terially impeded the flow of the liquid matter, and may thus probably have occasioned those great protuberances and irregularities of surface, which, although characteristic of all lavacurrents, are remarkably displayed in this. These peculiar features are several times repeated, for whenever an expansion of the valley offers itself, the lava is spread out into a broader and thinner sheet, and at those points where salient promonto-. ries of gneiss-rock seem to have barred its passage, the volcanic matter swells up to great heights. Since the mineralogical account of this current has been described by Mr Scrope *, we shall here merely state, that in general characters it is identical with most others of nearly a contemporaneous origin, in having an upper part light, cellular, and scoriaceous, which passes downwards into a basaltic mass, occasionally divided into irregular and converging columns, and frequently terminating at its base in most perfect vertical prisms. The deepest section of the lava occurs directly under the village of Chaluzet, where the left bank of the Sioule consists almost entirely of vertical cliffs 400 feet in depth, exposing in their higher parts black and red scoriæ, in their lower columnar basalt. These rocks abound in caverns, some occurring in scoriæ black and cellular, as those of Vesuvius, whilst others, situated at lower levels, are grotesquely hollowed out of the compact basalt. Amongst the latter, the arcades of Pranal are remarkably picturesque.

The lower ends of the prisms of basalt generally terminate at some height above the present bed of the Sioule, and whenever not obscured by debris, they are seen to rest upon a bed of pebbles, of varying thickness, below which the subjacent gneiss is cut through, down to the actual level of the river. Near Pranal this pebble-bed is observable in the cliff, about fifty feet above the stream, and the space thus denuded exposes a face of gneiss, nearly vertical, in which lead-mines are now extensively worked †. If, however, no other than these longitudinal sec

The greatest depth of gneiss worn into by the river under the basalt, appeared to us, after a careful examination, never to exceed fifty or sixty feet. Mr Scrope had conjectured the amount of excavation through gneiss (called mica-schist by him), at 200 feet.

tions had been afforded, there might have existed some difficulty in proving that this alluvium did at any time pass horizontally under the basalt, constituting a river bed, and not merely debris covering a mountain slope; but we obtained the most unanswerable evidence in support of this fact at Les Combres. There, within a very short distance of its termination, and when the lava-current has diminished to the thickness of thirty feet, it rests upon a pebble-bed, having between it and the stream about eighteen feet of gneiss*. Fortunately at this point there is an ancient excavation (said to have been a mine of the Romans), and a gallery has been driven in horizontally through the upper part of the gneiss and the superposed alluvium, so that the lower ends of the columns of basalt form the roof, resting on the pebble-bed, which can be traced for the space of fifty or sixty feet inwards on either side of the gallery, occupying, regularly throughout, the space between the basalt and the gneiss. In this alluvial bed are fragments of old basalt, with others of gneiss, and much white quartz and schist. Many of these boulders are quite rounded, whilst others are flat and slightly angular, as in the bed of the present river, and the surface of the gneiss on which they are deposited is slightly undulating.

The higher portion of the alluvial bed consists of fine brown sand, identical with bands occupying similar positions at Montpezat and other localities in the Vivarais, and the ends of the prisms of basalt terminating upon this sandy bed are here, as in that district, scoriaceous and cellular.

No remnant of the lava of Chaluzet is now found on the right bank of the Sioule, but it must once have reached it at many points, as the gorge is every where narrow, and at Pranal cannot far exceed fifty or sixty feet in width, although here the basalt, resting on gneiss, forms an impending and vertical cliff, about 100 feet in height. For two-thirds of this thickness the basalt is prismatic, and this structure implies its former liquidity and slow cooling on the spot, and that it did not form the exposed surface of a "cheire." It may indeed be contended, that the Sioule excavated the present chasm between the basalt and the gneiss, and chiefly through the latter, because this would

* See Plate I.

account for some ranges of basaltic prisms lying on the gneiss, nearly horizontal, and presenting their ends towards the gorge. For, in the valley of Antraigues in the Vivarais, where a section is afforded of basaltic current abutting against a precipice of gneiss, the prisms are nearly horizontal; whereas they are oblique on the slope beneath, and vertical where they rest on the level bed of the ancient river, in accordance with a fact now generally established, that the axes of the prisms are always at right angles to the cooling surfaces.

The comparative destructibility of the basalt and the gneiss is by no means unfavourable to this hypothesis; for although there are here, as usual, in the gneiss formation, some strata of a hard granitic texture, yet in general it is a slaty rock, containing much mica and decomposing felspar, and its high inclination, usually at an angle of about 65°, and numerous fissures parallel to the main course of the river, facilitate disintegration. This tendency to waste is evinced by the huge slaty masses encumbering the foot of the cliffs or the middle of the river-bed. The basalt, on the other hand, is extremely compact, resists decomposition, and is only perishable in some parts where the prisms happen to be vertical, and exposed to the mechanical action of an undermining current.

Mr Scrope has observed, that the undisturbed and perfect state of the cone of loose scoria whence the lava of Chaluzet proceeded, demonstrates that no denuding wave, or extraordinary body of water, has passed over this spot since the eruption; and we may add, that here as at Côme, the upper scoriaceous surface of the" cheire" preserves its original asperities, and is in no part strewed over with sand or alluvial pebbles. Now, if we compare the extent of waste at Chaluzet with that undergone by the "cheire" of Côme below Fontgibaud, it is difficult not to conclude with Mr Scrope, that the eruption of Chaluzet was considerably anterior. But we can have no hesitation in assuming that Chaluzet is at least as ancient a volcano as Côme. Whence it follows, that in times comparatively recent, there has passed through the narrow gorge of the Sioule near Chaluzet, 1st, The whole contents now removed from the higher valley before alluded to, connecting the Sioule and the Monges; 2dly, All that has

been excavated out of the "cheire" of Côme below Pontgibaud; 3dly, All the basalt and subjacent gneiss destroyed as above stated between Pranal and Combres; and, 4thly, The whole detritus of primary rocks brought down from countries higher on the Sioule than the above localities. This enormous quantity of rock and soil has been carried, we say, through this gorge, in times comparatively recent; for the products of Côme and Chaluzet are modern in comparison with the ancient basalts of Auvergne, as are the latter in reference to the lacustrine tertiary strata whereon they rest.

If, then, within a period of time inconsiderable with reference to the earth's history, the volume of matter carried down by a single tributary stream of the Allier be so prodigious, it is easy to perceive, that, were all the materials which have been transported during the same time by other rivers from the central granitic mountains of France, united and presented to us in one view, they would bear witness to the force of causes that have operated, and are still operating, in nature, to an extent far beyond the calculation of many geologists.

Volcano of Montpezat.

Without regard to geographical connection, we shall next describe some phenomena analogous to those of Chaluzet, as presented by the valley of Montpezat in the Vivarais. For the leading facts relating to this lava-current, we refer to Mr Scrope's work *. He, in common with M. Faujas de St Fond, has described the primary rock of this district as granite. But the whole country, from the source of the Ardêche in the Haut Vivarais to near Aubenas consists of gneiss, composed in general of the ordinary ingredients of granite, but distinctly laminated in strata, for the most part of great thickness, often vertical, and always highly inclined, and here rarely deviating from the same direction, viz. from NW. to SE. Some members of this formation are composed here, as in Scotland, chiefly of hornblende; others are, in mineralogical characters, true granites, though belonging to the same system, and granite veins are not uncommon. The verticality of the strata has contributed to

* Geology of Central France.

the serrated outline of the mountains, and to the deep and fissured character of the valleys.

The cone and crater of Montpezat have as perfect and fresh an appearance as the most recent in Auvergne. The volcanic bombs are extremely numerous, and the ashes are so loose, that there is no vegetation on the greater part of the northern slope. As there are no smooth rounded alluvial pebbles at the bottom of the crater, or on any part of the cone, the angular blocks of gneiss near the summit (some heavier than a man can lift) were not transported thither by water, but must have been ejected from below. These have undergone no visible alteration; a remark that will apply equally to many of the granitic fragments contained both in the scoria and basaltic lava of this volcano. The black cellular scoriæ that were thrown out from the crater towards Thueyts, attest, if possible, still more strongly than the state of the cone, that no general inundation has passed over the country since the eruption. These ashes and scoriæ are very light, the fragments varying from the size of walnuts to peas, covering, in beds from three to four feet thick, a rapid acclivity of gneiss. They have in great part disappeared, but, from the abrupt shape of the valley, and their feeble adhesion to the subjacent rocky slope, it is clear, that even the force of a very small body of water would have swept away the whole.

The lava of Montpezat descended from a breach on the NW. side of the crater into the bed of the Fontaulier. Now, it is natural, and indeed almost necessary, to suppose, that whenever a mountain stream is thus dammed up, a lake must have been the result, which would receive for a time all the sand and rocks brought down continually from the higher mountains. The river issuing from such a lake would exert but little power near its outlet, not being yet charged with foreign matter. These circumstances are still clearly exemplified in lakes caused by similar obstructions in Auvergne, and we shall enumerate some facts there observed, as illustrative of phenomena to be described in the sequel. At the lake of Chambon, for instance, where the lava of Tartaret blocked up the whole valley of the Couze, that river has been unable, near its outlet, to wear to the depth of more than a few feet through the barrier; whereas, not far below, at Sailhens, where the body of water is not greater, it has