'of lead, ebonite, and glass are 
•000029, '000077, and '000009 respectively. That of mica does not seem to have 
been determined; but it is probably less than that of lead. So that the combined 
effects of expansion with rise of temperature do not affect the capacity before the 
fifth decimal place; and, as the results do not profess to be correct to more than four 
decimal places, the expansion of the materials may be neglected. 
Some observations were also taken with ebonite by an electrostatic method based 
upon that described in Maxwell’s ‘‘Electricity and Magnetism,” §229 (second 
edition); the only difference being that there was no guard ring. The condenser was 
compared with a variable condenser by dividing a charge between them, separating 
them, and connecting opposite poles together and to the electrometer; then, if they 
were unequal, the difference deflected the electrometer. For this method a special 
key was required to make the connections quickly and to keep the electrometer to 
earth until everything except the charge to be measured had been discharged. If the 
electrometer were not kept to earth it would be deflected in spite of ordinary screens 
by induction inside the key on account of the high potentials required. 
When neither of the condensers being compared has any absorption, there is no 
difficulty with this method, and any ordinary cj^uadrant electrometer may be used. 
But when this is not the case the residual discharge will besrin to come out as soon as 
the condensers are connected, a'nd this makes the needle always go ultimately to the 
same side. The only way to get a balance is to make the variable condenser too 
large, and gradually diminish it until the initial motion of the needle due to it 
disappears. This requires a needle with a short time of swing, because with a slow 
