CHAPTER XXX.

THE PHYSICAL CAUSE OF THE MOTION OF GLACIERS.—THEORIES OF GLACIER-MOTION.

Why the Question of Glacier-motion has been found to be so difficult.—The Regelation Theory.—It accounts for the Continuity of a Glacier, but not for its Motion.—Gravitation proved by Canon Moseley insufficient to shear the Ice of a Glacier.—Mr. Mathew’s Experiment.—No Parallel between the bending of an Ice Plank and the shearing of a Glacier.—Mr. Ball’s Objection to Canon Moseley’s Experiment.—Canon Moseley’s Method of determining the Unit of Shear.—Defect of Method.—Motion of a Glacier in some Way dependent on Heat.—Canon Moseley’s Theory.—Objections to his Theory.—Professor James Thomson’s Theory.—This Theory fails to explain Glacier-motion.—De Saussure and Hopkins’s “Sliding” Theories.—M. Charpentier’s “Dilatation” Theory.—Important Element in the Theory.

The cause of the motion of glaciers has proved to be one of the most difficult and perplexing questions within the whole domain of physics. The main difficulty lies in reconciling the motion of the glacier with the physical properties of the ice. A glacier moves down a valley very much in the same way as a river, the motion being least at the sides and greatest at the centre, and greater at the surface than at the bottom. In a cross section scarcely two particles will be moving with the same velocity. Again, a glacier accommodates itself to the inequalities of the channel in which it moves exactly as a semifluid or plastic substance would do. So thoroughly does a glacier behave in the manner of a viscous or plastic body that Professor Forbes was induced to believe that viscosity was a property of the ice, and that in virtue of this property it was enabled to move with a differential motion and accommodate itself to all the inequalities of its channel without losing its continuity just as a mass of mud or putty would do. But experience proves that ice is a hard and brittle substance far more resembling glass than putty. In fact it is one of the most brittle and unyielding substances in nature. So unyielding is a glacier that it will snap in two before it will stretch to any perceptible extent. This is proved by the fact that crevasses resulting from a strain on the glacier consist at first of a simple crack scarcely wide enough to admit the blade of a penknife.

All the effects which were considered to be due to the viscosity of the ice have been fully explained and accounted for on the principle of fracture and regelation discovered by Faraday. The principle of regelation explains why the ice moving with a differential motion and accommodating itself to the inequalities of its channel is yet enabled to retain its continuity, but it does not account for the _cause_ of glacier motion. In fact it rather involves the question in deeper mystery than before. For it is far more difficult to conceive how the particles of a hard and brittle solid like that of ice can move with a differential motion, than it is to conceive how this may take place in the case of a soft and yielding substance. The particles of ice have all to be displaced one over another and alongside each other, and as those particles are rigidly fixed together this connection must be broken before the one can slide over the other. _Shearing-force_, as Canon Moseley shows, comes into play. Were ice a plastic substance there would not be much difficulty in understanding how the particles should move the one over the other, but it is totally different when we conceive ice to be a solid and unyielding substance. The difficulty in connection with glacier-motion is not to account for the continuity of the ice, for the principle of regelation fully explains this, but to show how it is that one particle succeeds in sliding over the over. The principle of regelation, instead of assisting to remove this difficulty, increases it tenfold. Regelation does not explain the cause of glacier-motion, but the reverse. It rather tends to show that a glacier should not move. What, then, is the cause of glacier-motion? According to the regelation theory, gravitation is the impelling cause. But is gravitation sufficient to _shear_ the ice in the manner in which it is actually done in a glacier?

I presume that few who have given much thought to the subject of glacier-motion have not had some slight misgivings in regard to the commonly received theory. There are some facts which I never could harmonize with this theory. For example, boulder clay is a far looser substance than ice; its shearing-force must be very much less than that of ice; yet immense masses of boulder clay will lie immovable for ages on the slope of a hill so steep that one can hardly venture to climb it, while a glacier will come crawling down a valley which by the eye we could hardly detect to be actually off the level. Again, a glacier moves faster during the day than during the night, and about twice as fast during summer as during winter. Professor Forbes, for example, found that the Glacier des Bois near its lower extremity moved sometimes in December only 11·5 inches daily, while during the month of July its rate of motion sometimes reached 52·1 inches per day. Why such a difference in the rate of motion between day and night, summer and winter? The glacier is not heavier during the day than it is during the night, or during the summer than it is during the winter; neither is the shearing-force of the great mass of the ice of a glacier sensibly less during day than night, or during summer than winter; for the temperature of the great mass of the ice does not sensibly vary with the seasons. If this be the case, then gravitation ought to be as able to shear the ice during the night as during the day, or during the winter as during the summer. At any rate, if there should be any difference it ought to be but trifling. It is true that, owing to the melting of the ice, the crevices of the glacier are more gorged with water during summer than winter; and this, as Professor Forbes maintains,[297] may tend to make the glacier move faster during the former than the latter season. But the advocates of the regelation theory cannot conclude, with Professor Forbes, that the water favours the motion of the glacier by making the ice more soft and plastic. The melting of the ice, according to the regelation theory, cannot very materially aid the motion of the glacier.

The theory which has led to the general belief that the ice of a glacier is sheared by the force of gravity appears to be the following. It is supposed that the only forces to which the motion of a glacier can be referred are _gravitation_ and _heat_; but as the great mass of a glacier remains constantly at the same uniform temperature it is concluded to be impossible that the motion of the glacier can be due to this cause, and therefore of course it must be attributed to gravitation, there being no other cause.

That gravitation is insufficient to shear the ice of a glacier has been clearly demonstrated by Canon Moseley.[298] He determined by experiment the amount of force required to shear one square inch of ice, and found it to be about 75 lbs. By a process of calculation which will be found detailed in the Memoir referred to, he demonstrated that to descend by its own weight at the rate at which Professor Tyndall observed the ice of the Mer de Glace to be descending at the Tacul, the unit of shearing force of the ice could not have been more than 1·31931 lbs. Consequently it will require a force more than 34 times the weight of the glacier to shear the ice and cause it to descend in the manner in which it is found to descend.

It is now six years since Canon Moseley’s results were laid before the public, and no one, as far as I am aware, has yet attempted to point out any serious defect in his mathematical treatment of the question. Seeing the great amount of interest manifested in the question of glacier-motion, I think we are warranted to conclude that had the mathematical part of the memoir been inconclusive its defects would have been pointed out ere this time. The question, then, hinges on whether the experimental data on which his calculations are based be correct. Or, in other words, is the unit of shear of ice as much as 75 lbs.? This part of Mr. Moseley’s researches has not passed unquestioned. Mr. Ball and Mr. Mathews, both of whom have had much experience among glaciers, and have bestowed considerable attention on the subject of glacier-motion, have objected to the accuracy of Mr. Moseley’s unit of shear. I have carefully read the interesting memoirs of Mr. Mathews and Mr. Ball in reply to Canon Moseley, but I am unable to perceive that anything which they have advanced materially affects his general conclusions as regards the commonly received theory. Mr. Mathews objects to Canon Moseley’s experiments on the grounds that extraneous forces are brought to bear upon the substance submitted to operation, and that conditions are thus introduced which do not obtain in the case of an actual glacier. “It would throw,” he says, “great light upon our inquiry if we were to change this method of procedure and simply to observe the deportment of masses of ice under the influence of no external forces but the gravitation of their own particles.”[299] A plank of ice six inches wide and 2⅜ inches in thickness was supported at each end by bearers six feet apart. From the moment the plank was placed in position it began to sink, and continued to do so until it touched the surface over which it was supported. Mr. Mathews remarks that with this property of ice, viz., its power to change its form under strains produced by its own gravitation, combined with the sliding movement demonstrated by Hopkins, we have an adequate cause for glacier-motion. Mr. Mathews concludes from this experiment that the unit of shear in ice, instead of being 75 lbs., is less than 1¾ lbs.

There is, however, no parallel between the bending of the ice-plank and the shearing of a glacier. Mr. Mathews’ experiment appears to prove too much, as will be seen from the following reply of Canon Moseley:—

“Now I will,” he says, “suggest to Mr. Mathews a parallel experiment and a parallel explanation. If a bar of wrought iron 1 inch square and 20 feet long were supported at its extremities, it would _bend_ by its weight alone, and would therefore shear. Now the weight of such a rod would be about 67 lbs. According to Mr. Mathews’s explanation in the case of the ice-plank, the unit of shear in wrought-iron should therefore be 67 lbs. per square inch. It is actually 50,000 lbs.”[300]

Whatever theory we may adopt as to the cause of the motion of glaciers, the deflection of the plank in the way described by Mr. Mathews _follows as a necessary consequence_. Although no weight was placed upon the plank, it does not necessarily follow that the deflection was caused by the weight of the ice alone; for, according to Canon Moseley’s own theory of the motion of glaciers by heat, the plank ought to be deflected in the middle, just as it was in Mr. Mathews’s experiment. A solid body, when exposed to variations of temperature, will expand and contract transversely as well as longitudinally. Ice, according to Canon Moseley’s theory, expands and contracts by heat. Then if the plank expands transversely, the upper half of the plank must rise and the lower half descend. But the side which rises has to perform work against gravity, whereas the side which descends has work performed upon it by gravity; consequently more of the plank will descend than rise, and this will, of course, tend to lower or deflect the plank in the middle. Again, when the plank contracts, the lower half will rise and the upper half will descend; but as gravitation, in this case also, favours the descending part and opposes the rising part, more of the plank will descend than rise, and consequently the plank will be lowered in the middle by contraction as well as by expansion. Thus, as the plank changes its temperature, it must, according to Mr. Moseley’s theory, descend or be deflected in the middle, step by step—and this not by gravitation alone, but chiefly by the motive power of heat. I do not, of course, mean to assert that the descent of the plank was caused by heat; but I assert that Mr. Mathews’s experiment does not necessarily prove (and this is all that is required in the meantime) that gravitation alone was the cause of the deflection of the plank. Neither does this experiment prove that the ice was deflected without shearing; for although the weight of the plank was not sufficient to shear the ice, as Mr. Mathews, I presume, admits, yet Mr. Moseley would reply that the weight of the ice, assisted by the motive power of heat, was perfectly sufficient.

I shall now briefly refer to Mr. Ball’s principal objections to Canon Moseley’s proof that a glacier cannot shear by its weight alone. One of his chief objections is that Mr. Moseley has assumed the ice to be homogeneous in structure, and that pressures and tensions acting within it, are not modified by the varying constitution of the mass.[301] Although there is, no doubt, some force in this objection (for we have probably good reason to believe that ice will shear, for example, more easily along certain planes than others), still I can hardly think that Canon Moseley’s main conclusion can ever be materially affected by this objection. The main question is this, Can the ice of the glacier shear by its own weight in the way generally supposed? Now the shearing force of ice, take it in whatever direction we may, so enormously exceeds that required by Mr. Moseley in order to allow a glacier to descend by its weight only, that it is a matter of indifference whether ice be regarded as homogeneous in structure or not. Mr. Ball objects also to Mr. Moseley’s imaginary glacier lying on an even slope and in a uniform rectangular channel. He thinks that an irregular channel and a variable slope would be more favourable to the descent of the ice. But surely if the work by the weight of the ice be not equal to the work by the resistance in a glacier of uniform breadth and slope, it must be much less so in the case of one of irregular shape and slope.

That a relative displacement of the particles of the ice is involved in the motion of a glacier, is admitted, of course, by Mr. Ball; but he states that the amount of this displacement is but small, and that it is effected with extreme slowness. This may be the case; but if the weight of the ice be not able to overcome the mutual cohesion of the particles, then the weight of the ice cannot produce the required displacement, however small it may be. Mr. Ball then objects to Mr. Moseley’s method of determining the unit of shear on this ground:—The shearing of the ice in a glacier is effected with extreme slowness; but the shearing in Canon Moseley’s experiment was effected with rapidity; and although it required 75 lbs. to shear one square inch of surface in his experiment, it does not follow that 75 lbs. would be required to shear the ice if done in the slow manner in which it is effected in the glacier. “In short,” says Mr. Ball, “to ascertain the resistance opposed to very slow changes in the relative positions of the particles, so slight as to be insensible at short distances, Mr. Moseley measures the resistance opposed to rapid disruption between contiguous portions of the same substance.”

There is force in this objection; and here we arrive at a really weak point in Canon Moseley’s reasoning. His experiments show that if we want to shear ice quickly a weight of nearly 120 lbs. is required; but if the thing is to be done more slowly, 75 lbs. will suffice.[302] In short, the number of pounds required to shear the ice depends, to a large extent, on the length of time that the weight is allowed to act; the longer it is allowed to act, the less will be the weight required to perform the work. “I am curious to know,” says Mr. Mathews, when referring to this point, “what weight would have sheared the ice if a _day_ had been allowed for its operation.” I do not know what would have been the weight required to shear the ice in Mr. Moseley’s experiments had a day been allowed; but I feel pretty confident that, should the ice remain unmelted, and sufficient time be allowed, shearing would be produced without the application of any weight whatever. There are no weights placed upon a glacier to make it move, and yet the ice of the glacier shears. If the shearing is effected by weight, the only weight applied is the weight of the ice; and if the weight of the ice makes the ice shear in the glacier, why may it not do the same thing in the experiment? Whatever may be the cause which displaces the particles of the ice in a glacier, they, as a matter of fact, are displaced without any weight being applied beyond that of the ice itself; and if so, why may not the particles of the ice in the experiment be also displaced without the application of weights? Allow the ice of the glacier to take its own time and its own way, and the particles will move over each other without the aid of external weights, whatever may be the cause of this; well, then, allow the ice in the experiment to take its own time and its own way, and it will probably do the same thing. There is something here unsatisfactory. If, by the unit of shear, be meant the pressure in pounds that must be applied to the ice to break the connection of one square inch of two surfaces frozen together and cause the one to slip over the other, then the amount of pressure required to do this will depend upon the time you allow for the thing being done. If the thing is to be done rapidly, as in some of Mr. Moseley’s experiments, it will take, as he has shown, a pressure of about 120 lbs.; but if the thing has to be done more slowly, as in some other of his experiments, 75 lbs. will suffice. And if sufficient time be allowed, as in the case of glaciers, the thing may be done without any weight whatever being applied to the ice, and, of course, Mr. Moseley’s argument, that a glacier cannot descend by its weight alone, falls to the ground. But if, by the unit of shear, be meant not the _weight_ or _pressure_ necessary to shear the ice, but the amount of _work_ required to shear a square inch of surface _in a given time or at a given rate_, then he might be able to show that in the case of a glacier (say the Mer de Glace) the work of all the resistances which are opposed to its descent at the _rate_ at which it is descending is greater than the work of its weight, and that consequently there must be some cause, in addition to the weight, urging the glacier forward. But then he would have no right to affirm that the glacier would not descend by its weight only; all that he could affirm would simply be that it could not descend by its weight alone at the _rate_ at which it is descending.

Mr. Moseley’s unit of shear, however, is not the amount of work performed in shearing a square inch of ice in a given time, but the amount of _weight_ or _pressure_ requiring to be applied to the ice to shear a square inch. But this amount of pressure depends upon the length of time that the pressure is applied. Here lies the difficulty in determining what amount of pressure is to be taken as the real unit. And here also lies the radical defect in Canon Moseley’s result. Time as well as pressure enters as an element into the process. The key to the explanation of this curious circumstance will, I think, be found in the fact that the rate at which a glacier descends depends in some way or other upon the amount of heat that the ice is receiving. This fact shows that heat has something to do in the shearing of the ice of the glacier. But in the communication of heat to the ice _time_ necessarily enters as an element. There are two different ways in which heat may be conceived to aid in shearing the ice: (1.) we may conceive that heat acts as a force along with gravitation in producing displacement of the particles of the ice; or (2.) we may conceive that heat does not act as a force in pushing the particles over each other, but that it assists the shearing processes by diminishing the cohesion of the particles of the ice, and thus allowing gravitation to produce displacement. The former is the function attributed to heat in Canon Moseley’s theory of glacier-motion; the latter is the function attributed to heat in the theory of glacier-motion which I ventured to advance some time ago.[303] It results, therefore, from Canon Moseley’s own theory, that the longer the time that is allowed for the pressure to shear the ice, the less will be the pressure required; for, according to his theory, a very large proportion of the displacement is produced by the motive power of heat entering the ice; and, as it follows of course, other things being equal, the longer the time during which the heat is allowed to act, the greater will be the proportionate amount of displacement produced by the heat; consequently the less will require to be done by the weight applied. In the case of the glacier, Mr. Moseley concludes that at least thirty or forty times as much work is done by the motive power of heat in the way of shearing the ice as is done by mere pressure or weight. Then, if sufficient time be allowed, why may not far more be done by heat in shearing the ice in his experiment than by the weight applied? In this case how is he to know how much of the shearing is effected by the heat and how much by the weight? If the greater part of the shearing of the ice in the case of a glacier is produced, not by pressure, but by the heat which necessarily enters the ice, it would be inconceivable that in his experiments the heat entering the ice should not produce, at least to some extent, a similar effect. And if a portion of the displacement of the particles is produced by heat, then the weight which is applied cannot be regarded as the measure of the force employed in the displacement, any more than it could be inferred that the weight of the glacier is the measure of the force employed in the shearing of it. If the weight is not the entire force employed in shearing, but only a part of the force, then the weight cannot, as in Mr. Moseley’s experiment, be taken as the measure of the force.

How, then, are we to determine what is the amount of force required to shear ice? in other words, how is the unit of shear to be determined? If we are to measure the unit of shear by the weight required to produce displacement of the particles of the ice, we must make sure that the displacement is wholly effected by the weight. We must be certain that heat does not enter as an element in the process. But if time be allowed to elapse during the experiment, we can never be certain that heat has not been at work. It is impossible to prevent heat entering the ice. We may keep the ice at a constant temperature, but this would not prevent heat from entering the ice and producing molecular work. True that, according to Moseley’s theory of glacier-motion, if the temperature of the ice be not permitted to _vary_, then no displacement of the particles can take place from the influence of heat; but according to the molecular theory of glacier-motion, which will shortly be considered, heat will aid the displacement of the particles whether the temperature be kept constant or not. In short, it is absolutely impossible in our experiments to be certain that heat is not in some way or other concerned in the displacement of the particles of the ice. But we can shorten the time, and thus make sure that the amount of heat entering the ice during the experiments is too small to affect materially the result. We cannot in this case say that all the displacement has been effected by the weight applied to the ice, but we can say that so little has been done by heat that, practically, we may regard it as all done by the weight.

This consideration, I trust, shows that the unit of shear adopted by Canon Moseley in his calculations is not too large. For if in half an hour, after all the work that may have been done by heat, a pressure of 75 lbs. is still required to displace the particles of one square inch, it is perfectly evident that if no work had been done by heat during that time, the force required to produce the displacement could not have been less than 75 lbs. It might have been more than that; but it could not have been less. Be this, however, as it may, in determining the unit of shear we cannot be permitted to prolong the experiment for any considerable length of time, because the weight under which the ice might then shear could not be taken as the measure of the force which is required to shear ice. By prolonging the experiment we might possibly get a unit smaller than that required by Canon Moseley for a glacier to descend by its own weight. But it would be just as much begging the whole question at issue to assume that, because the ice sheared under such a weight, a glacier might descend by its weight alone, as it would be to assume that, because a glacier shears without a weight being placed upon it, the glacier descends by its weight alone.

But why not determine the unit of shear of ice in the same way as we would the unit of shear of any other solid substance, such, as iron, stone, or wood? If the shearing force of ice be determined in this manner, it will be found to be by far too great to allow of the ice shearing by its weight alone. We shall be obliged to admit either that the ice of the glacier does not shear (in the ordinary sense of the term), or if it does shear, that there must, as Canon Moseley concludes, be some other force in addition to the weight of the ice urging the glacier forward.

The fact that the rate of descent of a glacier depends upon the amount of heat which it receives, proves that heat must be regarded either as a cause or as a necessary condition of its motion; what, then, is the necessary relationship between heat and the motion of the glacier? If heat is to be regarded as a cause, in what way does the heat produce motion? I shall now briefly refer to one or two theories which have been advanced on the subject. Let us consider first that of Canon Moseley.

_Canon Moseley’s Theory._—He found, from observations and experiments, that sheets of lead, placed upon an inclined plane, when subjected to variations of temperature, tend to descend even when the slope is far less than that which would enable it to slide down under the influence of gravitation. The cause of the descent he shows to be this. When the temperature of the sheet is raised, it expands, and, in expanding, its upper portion moves up the slope, and its lower portion down the slope; but as gravitation opposes the upward and favours the downward motion, more of the sheet moves down than up, and consequently the centre of gravity of the sheet is slightly lowered. Again, when the sheet is cooled, it contracts, and in contracting the upper portion moves downwards and the lower portion upwards, and here again, for the same reason, more of the sheet moves downwards than upwards. Consequently, at every change of temperature there is a slight displacement of the sheet downwards. “Now a theory of the descent of glaciers,” says Canon Moseley, “which I have ventured to propose myself, is that they descend, as the lead in this experiment does, by reason of the passage into them and the withdrawal of the sun’s rays, and that the dilatation and contraction of the ice so produced is the proximate cause of their descent, as it is of that of the lead.”[304]

The fundamental condition in Mr. Moseley’s theory of the descent of solid bodies on an incline, is, not that heat should maintain these bodies at a high temperature, but that the temperature should vary. The rate of descent is proportionate, not simply to the amount of heat received, but to the extent and frequency of the variations of temperature. As a proof that glaciers are subjected to great variations of temperature, he adduces the following:—“All alpine travellers,” he says, “from De Saussure to Forbes and Tyndall, have borne testimony to the intensity of the solar radiation on the surfaces of glaciers. ‘I scarcely ever,’ says Forbes, ‘remember to have found the sun more piercing than at the Jardin.’ This heat passes abruptly into a state of intense cold when any part of the glacier falls into shadow by an alteration of the position of the sun, or even by the passing over it of a cloud.”[305]

Mr. Moseley is here narrating simply what the traveller feels, and not what the glacier experiences. The traveller is subjected to great variations of temperature; but there is no proof from this that the glacier experiences any changes of temperature. It is rather because the temperature of the glacier is not affected by the sun’s heat that the traveller is so much chilled when the sun’s rays are cut off. The sun shines down with piercing rays and the traveller is scorched; the glacier melts on the surface, but it still remains “cold as ice.” The sun passes behind a cloud or disappears behind a neighbouring hill; the scorching rays are then withdrawn, and the traveller is now subjected to radiation on every side from surfaces at the freezing-point.

It is also a necessary condition in Mr. Moseley’s theory that the heat should pass easily into and out of the glacier; for unless this were the case sudden changes of temperature could produce little or no effect on the great mass of the glacier. How, then, is it possible that during the heat of summer the temperature of the glacier could vary much? During that season, in the lower valleys at least, everything, with the exception of the glacier, is above the freezing-point; consequently when the glacier goes into the shade there is nothing to lower the ice below the freezing-point; and as the sun’s rays do not raise the temperature of the ice above the freezing-point, the temperature of the glacier must therefore remain unaltered during that season. It therefore follows that, instead of a glacier moving more rapidly during the middle of summer than during the middle of winter, it should, according to Moseley’s theory, have no motion whatever during summer.

The following, written fifteen years ago by Professor Forbes on this very point, is most conclusive:—“But how stands the fact? Mr. Moseley quotes from De Saussure the following _daily ranges_ of the temperature of the air in the month of July at the Col du Géant and at Chamouni, between which points the glacier lies:

° At the Col du Géant 4·257 Réaumur. At Chamouni 10·092 〃

And he assumes ‘the same mean daily variation of temperature to obtain throughout the length’ [and depth?] ‘of the Glacier du Géant which De Saussure observed in July at the Col du Géant.’ But between what limits does the temperature of the air oscillate? We find, by referring to the third volume of De Saussure’s ‘Travels,’ that the mean temperature of the coldest hour (4 A.M.) during his stay at the Col du Géant was 33°·03 Fahrenheit, and of the warmest (2 P.M.) 42°·61 F. So that even upon that exposed ridge, between 2,000 and 3,000 feet above where the glacier can be properly said to commence, the air does not, on an average of the month of July, reach the freezing-point at any hour of the night. Consequently the _range of temperature attributed to the glacier is between limits absolutely incapable of effecting the expansion of the ice in the smallest degree_.”[306]

Again, during winter, as Mr. Ball remarks, the glacier is completely covered with snow and thus protected both from the influence of cold and of heat, so that there can be nothing either to raise the temperature of the ice above the freezing-point or to bring it below that point; and consequently the glacier ought to remain immovable during that season also.

“There can be no doubt, therefore,” Mr. Moseley states, “that the rays of the sun, which in those alpine regions are of such remarkable intensity, find their way into the depths of the glacier. They are a _power_, and there is no such thing as the loss of power. The mechanical work which is their equivalent, and into which they are converted when received into the substance of a solid body, accumulates and stores itself up in the ice under the form of what we call elastic force or tendency to dilate, until it becomes sufficient to produce actual dilatation of the ice in the direction in which the resistance is weakest, and by its withdrawal to produce contraction. From this expansion and contraction follows of necessity the descent of the glacier.”[307] When the temperature of the ice is below the freezing-point, the rays which are absorbed will, no doubt, produce dilatation; but during summer, when the ice is not below the freezing-point, no dilatation can possibly take place. All physicists, so far as I am aware, agree that the rays that are then absorbed go to melt the ice, and not to expand it. But to this Mr. Moseley replied as follows:—“To this there is the obvious answer that radiant heat does find its way into ice as a matter of common observation, and that it does not melt it except at its surface. Blocks of ice may be seen in the windows of ice-shops with the sun shining full upon them, and melting nowhere but on their surfaces. And the experiment of the ice-lens shows that heat may stream through ice in abundance (of which a portion is necessarily stopped in the passage) without melting it, except on its surface.” But what evidence is there to conclude that if there is no melting of the ice in the interior of the lens there is a portion of the rays “necessarily stopped” in the interior? It will not do to assume a point so much opposed to all that we know of the physical properties of ice as this really is. It is absolutely essential to Mr. Moseley’s theory of the motion of glaciers, during summer at least, that ice should continue to expand after it reaches the melting-point; and it has therefore to be shown that such is the case; or it need not be wondered at that we cannot accept his theory, because it demands the adoption of a conclusion contrary to all our previous conceptions. But, as a matter of fact, it is not strictly true that when rays pass through a piece of ice there is no melting of the ice in the interior. Experiments made by Professor Tyndall show the contrary.[308]

There is, however, one fortunate circumstance connected with Canon Moseley’s theory. It is this: its truth can be easily tested by direct experiment. The ice, according to this theory, descends not simply in virtue of heat, but in virtue of _change of temperature_. Try, then, Hopkins’s famous experiment, but keep the ice at a _constant temperature_; then, according to Moseley’s theory, the ice will not descend. Let it be observed, however, that although the ice under this condition should descend (as there is little doubt but it would), it would show that Mr. Moseley’s theory of the descent of glaciers is incorrect, still it would not in the least degree affect the conclusions which he lately arrived at in regard to the generally received theory of glacier-motion. It would not prove that the ice sheared, in the way generally supposed, by its weight only. It might be the heat, after all, entering the ice, which accounted for its descent, although gravitation (the weight of the ice) might be the impelling cause.

According to this theory, the glacier, like the sheet of lead, must expand and contract as one entire mass, and it is difficult to conceive how this could account for the differential motion of the particles of the ice.

_Professor James Thomson’s Theory._—It was discovered by this physicist that the freezing-point of water is lowered by pressure. The extent of the lowering is equal to ·0075° centigrade for every atmosphere of pressure. As glacier ice is generally about the melting-point, it follows that when enormous pressure is brought to bear upon any given point of a glacier a melting of the ice at that particular spot will take place in consequence of the lowering of the melting-point. The melting of the ice will, of course, tend to favour the descent of the glacier, but I can hardly think the liquefaction produced by pressure can account for the motion of glaciers. It will help to explain the giving way of the ice at particular points subjected to great pressure, but I am unable to comprehend how it can account for the general descent of the glacier. Conceive a rectangular glacier of uniform breadth and thickness, and lying upon an even slope. In such a glacier the pressure at each particular point would remain constant, for there would be no reason why it should be greater at one time than at another. Suppose the glacier to be 500 feet in thickness; the ice at the lower surface of the glacier, owing to pressure, would have its melting-point permanently lowered one-tenth of a degree centigrade below that of the upper surface; but the ice at the lower surface would not, on this account, be in the fluid state. It would simply be ice at a slightly lower temperature. True, when pressure is exerted the ice melts in consequence of the lowering of the melting-point, but in the case under consideration there would, properly speaking, be no exertion of pressure, but a constant statical pressure resulting from the weight of the ice. But this statical condition of pressure would not produce fluidity any more than a statical condition of pressure would produce heat, and consequently motion could not take place as a result of fluidity. In short, motion itself is required to produce the fluidity.

I need not here wait to consider the sliding theories of De Saussure and Hopkins, as they are now almost universally admitted to be inadequate to explain the phenomena of glacier-motion, seeing that they do not account for the displacement of the particles of the ice over one another.

According to the dilatation theory of M. Charpentier, a glacier is impelled by the force exerted by water freezing in the fissures of the ice. A glacier he considers is full of fissures into which water is being constantly infiltrated, and when the temperature of the air sinks below the freezing-point it converts the water into ice. The water, in passing into ice, expands, and in expanding tends to impel the glacier in the direction of least resistance. This theory, although it does not explain glacier-motion, as has been clearly shown by Professor J. D. Forbes, nevertheless contains one important element which, as we shall see, must enter into the true explanation. The element to which I refer is the expansive force exerted on the glacier by water freezing.