310 ‘Royal Astronomical Society. 
material and size. Maskelyne’s rough computation, made with 
such knowledge as he had of the composition of the mountain, gave 
the mean density of the earth nine-fifths of that of the mountain, or 
from four to five times that of water; Hutton’s subsequent and 
laborious investigation made four and a half for the ratio; but 
Playfair’s examination of the strata of the mountain led to the 
inference that the result of the experiment could only be considered 
as placing the same ratio between four and six-tenths and four and 
nine-tenths. It is worth noting that Newton (Principia, book iii. 
prop. 10) had ventured a conjecture, which, as happens so frequently 
with him, turns out to be true. He thinks that the mean density 
of the earth is between five and six times that of water. 
The method first proposed by Michell, and adopted and exe- 
cuted by Cavendish, was wholly independent both of astronomical 
data and of the uncertainty of the material of comparison. A 
horizontal pendulum, suspended by a wire, the torsion of which 
caused it to make slight oscillations about a position of equilibrium, 
was substituted for the gravitation pendulum, or plumb-line ; while 
a massive leaden ball took the place of the mountain. If the 
torsion pendulum had been as secure of its position of equilibrium 
as that of gravitation, the experiment would have been almost as 
identical in its details with that of Maskelyne as it is in its 
principle. The time of an oscillation would first have led to the 
settlement of the amount of torsion force in any given position of 
the pendulum: it is well known that nothing but the time of 
oscillation is wanted to give the means of determining the quantity 
of restitutive force which acts on the pendulum at any degree of 
departure from equilibrium. The degree of departure caused by 
submitting the ball at the end of the torsion pendulum to the 
lateral influence of the large mass would then have given the means 
of calculating the attraction of that mass, just as the amount of 
deviation of the plumb-line gave that of the mountain in Mas- 
kelyne’s experiment. Small as may be the leaden mass compared 
with the mountain, it was so much better known as to make the 
risk of error materially less; to which must be added, that it was 
submitted to an instrument of very much greater delicacy than the 
ordinary plumb-line. ‘ 
The torsion pendulum is, in fact, possessed of such an extreme 
susceptibility, that every detail of the experiment requires adapta- 
tion, to an extent which would make a superficial inquirer wonder 
how it could be in any way compared with that of the plumb-line 
and mountain. There is no position of equilibrium—the instru- 
ment is never at rest. The effect of presenting the large leaden 
ball is not to draw the torsion-rod from one position into another, 
but to change its motion from an oscillation of one extent about 
one point of rest to one of another extent about another point of 
rest; and a careful and peculiar mode (into which I cannot here 
enter) must be practised of determining this point of rest, both 
before and after the pendulum is placed under the influence of the 
