389 
on passing from the fluid to the solid crystalline state. After 
the first formation ofa solid shell, all the succeeding additions 
to its inner surface occurring by contraction from within out- 
wards, the tendency of the process of solidification must be to 
lessen the pressure on the nucleus. If the increasing density 
of the nucleus towards its centre be due to pressure, it must 
follow that its mean density will be diminished by the removal 
of pressure from its surface ; it will therefore tend to expand, 
and become more homogeneous. The moment of inertia of 
the earth will be thus directly increased ; and, supposing the 
shell and nucleus to move as one mass, which appears to re- 
sult from one of the following conclusions, as well as from the 
probable nature of the matter of which they are composed, a 
cause will exist for increasing the length of the day. But 
the changes referred to will also cause the surface of the nu- 
cleus to gradually become more oblate, and, consequently, 
each successive stratum of solidified matter added from it to 
the shell. The strata of the shell will therefore increase in 
oblateness from its outer to its inner surface, while those of 
the nucleus will still continue to decrease from its surface to 
its centre, although not always according to the same law. 
From this conclusion two others are deduced:—1. The ex- 
istence of great pressure and friction at the surface of con- 
tact of the solid and fluid; for otherwise, according to a result — 
obtained by Mr. Hopkins, the precession of the equinoxes 
would greatly differ from that which is observed.* 2. The 
moment of inertia corresponding to the earth’s axis of rotation 
would be increased independently of the cause already men- 
tioned, which would increase all the moments of inertia of the 
earth. This result had been already used by Mr. Hennessy, 
in the memoirs referred to, and also in a letter to Sir John 
Lubbock,{ to prove the continued stability of the earth’s axis 
* Phil. Trans. 1840, p. 207; also, Phil. Trans. 1851, vol. ii, p. 546. 
+ Proceedings of the Royal Society, February, 1852. 
