236 



NA TURE 



[July 5, i»8 3 



be in a melting state. Hut glaciers are of all sizes and thick- 

 nesses, and they move in wmter as well as summer. Bessels 

 (" Die Amerikanische Nordpul Expedition," p. 398) measured 

 the motion of an Arctic glacier (not apparently very thick), in 

 the month of April, which is just when the winter cold would 

 have sunk deepe-t, and found it considerable. Again in the 

 Zeitschrift des deutschen Ceologischen Gtsellscha/t, vol. xxxiii. 

 p. 693, is an account of measurements of a Greenland glacier, 

 both in winter and summer, which show (hat the motion in 

 winter is only 20 per cent, less than in summer. It has been 

 suggested to me that the interior heat of the earth may be suf- 

 ficient to keep the bottom of the ice from freezing ; but this 

 cannot apply near the sides, where the ice is shallow, and the 

 freezing of a very small strip on each side would be sufficient to 

 keep the whole mass from descending. Moreover, this cause 

 would apply to masses of snow as much as to ice. Hut it is 

 known that masses of snow, though lying on steep slopes, do 

 not descend in this way, even in summer, but melt away where 

 they lie. 



(5.) According to the theories of Tyndall, Croll, anl others, 

 the glacier moves not in the form of ice hut of water. These 

 theories are based on the known fact that the freezing point of 

 ice is lowered by pressure. Hence it is supposed that certain 

 parts of a glacier are c mtinually being exposed to so much 

 pressure that they melt. The water escapes downward, and the 

 pressure being relieved it freezes again. The continuity of the 

 glacier is further kept up by the process of regelation, according 

 to which two pieces of ice, if placed in contact, form into one 

 solid mass. 



The advocates of this theory hardly seem to consider how very 

 small the lowering of the freezing point is for any ordinary 

 pressure. It is only '0075° per atmosphere. In other words, 

 it will require a pressure ot 2000 lbs. per square inch to liquefy 

 ice at 31° instead of 32". This is equivalent to the weight of a 

 column of ice about 5000 feet high. It is needless to ask 

 whether such a pressure can exist within an ordinary glacier, 

 while on the other hand glaciers undoubtedly move at tempera- 

 tures far below freezing point — in the Arctic regions below zero. 



It seems to be generally supposed that the pressure in the 

 lower part of a glacier is due to the steeper upper portions : the 

 glacier channel is spoken of as a mould, through which the ice 

 is forced by pressure from behind. But in the upper glacier, 

 slopes of ice or nevi are not uncommon at angles of 30 or even 

 more. Such slopes Usually do not even touch the more level parts of 

 the glacier below them, but are separated from them by a wide, 

 deep crevasse calle 1 a Bergschrund. Of this the well-known 

 ice wall of the Strahleck is a conspicuous example. In other 

 cases such sloj.es do not end in a glacier at all, but die away 

 upon the mountain side. It is certain, therefore, that ice or 

 ntvi is able to maintain itself at a high angle upon its slope of 

 rocks, and therefore cannot possibly exercise pressure upon the 

 parts of the glacier far in advance of its foot. The fallacy of 

 1 his idea may be further illustrated by referring, not to modern 

 glaciers, but to those of the Great Ice Age. Can we suppose 

 that the pressure of the snows about the sources of the Rhone 

 was sufficient to drive that glacier down the valley to Martigny, 

 round a sharp angle to the Lake of Geneva, through the bed of 

 that lake, and on to the slopes of the Jura, a distance of more 

 than too miles, in which the average slope was about I in 200 ; 

 giving a propelling force per ton of ice of about n lbs. only? 



All these theories have this in common, that they regard 

 gravity as the sole and direct agent in the movement of glaciers, 

 and the above considerations seem to prove that it is an agent 

 far too weak for the work it has to do. 1 



The only other agent which has been suggested, or seems 

 likely to be suggested, to account for the motion of glaciers, is 

 heat. This suggestion, as is well known, is due to the late 

 Canon Moseley, F.R.S., and was to some extent worked out by 

 him in papers published in the Phil. Mag., 1869 and 1870. 



The mode of operation, on this theory, is well known. Ice 

 is here considered merely as a solid bidy, obeying the ordinary 

 laws of expansion and contraction under differences of tempera- 

 ture. TLis it is known to do, the coefficient of linear expansion, 

 for 1° F., being '00002856 (Moseley, Phil. Mag., January, 1870), 

 which is very high. When a mass of ice, such as a glacier, 

 suffers a rise in temperature, either through conduction or radia- 



1 Another evidence against pressure from behind as a cause of motion is 

 furnished by the very small size of many glaciers. Some of these, nitably 

 those of the class called " glaciers remanies," are only a few hundred yards 

 long, and cannjt be many fett deep. 



tion, it will expand ; this expansion will take place mainly in 

 the direction where movement is easiest, that is, clown the valley. 

 If from any cause the temperature falls, the glacier w ill again 

 contract ; but since the expansion is assisted by gravity whilst 

 the contraction is opposed by it, the latter will be somewhat less 

 in amount than the former, and when the ice has returned to its 

 original temperature, its centre of gravity will have moved a 

 certain small distance down the valley. By such alternate 

 expansions and contractions the glacier moves gradually from the 

 top to the bottom of its course. 



That variations of temperature d> take place in a glacier 

 cannot be doubted, whatever be the condition in which it lies. 

 This granted, the fact that it should move in the way 

 described appears to me no more surprising than that the sheets 

 of le :d on which Canon Moseley made his well-known experi- 

 ments did so move ; and that the motion thus produced is of the 

 character which answers to all the facts of the case, so far as 

 they are at present known, can, I believe, be established. 



The controversy occa i med by Canon Moseley's articles was 

 unfortunately terminated by his illness and death, before the 

 matter had been fully cleared up. The main objections urged 

 to his theory were two. The first was that a glacier is not one 

 continuous body (as assumed by Canon Moseley in his mathe- 

 matical investigation), but is broken up into many parts by 

 crevasses. But in the first place, the assumption above men- 

 tioned is merely one of convenience, and not in the least neces- 

 sary to the theory. A detached piece of ice would move in the 

 same way as a glacier, or as the sheet of lead did in Canon 

 Mo-eley's experiments. Secondly, if a glacier is anywhere 

 divided in its whole thickness by a crevasse, this is absolutely 

 fatal to the gravitation theories, since there can be no pressure 

 between the portions above and below this division. The only 

 possible explanation of crevasses, on these theories, is that they 

 are due to the glacier bending over a convex part of its bed. In 

 that case the bottom half will be in compression, and only the 

 top half in tension, so that the crevasse cannot possibly extend 

 more than half way through the thickness. 



The second objection was that the conductivity of ice is low ; 

 hence the effect of the heat would be confined to the layers near 

 the surface, and could not account for the motion of the glacier 

 as a wh de. This objection does not seem to be confirmed by 

 careful reflection upon the way in which such forces act. Let us 

 suppo-e a glacier 100 feet deep, of which each successive foot 

 expands and contracts alike throughout, but adheres with a 

 definite shearing resistance to the layers above and below. Let 

 there be a rise in temperature, which does not extend beyond the 

 uppermost 10 feet. This layer will expand, and if it were free 

 would expand to the full amount due to the increase in tempera- 

 ture. But its lower surface is not free. In expanding it will 

 therefore drag the next layer after it, or in other words will 

 cause it to expand also. The amount of expansion, however, 

 will not be so great, because there will be a certain shearing ex- 

 tension at the plane of division between the two. The second 

 layer will similarly cause an expansion in the third, and so on to 

 the bottom. In consequence the energy which would all have 

 been exerted on the top layer, had that been free, will be distri- 

 buted over the whole of the layers ; and the extension of the top 

 layers will of course be much smaller than it otherwise would 

 have been. Should the temperature then remain constant, the 

 layers will retain their position, and adapt themselves to the new 

 circumstances. If the temperature falls, the layers will contract, 

 but from the now opposing effect of gravity they will not return 

 to their original po ition. The top layer, which has extended 

 furthest, will be the furthest below its original position ; the 

 second layer next and so on. If we suppose the layers to be 

 indefinitely thin, we have the condition of things in an actual 

 glacier. The ice in any vertical section will, on the whole, move 

 down the slope, but the top will move faster than the middle, 

 and the middle than the bottom, exactly as it is known to do. 

 The same holds with regard to a horizontal section. At the sides 

 the ice vt ill be held back, not only by the friction, but also by 

 the protuberances of the rock, which compel the ice to shear 

 over them. Hence the velocity there will be retarded, and will 

 be less than that in the middle, which is comparatively free. 



A more important objection remains to be considered, which 

 is this. On the present theory the motion at any point on the 

 surface of a glacier will be not continuous, but oscillating alter- 

 nately downwards and upwards, and the net distance by which 

 it has descended, say. in a day, will be a mere fraction of the 

 total distance through which it has moved in that period. If 



