fires, now quenched, once blazed forth fiercely. The wonderful groups of volcanoes in Auvergne and the Eifel are as fresh as if they had not yet ceased to be active, and might break forth again at any moment; yet they have been quiescent ever since the beginning of authentic human history. But in the progress of the degradation which everywhere slowly changes the face of the land, it is impossible that volcanic hills should escape the waste which befalls every other kind of eminence. We can picture a time when the volcanic cones of Auvergne will have been worn away, and when the lava-streams that descend from them will be cut into ravines and isolated into separate masses by the streams that have even already deeply trenched them. Where all the ordinary and familiar signs of a volcano have been removed, how can we tell that any volcano ever existed? What enduring record do volcanoes inscribe in geological history? Now, it must be obvious that among the operations of an active volcano, many of the most striking phenomena have hardly any importance as aids in recognising the traces of long-extinct volcanic action. The earthquakes and tremors that accompany volcanic outbursts, the constant and prodigious out-rushing of steam, the abundant discharge of gases and acid vapours, though singularly impressive at the time, leave little or no lasting mark of their occurrence. It is not in phenomena, so to speak, transient in their effects, that we must seek for a guide in exploring the records of ancient volcanoes, but in those which fracture or otherwise affect the rocks below ground, and pile up heaps of material above. Keeping this aim before us, we may obtain from an examination of what takes place at an active volcano such durable proofs of volcanic energy as will enable us to recognise the former existence of volcanoes over many tracts of the globe where human eye has never witnessed an eruption, and where, indeed, all trace of what could be called a volcano has utterly vanished. A method of observation and reasoning has been established, from the use of which we learn that in some countries, Britain for example, though there is now no sign of volcanic activity, there has been a succession of volcanoes during many protracted and widely separated periods, and that probably the interval that has passed away since the last eruptions is not so vast as that which separated these from those that preceded them. A similar story has been made out in many parts of the continent of Europe, in the United States, India, and New Zealand, and, indeed, in most countries where the subject has been fully investigated. A little reflection on this question will convince us that the permanent records of volcanic action must be of two kinds: first and most obvious are the piles of volcanic materials which have been spread out upon the surface of the earth, not only round the immediate vents of eruption, but often to great distances from them; secondly, the rents and other openings in the solid crust of the earth caused by the volcanic explosions, and some of which have served as channels by which the volcanic materials have been expelled to the surface. Volcanic Products. We shall first consider those materials which are erupted from volcanic vents and are heaped up on the surface as volcanic cones or spread out as sheets. They may be conveniently divided into two groups: 1st, Lava, and 2d, Fragmentary materials. (1) Lava.-Under this name are comprised all the molten rocks of volcanoes. These rocks present many varieties in composition and texture, some of the more important of which will be described in Chapter XI. Most of them are crystalline—that is, are made up wholly or in greater part of crystals of two or more minerals interlocked and felted together into a coherent mass. Some are chiefly composed of a dark brown or black glass, while others consist of a compact stony substance with abundant crystals imbedded in it. Probably most of them when in completest fusion within the earth's crust existed in the condition of thoroughly molten glass, H the transition from that state to a stony or lithoid one being due to a process of "devitrification" (p. 144) consequent on cooling. During this process some of the component ingredients of the glass crystallise out as separate minerals, and this crystallisation sometimes proceeds so far as to use up all the glass and to transform it into a completely crystalline substance. In many cases lavas are strikingly cellular—that is to say, they contain a large number of spherical or almond-shaped cavities somewhat like those of a sponge or of bread, formed by the expansion of the steam absorbed in the molten rock (Figs. 35 and 36 and p. 146). Lavas vary much in weight and in colour. The heavier kinds are more than three times the weight of water; or, in other words, they have a specific gravity ranging up to 3.3; and are commonly dark grey to black. The lighter varieties, on the other hand, are little more than twice the weight of water, or have a specific gravity which may be as low as 2.3, while their colours are usually paler, sometimes almost white. When lava is poured out at the surface it issues at a white heat —that is, at a temperature sometimes above that of melting copper, or more than 2204° Fahr.; but its surface rapidly darkens, cools, and hardens into a solid crust which varies in aspect according to the liquidity of the mass. Some lavas are remarkably fluid, flowing along swiftly like melted iron; others move sluggishly in a stiff viscous stream. In many pasty lavas, the surface breaks up into rough cindery blocks or scoriæ like the slags of a foundry, which grind upon each other as the still molten stream underneath a FIG. 36.-Section of a lava-current. creeps forward (p. 146). In general, the upper part of a lava-stream is more cellular than the central portions, no doubt because the imprisoned steam can there more easily expand. The bottom, too, is often rough and slaggy, as the lava is cooled by contact with the ground, and portions of the chilled bottomcrust are pushed along or broken up and involved in the still fluid portion above. There are thus three more or less well-defined zones in a solidified lava-current-a cellular or slaggy upper part (c in Fig. 36), a more solid and jointed centre (6), embracing usually by much the largest proportion of the whole, and a cellular or slaggy bottom (a). A rock presenting these characters tells its story of volcanic action in quite unmistakable language. It remains as evidence of the existence of some neighbouring volcanic vent, now perhaps entirely covered up, whence it flowed. We may even be able to detect the direction in which the lava moved. The cells opened by the segregation and expansion of the steam entangled in the interstices of a mass of lava which is at rest are, on the whole, spherical. But if the rock is still moving, the cells will be drawn out and flattened into almond-shaped (amygdaloidal) vesicles, with their flat sides parallel to the surface of the lava, and their longer FIG. 37.-Elongation of cells in direction of flow of a lava-stream. axes ranged in one general direction, which is that of the motion of the molten stream (Fig. 37). At a volcanic vent, the mass of erupted lava is generally thickest, and it thins away as its successive streams terminate on the lower grounds surrounding the cone. But sometimes a lavacurrent may flow for 40 miles or more from its source, and may here and there attain locally a great thickness by rolling into a valley and filling it up, as has been witnessed among the Icelandic eruptions. As a rule, where ancient lava-streams are found to thicken in a certain direction, we may reasonably infer that in that direction lay the vent from which they flowed. Again, sheets of lava that solidify on the slopes of a volcanic cone are inclined; they may congeal on declivities of as much as 30° or 40°. If a series of ancient lavas were observed to slope upward to a common centre, we might search there for some trace of the funnel from which they were discharged. But, of course, in proportion to their antiquity, lava-streams, like every other kind of rock, have suffered from geological revolutions, among which those that involve upheaval and dislocation are especially important, so that the inclination of an ancient lava-bed must not be too hastily assumed as an indication of the slope of the cone of a volcano. It must be taken in connection with the rest of the evidence supplied by the whole district. Where lavas reach the lower grounds beyond the foot of a volcanic cone, they may spread out in wide nearly horizontal sheets. As current succeeds current, the original features of the plain may be entirely buried under a mass of lava many feet thick. If a section could be cut through such an accumulation, it might be possible to determine the thickness of each successive lavastream by means of the slaggy upper and lower surfaces. Here and there, too, where two eruptions were separated by an interval long enough to allow the surface of the older mass partially to crumble into soil and support some vegetation, the layer of burnt soil between the two sheets of lava would remain as a witness of this interval. In other instances, we can understand that in the larger hollows of a lava-plain, ponds or lakes might gather, on the floor of which there might be deposited layers of fine silt full of terrestrial leaves, insect remains, land and fresh-water shells, and other organic relics of a land-surface. If, now, a lacustrine accumulation of this kind were to be buried under a new outburst of lava, it would be sealed up and might preserve its record intact for vast ages. In any section cut through such a series of lava-beds by a river or the sea, or by man, the layers of silt with their organic remains intercalated between the lava-streams would prove the eruptions to have taken place on land, and to have been separated by a long interval, during which a lake was formed on the cold and decomposing surface of the earlier lava. The conditions under which the volcanic outbursts occurred may be thus inferred, not so much from the nature of the volcanic materials themselves, as from that of the layers of sediment that may happen to have been preserved among them. Seams of red baked soil, with charred remains of terrestrial vegetation interposed between the upper and under sides of successive lavas would point to subaerial eruptions. Bands of hardened clay or marl, with leaves and fresh-water shells, would show that the lavas had invaded a lake. Beds of limestone or other rock, containing corals, |