mass by the process of regelation. We are not here speaking of the nature of this process of the molecular actions which may be involved in it. We are appealing merely to the result of that process as an observed fact; and the fact itself may manifestly be made the base of our speculations, without our knowing the modus operandi of the process, just as we may reason upon the facts or results of crystallization, notwithstanding our ignorance of the physical process by which those results are produced. We are the more anxious to point out this distinction because we imagine that we discern a disposition on the part of some glacialists, in the application of regelation to the explanation of the motion of a glacier, to depart from the facts or results of regelation, with which we are acquainted, to the modus operandi, with which we are not acquainted. The term 'regelation' has been objected to as seeming to indicate the nature of the process by which the effect above described is produced; but it must be distinctly understood that when we speak of the property of regelation' as characterising ice at the particular temperature of 32° (Fahr.), we mean simply that property in virtue of which ice at that temperature is capable of being broken and fractured, and instantly reunited into a continuous mass, as above described. We shall see in the sequel the great importance of this property of ice, in the theory of the motion of glaciers.

We may add that Dr. Tyndall has varied the above experiment in several ways, as may be seen by referring to his 'Glaciers of the Alps,' p. 346, or to his Memoirs in the Transactions of the Royal Society.

The modus operandi in the conversion of snow into the compact ice of the lower glacier, is intimately connected with the internal temperature of the mass. In the colder glacial regions the falling snow is usually dry, and consists of fine granules; but when the atmosphere is more moist, and its temperature little exceeds that of freezing, the snow is flocculent. During the winter a thick covering of snow is deposited on the glacier; but, below the snow-line, the whole of this snow, together with the superficial portion of the pre-existing glacier beneath it, is dissolved by the heat of the following summer. Above the snow-line, on the contrary, a part only of the previous winter's snow is dissolved, and the other part remains as a permanent addition to the glacier, thus forming an annual stratum which may or may not be afterwards recognisable as distinct from similar strata above or below it. When the summer warmth begins to predominate in these higher regions, the superficial snow is melted by the sun's rays, though the atmospheric temperature may be considerably below 32°. The water thus

produced

produced sinks into the porous mass of snow, the temperature of which will necessarily be below-and in the highest regions considerably below-32°(Fahr.). This percolating water will therefore become partly frozen, as above intimated, the depth to which the infiltration proceeds depending on circumstances. The portion of the last winter's snow which remains at the end of the summer thus becomes changed into a granular mass, while the mass immediately below it will also be further modified in like manner. The more superficial portion of the whole mass thus transformed becomes granular, and is called névé; it becomes more and more consolidated as the depth increases, till it finally assumes the character of compact glacial ice. We should expect the mass thus formed to be stratified, but that its indications of stratification would be feeble. It is in this manner that the glacial mass increases above the snow-line, to compensate for the waste below it.

In the higher regions in which glaciers originate, the minimum superficial winter temperature will frequently be much less than that determined, as above stated, by M. Agassiz in the middle region of the Aar glacier, though the winter covering of snow will tend to equalise these temperatures in different localities. Whatever effect, however, may be produced by a lower atmospheric temperature in the higher glacial regions, the tendency of the infiltrated water, as above explained, must always be to raise the temperature to that of freezing in the lower and far greater part of the mass into which the winter cold never penetrates. Allowing this influence of infiltration, the lower portion of the glacial mass will have the same temperature in these higher and colder regions as in the milder middle and lower regions of the glacier; but the portion affected by the winter temperature will be generally colder and its depth greater where the mean external atmospheric temperature is the lowest, and especially in winter.

The conversion of snow into névé, and subsequently into consolidated ice, has been a subject of frequent discussion. The views of all the earlier glacialists, and of some also of the later 'ones, were founded on conceptions more or less erroneous respecting the internal temperature of glaciers. Pressure is a cause, as well as temperature, to which this conversion has been attributed (p. 78). M. Agassiz has described his experiments and stated his views more explicitly than any other glacialist in Chapter V. of his 'Système Glaciaire.' He probably erred in attributing too much importance to the interior temperature. Principal Forbes, in his earlier speculations, appears to have recognised congelation, due to the winter temperature, as the effective cause

in producing the transmutation we are speaking of; but he afterwards rejected this idea, and adopted the opinion that it was due to pressure alone; for, in 1846, he writes, "I am satisfied, then (and it is only after long doubt that I venture this confident expression), that the conversion of snow into ice is due to the effects of pressure on the loose and porous structure of the former.' To the operation of direct pressure he adds that of the 'kneading or working of the parts on one another,' due to a difference of motion of two contiguous particles and the consequent friction between them.

Dr. Tyndall has stated his views on this question in his 'Glaciers of the Alps' (p. 249-251). He appears to consider direct pressure as the principal cause of the solidification of the ice, aided, perhaps, by congelation in the colder portions of the

mass.

None of these views appear to be sufficiently based on determinate conceptions of the interior temperature of the glacial mass. If the mean annual atmospheric temperature be several degrees less than 32° (Fahr.), the temperature during the later winter and earlier spring months will be considerably below the freezing temperature generally, at depths not exceeding that to which the winter cold is able to penetrate. In that part of the mass, therefore, congelation must necessarily attend infiltration, and must probably be a more efficient cause than pressure, which, in the more superficial portion of the mass, must be comparatively small. In its lower portion, on the contrary (if we allow the full effect of infiltration there), the temperature must be very nearly that of freezing, and congelation will proceed very slowly, while the pressure will become comparatively large and efficient. It appears to us that both the causes here spoken of must be effective, but more especially in different parts of the

mass.

The process of regelation could not, of course, be even tacitly alluded to in any of the explanations above mentioned previous to that given by Dr. Tyndall, since he was the first to discover its importance in glacial questions; nor even in his own explanation do we see any explicit allusion to its probable efficiency in the consolidation of the névé into compact ice. But it does appear to us that it is by means of this process that pressure is enabled to produce a particular kind of consolidation in ice at the freezing temperature which it is incapable of producing at any lower temperature. In fact, we do not see how we can do otherwise than recognise the efficiency of this cause, so far as we recognise the temperature of 32° in the greater portion of the

mass.

When

When the glacial mass passes from the state of névé to that of the proper glacial ice, it does not necessarily become a homogeneous hard transparent mass, but is frequently found to consist of alternate layers of two apparently different kinds of ice, one of which is of a dark bluish colour, and transparent, the other of a dull white colour, and opaque. These layers usually vary in thickness from the fraction of an inch to one or two inches, or upwards. Their continuity is more or less perfect for considerable distances, and their position, in the great majority of cases in which their development is most complete, approximates to verticality. The colour of the whiter layers is found to be due to the presence of a great number of small air-bubbles contained in them; the blue layers derive their greater transparency from the comparative absence of these bubbles. The structure is usually designated as the ribboned, laminar, or veined structure of glacial ice. These lamina appear to be developed as the ice consolidates from its state of névé, and may be regarded as a general property of the ice in its consolidated form, however different its development may be in different parts of a glacier, and however much that development may seem to depend on local conditions.

Whatever may be the physical cause of this peculiar structure, there seems to be no doubt of its being, in many cases, gradually developed during the transmutation of the névé into compact ice; and it appears to be equally certain that the structure, so far as regards the positions of the bands and their degree of development, may be suddenly and entirely changed when the cause producing the change is sufficiently energetic. The most complete proof of this latter statement is found in the structure immediately at the bottom of the ice-falls which form such striking features in the external aspect of a glacier. The structure in such localities is always finely developed, the veins are nearly vertical and transverse, their intersections with the surface of the glacier running nearly in straight lines across it, in directions perpendicular to its axis. This appears to be universally true, whatever may have been the degree of development of the structure, or the positions of the veins in the glacier immediately above the fall. There can be no doubt, therefore, as to the structure originating at the bottom of the fall, so far as it is distinguished by the characteristic positions of the veins as above described. When we examine the glacier at points more or less remote from the fall, we find the nearly straight transverse lines of structure converted into elongated loops, with their vertices directed towards the lower end of the glacier, and the question arises whether these loops are the original transverse lines of struc

ture,

ture, distorted into lengthened curves by the more rapid motion of the axial portion of the glacier; or whether they are altogether new structural lines resulting from the action of causes similar to those at the foot of the fall, their effects being modified by the change of conditions under which they act? This is a question which we shall discuss in the sequel. It may here be sufficient to remark that the positions of the veins and structural curves on the face of the glacier are generally such as might be anticipated, supposing them to be transmitted from the locality in which they originated, but to be elongated and deformed, as above described, by the unequable motion of different parts of the glacier.

In a canal-shaped glacier the elongated curves of structure will thus become more nearly parallel to the sides of the glacier in its marginal portions, as they move onward from the fall. Dr. Tyndall has appropriately designated the structure in those portions, the marginal structure. The laminar structure is also strongly developed on large glaciers beneath their central moraines, which arise, as above explained, from the junction of two of the lateral moraines of two large tributaries, as on the glacier of the Aar. In such cases the veins are vertical and longitudinal, and such as would result in the united glacier from the marginal veins of the tributaries, when those veins should be nearly parallel to the sides of their respective tributaries. This has been called the longitudinal structure. From the foot of the great fall of the Rhone glacier, and in some other glaciers, the forms of the valleys are such that the ice moves from them in radiating lines, and the eurves of structure consequently expand into curves of an approximately circular form. Most Alpine travellers will have remarked the striking feature they form on the glacier of the Rhone, between the fall and the terminating circular contour of the glacier. The Mer de Glace is also one of the well-known glaciers which exhibits the different varieties of this structure in great perfection.

We have already indicated the way in which the névé may become more or less distinctly stratified, and all glacialists probably agree in the belief that stratification may be frequently recognised in that portion of a glacial mass. There has been, however, great difference of opinion as to the permanence of any visible stratification in the consolidated ice of the lower portions of glaciers. M. Agassiz regards it as a permanent and pervading character of all glacial ice, derived from the original stratification of the névé. Principal Forbes, on the contrary, considers it to exist only in the névé, all indication of it disappearing in the true glacial ice. He cites the Talèfre glacier in support of his assertion. But these two observers did not agree as to what appear

ances