Postponing to the next chapter the consideration of the Middle and Lower Eocene groups, I shall now speak of the Upper Eocene of Paris, and its foreign equivalents.

The upper freshwater marls and limestone (1. _a_) seem to have been formed in a great number of marshes and shallow lakes, such as frequently overspread the newest parts of great deltas. It appears that many layers of marl, tufaceous limestone, and travertin, with beds of flint, continuous or in nodules, acc.u.mulated in these lakes. _Charae_, aquatic plants, already alluded to (see p. 32.) left their stems and seed-vessels imbedded both in the marl and flint, together with freshwater and land sh.e.l.ls. Some of the siliceous rocks of this formation are used extensively for millstones. The flat summits or platforms of the hills round Paris, large areas in the forest of Fontainebleau, and the Plateau de la Beauce, between the Seine and the Loire, are chiefly composed of these upper freshwater strata.

The upper marine sands (1. _b_), consist chiefly of micaceous and quartzose sands, 80 feet thick. As they succeed throughout an extensive area deposit of a purely freshwater origin (2 _a_.), they appear to mark a subsidence of the subjacent soil, whether it had formed the bottom of an estuary or a lake. The sea, which afterwards took possession of the same s.p.a.ce, was inhabited by testacea, almost all of them differing from those found in the lower formations (2. _b_ and 2. _c_) and equally or still more distinct from the Miocene Faluns of subsequent date. One of these upper Eocene strata in the neighbourhood of Paris has been called the oyster bed, "couche a _Ostrea cyathula_, Lamk.," which is spread over a remarkably wide area. From the manner in which the oysters lie, it is inferred that they did not grow on the spot, but that some current swept them away from a bed of oysters formed in some other part of the bay. The strata of sand which immediately repose on the oyster-bed are quite dest.i.tute of organic remains; and nothing is more common in the Paris basin, and in other formations, than alternations of sh.e.l.ly beds with others entirely devoid of them. The temporary extinction and renewal of animal life at successive periods have been rashly inferred from such phenomena, which may nevertheless be explained, as M. Prevost justly remarks, without appealing to any such extraordinary revolutions in the state of the animate creation.

A current one day scoops out a channel in a bed of sh.e.l.ly sand and mud, and the next day, by a slight alteration of its course, ceases to prey upon the same bank. It may then become charged with sand unmixed with sh.e.l.ls, derived from some dune, or brought down by a river. In the course of ages an indefinite number of transitions from sh.e.l.ly strata to those without sh.e.l.ls may thus be caused.

Besides these oysters, M. Deshayes has described 29 species of sh.e.l.ls, in his work (Coquilles fossiles de Paris), as belonging to this formation, all save one regarded by him as differing from fossils of the calcaire grossier. Since that time the railway cuttings near Etampes have enabled M. Hebert to raise the number to 90. I have myself collected fossils in that district, where the sh.e.l.ls are very entire, and detachable from the yellow sandy matrix. M. Hebert first pointed out that most of them agree specifically with those of Kleyn Spauwen, Boom, and other localities of Limburg in Flanders, where they have been studied by MM. Nyst and De Koninck.[176-A]

The position in Belgium of this formation above the older Eocene group is well seen in the small hill of Pellenberg, rising abruptly from the great plain, half a mile south-east of the city of Louvain, where I examined it in company with M. Nyst in 1850. At the top of the hill, a thin bed of dark greyish green tile-clay is seen 1-1/2 foot thick, with casts of _Nucula Deshaysiana_. This clay rests on 12 feet of yellow sand, separated, by a band of flint and quartz pebbles, from a ma.s.s of subjacent white sand 15 feet thick, in which casts of the Kleyn Spauwen fossils have been met with.

Under this is a bed of yellow sand 12 feet thick, and, at a lower level, the railway cuttings have pa.s.sed through calcareous sands like those of Brussels, in which the _Nautilus Burtini_, and various sh.e.l.ls common to the older Eocene strata of the neighbourhood of London, have been obtained.

Every new fact which throws light on the true paleontological relations of the strata now under consideration, (the Upper Marine or Fontainebleau beds of the Paris basin, 1. _b_, p. 175.), deserves more particular attention, because geologists of high authority differ in opinion as to whether they should be cla.s.sed as Eocene or Miocene.

Professor Beyrich has lately described a formation of the same age, occurring within 7 miles of the gates of Berlin, near the village of Hermsdorf, where, in the midst of the sands of which that country chiefly consists, a ma.s.s of tile-clay, more than 40 feet thick, and of a dark blueish grey colour, is found, full of sh.e.l.ls, among which the genera _Fusus_ and _Pleurotoma_ predominate, and among the bivalves, _Nucula Deshaysiana_, Nyst, an extremely common sh.e.l.l in the Belgian beds above-mentioned. M. Beyrich has identified eighteen out of forty-five species of the Hermsdorf fossils with the Belgian species; and I believe that a much larger proportion agree with the Upper Eocene of Belgium, France, and the Rhine. On the other hand, eight of the forty-five species are supposed by him to agree with English Eocene sh.e.l.ls. Messrs. Morris, Edwards, and S. Wood have compared a small collection, which I obtained of these Berlin sh.e.l.ls, with the Eocene fossils of their museums, and confirmed the result of M. Beyrich, the species common to the English fossils belonging not simply to the uppermost of our marine beds, or those of Barton, but some of them to lower parts of the series, such as Bracklesham and Highgate. On the other hand, while these testacea, like those of Kleyn Spauwen and Etampes, present many a.n.a.logies to the Middle and Lower Eocene group, they differ widely from the Falun sh.e.l.ls,--a fact the more important in reference to Etampes, as that locality approaches within 70 miles of Pontlevoy, near Blois, and within 100 miles of Savigne, near Tours, where Falun sh.e.l.ls are found. It is evident that the discordance of species cannot be attributed to distance or geographical causes, but must be referred to time, or the different epoch at which the upper marine beds of the Paris basin and the Faluns of the Loire originated.

_Mayence._--The true chronological relation of many tertiary strata on the banks of the Rhine has always presented a problem of considerable difficulty. They occupy a tract from 5 to 12 miles in breadth, extending along the left bank of the Rhine from Mayence to the neighbourhood of Manheim, and are again found to the east, north, and south-west of Frankfort. In some places they have the appearance of a freshwater formation; but in others, as at Alzey, the sh.e.l.ls are for the most part marine. _Cerithia_ are in great profusion, which indicates that the sea where the deposit was formed was fed by rivers; and the great quant.i.ty of fossil land sh.e.l.ls, chiefly of the genus _Helix_, confirm the same opinion. The variety in the species of sh.e.l.ls is small, while the individuals are exceedingly numerous; a fact which accords perfectly with the idea that the formation may have originated in a gulf or sea which, like the Baltic, was brackish in some parts, and almost fresh in others. A species of _Paludina_ (fig. 154.), very nearly resembling the recent _Littorina ulva_, is found throughout this basin. These sh.e.l.ls are like grains of rice in size, and are often in such quant.i.ty as to form entire beds of marl and limestone. They are as thick as grains of sand, in stratified ma.s.ses from 15 to 30 feet in thickness.

[Ill.u.s.tration: Fig. 154. _Paludina._ Mayence.]

That these Rhenish tertiary formations agree more nearly with the Upper Eocene deposits above enumerated, than with any others, I have no doubt, since I had the advantage of comparing (August, 1850), with the a.s.sistance of M. De Koninck of Liege, the fossils from Kleyn Spauwen, Boom, and other Limburg localities, with those from Mayence, Alzey, Weinheim, and other Rhenish strata. Among the common Belgian and Rhenish sh.e.l.ls which are identical, I may mention _Ca.s.sidaria depressa_, _Tritonium flandric.u.m_ De Koninck, _Cerithium tricinctum_ Nyst, _Tornatella simulata_, _Rostellaria Sowerbyi_, _Nucula Deshaysiana_, _Corbula pisum_, and _Pectunculus terebratularis_.

From these Upper Eocene deposits of the Rhine M. H. von Meyer has obtained a great number of characteristic fossil mammalia, such as _Palaeomaeryx medius_, _Hyotherium Meissneri_, _Tapirus Helveticus_, _Anthracotherium Alsatic.u.m_, and others. The three first of these are species common to some of the lignite, or brown coal beds in Switzerland, commonly cla.s.sed with the mola.s.se, but of which the true age has not yet been distinctly made out.

The fossils of the sandy beds of Eppelsheim, comprising bones of the Deinotherium, Mastodon, and other quadrupeds, are regarded by H. von Meyer as belonging to a mammiferous fauna quite distinct from that of the Mayence basin, and they are probably referable to the Miocene period.

The upper freshwater strata (1. _a_, p. 175.), of the neighbourhood of Paris, stretch southwards from the valley of the Seine to that of the Loire, and in the last-mentioned region are seen to be older than the marine faluns, so that the perforating sh.e.l.ls of the Miocene sea have sometimes bored the hard compact freshwater limestones; and fragments of the Upper Eocene rocks are found at Pontlevoy and elsewhere, which have been rolled in the bed of the Miocene sea.

[Ill.u.s.tration: Fig. 155. Simplified geological map south of Paris.]

_Central France._--Lacustrine strata belonging, for the most part, to the same Upper Eocene series, are again met with in Auvergne, Cantal, and Velay, the sites of which may be seen in the annexed map. They appear to be the monuments of ancient lakes, which, like some of those now existing in Switzerland, once occupied the depressions in a mountainous region, and have been each fed by one or more rivers and torrents. The country where they occur is almost entirely composed of granite and different varieties of granitic schist, with here and there a few patches of secondary strata, much dislocated, and which have probably suffered great denudation. There are also some vast piles of volcanic matter (see the map), the greater part of which is newer than the freshwater strata, and is sometimes seen to rest upon them, while a small part has evidently been of contemporaneous origin.

Of these igneous rocks I shall treat more particularly in another part of this work.

Before entering upon any details, I may observe, that the study of these regions possesses a peculiar interest, very distinct in kind from that derivable from the investigation either of the Parisian or English tertiary strata. For we are presented in Auvergne with the evidence of a series of events of astonishing magnitude and grandeur, by which the original form and features of the country have been greatly changed, yet never so far obliterated but that they may still, in part at least, be restored in imagination. Great lakes have disappeared,--lofty mountains have been formed, by the reiterated emission of lava, preceded and followed by showers of sand and scoriae,--deep valleys have been subsequently furrowed out through ma.s.ses of lacustrine and volcanic origin,--at a still later date, new cones have been thrown up in these valleys,--new lakes have been formed by the damming up of rivers,--and more than one creation of quadrupeds, birds, and plants, Eocene, Miocene, and Pliocene, have followed in succession; yet the region has preserved from first to last its geographical ident.i.ty; and we can still recall to our thoughts its external condition and physical structure before these wonderful vicissitudes began, or while a part only of the whole had been completed. There was first a period when the s.p.a.cious lakes, of which we still may trace the boundaries, lay at the foot of mountains of moderate elevation, unbroken by the bold peaks and precipices of Mont Dor, and unadorned by the picturesque outline of the Puy de Dome, or of the volcanic cones and craters now covering the granitic platform. During this earlier scene of repose deltas were slowly formed; beds of marl and sand, several hundred feet thick, deposited; siliceous and calcareous rocks precipitated from the waters of mineral springs; sh.e.l.ls and insects imbedded, together with the remains of the crocodile and tortoise, the eggs and bones of water birds, and the skeletons of quadrupeds, some of them belonging to the same genera as those entombed in the Eocene gypsum of Paris. To this tranquil condition of the surface succeeded the era of volcanic eruptions, when the lakes were drained, and when the fertility of the mountainous district was probably enhanced by the igneous matter ejected from below, and poured down upon the more sterile granite. During these eruptions, which appear to have taken place after the disappearance of the Eocene fauna, and in the Miocene epoch, the mastodon, rhinoceros, elephant, tapir, hippopotamus, together with the ox, various kinds of deer, the bear, hyaena, and many beasts of prey, ranged the forest, or pastured on the plain, and were occasionally overtaken by a fall of burning cinders, or buried in flows of mud, such as accompany volcanic eruptions. Lastly, these quadrupeds became extinct, and gave place to Pliocene mammalia, and these, in their turn, to species now existing. There are no signs, during the whole time required for this series of events, of the sea having intervened, nor of any denudation which may not have been accomplished by currents in the different lakes, or by rivers and floods accompanying repeated earthquakes, during which the levels of the district have in some places been materially modified, and perhaps the whole upraised relatively to the surrounding parts of France.

_Auvergne._--The most northern of the freshwater groups is situated in the valley-plain of the Allier, which lies within the department of the Puy de Dome, being the tract which went formerly by the name of the Limagne d'Auvergne. It is inclosed by two parallel mountain ranges,--that of the Forez, which divides the waters of the Loire and Allier, on the east; and that of the Monts Domes, which separates the Allier from the Sioule, on the west.[181-A] The average breadth of this tract is about 20 miles; and it is for the most part composed of nearly horizontal strata of sand, sandstone, calcareous marl, clay, and limestone, none of which observe a fixed and invariable order of superposition. The ancient borders of the lake, wherein the freshwater strata were acc.u.mulated, may generally be traced with precision, the granite and other ancient rocks rising up boldly from the level country. The actual junction, however, of the lacustrine and granitic beds is rarely seen, as a small valley usually intervenes between them. The freshwater strata may sometimes be seen to retain their horizontality within a very slight distance of the border-rocks, while in some places they are inclined, and in few instances vertical. The princ.i.p.al divisions into which the lacustrine series may be separated are the following:--1st, Sandstone, grit, and conglomerate, including red marl and red sandstone.

2dly, Green and white foliated marls. 3dly, Limestone or travertin, often oolitic. 4thly, Gypseous marls.

1. _a_. _Sandstone and conglomerate._--Strata of sand and gravel, sometimes bound together into a solid rock, are found in great abundance around the confines of the lacustrine basin, containing, in different places, pebbles of all the ancient rocks of the adjoining elevated country; namely, granite, gneiss, mica-schist, clay-slate, porphyry, and others. But these strata do not form one continuous band around the margin of the basin, being rather disposed like the independent deltas which grow at the mouths of torrents along the borders of existing lakes.

At Chamalieres, near Clermont, we have an example of one of these deltas, or littoral deposits, of local extent, where the pebbly beds slope away from the granite, as if they had formed a talus beneath the waters of the lake near the steep sh.o.r.e. A section of about 50 feet in vertical height has been laid open by a torrent, and the pebbles are seen to consist throughout of rounded and angular fragments of granite, quartz, primary slate, and red sandstone; but without any intermixture of those volcanic rocks which now abound in the neighbourhood, and which could not have been there when the conglomerate was formed. Partial layers of lignite and pieces of wood are found in these beds.

At some localities on the margin of the basin quartzose grits are found; and, where these rest on granite, they are sometimes formed of separate crystals of quartz, mica, and felspar, derived from the disintegrated granite, the crystals having been subsequently bound together by a siliceous cement. In these cases the granite seems regenerated in a new and more solid form; and so gradual a pa.s.sage takes place between the rock of crystalline and that of mechanical origin, that we can scarcely distinguish where one ends and the other begins.

In the hills called the Puy de Jussat and La Roche, we have the advantage of seeing a section continuously exposed for about 700 feet in thickness.

At the bottom are foliated marls, white and green, about 400 feet thick; and above, resting on the marls, are the quartzose grits, cemented by calcareous matter, which is sometimes so abundant as to form imbedded nodules. These sometimes const.i.tute spheroidal concretions 6 feet in diameter, and pa.s.s into beds of solid limestone, resembling the Italian travertins, or the deposits of mineral springs. This section is close to the confines of the basin; so that the lake must here have been filled up near the sh.o.r.e with fine mud, before the coa.r.s.e superinc.u.mbent sand was introduced. There are other cases where sand is seen below the marl.

1. _b._ _Red marl and sandstone_.--But the most remarkable of the arenaceous groups is one of red sandstone and red marl, which are identical in all their mineral characters with the secondary _New Red sandstone_ and marl of England. In these secondary rocks the red ground is sometimes variegated with light greenish spots, and the same may be seen in the tertiary formation of freshwater origin at Coudes, on the Allier. The marls are sometimes of a purplish-red colour, as at Champheix, and are accompanied by a reddish limestone, like the well-known "cornstone," which is a.s.sociated with the Old Red sandstone of English geologists. The red sandstone and marl of Auvergne have evidently been derived from the degradation of gneiss and mica-schist, which are seen _in situ_ on the adjoining hills, decomposing into a soil very similar to the tertiary red sand and marl. We also find pebbles of gneiss, mica-schist, and quartz in the coa.r.s.er sandstones of this group, clearly pointing to the parent rocks from which the sand and marl are derived. The red beds, although dest.i.tute themselves of organic remains, pa.s.s upwards into strata containing Eocene fossils, and are certainly an integral part of the lacustrine formation.

From this example the student will learn how small is the value of mineral character alone, as a test of the relative age of rocks.

2. _Green and white foliated marls._--The same primary rocks of Auvergne, which, by the partial degradation of their harder parts, gave rise to the quartzose grits and conglomerates before mentioned, would, by the reduction of the same materials into powder, and by the decomposition of their felspar, mica, and hornblende, produce aluminous clay, and, if a sufficient quant.i.ty of carbonate of lime was present, calcareous marl. This fine sediment would naturally be carried out to a greater distance from the sh.o.r.e, as are the various finer marls now deposited in Lake Superior. And, as in the American lake, shingle and sand are annually ama.s.sed near the northern sh.o.r.es, so in Auvergne the grits and conglomerates before mentioned were evidently formed near the borders.

[Ill.u.s.tration: Fig. 156. _Cypris unifasciata_, a living species, greatly magnified.

_a._ Upper part.

_b._ Side view of the same.]

[Ill.u.s.tration: Fig. 157. _Cypris vidua_, a living species, greatly magnified.[183-A]]

The entire thickness of these marls is unknown; but it certainly exceeds, in some places, 700 feet. They are, for the most part, either light-green or white, and usually calcareous. They are thinly foliated,--a character which frequently arises from the innumerable thin sh.e.l.ls, or carapace-valves, of that small animal called _Cypris_; a genus which comprises several species, of which some are recent, and may be seen swimming swiftly through the waters of our stagnant pools and ditches. The antennae, at the end of which are fine pencils of hair, are the princ.i.p.al organs of motion, and are seen to vibrate with great rapidity. This animal resides within two small valves, not unlike those of a bivalve sh.e.l.l, and moults its integuments periodically, which the conchiferous mollusks do not. This circ.u.mstance may partly explain the countless myriads of the sh.e.l.ls of _Cypris_ which were shed in the ancient lakes of Auvergne, so as to give rise to divisions in the marl as thin as paper, and that, too, in stratified ma.s.ses several hundred feet thick. A more convincing proof of the tranquillity and clearness of the waters, and of the slow and gradual process by which the lake was filled up with fine mud, cannot be desired.

But we may easily suppose that, while this fine sediment was thrown down in the deep and central parts of the basin, gravel, sand, and rocky fragments were hurried into the lake, and deposited near the sh.o.r.e, forming the group described in the preceding section.

Not far from Clermont, the green marls, containing the _Cypris_ in abundance, approach to within a few yards of the granite which forms the borders of the basin. The occurrence of these marls so near the ancient margin may be explained by considering that, at the bottom of the ancient lake, no coa.r.s.e ingredients were deposited in s.p.a.ces intermediate between the points where rivers and torrents entered, but finer mud only was drifted there by currents. The _verticality_ of some of the beds in the above section bears testimony to considerable local disturbance subsequent to the deposition of the marls; but such inclined and vertical strata are very rare.

[Ill.u.s.tration: Fig. 158. Vertical strata of marl, at Champradelle, near Clermont.

A. Granite.

B. s.p.a.ce of 60 feet, in which no section is seen.

C. Green marl, vertical and inclined.

D. White marl.]

3. _Limestone, travertin, oolite._--Both the preceding members of the lacustrine deposit, the marls and grits, pa.s.s occasionally into limestone. Sometimes only concretionary nodules abound in them; but these, where there is an increase in the quant.i.ty of calcareous matter, unite into regular beds.

On each side of the basin of the Limagne, both on the west at Gannat, and on the east at Vichy, a white oolitic limestone is quarried. At Vichy, the oolite resembles our Bath stone in appearance and beauty; and, like it, is soft when first taken from the quarry, but soon hardens on exposure to the air. At Gannat, the stone contains land-sh.e.l.ls and bones of quadrupeds, resembling those of the Paris gypsum. At Chadrat, in the hill of La Serre, the limestone is pisolitic, the small spheroids combining both the radiated and concentric structure.

_Indusial limestone._--There is another remarkable form of freshwater limestone in Auvergne, called "indusial," from the cases, or _indusiae_, of caddis-worms (the larvae of _Phryganea_); great heaps of which have been incrusted, as they lay, by carbonate of lime, and formed into a hard travertin. The rock is sometimes purely calcareous, but there is occasionally an intermixture of siliceous matter. Several beds of it are frequently seen, either in continuous ma.s.ses, or in concretionary nodules, one upon another, with layers of marl interposed. The annexed drawing (fig. 159.) will show the manner in which one of these indusial beds (_a_) is laid open at the surface, between the marls (_b b_), near the base of the hill of Gergovia; and affords, at the same time, an example of the extent to which the lacustrine strata, which must once have filled a hollow, have been denuded, and shaped out into hills and valleys, on the site of the ancient lakes.

[Ill.u.s.tration: Fig. 159. Bed of indusial limestone, interstratified with freshwater marl, near Clermont (Kleinschrod.)]

[Ill.u.s.tration: Fig. 160. Larva of recent Phryganea.[185-A]]

[Ill.u.s.tration: Fig. 161.

_a_. Indusial limestone of Auvergne.

_b_. Fossil _Paludina_ magnified.]

We may often observe in our ponds the _Phryganea_ (or Caddis-fly), in its caterpillar state, covered with small freshwater sh.e.l.ls, which they have the power of fixing to the outside of their tubular cases, in order, probably, to give them weight and strength. The individual figured in the annexed cut, which belongs to a species very abundant in England, has covered its case with sh.e.l.ls of a small _Planorbis_. In the same manner a large species of caddis-worm, which swarmed in the Eocene lakes of Auvergne, was accustomed to attach to its dwelling the sh.e.l.ls of a small spiral univalve of the genus _Paludina_. A hundred of these minute sh.e.l.ls are sometimes seen arranged around one tube, part of the central cavity of which is often empty, the rest being filled up with thin concentric layers of travertin. The cases have been thrown together confusedly, and often lie, as in fig. 161., at right angles one to the other. When we consider that ten or twelve tubes are packed within the compa.s.s of a cubic inch, and that some single strata of this limestone are 6 feet thick, and may be traced over a considerable area, we may form some idea of the countless number of insects and mollusca which contributed their integuments and sh.e.l.ls to compose this singularly constructed rock. It is unnecessary to suppose that the _Phryganeae_ lived on the spots where their cases are now found; they may have multiplied in the shallows near the margin of the lake, or in the streams by which it was fed, and their cases may have been drifted by a current far into the deep water.

In the summer of 1837, when examining, in company with Dr. Beck, a small lake near Copenhagen, I had an opportunity of witnessing a beautiful exemplification of the manner in which the tubular cases of Auvergne were probably acc.u.mulated. This lake, called the Fuure-Soe, occurring in the interior of Seeland, is about twenty English miles in circ.u.mference, and in some parts 200 feet in depth. Round the shallow borders an abundant crop of reeds and rushes may be observed, covered with the indusiae of the _Phryganea grandis_ and other species, to which sh.e.l.ls are attached. The plants which support them are the bullrush, _Scirpus lacustris_, and common reed, _Arundo phragmitis_, but chiefly the former. In summer, especially in the month of June, a violent gust of wind sometimes causes a current by which these plants are torn up by the roots, washed away, and floated off in long bands, more than a mile in length, into deep water. The _Cypris_ swarms in the same lake; and calcareous springs alone are wanting to form extensive beds of indusial limestone, like those of Auvergne.

4. _Gypseous marls._--More than 50 feet of thinly laminated gypseous marls, exactly resembling those in the hill of Montmartre, at Paris, are worked for gypsum at St. Romain, on the right bank of the Allier. They rest on a series of green cypriferous marls which alternate with grit, the united thickness of this inferior group being seen, in a vertical section on the banks of the river, to exceed 250 feet.

_General arrangement, origin, and age of the freshwater formations of Auvergne._--The relations of the different groups above described cannot be learnt by the study of any one section; and the geologist who sets out with the expectation of finding a fixed order of succession may perhaps complain that the different parts of the basin give contradictory results. The arenaceous division, the marls, and the limestone, may all be seen in some places to alternate with each other; yet it can, by no means, be affirmed that there is no order of arrangement. The sands, sandstone, and conglomerate, const.i.tute in general a littoral group; the foliated white and green marls, a contemporaneous central deposit; and the limestone is for the most part subordinate to the newer portions of both. The uppermost marls and sands are more calcareous than the lower; and we never meet with calcareous rocks covered by a considerable thickness of quartzose sand or green marl. From the resemblance of the limestones to the Italian travertins, we may conclude that they were derived from the waters of mineral springs,--such springs as even now exist in Auvergne, and which may be seen rising up through the granite, and precipitating travertin. They are sometimes thermal, but this character is by no means constant.

It seems that, when the ancient lake of the Limagne first began to be filled with sediment, no volcanic action had yet produced lava and scoriae on any part of the surface of Auvergne. No pebbles, therefore, of lava were transported into the lake,--no fragments of volcanic rocks embedded in the conglomerate. But at a later period, when a considerable thickness of sandstone and marl had acc.u.mulated, eruptions broke out, and lava and tuff were deposited, at some spots, alternately with the lacustrine strata. It is not improbable that cold and thermal springs, holding different mineral ingredients in solution, became more numerous during the successive convulsions attending this development of volcanic agency, and thus deposits of carbonate and sulphate of lime, silex, and other minerals were produced. Hence these minerals predominate in the uppermost strata. The subterranean movements may then have continued until they altered the relative levels of the country, and caused the waters of the lakes to be drained off, and the farther acc.u.mulation of regular freshwater strata to cease.

We may easily conceive a similar series of events to give rise to a.n.a.logous results in any modern basin, such as that of Lake Superior, for example, where numerous rivers and torrents are carrying down the detritus of a chain of mountains into the lake. The transported materials must be arranged according to their size and weight, the coa.r.s.er near the sh.o.r.e, the finer at a greater distance from land; but in the gravelly and sandy beds of Lake Superior no pebbles of modern volcanic rocks can be included, since there are none of these at present in the district. If igneous action should break out in that country, and produce lava, scoriae, and thermal springs, the deposition of gravel, sand, and marl might still continue as before; but, in addition, there would then be an intermixture of volcanic gravel and tuff, and of rocks precipitated from the waters of mineral springs.

Although the freshwater strata of the Limagne approach generally to a horizontal position, the proofs of local disturbance are sufficiently numerous and violent to allow us to suppose great changes of level since the lacustrine period. We are unable to a.s.sign a northern barrier to the ancient lake, although we can still trace its limits to the east, west, and south, where they were formed of bold granite eminences. Nor need we be surprised at our inability to restore entirely the physical geography of the country after so great a series of volcanic eruptions; for it is by no means improbable that one part of it, the southern, for example, may have been moved upwards bodily, while others remained at rest, or even suffered a movement of depression.

Whether all the freshwater formations of the Limagne d'Auvergne belong to one period, I cannot pretend to decide, as large ma.s.ses both of the arenaceous and marly groups are often devoid of fossils. Much light has been thrown on the mammiferous fauna by the labours of MM. Bravard and Croizet, and by those of M. Pomel. The last-mentioned naturalist has pointed out the specific distinction of all, or nearly all, the species of mammalia, from those of the gypseous series near Paris. Nevertheless, many of the forms are a.n.a.logous to those of Eocene quadrupeds. The _Cainotherium_, for example, is not far removed from the _Anoplotherium_, and is, according to Waterhouse, the same as the genus _Microtherium_ of the Germans. There are two species of marsupial animals allied to _Didelphys_, a genus also found in the Paris gypsum. The _Amphitragulus elegans_ of Pomel, has been identified with a Rhenish species from Weissenau near Mayence, called by Kaup _Dorcatherium nanum_; and other Auvergne fossils, e.g., _Microtherium Reuggeri_, and a small rodent, _t.i.tanomys_, are specifically the same with mammalia of the Mayence basin.