connected with the Theory of the Universe. 30 
that, in the rubbing together of portions of matter built up 
in the manner above described, a small cluster of grains may 
be torn away from its place, and shifted to another neigh- 
bouring place. Two opposite disarrangements are thus 
produced. At the place from which the cluster has been 
removed, surrounding grains are pushed inwards to fill the 
gap. At the place to which the cluster has been removed, 
surrounding grains are pushed outwards to make room for it. 
Hach of these two opposite disarrangements produces 
dilatation, by rendering the packing abnormal; though 
Prof. Reynolds calls one of them a positive inequality, and uses 
language which conveys the impression that the intrusion 
of the cluster of grains produces an increase of compactness. 
By reasoning which is difficult to follow, the attempt is 
made to prove that these two dissimilar centres of disturbance 
will exhibit mutual attraction. They are regarded as con- 
stituting the two opposite kinds of electricity. 
Another kind of complex inequality is introduced to 
explain magnetism. Suppose a spherical cluster of grains 
to receive a forcible twist, large enough to make portions 
near its equator (where the motion is largest) change their 
external neighbours ; while portions near the poles have not 
moved far enough to change their neighbours. These latter 
portions of the sphere will tend to return to their original 
position, while portions which have passed a little beyond 
the position of unstable equilibrium will tend to move further 
away. Rotational stresses will thus exist between the 
spherical cluster and the surrounding grains, the stresses 
being opposite-ways round in different portions. The strains 
accompanying these stresses involve dilatation ; andattractions 
and repulsions will be exhibited, governed by the criterion 
that the tendency is for less space to be occupied. The 
reasoning is not worked out in detail ; but the brief state- 
ment is made that permanent magnetism is thus explained. 
The propagation of light is ascribed to transverse dis- 
placement of the granular medium by successive impacts ; 
and Roéntgen rays, to normal displacement, similarly pro- 
pagated. Calculations are given to show that, when proper 
values are assigned to the diameters, masses, velocities, and 
mean free paths of the grains, the diminution of the energy 
of a wave, to the fraction - of its original amount, will 
occupy some millions of years in the case of transverse 
waves, and less than a millionth part of a second for normal 
waves, the distance traversed by the normal waves in under- 
going this diminution being 2800 metres, 
D 2 
