its carbonic acid with so large a proportion of carbonate of lime, which it deposits on issuing into the air, that its incrustations have formed an elevated natural aqueduct 240 feet in length, and terminating in an arch thrown across the stream it originally

2.

flowed into, 16 feet high and 12 wide. Near it are the rudiments of a similar arch, the construction of which is still going on, and aptly illustrates the formation of the other. Like many other similar incrusting springs in different countries, it has been turned into a source of emolument by the proprietor, who breaks the fall of the water in such a manner that its stony particles may be deposited on various natural objects exposed to its spray. At the time of my visit the stuffed skins of a horse and a cow were undergoing this petrifying process, as well as the usual proportion of birds, fruit, flowers, medals, cameos, &c.

There are several other springs in Auvergne possessed in a high degree of the same qualities. One at Chalucet, near Pontgibaud; another, called La Fontaine de Rambon, on the banks of the Crouze, near St. Floret. On both sides of this river, and for a considerable distance down the gorge it flows in, are seen colossal fragments of calcareous travertin, which by their position prove this mineral spring to have once erected a bridge similar to that of St. Alyre at Clermont, but exceeding it prodigiously in dimensions, and probably choking up the whole valley, since the source itself is elevated more than 100 feet above the river.

It is worthy of remark that the three springs mentioned above, whose deposits are, except in a greater or less proportion of iron, exactly alike, rise from rocks of different kinds: the first, from a calcareous peperino; the second, from the foot of a regular volcanic cone, at least twenty miles from any calcareous rock; the third, from granite. It is apparent from this that all have their origin in or below the granitic rocks which form the basis of the whole territory, and which include or cover the volcanic focus whence in reality these mineral springs in all probability ultimately derive. The same observation applies to the many thermal sources which occur on various points of the platform, springing indifferently from primitive or volcanic rocks; as at Mont Dore les Bains, La Bourboule, St. Nectaire, Châtelguyon, Gimeaux, Néris, Vichy, Vic en Carladèz, Chaudesaigues, &c. Some of these deposit a travertin having much silex as well as carbonate of lime in its composition, and arragonite is occasionally found to have crystallized in its fissures. There appears reason to believe that the quantity of mineral matter brought to the surface by such springs was much greater in earlier times, and is still annually diminishing.

II. BASIN OF THE CANTAL.

A freshwater formation, probably of the same age and very similar to that of the Limagne, occurs in the department of the Cantal, particularly in the neighbourhood of the chief town, Aurillac. Its principal distinction consists in the far greater abundance of silex associated with its calcareous marls and limestone. As in the Limagne formation, the lower beds are arenaceous, and derived apparently from the detritus of the rocks of gneiss and mica-schist within which the freshwater basin lay, and upon which these beds rest. The upper series consists of calcareous and siliceous marls, containing subordinate beds of silex, gypsum, and limestone. The bands of silex often very much resemble in position and aspect the flints of our English chalk, being sometimes continuous, sometimes in layers of concretionary nodules, assuming much the same forms as our flints, and, like them, having a white surface. Their substance is, however, generally more vitreous, like resinite, and often approaches to opal or agate. Its colours are usually yellowishbrown or greyish-blue, and it is sometimes ribboned delicately with stripes of these colours. The marly limestone, when pure from flint, is white or yellowish white, of an earthy and coarse fracture, full of shapeless tubular cavities and small filamentous perforations, such as would be left by reeds, grasses, or other weeds enveloped as they grew in a calcareous sediment or incrustation. It contains numerous shells of the genera Potamides, Helix, Limneus, Bulimus, Planorbis, &c.; with minute cypris and the seed-vessels of charæ. Sometimes the interior of the shell is lined with chalcedony, while the shell itself is of carbonate of lime; at others, the shell is of flint; the interior, limestone. The foliated marls are as thinly laminated as paper, owing probably to the myriads of small shells, or flattened stems

of charæ, preserved in them. Several hills are seen in the neighbourhood of Aurillac composed of such strata two or three hundred feet in thickness. They are generally covered by massive beds of volcanic breccia, trachyte, and basalt. In some places, as between Aurillac and Polminhac, there appear to be a confusion, and occasional alternations, of the volcanic and calcareous beds, similar to what has been described in the Limagne formation.

The original limits of the lake-basin of the Cantal can with difficulty be ascertained, owing to the colossal proportions of the volcanic mountain in the vicinity, by whose products its deposits have been overwhelmed. They are found, however, beneath the volcanic beds wherever any torrent discloses the inferior strata, within a space limited by lines passing through Jussac, Vic en Carladez, Mur de Barrèz, and Panet. But as they reappear on the north-east side of the central heights of the Cantal, near Murat, bearing the same characters as in the valleys of the Cère, the Goule, and the Jourdanne, it seems probable that the lake was continuous through the intervening area.

III. BASIN OF THE HAUTE LOIRE.

The freshwater formation of the basin of the Upper Loire which surrounds Le Puy differs but slightly from the two already described, and, like these, has been covered in part by prodigious and repeated extravasations of volcanic matter, which have loaded its generally-horizontal strata with massive coverings of basalt and basaltic breccia three or four hundred feet in thickness. It is, therefore, in general only by tracing up the deep waterworn ravines which furrow these superimposed rocks that the extent of the lacustrine deposits can be observed: the main valley of the Loire, however, and some of its tributaries, offer an ampler view of the beds through which they have been excavated.

The limits of the original basin show themselves on the west, in the base of the granitic chain which separates the waters of the Loire and the Allier. The rise of the granite platform towards the Haut Vivarais bounded it on the south; and some irregular embranchments from the heights of St. Bonnet and La Chaise Dieu on the east and north. One of these granitic spurs indeed stretches across the lacustrine formation, and separates it into two, cutting off the small basin of Emblaves from the upper and larger one, in which lies the town of Le Puy. The river Loire passes now from the latter basin into that of Emblaves through the narrow, deep, and sinuous gorge of La Voute, and finally issues again from this by means of the similar defile of Chamalières. Both of these outlets appear to owe their origin to some of the most recent changes to which this singular district has been subjected. Certainly neither could have been in existence at the time that the massive sheets of basalt and clinkstone which severally cap the cliff-ranges above them were in a state of igneous liquefaction.

The lower series of lacustrine beds consist, as in the formations already described, of sandstone, blue, green, and red variegated sandy marls, and clays. The sandstone is an excellent buildingstone; the clays are used for pottery.* The upper beds are chiefly of marly limestone, often highly siliceous, and enclosing

Doubts have been expressed with respect to these sandstones, &c., as well as the analogous beds which underlie the unquestionably freshwater and tertiary marls and limestone of the Limagne, whether they do not belong to the secondary formation of new red or variegated sandstone. The remains of plants which they occasionally contain indicate a marshy soil and atmosphere, consisting of large reeds, casts of cyclop. teris and pecopteris, with some small seeds and fruits which appear referable

to dicotyledonous plants. No shells have yet been found. In the corresponding sands of Auvergne a species of Cyrena is met with, and M. l'Abbé Croizet is, indeed, said to have found a bivalve apparently of marine origin; but the evidence of this isolated fact seems obscure, and not to be depended on, unless confirmed by future discoveries of the same character. See Appendix for the Fauna of the Tertiary strata.