Action of Dielectric Polarization. 387 
back half. The question is,.then, whether these vertical com- 
ponents do actually exert such an effect. 
In order to determine this, the upper glass plate was sur- 
mounted by a metallic case, connected to earth, which con- 
tained an extremely sensitive astatic pair of needles; the 
lower needle was about 0°6 centim. distant from the ebonite 
plate, its centre being in the prolongation of the axis of 
rotation of the disk, and its direction parallel to the dividing- 
line of the two half-rings; the length of the needle was 
about 4°8 centim., a little more than the inner diameter of the 
half-rings. The second needle was situated 21°5 centim. 
above the lower one. The deviations were read by means of a 
telescope and scale at a distance of 8 metres. All the neces- 
sary precautions were taken that the needle should not be 
affected by external statical electricity; and a special con- 
struction of the axis was adopted so that the deviations pro- 
duced by rotational magnetism should be as small as possible 
(2 to 3 divisions of the scale). Notwithstanding, upon rapid 
rotation of the disk the needle changed its position of rest 
continually, which made the observations much more difficult. 
The reason of these motions lay, as I satisfied myself, in 
currents of air, and more particularly in small vibrations to 
which the apparatus was exposed, in consequence of the 
defective arrangements of the Institute here. 
The experiments were arranged so that one observer sat 
at the telescope whilst an assistant revolved the disk, and 
another changed the commutator upon a signal from the 
observer. 
The observer intentionally remained ignorant of the direc- 
tion in which the commutator was moved until the end of the 
series of experiments; generally the commutator was changed 
eight times during each series of experiments. There was 
no possibility of accurately determining the magnitude of the 
deflection produced by reversal of the commutator, since in 
all cases it was very small and under the best conditions only 
amounted to 1°5 division of the scale, generally amounting 
only to a fraction of a division. The observer therefore con- 
fined his attention to determining the direction of the deflec- 
tion each time, and for this a certain amount of practice was 
necessary in consequence of the continual small motions of 
the needle. As the result of more than 1000 observations I 
have obtained so much practice that in the later experiments 
I have been able almost without mistake to determine every 
time the direction of deflection. 
From these experiments, which were varied in many ways, 
the result was obtained that the deflection always agreed with 
that given by Faraday’s theory. The change in the dielectric 
