feet from the plain around, whence it rises majestically at an angle of about 35°; and, above all, for the prodigious dimensions of the lava-current it has poured forth. This current is also one of the most interesting of the whole chain, from the obstacles it has met with in its course, and the changes it has effected in the surface of the ground it overflowed. The sides of this hill are covered with forest-wood; its summit presents two distinct and very regular craters; one of them, with a vertical depth of 250 feet, is of considerable diameter.

The cone was apparently thrown up after or during the eruption of its stream of lava; which, instead of issuing from either crater, takes its rise at the western base of the hill.

At no great distance from its source the lava encountered an angular protuberance of granite, which evidently caused the current to separate into two branches. That to the right, the most considerable of the two, spread itself over a vast surface, and, aided by the gradual inclination of the granite platform, pursued its course to the west till it found an obstacle in a long line of hill, consisting of alluvial tuff from the Mont Dore covered by an ancient plateau of basalt. Thus impeded in its progress, the lava followed the sweep of the hill in a northeast direction; and finding an issue at length between this and a granitic knoll which obstructed its progress towards the north, poured down on the present site of the castle and town of Pont Gibaud; immediately above which it seems to have met and flowed over a more ancient stream from the Puy de Loucha-· dière (No. 21).

Thence both together poured in a broad sheet of lava down the steep side of a granitic hill which formed the eastern border of the valley of the Sioule, dashing forward against the rocks on the opposite side, and usurping the channel of the river, down which they pursued their course to the distance of more than a mile.

3.

The Sioule, thus dispossessed of its bed, has been constrained to work out a fresh one between the lava and the granite of its western bank, which in consequence is extremely precipitous. But, before this was accomplished, there is every appearance of its waters having been so far obstructed as to create a lake, covering the flat and alluvial surface now forming the meadows of Pont Gibaud.

In one part of this new channel, where the valley has a slight bend, and the torrent of lava being opposed by a salient rock accumulated to a considerable thickness, the excavation effected by the river has disclosed its internal division into vertical jointed columns, the lower portions of which are straight and well formed, the upper twisted into various curves, and less regularly polygonal. The wall of lava is about fifty feet high, and the columnar division is prolonged incompletely to the extent of between 200 and 300 yards. (See the woodcut overleaf.)

The rest of the lava filling the valley where it is about the third of a mile in breadth, is split by fissures, mostly vertical, into amorphous masses which still at various points evince a tendency to the prismatic form. This is perhaps the most marked instance of all Auvergne in which the lava of one of the very recent volcanos assumes a decided columnar division. In the Vivarais, as will be seen in a future chapter, this circumstance is of frequent occurrence.

The sketch (Plate II.) was taken from a basaltic plateau on the summit of the granitic range of hills, west of the Sioule, from whence the greater number of puys may be observed; that of Côme being the most conspicuous as well as the nearest. The whole course of this branch of its lava, as well as that of the Puy de Louchadière, is seen from this point.

But the Sioule was not to suffer from this invasion alone. The other branch of the lava-current of Côme, called the Cheire de l'Aumône, which flowed on from the point of separation in a direction west-south-west, soon reached the bed of this river, about three miles above the spot of the other irruption, and pouring over its banks filled up the entire valley with an immense causeway more than 100 feet high. Exhausted by this effort, it proceeded but a short way down the bed of the stream towards the north, and stopped where the village of Mazayes now stands.

The baffled waters of the Sioule here, as at Pont Gibaud, obstructed by the rocky dyke thus suddenly thrown across their channel, must have given birth to a lake by their stagnation, and would probably have ended, as in the other instance, by wearing away a passage parallel to their former one, had not the hill forming their western bank, not in this instance composed of granite, but of a soft alluvial tuff, yielded, at some distance up the stream, to the excessive pressure of the dammedup waters. An immense excavation, still subsisting, was broken across this hill, through which the lake emptied itself into the bed of the Monges at no great distance, and through which the Sioule still joins this latter stream about three miles above their former confluence.

A considerable body of back-water still remained behind in the part of the valley of the Sioule intercepted between the dyke of lava and the emissory thus forcibly created, which, from the inclination of the ground to the north, could not vent itself by this opening; and here a stagnant piece of water called the Etang de Fung, and used by the Seigneurs of Pont Gibaud as a stock-pond, existed till within a few years, when its drainage was artificially effected. The high banks which rise on each side of the long marshy meadow now occupying the site of the old pond still present the correspondence of angles so generally observable along channels of running water.

On the opposite side of the enormous causeway of lava to which this diversion is owing, another small piece of water, occupying the bed of a tributary rivulet choked up in the same manner, is dignified by the name of the Lake of Mazayes, and its insignificant drain now runs into the Sioule through the remainder of the wide and deep valley which this river itself once excavated and possessed.

The changes thus effected do not only present themselves to

the eye of a nice observer, but are exhibited in a manner not to be mistaken by the most casual; and in fact they result so simply and necessarily from the causes brought into action, that I should not have dwelt so long upon their details, but that they serve to exemplify the mode of formation of other lakes in the Auvergne and Velay, and the origin of other changes in the surface of a country or direction of its rivers, where every link in the chain of causes and effects is not quite so palpable.

The whole superficies of the plateau covered by the lava of Côme cannot be estimated under ten square miles. Its thickness is not to be ascertained, since no excavation has been made through it; where the current has met with any obstacle, it must of necessity be considerable, and thirty feet may be taken as the probable average.

It is one of the most rugged Cheires of the Monts Dôme, presenting a succession of continual asperities, following one another like the waves of an ocean, with similar depressions between. Upon walking over, its surface,-no easy task,-it appears to consist of chaotic heaps of rocky and angular blocks of compact basalt, tossed together in every variety of disorder; yet, in the deep and narrow intervals between these heaps, occur little patches of fresh and flowery turf, and knots of underwood spring from their clefts, contrasting strangely with the horrid desolation which prevails over this extensive wilderness.

Near the limit of the northern current, at some distance from Pont Gibaud, is a natural grotto in the basalt; from its interior gushes a small spring, which is partly frozen during the greatest heats of summer, and is said to be warm in winter; probably, however, only seeming warm by contrast with the external temperature.

The water is apparently frozen by means of the powerful