156 
quantity of ice. There is, therefore, probably sufficient cause 
for the descent of a glacier in its own weight, though certain 
objections may be urged against this theory. 
Whenever ice is crushed and the fragments are forced 
together into a new shape, they reunite, forming again a solid 
block ; or, if two pieces of melting ice be only gently pressed 
together, they gradually freeze together into one piece. The 
explanation of this phenomenon of regelation is probably also 
to be sought in the effect of pressure on the melting point- 
When the pressure is greatest, then a small part of the ice 
will melt and the rest will be cooled ; the water escaping out 
of the way forms a thin film round the point of contact, which 
is gradually frozen by contact with the cooled ice, and though 
the pressure be very sKght, yet, if time be allowed, the effect 
will become sensible. 
An interesting experiment of Helmholtz shows how small 
a difference in pressure is sufficient to decide whether ice shall 
melt or water shall freeze, when they are in contact at 0° 0. 
He surrounded a flask containing water, but no air, with 
melting ice, and found that the water in the flask gradually 
froze. The freezing-point of the water in the flask was higher 
by 0'0075° C, in consequence of the defect of pressure, and 
the melting ice, by which it was surrounded, was colder by 
this amount than the temperature at which the water in the 
flask passed into ice, and therefore acted like a freezing 
mixture to it. The earlier regelation theory of glacier motion 
supposed that the ice was crushed against obstacles. We 
have already seen that a column of ice must be 748 feet high, 
in order to crush at its base by its own weight, and this is a 
depth greater than that of our present glaciers. 'No doubt, 
the pressure may be much greater at particular points than 
that due merely to the depth of ice, but the motion which 
takes place would hardly be intelligible, if it were not for the 
slow melting which goes on under pressure. 
