ME. HOPKINS ON THE THEOET OP THE MOTION OP GLACIEES. 
713 
taneously takes place after the tangential dislocation is a motion of rotation about G, 
the centre of gravity of the element, by which that point is not affected ; whereas the 
motion with which we are here concerned, as that which alone can possibly produce any 
finite differential motion, is the motion of the centre of gravity (G) itself, arising from 
the general onward progress of the whole mass. Any real differential motion between 
two particles must, in all cases (whatever, in fact, may be the directions of tangential 
dislocation), take place in the actual direction of the motion of the particles. Thus, if 
the directions of tangential dislocation be not parallel and perpendicular to the axis of 
the glacier, as in the above case, the true differential motion of two particles must still 
take place in the actual direction of their motion. 
38. In the above case we have assumed the rupture to be complete and simultaneous 
on every side of the element, and also the absence of friction between contiguous 
elements. If it be otherwise, as it doubtless vnll be, the same reasoning will be appli- 
cable ; but the relief of the constraint of the mass, or of any element, at each disloca- 
tion will be only partial, and the consequent differential motion will be somewhat less. 
In such case the facility with which the central portions of the mass move faster than 
the other parts will be diminished. 
39. I am not here supposing that this tangential dislocation is the most probable 
mode by which the constraint of the glacial mass is commonly relieved under great 
pressure. The crushing of its elementary portions (art. 35) would appear, perhaps, a 
more likely modus 0 ]perandi ; but possibly both these processes may contribute to the 
actual dislocations (without finite fissures) by which the constraint of the mass is instan- 
taneously relieved. If the dislocations were entirely tangential, it is easily seen that 
two contiguous elements, like G and G' (fig. 6), would ultimately be separated, while 
each should preserve its physical identity. Consequently, identical particles forming at 
one time the continuous transverse linear element, MN, might subsequently be con- 
verted into an elongated loop which should be discontinuous in the marginal parts of 
the glacier, where the differential motion would be greatest, and continuous in the cen- 
tral portions, where that motion would be least. If, on the contrary, each element 
should be crushed in the dislocation, it would manifestly never regain its primitive form, 
but would become compressed or extended as if it were plastic or viscous (art. 4, &c.). 
In such case the original transverse element M N would be converted into a continuous 
loop. But, at all events, whether the real modus operandi consist of only one of the 
above processes, or of a combination of both, it is of the first importance for a just 
appreciation of the principle of regelation, that we should see distinctly how the conti- 
nuity of the glacial mass is broken and again restored by it, in contradistinction to the 
effect of real viscosity, which would prevent that continuity from being broken at all. 
In the latter case also there would be no particular structure, such as the crystalline 
structure, to be destroyed, and no necessity for the power of regelation to restore it. It 
is in the difference between the modus operandi when the mass is viscous, and that when 
