MU. HOPKINS ON THE THEOET OE THE MOTION OE OLACIEES. 
737 
63. It appears, then, from these explanations, that there will be at every point (P) 
of every duct, two internal forces, the fluid pressure p acting normally, and the friction 
acting tangentially at that point on the sides of the tube. But assuming these small 
ducts to pervade the superficial as well as the inner portions of the mass, the fluid 
pressure p can never much exceed the atmospheric pressure, which, acting externally, 
will very nearly counterbalance it; and this will be true whether the motion of the 
water in the ducts be impeded by friction or not. Also the whole resolved part of fric- 
tion acting on the ducts, in the direction of the general motion of the mass, must be 
very small compared with the weight of the whole percolating fluid. Neither of these 
forces, therefore, can have any sensible influence in augmenting the onward motion of 
a glacier, compared with other forces which, as I shall shortly explain, tend to produce 
that effect. 
64. In the preceding explanations I have spoken of the retardation of the motion 
of a fluid in small ducts, as due to friction rather than capillary action. Capillarity, 
when it exhibits itself in the form under which it is more usually contemplated, acts in 
a manner totally different to that in which friction acts. We know that a column of 
water, or of other fluids, of a certain length may be supported in a vertical tube of 
sufficiently small bore, by the attractions between the particles of the fluid, and the 
attraction between the fluid and the tube. These attractions produce the capillarity of 
the tube. If, however, the tube be perfectly full, whether the fluid be at rest or in 
motion, its capillary action on the fluid will counteract itself, and will produce no effect 
sensible to observation. It is only when the tube is partially empty that a part of the 
capillary action on the contained fluid is uncompensated, and exhibits itself in the 
column of fluid which it is capable of supporting. These partially empty tubes cannot 
properly belong to a system of tubes, like those of a glacier, which serve as ducts through 
which the water is constantly running, and which must therefore be constantly full ; at 
all events, the number of tubes exhibiting the effect of capillarity as above described 
must probably be extremely rare in a system of ducts like those of a glacier. Supposing 
such tubes, however, to exist, we proceed to explain their effect in producing internal 
pressure. 
Let A E F B be a section of the capillary tube made by a plane 
through its axis, the tube being considered cylindrical and vertical. 
Let E D F be the level of the surface of the external fluid, which 
we may suppose to be water, and into which the lower end of the 
tube is placed. Also let A C B be a section, by the above plane, 
of the surface of the water maintained in the tube by capillary 
action. Our object is to ascertain whether this water will produce 
any pressure on the sides of the tube tending to thrust them outwards, and thus pro- 
duce an expansion of a solid mass in which any number of such capillary tubes might 
exist. 
For this purpose we must consider how the column of water in the tube is supported. 
Eig. 18. 
A 
B 
c 
Q 
M 
P 
JQ 
D 
I' 
