(a.) General Structure of the Auvergne District.—From a granitic and gneissose platform situated near the centre of France, and separated from the western spurs of the Alps by the wide valley of the Rhone, there rises a group of volcanic mountains surpassing in variety of form and structure any similar mountain group in Europe, and belonging to an epoch ranging from the Middle Tertiary down almost to the present day. This volcanic group of mountains gives rise to several important rivers, such as the Loire, the Allier, the Soule (a branch of the Loire), the Creuse, the Dordogne, and the Lot; and in the Plomb du Cantal attains an elevation of 6130 feet above the sea. Its southern section, that of Mont Dore, the Cantal, and the Haute Loire, is characterised by magnificent valleys, traversing plateaux of volcanic lava, and exhibiting the results of river erosion on a grand scale; while its northern section, that of the Puy de DÔme, presents to us a varied succession of volcanic crater-cones and domes, with their extruded lava-streams, almost as fresh and unchanged in form as if they had only yesterday become extinct. A somewhat similar, but less important, chain of extinct volcanoes also occurs in the Velay and Vivarais, between the upper waters of the Loire and the Allier, in the vicinity of the town of Le Puy.[1] The principal city in this region is Clermont-Ferrand, lying near the base of the Puy de DÔme, and ever memorable as the birthplace of Blaise Pascal.[2]

The physical structure of this region is on the whole very simple. The fundamental rocks consist of granite and gneiss passing into schist, all of extreme geological antiquity, forming a vast platform gradually rising in a southerly direction towards the head waters of the Loire and the Allier in the Departments of Haute Loire, LozÈre, and ArdÈche. On this platform are planted the whole of the volcanic mountains. (See Fig. 17.)

The granitic plateau is bounded on the east, throughout a distance of about 50 miles, by the wide and fertile plain of Clermont, watered by the Allier and its numerous branches descending from the volcanic mountains, and is about 25 miles in width from east to west in the parallel of Clermont, but gradually narrowing in a southerly direction, till at Brioude it becomes an ordinary mountain ravine. The eastern margin of the plain is formed by another granitic ridge expanding into a plateau towards the south, and joining in with that already described; but towards the north and directly east of Clermont it forms a high ridge traversed by the railway to St. Étienne and Lyons, and descending towards the east into the valley of the Loire. No more impressive view is to be obtained of the volcanic region than that from the summit of this second ridge, on arriving there towards evening from the city of Lyons. At your feet lies the richly-cultivated plain of Clermont, dotted with towns, villages, and hamlets, and decorated with pastures, orchards, vineyards, and numerous trees; while beyond rises the granitic plateau, breaking off abruptly along the margin of the plain, and deeply indented by the valleys and gorges along which the streams descend to join the Allier. But the chief point of interest is the chain of volcanic crater-cones and dome-shaped eminences which rise from the plateau, amongst which the Puy de DÔme towers supreme. Their individual forms stand out in clear and sharp relief against the western sky, and gradually fade away towards the south into the serried masses of Mont Dore and Cantal, around whose summits the evening mists are gathering. Except the first view of the Mont Blanc range from the crest of the Jura, there is no scene perhaps which is calculated to impress itself more vividly on the memory than that here faintly described.[3]

(b.) The Vale of Clermont.—The plain upon which we look down was once the floor of an extensive lake, for it is composed of various strata of sand, clay, marl, and limestone, containing various genera and species of fresh-water shells. These strata are of great thickness, perhaps a thousand feet in some places; and along with such shells as Paludina, Planorbis, and LimnÆa are also found remains of various other animals, such as fish, serpents, batrachians, crocodiles, ruminants, and those of huge pachyderms, as Rhinoceros, Dinotherium, and CÆnotherium. This great lake, occupying a hollow in the old granitic platform of Central France, must have been in existence for an extensive period, which MM. Pomel, Aymard, and Lyell all unite in referring to that of the Lower Miocene. But what is to us of special interest is the fact that, in the deposits of this lake of the Haute Loire, with the exception of the very latest, there is no intermixture of volcanic products such as might have been expected to occur if the neighbouring volcanoes had been in activity during its existence. Hence it may be supposed that, as Scrope suggested, the waters of the lake were drained off owing to the disturbance in the levels of the country caused by the first explosions of the Auvergne volcanoes.[4] If this be so, then we possess a key by which to determine the period of the first formation of volcanoes in Central France; for, as the animal remains enclosed in the lacustrine deposits of the Vale of Clermont belong to the early Miocene stage, and the earliest traces of contemporaneous volcanic ejecta are found only in the uppermost deposits, we may conclude that the first outburst of volcanic action occurred towards the close of the Miocene period—a period remarkable for similar exhibitions of internal igneous action in other parts of the world.

(c.) Successive Stages of Volcanic Action in Auvergne.—The volcanic region here described, which has an area of about one hundred square miles, does not appear to have been at one and the same period of time the theatre of volcanic action over its whole extent. On the contrary, this action appears to have commenced at the southern border of the region in the Cantal, and travelling northwards, to have broken out in the Mont Dore region; finally terminating its outward manifestations among the craters and domes of the Puy de DÔme. In a similar manner the volcanic eruptions of the Haute Loire and ArdÈche, lying to the eastward, and separated from those of the Cantal by the granitoid ridge of the Montagnes de Margeride, belong to two successive periods referable very closely to those of the Mont Dore and the Puy de DÔme groups.[5] The evidence in support of this view is very clear and conclusive; for, while the volcanic craters formed of ash, lapilli, and scoriÆ, together with the rounded domes of trachytic rock of which the Puy de DÔme group is composed, preserve the form and surface indications of recently extinguished volcanoes, those which we may assume to have been piled up in the region of Mont Dore and Cantal have been entirely swept away by prolonged rain and river action, and the sites of the ancient craters and cones of eruption are only to be determined by tracing the great sheets of lava up the sides of the valleys to their sources, generally situated at the culminating points of their respective groups. Other points of evidence of the great antiquity of the latter groups might be adduced from the extent of the erosion which has taken place in the sheets of lava having their sources in the vents of the Plomb du Cantal and of Mont Dore, owing to which, magnificent valleys, many miles in length and hundreds of feet in depth, have been cut out of these sheets of lava and their supporting rocks, whether granitic or lacustrine, and the materials carried away by the streams which flow along their beds. These points will be better understood when I come to give an account of the several groups; and in doing so I will commence with that of the Cantal.[6]

(d.) The Volcanoes of the Cantal.—The original crater-cones of this group have entirely disappeared throughout the long ages which have elapsed since the lava-streams issued forth from their internal reservoirs. The general figure of this group of volcanic mountains is that of a depressed cone, whose sides slope away in all directions from the central heights, which are deeply eroded by streams rising near the apex and flowing downwards in all directions towards the circumference of the mountain, where they enter the Lot, the Dordogne, and the Allier. The orifice of eruption was situated at the Plomb du Cantal, formed of solid masses of trachyte, which, owing, as Mr. Scrope supposes, to a high degree of fluidity, were able to extend to great distances in extensive sheets, and were afterwards covered by repeated and widely-spread flows of basalt; so that the trachyte towards the margin of the volcanic area becomes less conspicuous than the basalt by which it is more or less concealed from view, or overlapped. Extensive beds of tuff and breccia accompany the trachytic masses.

Magnificent sections of the rocks are laid open to view along the sides of the valleys, which are steep and rock-bound. Except towards the south-west, about Aurillac, where lacustrine strata overlie the granite, the platform from which rises the volcanic dome is composed of granitic or gneissose rocks. Accompanying the lava-streams are great beds of volcanic agglomerate, which Mr. Scrope considers to have been formed contemporaneously with the lava which they envelop, and to be due to torrents of water tumultuously descending the sides of the volcano at periods of eruption, and bearing down immense volumes of its fragmental ejecta in company with its lava-streams.[7] Nowhere throughout this region do beds of trachyte and basalt alternate with one another; in all cases the basalt is the newer of the two varieties of rock, and this is generally the case throughout the region here described.

(e.) Volcanoes of Mont Dore.—This mountain lies to the north of that of Cantal, and somewhat resembles it in general structure and configuration. Like Cantal, it is destitute of any distinct crater; all that is left of the central focus of eruption being the solidified matter which filled the throat of the original volcano, and which forms a rocky mass of lava, rising in its highest point, the Pic de Saucy, to an elevation (as given by Ramond) of 6258 feet above the level of the sea, thus exceeding that of the Plomb du Cantal by 128 feet. Its figure will be best understood by supposing seven or eight rocky summits grouped together within a circle of about a mile in diameter, from whence, as from the apex of an irregular and flattened cone, all the sides slope more or less rapidly downwards, until their inclination is gradually lost in the plain around. This dome-shaped mass has been deeply eroded on opposite sides by the valleys of the Dordogne and Chambon; while it is further furrowed by numerous minor streams.[8]

The great beds of volcanic rock, disposed as above stated, consist of prodigious layers of scoriÆ, pumice-stones, and detritus, alternating with beds of trachyte and basalt, which often descend in uninterrupted currents till they reach the granite platform, and then spread themselves for miles around. The sheets of basalt are found to stretch to greater distances than those of trachyte, and have flowed as far as 15 or 20 miles from their orifices of eruption; while in some cases, on the east and north sides, they have extended as far as 25 or 30 miles from the central height. On the other hand, a radius of about ten miles from the centre would probably include all the streams of trachyte;—so much greater has been the viscosity of the basalt over the latter rock. Some portions of these great sheets of lava, cut off by river valleys or eroded areas from the main mass of which they once formed a part, are found forming isolated terraces and plateaux either on the granitic platform, or resting on the fresh-water strata of the valley of the Allier, while in a northern direction they overspread a large portion of the granitic plateau from which rise the Puy de DÔme and associated volcanic mountains. Still more remarkable are the cases in which these lava-streams have descended into the old river channels which drained the granitic plateau. Thus the current which took its origin in the Puy Gros descended into the valley of the Dordogne, while another stream invaded the gorge of Champeix on the eastern side.

The more ancient lava-streams just described are invaded by currents and surmounted by cones of eruption of more recent date, similar to those of the Puy de DÔme group lying to the northward. Such cones and currents, amongst which are the Puy de Tartaret and that of Montenard, present exactly the same characters as those of this group, to which we shall return further on.

(f.) Volcanoes of the Haute Loire and ArdÈche.—Separated by the valley of the Allier and the granitic ridge of La Margeride from the volcanic regions of Cantal and Mont Dore is another volcanic region of great extent, which reaches its highest elevation in the central points of Mont Mezen, attaining an elevation (according to Cordier) of 5820 feet, and formed of "clinkstone." The volcanic products of Mezen have been erupted from one central orifice of vast size, and consist mainly of extensive sheets of "clinkstone," a variety of trachytic lava, which have taken courses mainly towards the north-west and south-east. These great sheets, one of which appears to have covered a space more than 26 miles in length with an average breadth of 6 miles, thus overspreading an estimated area of 156 square miles, has been deeply eroded by streams draining into the Loire, along whose banks the rocks tower in lofty cliffs; while it has also suffered enormous denudation, by which outlying fragments are disconnected from the main mass, and form flat-topped hills and plateaux as far distant as Roche en Reigner and Beauzac, at the extreme distance (as stated above) of 26 miles from the source of eruption.

But even more remarkable than the above are the vast basaltic sheets which stretch away for a distance of 30 miles by Privas almost to the banks of the RhÔne, opposite Montlimart. These have their origin amongst the clinkstone heights of Mont Mezen, and taking their course along the granitic plateau in a south-easterly direction, ultimately pass over on to the Jurassic and Cretaceous formations composing the plateau of the Coiron, which break off in vertical cliffs from 300 to 400 feet in height, surmounting the slopes that rise from the banks of the ArdÈche and Escourtais rivers near Villeneuve de Bere. This is probably one of the most extensive sheets of basalt with which we are acquainted in the European area, and it is only a remnant of a vastly greater original sheet.[9]

(g.) Newer Volcanoes of the Haute Loire (the Velay and Vivarais).—Subsequently to the formation of the lava-streams above described, and probably after the lapse of a lengthened period, the region of the Haute Loire and ArdÈche became the scene of a fresh outburst of volcanic action, during which the surface of the older lavas, or of the fundamental granite, was covered by numerous crater-cones and lava-streams strewn along the banks of the Allier and of the Loire for many miles. These cones and craters are not quite so fresh as those of the Mont DÔme group; those of the Haute Loire being slightly earlier in point of time, and, as Daubeny shows, belonging to a different system. So numerous are these more recent cones and craters that Scrope counted more than 150 of them, and probably omitted many.

The volcanic phenomena now described have a special interest as bearing on the question whether man was an inhabitant of this region at the time of these later eruptions. The question seems to be answered in the affirmative by the discovery of a human skull and several bones in the volcanic breccia of Mont Demise, in company with remains of the elephant (E. primigenius), rhinoceros (R. tichorhinus), stag, and other large mammifers. The discovery of these remains was made in the year 1844, and the circumstances were fully investigated and reported upon by M. Aymard, and afterwards by Mr. Poulett Scrope, upon whose mind no possible doubt of the fact remained. From what we now know of the occurrence of human remains and works of art in other parts of France and Europe, no surprise need be felt at the occurrence of human remains in company with some extinct mammalia in these volcanic tuffs, which belong to the Post-Pliocene or superficial alluvia antecedent to the historic period.[10]

(h.) Mont DÔme Chain.—We now come to the consideration of the most recent of all the volcanic mountain groups of the region of Central France, that of the Puy de DÔme, lying to the north of Mont Dore and Cantal. We have seen that there is almost conclusive evidence that man was a witness to the later volcanic outbursts of the Vivarais, and as these craters seem to be of somewhat earlier date than those of the Puy de DÔme group, we cannot doubt that they were in active eruption when human beings inhabited the country, and not improbably within what is known as the Historic Period. No mention, however, is made either by CÆsar, Pliny, or other Roman writers of the existence of active volcanoes in this region. CÆsar, who was a close observer, and who carried the Roman arms into Auvergne, makes no mention of such; nor yet does the elder Pliny, who enumerated the known burning mountains of his day all over the Roman Empire. It is not till we come down to the fifth century of our era that we find any notices which might lead us to infer the existence of volcanic action in Central France. This is the well-known letter written by Sidonius Apollonarius, bishop of Auvergne, to Alcinus Avitus, bishop of Vienne, in which the former refers to certain terrific terrestrial manifestations which had occurred in the diocese of the latter. But, as Dr. Daubeny observes, this is no evidence of volcanic action in Auvergne, where Sidonius himself resided; the terrestrial disturbances above referred to may have been earthquake shocks of extreme severity.[11]

But although we have no reliably historical record of volcanic action amongst the mountains of the Mont DÔme group, the fact that these are, comparatively, extremely recent will be evident to an observer visiting this district, and this conclusion is based on three principal grounds: first, because of the well-preserved forms of the original craters, though generally composed of very loose material, such as ashes, lapilli, and slag; secondly, because of the freshness of the lava-streams over whose rugged surfaces even a scanty herbage has in some places scarcely found a footing;[12] and thirdly, because the lava from the crater-cones has invaded channels previously occupied by the earlier lavas, or those which had been eroded since the overflow of the great basaltic sheets of Mont Dore. Still, as in the case of the valleys of Lake Aidot, of Channonat, and of Royat, these streams are sufficiently ancient to have given time for the existing rivers to have worn out in them channels of some depth, but bearing no comparison to the great valleys which had been eroded out of the more ancient lavas, such as those of the Coiron, of the ArdÈche, and of the Dordogne and Chambon in the district of Mont Dore.

(i.) Dome-shaped Volcanic Hills.—I have previously (page 15) referred to the two classes of volcanic eminences to be found in the chain of the Puy de DÔme; one indicated by the name itself, formed of a variety of trachytic lava called "domite," and of the form of a dome; the other, composed of fragmental matter piled up in the form of a crater or cup, often ruptured on one side by a stream of lava which has burst through the side, owing to its superior density. Of the former class the Puy de DÔme and the Grand Sarcoui (see Fig. 18) are the most striking examples out of the five enumerated by Scrope, while there is a large number, altogether sixty-one, belonging to the latter class. These domes and crater-cones, as already stated, rise from a platform of granite, either directly or from one formed of the lava-sheets of the Mont Dore region, which in turn overlies the granitic platform. Of the nearly perfect craters there are the Petit Puy de DÔme, lying partially against the northern flank of the greater eminence; the Puy de Cone, remarkable for the symmetry of its conical form, rising to a height of 900 feet from the plain; and the Puys de Chaumont and Thiolet lying to the north of the Puy de DÔme. Of those to the south of this mount, two out of the three craters of the Puy de Barme and the Puy de Vichatel are perfect; but most of the crater-cones south of the Puy de DÔme are breached. Some of the lava streams by which these craters were broken down flowed for long distances. That the lava followed the showers of ashes and lapilli forming the walls of the craters is rendered very evident in the case of the Puy de la Vache, whose lava-stream coalescing with those from the Puy de la Solas and Puy Noir, deluged the surrounding tracts and flowed down the Channonat Valley as far as La Roche Blanc in the Vale of Clermont. In the interior of the upper part of the crater still remaining may be seen the level (so to speak) to which the molten lava rose before it burst its barrier. This level is marked by a projecting platform of reddish or yellow material, rich in specular iron, apparently part of the frothy scum which formed on the surface of the lava and adhered to the side of the basin at the moment of its being emptied.

Space does not permit a fuller description of this remarkable assemblage of extinct volcanoes, and the reader must be referred for further details to the work of Mr. Scrope. I shall content myself with some further reference to the central figure in this grand chain, the Puy de DÔme itself.

Ascent of the Puy de DÔme.—On ascending by the winding path up the steep side of the mount, and on reaching the somewhat flattened summit, the first objects which strike the eye are the massive foundations of the Roman temple of Mercury; they are hewn out of solid grey lava, altogether different from the rock of the Puy de DÔme itself, which must have been obtained from one of the lava-sheets of the Mont Dore group. To have carried these large blocks to their present resting-place must have cost no little labour and effort. The temple is supposed to have been surmounted by a colossal statue of the winged deity, visible from all parts of the surrounding country which was dedicated to his honour, and the foundations were only discovered a few years ago when excavating for the foundation of the observatory, which stands a little further on under the charge of Professor Janssen. On proceeding to the northern crest of the platform a wonderful view of the extinct craters and domes—about forty in number, and terminating in the Puy de Beauny, the most northerly member of the chain—is presented to the spectator. To the right is the Vale of Clermont and the rich valley of the Allier merging into the great plain of Central France. On the south side of the platform a no less remarkable spectacle meets the eye. The chain of Puys and broken craters stretches away southwards for a distance of nearly ten miles, while the horizon is bounded in that direction by the lofty masses of the Mont Dore, Cantal, and Le Puy ranges. Nor does it require much effort of the imagination to restore the character of the region when these now dormant volcanoes were in full activity, projecting showers of ashes and stones high into the air amidst flames of fire and vast clouds of incandescent gas and steam.

The material of which the Puy de DÔme is formed consists of a light grey, nearly white, soft felsitic lava, containing crystals of mica, hornblende, and specular iron-ore. It is highly vesicular, and was probably extruded in a pasty condition from a throat piercing the granitic plateau and the overlying sheet of ancient lava of Mont Dore. It has been suggested that such highly felsitic and acid lavas as that of which the Puy de DÔme, the Grand Sarcoui, and Cliersou are composed, may have had their origin in the granite itself, melted and rendered viscous by intense heat. Dr. E. Gordon Hull has suggested that the domite hills (owing to their low specific gravity) may have filled up pre-existing craters of ashes and scoriÆ without rupturing them, as in the case of the heavier basaltic lavas, and then still continuing to be extruded, may have entirely enveloped them in its mass; so that each domite hill encloses within its interior a crater formed of ashes, stones, and scoriÆ. In the case of the Puy de DÔme there is some evidence that the domite matter rests on a basis of ashes and scoriÆ, which may be seen in a few places around the base of the cone. It is difficult without some such theory as this to explain how a viscous mass was able to raise mountains some 2000 or 3000 feet above the surrounding plain.[13]

(j.) Sketch of the Volcanic History of Central France.—It now only remains to give a brief resumÉ of the volcanic history of this region as it may be gathered from the relations of the rocks and strata to the volcanic products, and of these latter to each other.

1st Stage.—It would appear that at the close of the Eocene period great terrestrial changes occurred. The bed of the sea was converted into dry land, the strata were flexured and denuded, and a depression was formed in the granitic floor of Central France, which, in the succeeding Miocene period, was converted into an extensive lake peopled by molluscs, fishes, reptiles, and pachyderms of the period.

2nd Stage.—Towards the close of the Miocene epoch volcanic eruptions commenced on a grand scale over the granitic platform in the districts now called Mont Dore, Cantal, and the Vivarais. Vast sheets of trachytic and basaltic lavas successively invaded the tracts surrounding the central orifices of eruption, now constituting the more ancient of the lava-sheets of the Auvergne region, and, invading the waters of the neighbouring lake, overspread the lacustrine deposits which were being accumulated therein. These volcanic eruptions probably continued throughout the Pliocene period, interrupted by occasional intervals of inactivity, and ultimately altogether ceased.

3rd Stage.—Towards the close of the Pliocene period terrestrial movements took place, owing to which the waters of the lake began to fall away, and the sheets of lava were subjected to great denudation. This process, probably accelerated by excessive rainfall during the succeeding Post-Pliocene and Pluvial periods, was continued until plains and extensive river-valleys were eroded out of the sheets of lava and their supporting granitic rocks and the adjoining lacustrine strata.

4th Stage.—A new outburst of volcanic forces marks this stage, during which the chain of the Puy de DÔme was thrown up on the west, and that of the newer cones of the Vivarais on the south-east of the lacustrine tract. The waters of the lake were now completely drained away through the channel of the Allier, and denudation, extending down to the present day, began over the area now forming the Vale of Clermont and adjoining districts. The volcanic action ultimately spent its force; and somewhere about the time of the appearance of man, the mammoth, rhinoceros, stag, and reindeer on the scene, eruptions entirely ceased, and gradually the region assumed those conditions of repose by which it is now physically characterised.