ish bands on the glacier, the curves pointing downwards, and the two branches mingling indiscriminately with the (lateral) moraines, presenting an appearance of waves some hundred feet apart, and having, opposite to the Montanvert, the form which I have attempted to show upon the map, where they are represented in the exact figure and number in which they occur.

6

. I was satisfied, from the general knowledge which I then had of the veined structure' of the ice, that these coloured bands probably followed that direction."† Farther examination confirmed this conjecture, and showed that these superficial discolorations in the form of excessively elongated hyperbolas are due to the recurrence (at intervals of some hundred feet along the course of the glacier) of portions of ice in which the veined structure is more energetically developed than elsewhere, and where, by the decomposition of the softer laminæ, portions of sand and dirt become entangled in the superficial ice, and give rise to the phenomena of dirt-bands, which thus at a distance display (though in a manner requiring some attention to discover) the exact course of this singular structure on the surface of the glacier. The annexed figure 26 displays the superficial form of the dirt-bands, and the course of the structural laminæ projected

Fig. 27.

Fig. 28.

Fig. 26. horizontally. Fig. 27 shows an ideal transverse section of the glacier; and Fig. 28 another vertical section parallel to its length. These three sections in rectangular planes will serve to give a correct idea of the course of this remarkable structure within the ice, but a more popular conception will be formed of it from the imaginary sections of a canal-shaped glacier in the annexed woodcut, Fig. 29. The structure of the compound glacier, originally double, becomes gradually single; and the frontal Map of the Mer de Glace of Chamouni, etc., in Forbes's Travels in the Alps, or in the Tour of Mont Blanc, etc.

t Travels in the Alps of Savoy, etc., 2d edit., p. 162.

COURSE OF VEINED STRUCTURE-MOTION OF GLACIERS. 247

dip of the lamina at the loop of the horizontal curves, which in the upper region of the glacier is nearly vertical, gradually slopes forwards, until, at the lower termination, it has a very slight dip inwards, or, indeed, may be reversed, and fall outwards and forwards. The general form of a structural lamina of a glacier rudely resembles that of a spoon.

This structure and the accompanying dirt-bands have been recognised by different observers in almost all glaciers, including those of Norway and of India. The interval between the dirtbands has been shown in the case of the Mer de Glace (and therefore probably in other cases) to coincide with annual rate of progression, and in the higher parts of the glacier (towards the névé) to be accompanied by wrinkles or inequalities of the surface, which are well marked by the snow lying in them during the period of its partial disappearance.*

The Motion of Glaciers, and its causes.-The most characteristic and remarkable feature of glaciers is their motion down

wards from the névé towards the lower valley. The explanation

* Fifth Letter on Glaciers. Edinburgh Philosophical Journal, 1844 [reprinted in the present volume]; and Travels in the Alps, 2d edit.

of it is by far the most important application of mechanical physics connected with the subject.

Obvious as the fact itself must appear by what has been already stated, manifest confusion has obtained in the minds of intelligent persons regarding it. Thus Ebel, in his well-known Swiss guide-book, affirms the motion of the glaciers of Chamouni to be 14 feet, and those of Grindelwald 25 feet in a year; quantities which, if they have any meaning, must refer to the apparent advance of the lower termination of those glaciers into the valley, which therefore only indicate the difference of the real motion, and of the waste in any particular season, and which may become null, or even negative, if the summer be more than usually warm. The peasants, however-who are inevitably made aware of the progressive motion of the ice by observing the progressive advance of conspicuous blocks on its surface-commonly ascribe to the glaciers the more correct measure of several hundred feet per annum.

M. Hugi, of Soleure, measured, with some accuracy, year by year, the progress of a conspicuous block on the glacier of the Aar, which he found to be 2200 feet in nine years, or about 240 feet per annum.* M. Agassiz continued some of these annual measures, but only in a rough way, by causing his guides to reckon the distance of a block on the moraine by lengths of a pole or rod from a fixed rock some thousand feet off. These measures appear not to have been altogether trustworthy.

The principal theories to account for the progressive motion of glaciers which were prevalent previous to 1842, may be briefly characterised as De Saussure's and De Charpentier's, though each had been maintained in times long antecedent by the earlier Swiss writers. The first may for brevity be called the gravitation theory, the latter the dilatation theory. Both suppose that the motion of the ice takes place by its sliding bodily over its rocky bed, but they differ as to the force which urges Agassiz, Etudes sur les Glaciers, p. 150.

GRAVITATION THEORY-DILATATION THEORY.

249

it over the obstacles opposed by friction and the irregularities of the surface on which it moves.

The following quotation from De Saussure explains his views with his usual precision:-"These frozen masses, carried along by the slope of the bed on which they rest, disengaged by the water (arising from their fusion owing to the natural heat of the earth) from the adhesion which they might otherwise contract to the bottom-sometimes even elevated by the water -must gradually slide and descend along the declivity of the valleys or mountain slopes (croupes) which they cover. It is this slow but continual sliding of the icy masses (des glaces) on their inclined bases which carries them down into the lower valleys, and which replenishes continually the stock of ice in valleys warm enough to produce large trees and rich harvests.' Very sufficient objections have been urged against this theory. It is evident that De Saussure considered a glacier as an accumulation of icy fragments, instead of a great and continuous mass, throughout which the fissures and crevasses bear a small proportion to the solid portion; and that he has attributed to the subglacial water a kind and amount of action for which there exists no sufficient or even probable evidence. The main objection, however, is this, that a sliding motion of the kind supposed, if it commence, must be accelerated by gravity, and the glacier must slide from its bed in an avalanche. The small slope of most glacier-valleys, and the extreme irregularity of their bounding walls, are also great objections to this hypothesis.

The Dilatation theory ingeniously meets the difficulty of the want of a sufficient moving power to drag or shove a glacier over its bed, by calling in the well-known force with which water expands on its conversion into ice. The glacier being traversed by innumerable capillary fissures, and being in summer saturated with water in all its parts, it was natural to invoke the freezing action of the night to convert this water into ice, and by the amount of its expansion to urge the glacier onwards in the direction of its greatest slope. In answer to

* Voyages dans les Alps, sec. 535.

this, it is sufficient to observe, in the first place, that during the height of summer the portions of those glaciers which move fastest are never reduced below the freezing point, and that even in the most favourable cases of nocturnal radiation producing congelation at the surface, it cannot (by well-known laws of conduction) penetrate above a few inches into the interior of the glacier. Again, the ascertained laws of glacier motion are (as will be immediately seen) entirely adverse to this theory, as it is always accelerated by hot weather and retarded by cold, yet does not cease even in the depth of winter.*

It is singular how slow observers were to perceive the importance to the solution of the problem of glacier motion of ascertaining with geometrical precision the amount of motion of the ice, not only from year to year, but from day to day; whether constant or variable at the same point, whether continuous or by starts; if variable, on what circumstances it depended, and in what manner it was affected at different points of the length and breadth of a glacier.

This method of studying the question was taken up by the writer of this paper. His observations were commenced on the Mer de Glace of Chamouni, in June 1842. Between the 26th and 27th of that month the motion of the ice opposite a point called the "Angle" was found, by means of a theodolite, to be 16.5 inches in 26 hours; between the 27th and 28th, 17.4 inches in 25 hours; and from about 6 A.M. to 6 P.M. on the 28th the motion was 9.5 inches, or 17.5 inches in 24 hours; whilst even the proportional motion during an hour and a half was observed. No doubt could therefore remain that the motion of the ice is continuous and tolerably uniform-in short, that it does not move by jerks. He also ascertained about the same time that the motion of the ice is greatest towards the centre of a glacier and slower at the sides, contrary to an opinion then maintained on high authority. He next found that the rate of

*The fullest exposition of the Dilatation Theory is to be found in De Charpentier, Essai sur les Glaciers (1841), and Agassiz, Etudes sur les Glaciers (1840).