84 
PROFESSOR R. THRELFALL AND MR. J. H. D. BREARLEY 
Table IV. — Comparison of Magnetic Moments of Discs and Cylinders. 
1 
Disc. 
Cylinder. 
Mass. 
"0052 gram ! 
•0045 gram 
Diameter. 
5 centim. 
13 centim. 
Length. 
. , 
•46 „ 
Time of complete vibration . 
•333 second 
•385 second 
Moment of inertia. 
•00008 
•000088 
Intensity of field. 
•26 
•26 
Magnetic moment. 
•10957 
•090137 
Intensity of magnetization 
21-07 
20-03 
Condition (3) has been already treated. 
(4.) The optical magnification should he a maximum. No part of our work has 
given us more trouble than the production of good scale images, and, in the end, in 
no part have we had more success. We began by using the telescope and scale set 
up in the usual manner, and, as usual, were troubled by want of illumination over 
some parts of the scale, and excess of illumination over others. The sources of light 
were also sources of heat, and caused the galvanometer to be unsteady, though it was 
w 7 ell boxed in. We will not describe the many ineffectual experiments w T e made. 
Finally, the following plan was hit upon. A minute point of light (a kerosene lamp 
turned down very low 7 ) was placed edge-on about 2'9 metres away from the galvano¬ 
meter. An image of this source was focussed on the centre of the galvanometer 
mirror by a large lens, 19'5 centims. in diameter, placed at a distance of 55 centims. 
from the mirror. Directly behind the lens, and between it and the mirror, was 
placed a scale, which, in our experiments, and after many trials, was made by coating 
a slip of glass with a film of Canada balsam stained with “nigrosine,” and cutting 
divisions on this film on the dividing engine. The divisions were transparent, the 
general field being dark, and carefully kept so by placing the engraved scale in a 
proper recess cut in a black wooden screen. The advantage of the Canada balsam 
film is not obvious in this case, but the scale had been used in other experiments 
where double images gave trouble, and then the Canada balsam was of the greatest 
advantage in helping to prevent their formation, much in the same manner as 
M. Cornu’s varnish cures photographic plates of giving a halo round the image of a 
bright star (Mourn, de Phys.,’ vol. 9, p. 275). 
The film was, of course, turned towards the mirror, and the light fell upon it 
normally. It will be understood that by this w 7 e mean that the scale was a tangent 
to a circle described about the centre of the mirror. The object was, of course, to let 
the light pass through each division of the scale normally ; and, since the scale was 
only about 10 centims. long, this condition was sufficiently fulfilled, and the scale did 
not require to be bent. Real images of the scale divisions were formed at about 
