200 PROFESSOR FORBES ON THE VISCOUS THEORY OF GLACIER MOTION. 
point where the chasm opens. On the contrary, there is a fully greater probability 
that under that very spot the ice is compressed. If one cause of a crevasse be, as is 
universally acknowledged, a protuberance or inequality in the bed over which the ice 
is impelled, for the same reason that a beam, broken by means of weights, is in a 
state of longitudinal compression below, where its surface is concave, and of disten- 
sion above, where its surface is convex, the cracks in the glacier may be due solely 
to this last and partial cause. Superficial crevasses may consequently be occasioned 
where there is no general distension of the mass, either (1) by the shoving of the 
semi-rigid glacier as a whole, over a convex declivity, or (2) from an internal tur- 
gescence arising from hydrostatic pressure, resisted by the intense friction of the ante- 
rior or more advanced parts of the glacier, which, causing the line of least resistance 
to be upwards and forwards, forces the pasty mass to tumefy or increase in thickness, 
exactly as it has been seen in § 2, p. 153, that sluggish lava streams do in a similar 
case. But if the tumefaction be pushed beyond the limits of plasticity of the superior 
and more distended portions, they must burst and assume the crevassed forms ac- 
tually observed in the plastic models described in p. 144. Hence the existence of 
crevasses not only does not always result from a state of general distension in the 
glacier, but may arise from the precisely contrary condition of great internal compres- 
sion. This argument is well illustrated by the recent observations of M. Agassiz’s 
co-operators on the glacier of the Aar, whose observations I have elsewhere shown* 
to be incompatible with any other view than that of intense longitudinal compression 
in the mass generally, and yet the surface abounds in crevasses of the usual form 
and dimensions. 
The manner of formation of crevasses generally, including such as may betoken a 
real distending force acting on any part of a glacier throughout its thickness^ is not 
only a most curious question in itself, but suggests others which a correct theory of 
glacier motion can alone answer. If a crevasse once formed remain a fissure in the 
ice for ever after, why is the horizontal projection or ground plan of the crevasses of 
a canal-shaped glacier convex towards the origin of the glacier, and not protuberant 
in the direction of its motion, as the ascertained greater velocity of the centre would 
assign? Why are the crevasses for the most part vertical and not inclined forwards, 
or at least not notably so, on the same account ? Why, if the glacier be urged down- 
wards by a longitudinal force distending it, do not the crevasses continually widen 
in proportion as they are further from the origin? These questions seem incapable 
of a sound answer except by supposing that the crevasses are, at least in a great de- 
gree, the fresh production of every spring, and arise from the sudden start which the 
glacier makes when that extremity which descends into the valley begins to experi- 
ence the thawing effects of returning summer. I should not wish to speak positively 
upon what involves a difficult if not impossible observation, — the state of the glacier 
with respect to crevasses whilst still under the winter’s covering of snow. But the fact 
* Ninth Letter on Glaciers. Appendix to Travels, 2nd edit., p. 443 
