than by day. This seems to be a crucial test which reduces the
hypothesis to an absurdity, since we know that glaciers flow more
rapidly in summer than in winter, and that their motion is greater by
day than by night.
If additional evidence of the inadequacy of the hypothesis of dilatation
and contraction was desired, the slow conductivity of both ice and snow,
and the manner in which glaciers are invariably blanketed with snow
throughout a large part of the year might be considered. For example,
in neve regions, the loose granular snow is frequently hundreds of feet
deep, and is not only an exceedingly poor conductor of heat, but, on account
of its open texture, would undergo but slight changes in mass on account
of changes in temperature, since slight movements of the granules would
be taken up by the adjacent interspaces. It does not require accurate
observations to show that in such regions changes of temperature are too
brief to be felt at any considerable depth, and even under the most ex-
178 GLACIERS OF NORTH AMERICA.
trerne conditions could not cause sufficient change to account for the flow
known to occur in neves. The considerations here suggested will be
again referred to in connection with certain " molecular hypotheses " that
have been proposed to account for glacier flow.
From the considerations offered above it seems evident that the
hypothesis suggested by Moseley cannot be accepted as a final explana-
tion of the flow of glaciers.
It is equally plain, however, that it does contain an element of truth,
since it cannot be denied that ice, like most other substances, does contract
and expand with changes of temperature, or that glaciers are exposed to
conditions which bring about such changes. Some fraction of glacier
motion must, therefore, be due to those causes.
The Hypothesis of Liquefaction under Pressure. The fact that
the freezing-point of water is lowered by pressure, discovered by James
Thompson 1 and confirmed by his brother William, 1 was at once applied in
explanation of glacier motion.
As stated in the papers just referred to, if ice at 32 F. is subjected
to pressure, pores occupied by liquid water must instantly be formed in
the compressed parts, for the reason that ice cannot exist at the above tem-
perature under a pressure exceeding one atmosphere. If the conditions
permit, the water formed by the melting of the parts under pressure will
be forced to where the pressure is less and at once refreeze. The parts
recongealed after being melted must in turn, through the yielding of
other parts, receive pressure from the applied force, thereby to be again
liquefied and to enter again into a similar cycle.
In applying this principle to glaciers it is claimed that the water
formed by liquefaction may in part descend, and on refreezing occupy a
lower position. (It might be asked, however, why the water, if under
pressure, should descend rather than move in any other direction.) A
continuation of this process, it must be admitted, would tend to crowd an
alpine glacier down the valley it occupies, but the amount of movement
thus produced would be small.
In opposition to this hypothesis it is evident that the greatest pressure,
at least in the case of a glacier flowing through an even channel, is at the
bottom, while the surface sustains the pressure of only one atmosphere,
!"0n the Plasticity of Ice, as manifested in Glaciers," Roy. Soc., Proc., vol. 8, 1857, pp.
455-458. " On Recent Theories and Experiments regarding Ice at or near its Melting-point,"
Roy. Soc., Proc., vol. 10, 1859, pp. 152-160.
HOW AND WHY GLACIERS MOVE. 179
and in the great majority of cases of less than one atmosphere, yet the
maximum flow is always at the surface. Additional weight is given to
this objection when w^e recall the fact that the flowing motion observed
in glaciers is greatest at the surface even in cold weather ; at such times
the surface ice may reasonably be concluded to have a lower temperature
than the bottom ice, and therefore require a greater amount of pressure
to cause it to liquefy.
It does not seem as if further argument was necessary to show that the
lowering of the melting-point of ice by pressure does not furnish a com-
plete and satisfactory explanation of glacier motion ; but that it may play
a part in the phenomena for which an explanation is sought, especially in
the case of excessively thick ice bodies or of glaciers flowing through
irregular channels, cannot be refuted.
The Hypothesis of Molecular Change. An ingenious and highly
suggestive hypothesis advanced by James Croll to account for the flow of
glaciers is based, in part, on the well-known fact that water in freezing
gives out heat and expands on passing to the solid state ; and conversely,
when ice melts heat is absorbed, and in passing to the liquid state occupies
less space than before the change. The difference in volume between ice
at 32 F. and the water formed from its melting, at the same temperature, is
about one-tenth, that is, the ice occupies one-tenth more space than the water.
Another principle that enters into this hypothesis is that heat above
32 may be transmitted through ice, and yet the ice as a mass remain
solid. This has been demonstrated by Tyndall and others by placing a
delicate heat-measuring apparatus on one side of a slab of ice and bringing
heat to the opposite side. It may thus be shown that some of the heat
will pass through the ice, but portions of it are retained and produce
changes within the mass. As demonstrated by Tyndall, in producing
"liquid flowers " in ice by passing heat through it, a portion of the interior
of the mass of ice may be melted by heat that passes through other
portions which remain solid.
The hypothesis before us assumes that the heat of the sun on reaching
the surface of a glacier is partially expended in melting its surface, and
that a part of the water thus formed is transferred to the ice below and is
refrozen, and that the heat liberated melts other portions, which again re-
freeze, and so on. The essential feature of the hypothesis is that the heat
which enters the ice directly also leads to molecular changes, which cause
the ice to descend. Molecules of ice are assumed to be melted, and on
180 GLACIERS OF NORTH AMERICA.
refreezing pass their heat to other molecules. But the molecules lique-
fied occupy less space than before melting and change their position in
response to gravity or pressure, and on refreezing again expand and exert
a force on their confining walls.
In the hope of stating this hypothesis more definitely I will use Croll's
own words : " Let us observe what takes place, say, at the lower end of
the glacier. A molecule A at the lower end, say at its surface, receives
heat from the sun's rays ; it melts and in melting not only loses its shear-
ing force and descends by its own weight, but it contracts also. B, imme-
diately above it, is now, so far as A is concerned, at liberty to descend, and
will do so the moment that it assumes the liquid state. A by this time
has become solid and again fixed by shearing force, but is not fixed in its
old position, but a little below where it was before. If B has not already
passed into the fluid state in consequence of heat derived from the sun,
the additional supply which it will receive from the solidifying of A will
melt it. The moment that B becomes fluid it will descend till it reaches
A. B then is solidified a little below its former position. The same
process of reasoning is in a similar manner applicable to every molecule of
the glacier. Each molecule of the glacier consequently descends step by
step as it melts and solidifies, and hence the glacier, considered as a mass,
is in a state of constant motion downwards."
The heat that reaches a glacier, as stated by Croll, is (1) from the sun,
either directly or through the medium of the atmosphere, rain, etc. ; (2)
earth-heat from the rocks over which the glacier passes ; and (3) the heat
produced by friction. Of these, it seems to the present writer, account
need only be taken of the heat derived from the sun. The earth-heat is
certainly small, and for the present at least can be considered as having no
practical bearing on the question in hand. The heat of friction, if the
movements of a glacier are caused solely by the molecular changes con-
sidered by Croll, is due to the arrest of motion produced by the energy of
the sun, and to admit it as a source of energy to be used in explaining
glacier movement would be utilizing the same energy twice. That is, if
the sun's heat produces glacial movement, and by the arrest of this move-
ment a part of the energy which caused it is reconverted into heat, which
in its turn causes glacial motion, there is no end to the circle. A glacier
would then be an example of perpetual motion.
The above is confessedly an imperfect statement of the molecular
hypothesis, and it is perhaps unjust to suggest obstacles to its acceptance
without a more complete presentation, but as space will not admit of this,
HOW AND WHY GLACIERS MOVE. 181
1 must refer the reader to Croll's papers and books for a full exposition of
his case. 1
Neglecting the purely theoretical discussion of the manner in which
the molecules of ice are supposed to be influenced by the passage of heat
through it, since this is a question for physicists, let us, as geographers,
see how the hypothesis before us meets the actual conditions with which
we are familiar from field observations.
Assuming, what must be practically true so far as this hypothesis is
concerned, that all the heat which reaches a glacier comes from the sun, it
follows that the energy requisite to loosen the molecules of the ice and to
allow gravity to act in the manner postulated, can only be transmitted to
the glacier when a thermometer at its surface at the point where the heat
enters stands above 32 F. The higher the temperature indicated by such
a thermometer the more rapid would be the ice movement. Also, from
the fact that a large amount of energy is consumed in changing the
molecular condition of ice before it can melt physicists tell us that the
amount of heat absorbed by ice at 32 F., in changing to water at the same
temperature, is equal to the amount of heat required to raise the water
thus formed from 32 F. to the boiling-point it follows that a thermom-
eter at the surface of a glacier would have to rise well above 32 F. or
remain somewhat above that temperature for a considerable time in order
that the ice might receive the requisite amount of heat to imitate the process
described by the author of the molecular hypothesis.
In attempting to apply this hypothesis, however, we are met at the
outset with the conclusion, not yet successfully controverted, that the in-
ternal temperature of a glacier is always 32 F. or lower. This conclu-
sion is based on several facts, as for example : first, direct observations show
that as nearly as can be determined the internal temperature of a glacier
in summer is 32 F. ; second, ice in melting under atmospheric pressure
changes to water with a temperature of 32 F., and a mixture of ice and
water has this temperature ; third, a mass of ice in contact with air below
32 F. will have its temperature lowered.
Turning now to a typical alpine glacier we find that near its source the
temperature of the air in contrast with it is always low. In the Mount
St. Elias region surface melting does not occur at elevations in excess of
about 13,000 feet. Above that elevation the snow is always light and dry.
lrt On the Physical Cause of the Motion of Glaciers," Philosophical Magazine, vol. 38,
1869, pp. 201-206. "Climate and Time," 1875, pp. 1166-1195. Also, "The Great Ice
Age," by James Geikie, 2d ed., 1877, pp. 21-31.
182 GLACIERS OF NOKTH AMERICA.
At noon on a cloudless August day, at an elevation of 14,000 feet on the
side of Mount St. Elias, I found the temperature of the snow at a depth
of two or three inches, where the surface was directly exposed to the sun,
to be sixteen degrees below freezing. This, it must be remembered, is the
most favorable condition for melting that occurs throughout the entire
year at the locality referred to. At night, even in summer, the tempera-
ture of the air falls far below freezing. For probably nine months or
more each year the temperature of the snow at the surface of the neves in
the St. Elias region is continuously below the freezing-point. It is im-
possible to see how under these conditions the molecular hypothesis can
be applied, yet it is in regions of the nature referred to that glaciers have
their birth. The snow accumulating on neves must move downward be-
fore trunk glaciers can foe formed, but if the flow of the lower portions of
glaciers is to be accounted for by molecular changes the same explanation
should be applicable to neve regions as well.
In some respects the impressions conveyed by what is stated above will
be incorrect, for the reason that few neves are so circumstanced that melt-
ing does not occur on them during a portion of each year. I shall endeavor
to show, however, that the other extreme of conditions to which neves are
subjected is no more favorable to the molecular hypothesis. As is well
known, the surfaces of neves for a large part of each year are composed of
light, dry snow. The consolidation of this snow, it would seem, must take
place before a pushing force of the nature postulated by Croll could act
efficiently in producing a flowing movement in the mass. Consolidation
of neve snow does not occur when the temperature of the air in contact
with it is above the freezing-point, as it is then partially melted and
v frequently so completely saturated with water as to be soft and slushy,
and, many times, holds shallow lakes in depressions of its surface. No
one will claim, I fancy, that a mixture of snow and water of such a con-
sistency that one will sink knea-deep into it at every step a condition
frequently present on the neve- of Alaska is favorable to the passage of
heat through it so as to produce molecular changes in the ice below. .
It appears; then, that the surface of a neve, both when below 32 F.,
or when it is open and porous, and when it is exposed to a greater tempera-
ture and becomes saturated with water, is unfavorable for the transmission
of solar energy. Again, the surface of a neve is renewed each winter, and
in many instances by summer storms as well ; and at such times is at 32 F.
or lower ; and when surface melting is in progress the snow is saturated with
water and consequently has a temperature of 32 F., hence it is impossible
HOW AND WHY GLACIERS MOVE. 183
to conceive how a temperature in excess of -32 F. could be transmitted to
the solid ice beneath so as to give it motion.
Nor are the difficulties in the way of applying the molecular hypothesis
confined to neve regions. The portion of a glacier that protrudes below
its neve is blanketed with snow in winter ; the air in contact with it
is then, also, normally below freezing. In summer the surface of the
ice is frequently covered with a porous, coral-like crust, which is almost
as perfect a non-conductor as dry snow ; and when this crust is in process
of melting it is saturated with water and consequently has a general tem-
perature of 32 F., and will not allow heat in excess of that temperature
to pass. Glaciers are also frequently covered more or less completely
with moraines, which, when over a few inches in thickness, still more
effectually shield the ice beneath from solar energy. When we consider
the nature of the surface presented by a glacier from its clear-white, snow-
covered neve to its dark and frequently moraine-covered terminus, it is
apparent that the positions where hard, blue ice is exposed to the sky are
relatively few. It may be said in general that clear ice is only exposed
for any considerable time when its surface gradient is sufficient to insure
the quick escape of the water formed by superficial melting. Under the
molecular hypothesis, other conditions being the same, motion should be
most rapid where the surface of a glacier is composed of clear ice. So far
as now known observations do not harmonize with this postulate.
Although heat may be transmitted through ice in laboratory experi-
ments and cause melting within its mass, yet the conditions, as shown
above, when this can occur in the case of glaciers are so infrequent that
the application of this principle in explanation of glacier motion, even
if we knew the nature of the molecular changes that occur, meets what
seem insuperable difficulties.
The Hypothesis of Granular Change. Recent observations on the
structure of certain Greenland glaciers by T. C. Chamberlin 1 have led
him to conclude, as previously noted, that they behave as rigid rather than
as plastic bodies. The impression gained was that the ice is thrust for-
ward by an expansive force acting from within rather than pulled by
gravity alone after the manner of plastic substances. In seeking for an
explanation of glacier flow with this idea in mind Chamberlin again di-
rected attention to the changes which occur in the granulation of glacier
ice when traced from its source to the extremity of a glacier, and, as others
!" Recent Glacial Studies in Greenland," Geol. Soc. Am., Bull., vol. 6, 1895, pp. 199-220.
184 GLACIERS OF NORTH AMERICA.
have done, suggested that the principal cause of movement may lie in the
growth of the granules. Through a process of partial melting and re-
freezing it is assumed that a granule may continually change its shape by
loss in one part and gain in another, and thus either move itself or permit
motion in a neighbor. To bring about this change the author states that
" every warm day sends down into the glacier a wave of heat energy, sen-
sible or potential, and that every night sends after it a wave of reverse
energy. These waves follow each other indefinitely, until by intercurrent
agencies they become vanishing quantities. Each season sends through
the mass a greater and more complex wave. The problem, therefore, in
simplified form postulates a mass of ice granules predisposed to melt at
certain points and to freeze or to promote freezing at others, acted upon
by the ever-present but differential force of gravity and swept by succes-
sive waves of heat energy competent to cause melting where predisposition
to melting exists and to cause growth by freezing where predisposition to
freezing exists. Out of this it would seem that localized freezing and
thawing, growths and decadences, innumerable and constantly changing,
must result, and with them motion of the granules themselves and of the
Although, in fairness to my readers, I must confess that I am unable
to discuss either the molecular hypothesis or the recent modification of it
from the standpoint of the physicist, yet the adverse bearing of certain
facts on these hypotheses and difficulties in the way of applying them
to explain well-known glacial phenomena suggest themselves even to a
Present knowledge of the physical properties of ice and of water seems
to show that when they are in contact at the same temperature, at least
when not under pressure exceeding one atmosphere, there is no reason to
suppose that there would be a mutual change of condition. That is, so
far as I am aware, there is no reason to suppose that a molecule of water
at 32 and a molecule of ice at- the same temperature, placed side by side,
would undergo a mutual interchange of their physical properties, the
water becoming ice and the ice changing to water. But this seems
to be an essential feature, although not stated in these words, of both
the molecular hypothesis and the modified version of it.
Another consideration is that molecular changes postulated cannot take
place in ice that is below 32 F. unless pressure is greatly increased. If
pressure is the controlling condition, then the movements supposed to occur
would increase with depth of ice, and the bottom of a glacier should flow
HOW AND WHY GLACIERS MOVE. 185
more rapidly than its surface. This, we know, is the opposite of what
really does take place.
Chamberlin speaks of waves of sensible or potential heat energy pass-
ing through a glacier every warm day, but it is to be remembered that
warm days on the upper portions of glaciers, especialty, are rare ; and, as
already pointed out, the nev of a glacier is constantly blanketed by a
non-conducting layer such as no wave of sensible heat, at least, can
The explanation suggested by Chamberlin, like its predecessor, is
based on theoretical deductions which have not been proven. Before
claiming that solar energy causes motion in glaciers by the melting
and refreezing of molecules, or by changes in the size and shape of
granules, it would be more consistent to determine if these changes do
occur in ice under the most favorable conditions. But even if it could be
shown that a " wave of heat energy, sensible or potential," could lead to a
change in the form and size of granules, the same objections to the trans-
fer of such energy from the surface to the inner positions of a glacier,
suggested in reference to Croll's hypothesis, would have to be met.
While it is apparently impossible to demonstrate that the changes in
ice assumed by Croll and Chamberlin do not take place, it is logical to
wait until sufficient reasons have been advanced to prove that they do
occur even under the most favorable conditions, before making the
assumption the basis of a still more extended hypothesis. At present the
postulates on which the molecular hypothesis and its recent modification
rest, may be said to be in the hands of the physicist. If their truth is
ultimately demonstrated, they may be passed on to the geologist and
geographer to be tested as a means of explaining glacial motion.
What seems intended as a modification of the molecular hypothesis
has been proposed by R. M. Deeley l and can be studied with profit, but
so far as I have been able to determine it does not differ materially from
the explanation proposed by Croll.
From the brief account given above of various hypotheses that have
been advanced to account for the movements of glaciers, it will be seen
that no one of them meets all the conditions of the problem. No one ex-
planation has been generally accepted, although the hypothesis of plas-
!" A Theory of Glacial Motion," Philosophical Magazine, vol. 25, 1888, pp. 136-164.
186 GLACIERS OF NORTH AMERICA.
ticity probably has more adherents than any other. From what has been
stated, however, I think it will appear that several of the explanations
offered are based on one or more well-established laws and furnish an
explanation of some phase of glacial motion. Even the earliest and long-
since abandoned hypothesis of Charpentier, which assumes that glaciers
slide as rigid bodies over their beds, contains, as we have seen, an element
of truth. If this can be said of the crudest of all the explanations
advanced, the later and more elaborate hypotheses should certainly not
be discarded without careful scrutiny in order to obtain from them all the
assistance in arriving at a final theory that is possible. To the discredit
of men of science, it must be acknowledged, that in discussing the problem
of glacier motion, the practice has too frequently prevailed of rejecting all
previous hypotheses in order to make room for some newer idea. The
attitude of scientific men in this connection has frequently been that of
an advocate pleading for his client, rather than a judicial balancing of
An Eclectic Hypothesis. The review just attempted leads to at least
one conclusion which seems well founded. That is, the phenomena to be
accounted for are complex, and no single law governing the behavior of ice
can be made to explain all phases of glacier motion.
The principal laws and the leading physical properties of ice which are
concerned in modifying the form of glaciers at one time or another, or at
one locality or another, may be briefly enumerated as follows :
1. Gravity is ever present and tends continually to change the form
of a glacier. This fundamental fact is recognized in every hypothesis that
has been advanced.
2. Ice, although brittle under a force quickly applied, yields con-
tinuously under its own weight to both pressure and tension in a dimin-
ishing ratio from 32 F. to lower temperatures.
3. Fragments of ice when brought in contact at or near a tempera-
ture of 32 F. will freeze together, but without pressure this does not