352 Kohlrausch and Loomis—Influence of Temperature 
between brass plates, of which the lower pair could be united 
to the weight by means of a upsilon rest, and the upper pair 
2. in like manner to a brass rod screwed into a 
wooden pivot. The latter was so secured ina 
solid support fixed to the wall as to be capable 
of rotation. The torsion vibrations were com- 
For the determination of the temperature, three 
th t prepared, such that the quick- 
silver reservoirs are at different distances from 
the zero point. When the graduated tubes pro- 
ject from the apparatus from the division 0°, 
the quicksilver reservoir of one of the thermometers is at a 
height corresponding to the middle of the wire and close beside 
it. The other two quicksilver reservoirs are at each end of the 
wire, and arranged symmetrically around it. The two latter 
thermometers alone are represented in the figure. The readings 
were made by means of a telescope. 
The tin cylinder was tightly inclosed on all sides with a cov- 
ering of felt of about a centimeter in thickness. It stood upon 
a support attached to the wall. The weight and mirror were 
contained in a box closed in front with a glass slide. These 
different temperatures, but a simpler method, at first only em- 
ployed to test what results oie be expected, proved better 
from the beginning to the end of a series of observations, eX- 
cepting only a slight turning of the wooden pivot, when the 
amplitude of the vibrations becomes too small. Several days 
a 
