Observations On The Mer De Glace
( Originally Published 1897 )
THE law established by Forbes and Agassiz, that the central portions of a glacier moved faster than the sides, was amply illustrated and confirmed by the deportment of lines of stakes placed across the Mer de Glace and its tributaries in 1857. The portions of the trunk glacier derived from these' tributaries were easily traceable throughout the glacier by means of the moraines. Thus, for example, the portion of the trunk stream derived from the Glacier du Géant might be distinguished in a moment from the other portions by the absence of débris upon its surface. Attention was drawn by Prof. Forbes to the fact that the eastern side of the Mer de Glace in particular is ' excessively crevassed ;' and he accounted for this crevassing by supposing that the Glacier du Géant moves most swiftly, and in its effort to drag its more sluggish companions along with it tears them asunder, thus producing the fissures and dislocation for which the eastern side of the glacier is remarkable. Too much weight must not be attached to this explanation.
It was one of those suggestions which are perpetually thrown out by men of science during the course of an investigation, and the fulfilment or non-fulfilment of which cannot materially affect the merits of the investigator. Indeed, the merits of Forbes must be judged on far broader grounds. The qualities of mind and the physical culture invested in his 'Travels in the Alps' are such as to make it, in the estimation of the physical investigator at least, outweigh all other books upon the subject.
While thus acknowledging its merits, however, let a free and frank comparison of its statements with facts be instituted. To test whether the Glacier du Géant moved more quickly than its' fellows, five different lines were set out across the Mer de Glace, in the vicinity of the Montanvert. In each case it was found that the point of swiftest motion did not lie upon the Glacier du Géant at all, but was displaced so as to bring it comparatively close to the eastern side of the glacier. But though the special opinion of Forbes just referred to here falls to the ground, the deviation of the point of swiftest motion from the centre of the glacier will probably, when its cause is pointed out, be regarded as of special importance to his theory.
At the place where these five lines were run across it the glazier turns its convex curvature to the eastern side of the valley, being concave towards the Montanvert. Let us then take a bolder analogy than even that suggested in the explanation of Forbes, where he compares the Glacier du Géant to a strong and swiftly flowing river. Let us enquire how a river would behave in sweeping round a curve similar to that here existing. The point of swiftest motion would undoubtedly lie on that side of the centre of the stream towards which it turns its convex curvature. Can this be the case with the trunk of the Mer de Glace ? If so, then we ought to have a shifting of the point of maximum motion towards the eastern side of the valley, when the curvature of the glacier so changes as to turn its convexity to the western side.
Now, such a change of flexure actually occurs opposite the passages called Les Ponts, and at this place the view just enunciated was tested. It was immediately ascertained that the point of swiftest motion here lay at a different side of the axis from that observed lower down. But to confer strict numerical accuracy upon the result, stakes were fixed at certain distances from the eastern side. The results of this measurement are given in the following table, the numbers denoting inches :
1st pair 2nd pair 3rd pair 4th pair 5th pair
West 15 17 22 23 234
It is here seen that in each case the western stake moved more swiftly than its eastern fellow stake ; thus proving, beyond a doubt, that opposite the Ponts the western side of the Mer de Glace moves swiftest—a result precisely the reverse of that observed where the curvature of the valley was different.
But an additional test of the explanation is possible. Between the Ponts and the promontory of Trélaporte the glacier passes another point of contrary flexure, its convex curvature opposite to Trélaporte being turned towards the base of the Aiguille du Moine, on the eastern side. A series of stakes was placed across the glacier here; and the velocities of those placed at certain distances from the western side were compared, as before, with those of stakes placed at the same distances from the eastern side. The following table shows the result of these measurements ; the numbers, as before, denote inches :
1st pair 2nd pair 3rd pair
West . 12 15 17
Here we find that in each case the eastern stake moved faster than its fellow. The point of maxi-mum motion has therefore once more crossed the axis of the glacier.
Determining the point of maximum motion for a great number of transverse sections of the Mer de Glace, and uniting these points, we have what is called the locus of the point. The dotted line in the annexed figure represents the centre of the Mer de Glace ; the hard line which crosses the axis of the glacier at the points A A is then the locus of the point of swiftest motion. It is a curve more deeply sinuous than the valley itself, and it crosses the central line of the valley at each point of contrary flexure. The position of towns upon the banks of rivers is usually on the convex side of the stream, where the rush of the water renders silting-up impossible ; and the same law which regulated the flow of the Thames, and determined the position of the towns upon its banks, is at this moment operating with silent energy among the Alpine glaciers.
Another peculiarity of glacier motion is now to be noticed.
Before any observations had been made upon the subject, it was surmised by Prof. Forbes that the portions of a glacier near its bed were retarded by friction against the latter. This view was afterwards confirmed by his own observations, and by those of M. Martins. Nevertheless the state of our knowledge upon the subject rendered further confirmation of the fact highly desirable. A rare opportunity for testing the question was furnished in 1857 by an almost vertical precipice of ice, constituting the side of the Glacier du Géant, exposed near the Tacul. The precipice was about 140 feet in height. At the top and near the bottom stakes were fixed, and by hewing steps in the ice I succeeded in fixing a stake in the face of the precipice at a point about forty feet above the base.' After the lapse of a sufficient number of days, the progress of the three stakes was measured ; reduced to the diurnal rate, the motion was as follows':
Top stake 6.00 inches.
We thus see that the top stake moved with more than twice the velocity of the bottom one, while the velocity of the middle stake lies between the two. But it also appears that the augmentation of velocity upwards is not proportional to the distance from the bottom, but increases in a quicker ratio. At a height of 100 feet from the bottom, the velocity would undoubtedly be practically the same as at the surface. Measurements made upon an adjacent ice-cliff proved this. We thus see the perfect validity of the reason assigned by Forbes for the continued verticality of the walls of transverse crevasses. Indeed a comparison of the result with his anticipations and reasonings will prove alike their sagacity and their truth.
The most commanding view of the Mer de Glace and its tributaries is obtained from a point above the remarkable cleft in the mountain-range underneath the Aiguille de Charmoz, which is sure to attract the attention of an observer standing at the Montanvert. This point, marked G on the map of Forbes, I succeeded in attaining. A Tubingen Professor once visited the glaciers of Switzerland, and seeing these apparently rigid masses enclosed in sinuous valleys, went home and wrote a book, flatly denying the possibility of their motion. An inspection from the point now referred to would have doubtless confirmed him in his opinion ; and indeed nothing can be more calculated to impress the mind with the magnitude of the forces brought into play than the squeezing of the three tributaries of the Mer de Glace through the neck of the valley at Trélaparte.
But let me state numerical results. Previous to its junction with its fellows, the Glacier du Géant measures 1,134 yards across. Before it is influenced by the thrust of the Talèfre, the Glacier de Léchaud has a width of 825 yards; while the width of the Talèfre branch across the base of the cascade, before it joins the Léchaud, is approximately 638 yards. The sum of these widths is 2,597 yards. At Trélaporte those three branches are forced through a gorge 893 yards wide, with a central velocity of 20 inches a day ! The result is still more astonishing if we confine our attention to one of the tributaries—that of the Léchaud. This broad ice-river, which before its junction with the Talèfre has a width of 825 yards, at Trélaporte is squeezed to a driblet of less than 88 yards in width, that is to say, to about one-tenth of its previous horizontal trans-verse dimension.
Whence is the force derived which drives the glacier through' the gorge? No doubt pressure from behind. Other facts also suggest that the Glacier du Géant is throughout its length in a state of forcible longitudinal compression. Taking a series of points along the axis of this glacier--if these points, during the descent of the glacier, preserved their distances asunder perfectly constant, there could be no longitudinal compression. The mechanical meaning of this term, as applied to a substance capable of yielding like ice, must be that the hinder points are incessantly advancing upon the forward ones. I was particularly anxious to test this view, which first occurred to me on à priori grounds. Three points, A, B, C, were therefore fixed upon the axis of the Glacier du Géant, A being the highest up the glacier. The distance between A and B was 545 yards, and that between B and C was 487 yards. The daily velocities of these three points, determined by the theodolite, were as follows : .
A. 20'55 inches.
The result completely corroborates the foregoing anticipation. The hinder points are incessantly advancing upon those in front, and that to an extent sufficient to shorten a segment of this glacier, measuring 1,000 yards in length, at the rate of 8 inches a day. Were this rate uniform at all seasons, the shortening would amount to 240 feet in a year. When we consider the compactness of this glacier, and the uniformity in the width of the valley which it fills, this result cannot fail to excite surprise ; and the exhibition of force thus rendered manifest must be mainly instrumental in driving the glacier through the jaws of the granite vice at Trélaporte.
When the Glacier du Géant is observed ,from a sufficient distance, a remarkable system of seams of white ice appears to sweep across it, in the direction of the ' dirt-bands.' These seams are more resistant than the ordinary ice of the glacier, and sometimes protrude above the surface to a height of three or four feet. Their origin was for some time a difficulty, and it was at the base of the ice-cascade which descends from the basin of the Talèfre that the key to their solution first presented itself. It was well known that the ice of a glacier is not of homogeneous structure, but that the general more or less milky mass is traversed by blue veins of a more compact and transparent texture. In the upper portions of the Mer de Glace these veins sweep across the glacier in gentle curves, leaning forward—to which leaning forward Prof. Forbes gave the name of the 'frontal dip.' A case of ' backward dip' has never been described. But at the base of the ice-cascade referred to I had often noticed the veins exposed upon the walls of a longitudinal crevasse leaning backwards and for-wards on both sides of a vertical line, like the joints of stones used to turn an arch.
This fact was found to connect itself in the following way with the general state of the glacier. At the base of the ice-fall a succession of protuberances, with steep frontal slopes, followed each other, and were intersected by crevasses. Let the hand be placed flat upon the table, with the palm down-wards; let the fingers be bent so as to render the space between the joints nearest the nails and the ends of the fingers nearly vertical. Let the second hand be now placed upon the back of the first, with its fingers bent as in the former case, and their ends resting upon the roots of the first fingers. The crumpling of the hands fairly represents the crumpling of the ice, and the spaces between the fingers represent the crevasses by which the protuberances are intersected. On the walls of these crevasses the change of dip of the veined structure above referred to was always observed, and at the base of each protuberance a vein of white ice was found firmly wedged into the mass of the glacier.
The next figure represents a series of these crumples with the veins of white ice i i i at their bases.
It was soon observed that the water which trickled down the protuberances, and gushed here and there from glacier orifices, collected at the bases of the crumples, and formed streams which cut for them-selves deep channels in the ice. These streams seemed to be the exact matrices or moulds of the veins of white ice, and the latter were finally traced to the gorging up of the channels of glacial rivulets by winter snow. The same explanation applies to the system of bands upon the Glacier du Géant. I was enabled to trace the little arms of white ice which once were the tributaries of the streams, to see a trunk vein of the ice dividing into branches, and uniting again so as to enclose glacial islands. I finally traced them to the region of their formation, and by sketches of existing streams taken near the base of the séracs, and of bands of white ice taken lower down, a resemblance so striking was exhibited as to leave no doubt of their relationship. On the walls of some deep crevasses, moreover, which intersected the white ice-seams, I found that the latter penetrated the glacier only to a limited depth, having the appearance of a kind of glacial 'trap' intruded from above.
But how is the backward dip of the blue veins to be accounted for ? Doubtless in the following way : At the base of the cascade the glacier is forcibly compressed by the thrust of the mass behind it ; besides this, it changes its inclination suddenly and considerably ; it is bent upwards, and the consequence of this bending is a system of wrinkles, such as those represented in the next figure. The interior of a bent umbrella-handle sometimes presents wrinkles which are the representatives, in little, of the protuberances upon the glacier. The coat-sleeve is an equally instructive illustration : when the arm is bent at the elbow the sleeve wrinkles, and as the places where these wrinkles occur in the cloth are determined, to some extent, by the previous creasing, so also the places where the wrinkles are formed upon the glacier are determined by the previous scarring of the ice during its descent down the cascade. The manner in which these crumples tend to scale off speaks strongly in favour of the ex-planation given. The following figure represents a type of numerous instances of scaling off. By means of a hydraulic press it is easy to produce a perfectly similar scaling in small masses of ice. One consequence of this crumpling of the glacier would be the backward and forward inclination of the veins as actually observed. The same appearance was noticed on the wrinkles of the Glacier du Géant. It was also proved, by measurements, that these wrinkles shorten as they descend.
In virtue of what quality, then, can ice be bent and squeezed, and have its form changed in the manner indicated in the foregoing observations? The only theory worthy of serious consideration at the pre-sent day is the celebrated Viscous Theory of glacial motion. Numerous appearances, as we have seen, favour the idea that ice is a viscous or ' semi-fluid' substance, and that it flows as such in the glaciers of the Alps. The aspect of many glaciers, as a whole —their power of closing up crevasses, and of reconstructing themselves after having been precipitated down glacial gorges—the obvious bendings and contortions of various portions of the ice, are all in harmony with the notion. The laminar structure of the glacier has also been regarded by eminent authorities as a crucial test in favour of the viscous theory, and affirmed to be impossible of explanation on any other hypothesis.
Nevertheless, this theory is so directly opposed to our ordinary experience of the nature of ice as to leave upon the mind a lingering doubt of its truth. Can we imitate the phenomena without invoking the explanation ? We can. Moulds of various forms were hollowed out in boxwood, and pieces of ice were placed in these moulds and subjected to pressure. In this way spheres of ice were flattened into cakes, and cakes formed into transparent lenses. A straight bar of ice, six inches long, was passed through a series of moulds augmenting in curvature, and was finally bent into a semiring. A small block of ice was placed in a hemispherical cavity, and was pressed upon by a hemispherical protuberance, not large enough to fill the cavity; the ice yielded and filled the space between both, thus forming itself into a transparent cup. The specimens of ice here employed were so exceedingly brittle that a pricker driven into the ice was competent to split blocks of the substance eight cubic feet in volume, the surface of fracture being in all cases as clean and sharp as that of glass.
These experiments, then, demonstrate a capacity on the part of small masses of ice which they have not been hitherto known to possess. They prove, to all appearance, that the substance is much more plastic than it was ever imagined to be. But the real germ from which these results have sprung is to be found in a lecture given at the Royal Institution in June 1850, and reported in the ' Athenaeum' and ' Literary Gazette' for that year. Faraday then showed that when two pieces of ice, at a temperature of 32° Fahr., are placed in contact with each other, they freeze together, by the conversion of the film of moisture between them into ice. The case of a snowball is a familiar illustration of the principle. When the snow is below 32°, and therefore dry, it will not cohere, whereas when it is in a thawing condition it can be squeezed into a hard mass. During one of the hottest days of July 1857, when the thermometer was upwards of 100° Fahr. in the sun, and more than 80° in the shade, I observed a number of blocks of ice, which had been placed in a heap, frozen together at their places of contact ; and I afterwards caused them to freeze together under water as hot as the hand could bear. Facts like these suggested the thought that if a piece of ice.—a straight prism, for example—were placed in a bent mould and subjected to pressure it would break, but that the force would also bring its ruptured surfaces into contact, and thus the continuity of the mass might be re-established. Experiment, as we have seen, completely confirmed this surmise : the ice passed from a continuous straight bar to a continuous bent one, the transition being effected, not by a viscous movement of the particles, but through fracture and regelation.
Let the transition from curve to curve be only gradual enough, and we have the exact case of a transverse slice of a glacier.
All the phenomena of motion, on which the idea of viscosity has been based, are brought by such experiments as the above into harmony with the demonstrable properties of ice. In virtue of this property, the glacier accommodates itself to its bed while pre-serving its general continuity, crevasses are closed up, and the broken ice of a cascade, such as that of the Talèfre or the Rhone, is recompacted to a solid continuous mass.
The very essence of viscosity is the ability of yielding to a force of tension, the texture of the substance, after yielding, being in a state of equilibrium, so that it has no strain to recover from ; and the substances chosen by Prof. Forbes as illustrative of the physical condition of a glacier possess this power of being drawn out in a very eminent degree. But it has been urged, and justly urged, that we ought not to conclude that viscosity is absent because hand specimens are brittle, any more than we ought to conclude that ice is not blue because small fragments of the substance do not exhibit this colour. To test the question of viscosity, then, we must appeal to the glacier itself. Let us do so.
An analogy between the motion of a glacier through a sinuous valley and of a river in a sinuous channel has been already pointed out. But the analogy fails in one important particular : the river, and much more so a mass of flowing treacle, honey, tar, or melted caoutchouc, sweeps round its curves. without rupture of continuity. The viscous mass stretches, but the icy mass breaks, and the ' excessive crevassing' pointed out by Prof. Forbes himself is the consequence. The inclinations of the Mer de Glace and its three tributaries were, moreover, taken, and the association of transverse crevasses with the changes of inclination were accurately noted. Every traveller knows the utter dislocation and confusion produced by the descent of the Mer de Glace from the Chapeau downwards. A similar state of things exists in the ice-cascade of the Talèfre. Descending from the Jardin, as the ice approaches the fall, great transverse chasms are formed, which at length follow each other so speedily as to reduce the ice-masses between them to mere plates and wedges, along which the explorer has to creep cautiously. These plates and wedges are in some cases bent and crumpled by the lateral pressure, and some large pyramids are turned 90° round, so as to have their veins at right angles to the normal position. The ice afterwards descends the fall, the portions exposed to view being- a fantastic assemblage of frozen boulders, pinnacles, and towers, some erect, some leaning, falling at intervals with a sound like thunder, and crushing the ice-crags on which they fall to powder. The descent of the ice through this fall has been referred to as a proof of its viscosity; but the description just given does not harmonise with our ideas of a viscous substance.
But the proof of the non-viscosity of the substance must be sought at places where the change of inclination is very small. Nearly opposite l'Angle there is a change from four to nine degrees, and the consequence is the production of transverse fissures which render the glacier here perfectly impassable. Further up the glacier transverse crevasses are produced by a change of inclination from three to five degrees. This change of inclination is protracted Again, the crevasses being due to a state of strain from which the ice relieves itself by breaking, the rate at which they widen may be taken as a measure of the amount of relief demanded by the ice. Both the suddenness of their formation and the slowness with which they widen are demonstrative of the non-viscosity of the ice. For were the substance capable of stretching, even at the small rate at which they widen, there would be no necessity for their formation.
Further, the marginal crevasses of a glacier are known to be a consequence of the swifter flow of its central portions, which throws the sides into a state of strain from which they relieve themselves by breaking. Now it is easy to calculate the amount of stretching demanded of the ice in order to accommodate itself to the speedier central flow. Take the case of a glacier half a mile wide. A straight transverse element, or slice, of such a glacier, is bent in twenty-four hours to a curve. The ends of the slice move a little, but the centre moves more : let us suppose the versed side of the curve formed by the slice in twenty-four hours to be a foot, which is a fair average. Having the chord of this arc, and its versed side, we can calculate its length. In the case of the Mer de Glace, which is about half a mile wide, the amount of stretching demanded would be about the eightieth of an inch in twenty-four hours. Surely, if the glacier possessed a property which could with any propriety be called viscosity, it ought to be able to respond to this moderate demand ; but it is not able to do so : instead of stretching as a viscous body, in obedience to this slow strain, it breaks as an eminently fragile one, and marginal crevasses are the consequence. It may be urged that it is not fair to distribute the strain over the entire length of the curve : but reduce the distance as we may, a residue must remain, which is demonstrative of the non-viscosity of the ice.
To sum up, then, two classes of facts present themselves to the glacier investigator—one class in harmony with the idea of viscosity, and another as distinctly opposed to it. Where pressure comes into play we have the former ; where tension comes into play we have the latter. Both classes of facts are reconciled by the assumption, or rather the experimental verity, that the fragility of ice and its power of regelation render it possible for it to change its form without prejudice to its continuity.