to determine the Density of the Earth. 363 
ticular, was a most comfortable room or rather suite of rooms. 
temperature. In each (the upper and the lower) was mounted 
an invariable brass pendulum, vibrating by means of a steel 
knife edge upon plates of agate, carried by a very firm iron stand. 
Close behind it, upon an independent stand, was a clock, carry- 
ing upon the bob of its pendulum an illuminated disk, of diam- 
eter nearly equal to the breadth of the tail of the invariable pen- 
dulnm ; and between the two pendulums was a chink or opening 
of two plates of metal, which admitted of adjustment, and was 
opened very nearly to the same breadth as the disk. To view 
these a telescope was fixed in a wall, and the observer was seated 
in another room. When the invariable pendulum and the clock 
pendulum pass the central points of vibration at the same instant, 
the invariable pendulum hides the illuminated disk as it passes 
the chink, and it is not seen at all. At other times it is seen in 
passing the chink. The observation, then, of this disappearance 
determines a coincidence with great precision. Suppose the next 
coincidence occurs after 400 seconds. Then the invariable pen- 
ulum (swinging more slowly), has lost exactly two swings upon 
the clock pendulum, or the proportion of its swings to those of 
the clock pendulum is 398:400. If an error of a second has 
been committed, the proportion is only altered to 397 : 399, which 
differs by an almost insignificant quantity. Thus the observa- 
tion, in itself extremely rude, gives results of very great accu- 
Tacy. As the proportion of invariable-pendulum-swings to clock- 
pendulum-swings is thus found, and as the clock-pendulum- 
swings in any required time are counted by the clock dial, the 
corresponding number of invariable-pendulum-swings is at once 
found. Corrections are then required for the expansion of the 
metal (depending on the thermometer-reading ), for the are of vi- 
wien, and for the buoyancy in air (depending on the barometer- 
ing), 
But when the corrected proportion of upper-invariable-pendu- 
lam-swings to upper-clock-pendulum-swings is found, and the 
Wire of this circuit, passing from a small battery through the 
auxiliary clock, then went throngh the upper galvanometer, then 
Passed down the shaft of the mine to the lower galvauometer, 
and then returned to the batiery. At each galyanometer there 
