Sir Charles Lyell

Trap dikes — sometimes project — sometimes leave fissures vacant by decomposition — Branches and veins of trap — Dikes more crystalline in the centre — Foreign fragments of rock imbedded — Strata altered at or near the contact — Obliteration of organic remains — Conversion of chalk into marble — and of coal into coke — Inequality in the modifying influence of dikes — Trap interposed between strata — Columnar and globular structure — Relation of trappean rocks to the products of active volcanos — Submarine lava and ejected matter corresponds generally to ancient trap — Structure and physical features of Palma and some other extinct volcanos.

Having in the last chapter spoken of the composition and mineral characters of volcanic rocks, I shall next describe the manner and position in which they occur in the earth's crust, and their external forms. Now the leading varieties, such as basalt, greenstone, trachyte, porphyry, and the rest, are found sometimes in dikes penetrating stratified and unstratified formations, sometimes in shapeless masses protruding through or overlying them, or in horizontal sheets intercalated between strata.

Volcanic dikes.—Fissures have already been spoken of as occurring in all kinds of rocks, some a few feet, others many yards in width, and often filled up with earth or angular pieces of stone, or with sand and pebbles. Instead of such materials, suppose a quantity of melted stone to be driven or injected into an open rent, and there consolidated, we have then a tabular mass resembling a wall, and called a trap dike. It is not uncommon to find such dikes passing through strata of soft materials, such as tuff or shale, which, being more perishable than the trap, are often washed away by the sea, rivers, or rain, in which case the dike stands prominently out in the face of precipices, or on the level surface of a country. (See the annexed figure.[378-A])

In the islands of Arran, Skye, and other parts of Scotland, where sandstone, conglomerate, and other hard rocks are traversed by dikes of trap, the converse of the above phenomenon is seen. The dike having decomposed more rapidly than the containing rock, has once more left open the original fissure, often for a distance of many yards inland from the sea-coast, as represented in the annexed view (fig. 440.). In these instances, the greenstone of the dike is usually more tough and hard than the sandstone; but chemical action, and chiefly the oxidation of the iron, has given rise to the more rapid decay.

There is yet another case, by no means uncommon in Arran and other parts of Scotland, where the strata in contact with the dike, and for a certain distance from it, have been hardened, so as to resist the action of the weather more than the dike itself, or the surrounding rocks. When this happens, two parallel walls of indurated strata are seen protruding above the general level of the country, and following the course of the dike.

As fissures sometimes send off branches, or divide into two or more fissures of equal size, so also we find trap dikes bifurcating and ramifying, and sometimes they are so tortuous as to be called veins, though this is more common in granite than in trap. The accompanying sketch (fig. 441.) by Dr. MacCulloch represents part of a sea-cliff in Argyleshire, where an overlying mass of trap, b, sends out some veins which terminate downwards. Another trap vein, a a, cuts through both the limestone, c, and the trap, b.

In fig. 442., a ground plan is given of a ramifying dike of greenstone, which I observed cutting through sandstone on the beach near Kildonan Castle, in Arran. The larger branch varies from 5 to 7 feet in width, which will afford a scale of measurement for the whole.

In the Hebrides and other countries, the same masses of trap which occupy the surface of the country far and wide, concealing the subjacent stratified rocks, are seen also in the sea cliffs, prolonged downwards in veins or dikes, which probably unite with other masses of igneous rock at a greater depth. The largest of the dikes represented in the annexed diagram, and which are seen in part of the coast of Skye, is no less than 100 feet in width.

Every variety of trap-rock is sometimes found in these dikes, as basalt, greenstone, felspar-porphyry, and more rarely trachyte. The amygdaloidal traps also occur, and even tuff and breccia, for the materials of these last may be washed down into open fissures at the bottom of the sea, or during eruptions on the land may be showered into them from the air.

Some dikes of trap may be followed for leagues uninterruptedly in nearly a straight direction, as in the north of England, showing that the fissures which they fill must have been of extraordinary length.

Dikes more crystalline in the centre.—In many cases trap at the edges or sides of a dike is less crystalline or more earthy than in the centre, in consequence of the melted matter having cooled more rapidly by coming in contact with the cold sides of the fissure; whereas, in the centre, the matter of the dike being kept long in a fluid or soft state, the crystals are slowly formed. In the ancient part of Vesuvius, called Somma, a thin band of half-vitreous lava is found at the edge of some dikes. At the junction of greenstone dikes with limestone, a sahlband, or selvage, of serpentine is occasionally observed.

On the left shore of the fiord of Christiania, in Norway, I examined, in company with Professor Keilhau, a remarkable dike of syenitic greenstone, which is traced through Silurian strata, until at length, in the promontory of NÆsodden, it enters mica-schist. Fig. 444. represents a ground plan, where the dike appears 8 paces in width. In the middle it is highly crystalline and granitiform, of a purplish colour, and containing a few crystals of mica, and strongly contrasted with the whitish mica-schist, between which and the syenitic rock there is usually on each side a distinct black band, 18 inches wide, of dark greenstone. When first seen, these bands have the appearance of two accompanying dikes; yet they are, in fact, only the different form which the syenitic materials have assumed where near to or in contact with the mica-schist. At one point, a, one of the sahlbands terminates for a space; but near this there is a large detached block, b, having a gneiss-like structure, consisting of hornblende and felspar, which is included in the midst of the dike. Round this a smaller encircling zone is seen, of dark basalt, or fine-grained greenstone, nearly corresponding to the larger ones which border the dike, but only 1 inch wide.

It seems, therefore, evident that the fragment, b, has acted on the matter of the dike, probably by causing it to cool more rapidly, in the same manner as the walls of the fissure have acted on a larger scale. The facts, also, illustrate the facility with which a granitiform syenite may pass into ordinary rocks of the volcanic family.

The fact above alluded to, of a foreign fragment, such as b, fig. 444., included in the midst of the trap, as if torn off from some subjacent rock or the walls of a fissure, is by no means uncommon. A fine example is seen in another dike of greenstone, 10 feet wide, in the northern suburbs of Christiania, in Norway, of which the annexed figure is a ground plan. The dike passes through shale, known by its fossils to belong to the Silurian series. In the black base of greenstone are angular and roundish pieces of gneiss, some white, others of a light flesh-colour, some without lamination, like granite, others with laminÆ, which, by their various and often opposite directions, show that they have been scattered at random through the matrix. These imbedded pieces of gneiss measure from 1 to about 8 inches in diameter.

Rocks altered by volcanic dikes.—After these remarks on the form and composition of dikes themselves, I shall describe the alterations which they sometimes produce in the rocks in contact with them. The changes are usually such as the intense heat of melted matter and the entangled gases might be expected to cause.

Plas-Newydd.—A striking example, near Plas-Newydd, in Anglesea, has been described by Professor Henslow.[381-A] The dike is 134 feet wide, and consists of a rock which is a compound of felspar and augite (dolerite of some authors). Strata of shale and argillaceous limestone, through which it cuts perpendicularly, are altered to a distance of 30, or even, in some places, to 35 feet from the edge of the dike. The shale, as it approaches the trap, becomes gradually more compact, and is most indurated where nearest the junction. Here it loses part of its schistose structure, but the separation into parallel layers is still discernible. In several places the shale is converted into hard porcellanous jasper. In the most hardened part of the mass the fossil shells, principally Producti, are nearly obliterated; yet even here their impressions may frequently be traced. The argillaceous limestone undergoes analogous mutations, losing its earthy texture as it approaches the dike, and becoming granular and crystalline. But the most extraordinary phenomenon is the appearance in the shale of numerous crystals of analcime and garnet, which are distinctly confined to those portions of the rock affected by the dike.[382-A] Some garnets contain as much as 20 per cent. of lime, which they may have derived from the decomposition of the fossil shells or Producti. The same mineral has been observed, under very analogous circumstances, in High Teesdale, by Professor Sedgwick, where it also occurs in shale and limestone, altered by basalt.[382-B]

Antrim.—In several parts of the county of Antrim, in the north of Ireland, chalk with flints is traversed by basaltic dikes. The chalk is there converted into granular marble near the basalt, the change sometimes extending 8 or 10 feet from the wall of the dike, being greatest near the point of contact, and thence gradually decreasing till it becomes evanescent. "The extreme effect," says Dr. Berger, "presents a dark brown crystalline limestone, the crystals running in flakes as large as those of coarse primitive (metamorphic) limestone; the next state is saccharine, then fine grained and arenaceous; a compact variety, having a porcellanous aspect and a bluish-grey colour, succeeds: this, towards the outer edge, becomes yellowish-white, and insensibly graduates into the unaltered chalk. The flints in the altered chalk usually assume a grey yellowish colour."[382-C] All traces of organic remains are effaced in that part of the limestone which is most crystalline.

The annexed drawing (fig. 446.) represents three basaltic dikes traversing the chalk, all within the distance of 90 feet. The chalk contiguous to the two outer dikes is converted into a finely granular marble, m m, as are the whole of the masses between the outer dikes and the central one. The entire contrast in the composition and colour of the intrusive and invaded rocks, in these cases, renders the phenomena peculiarly clear and interesting.

Another of the dikes of the north-east of Ireland has converted a mass of red sandstone into hornstone.[382-E] By another, the slate clay of the coal measures has been indurated, and has assumed the character of flinty slate[383-A]; and in another place the slate clay of the lias has been changed into flinty slate, which still retains numerous impressions of ammonites.[383-B]

It might have been anticipated that beds of coal would, from their combustible nature, be effected in an extraordinary degree by the contact of melted rock. Accordingly, one of the greenstone dikes of Antrim, on passing through a bed of coal, reduces it to a cinder for the space of 9 feet on each side.[383-C]

At Cockfield Fell, in the north of England, a similar change is observed. Specimens taken at the distance of about 30 yards from the trap are not distinguishable from ordinary pit coal; those nearer the dike are like cinders, and have all the character of coke; while those close to it are converted into a substance resembling soot.[383-D]

As examples might be multiplied without end, I shall merely select one or two others, and then conclude. The rock of Stirling Castle is a calcareous sandstone, fractured and forcibly displaced by a mass of greenstone which has evidently invaded the strata in a melted state. The sandstone has been indurated, and has assumed a texture approaching to hornstone near the junction. In Arthur's Seat and Salisbury Craig, near Edinburgh, a sandstone which comes in contact with greenstone is converted into a jaspideous rock.[383-E]

The secondary sandstones in Skye are converted into solid quartz in several places, where they come in contact with veins or masses of trap; and a bed of quartz, says Dr. MacCulloch, found near a mass of trap, among the coal strata of Fife, was in all probability a stratum of ordinary sandstone, having been subsequently indurated and turned into quartzite by the action of heat.[383-F]

But although strata in the neighbourhood of dikes are thus altered in a variety of cases, shale being turned into flinty slate or jasper, limestone into crystalline marble, sandstone into quartz, coal into coke, and the fossil remains of all such strata wholly and in part obliterated, it is by no means uncommon to meet with the same rocks, even in the same districts, absolutely unchanged in the proximity of volcanic dikes.

This great inequality in the effects of the igneous rocks may often arise from an original difference in their temperature, and in that of the entangled gases, such as is ascertained to prevail in different lavas, or in the same lava near its source and at a distance from it. The power also of the invaded rocks to conduct heat may vary, according to their composition, structure, and the fractures which they may have experienced, and perhaps, also, according to the quantity of water (so capable of being heated) which they contain. It must happen in some cases that the component materials are mixed in such proportions as prepare them readily to enter into chemical union, and form new minerals; while in other cases the mass may be more homogeneous, or the proportions less adapted for such union.

We must also take into consideration, that one fissure may be simply filled with lava, which may begin to cool from the first; whereas in other cases the fissure may give passage to a current of melted matter, which may ascend for days or months, feeding streams which are overflowing the country above, or are ejected in the shape of scoriÆ from some crater. If the walls of a rent, moreover, are heated by hot vapour before the lava rises, as we know may happen on the flanks of a volcano, the additional caloric supplied by the dike and its gases will act more powerfully.

Intrusion of trap between strata.—In proof of the mechanical force which the fluid trap has sometimes exerted on the rocks into which it has intruded itself, I may refer to the Whin-Sill, where a mass of basalt, from 60 to 80 feet in height, represented by a, fig. 447., is in part wedged in between the rocks of limestone, b, and shale, c, which have been separated from the great mass of limestone and shale, d, with which they were united.

The shale in this place is indurated; and the limestone, which at a distance from the trap is blue, and contains fossil corals, is here converted into granular marble without fossils.

Masses of trap are not unfrequently met with intercalated between strata, and maintaining their parallelism to the planes of stratification throughout large areas. They must in some places have forced their way laterally between the divisions of the strata, a direction in which there would be the least resistance to an advancing fluid, if no vertical rents communicated with the surface, and a powerful hydrostatic pressure was caused by gases propelling the lava upwards.

Columnar and globular structure.—One of the characteristic forms of volcanic rocks, especially of basalt, is the columnar, where large masses are divided into regular prisms, sometimes easily separable, but in other cases adhering firmly together. The columns vary in the number of angles, from three to twelve; but they have most commonly from five to seven sides. They are often divided transversely, at nearly equal distances, like the joints in a vertebral column, as in the Giant's Causeway, in Ireland. They vary exceedingly in respect to length and diameter. Dr. MacCulloch mentions some in Skye which are about 400 feet long; others, in Morven, not exceeding an inch. In regard to diameter, those of Ailsa measure 9 feet, and those of Morven an inch or less.[385-A] They are usually straight, but sometimes curved; and examples of both these occur in the island of Staffa. In a horizontal bed or sheet of trap the columns are vertical; in a vertical dike they are horizontal. Among other examples of the last-mentioned phenomenon is the mass of basalt, called the Chimney, in St. Helena (see fig. 448.), a pile of hexagonal prisms, 64 feet high, evidently the remainder of a narrow dike, the walls of rock which the dike originally traversed having been removed down to the level of the sea. In fig. 449. a small portion of this dike is represented on a less reduced scale.[385-B]

It being assumed that columnar trap has consolidated from a fluid state, the prisms are said to be always at right angles to the cooling surfaces. If these surfaces, therefore, instead of being either perpendicular, or horizontal, are curved, the columns ought to be inclined at every angle to the horizon; and there is a beautiful exemplification of this phenomenon in one of the valleys of the Vivarais, a mountainous district in the South of France, where, in the midst of a region of gneiss, a geologist encounters unexpectedly several volcanic cones of loose sand and scoriÆ. From the crater of one of these cones called La Coupe d'Ayzac, a stream of lava descends and occupies the bottom of a narrow valley, except at those points where the river Volant, or the torrents which join it, have cut away portions of the solid lava. The accompanying sketch (fig. 450.) represents the remnant of the lava at one of the points where a lateral torrent joins the main valley of the Volant. It is clear that the lava once filled the whole valley up to the dotted line d a; but the river has gradually swept away all below that line, while the tributary torrent has laid open a transverse section; by which we perceive, in the first place, that the lava is composed, as usual in this country, of three parts: the uppermost, at a, being scoriaceous; the second, b, presenting irregular prisms; and the third, c, with regular columns, which are vertical on the banks of the Volant, where they rest on a horizontal base of gneiss, but which are inclined at an angle of 45° at g, and then horizontal at f, their position having been every where determined, according to the law before mentioned, by the concave form of the original valley.

In the annexed figure (451.) a view is given of some of the inclined and curved columns which present themselves on the sides of the valleys in the hilly region north of Vicenza, in Italy, and at the foot of the higher Alps.[386-A] Unlike those of the Vivarais, last mentioned, the basalt of this country was evidently submarine, and the present valleys have since been hollowed out by denudation.

The columnar structure is by no means peculiar to the trap rocks in which hornblende or augite predominate; it is also observed in clinkstone, trachyte, and other felspathic rocks of the igneous class, although in these it is rarely exhibited in such regular polygonal forms.

It has been already stated that basaltic columns are often divided by cross joints. Sometimes each segment, instead of an angular, assumes a spheroidal form, so that a pillar is made up of a pile of balls, usually flattened, as in the Cheese-grotto at Bertrich-Baden, in the Eifel, near the Moselle (fig. 452.). The basalt, there, is part of a small stream of lava, from 30 to 40 feet thick, which has proceeded from one of several volcanic craters, still extant, on the neighbouring heights. The position of the lava bordering the river in this valley might be represented by a section like that already given at fig. 450. p. 385., if we merely supposed inclined strata of slate and the argillaceous sandstone called greywackÉ to be substituted for gneiss.

In some masses of decomposing greenstone, basalt, and other trap rocks, the globular structure is so conspicuous that the rock has the appearance of a heap of large cannon balls.

A striking example of this structure occurs in a resinous trachyte or pitchstone-porphyry in one of the Ponza islands, which rise from the Mediterranean, off the coast of Terracina and Gaeta. The globes vary from a few inches to three feet in diameter, and are of an ellipsoidal form (see fig. 453.). The whole rock is in a state of decomposition, "and when the balls," says Mr. Scrope, "have been exposed a short time to the weather, they scale off at a touch into numerous concentric coats, like those of a bulbous root, inclosing a compact nucleus. The laminÆ of this nucleus have not been so much loosened by decomposition; but the application of a ruder blow will produce a still further exfoliation."[387-A]

A fissile texture is occasionally assumed by clinkstone and other trap rocks, so that they have been used for roofing houses. Sometimes the prismatic and slaty structure is found in the same mass. The causes which give rise to such arrangements are very obscure, but are supposed to be connected with changes of temperature during the cooling of the mass, as will be pointed out in the sequel. (See Chaps. XXXV. and XXXVI.)

Relation of Trappean Rocks to the products of active Volcanos.

When we reflect on the changes above described in the strata near their contact with trap dikes, and consider how great is the analogy in composition and structure of the rocks called trappean and the lavas of active volcanos, it seems difficult at first to understand how so much doubt could have prevailed for half a century as to whether trap was of igneous or aqueous origin. To a certain extent, however, there was a real distinction between the trappean formations and those to which the term volcanic was almost exclusively confined. The trappean rocks first studied in the north of Germany, and in Norway, France, Scotland, and other countries, were either such as had been formed entirely under deep water, or had been injected into fissures and intruded between strata, and which had never flowed out in the air, or over the bottom of a shallow sea. When these products, therefore, of submarine or subterranean igneous action were contrasted with loose cones of scoriÆ, tuff, and lava, or with narrow streams of lava in great part scoriaceous and porous, such as were observed to have proceeded from Vesuvius and Etna, the resemblance seemed remote and equivocal. It was, in truth, like comparing the roots of a tree with its leaves and branches, which, although they belong to the same plant, differ in form, texture, colour, mode of growth, and position. The external cone, with its loose ashes and porous lava, may be likened to the light foliage and branches, and the rocks concealed far below, to the roots. But it is not enough to say of the volcano,

for its roots do literally reach downwards to Tartarus, or to the regions of subterranean fire; and what is concealed far below, is probably always more important in volume and extent than what is visible above ground.

We have already stated how frequently dense masses of strata have been removed by denudation from wide areas (see Chap. VI.); and this fact prepares us to expect a similar destruction of whatever may once have formed the uppermost part of ancient submarine or subaerial volcanos, more especially as those superficial parts are always of the lightest and most perishable materials. The abrupt manner in which dikes of trap usually terminate at the surface (see fig. 454.), and the water-worn pebbles of trap in the alluvium which covers the dike, prove incontestably that whatever was uppermost in these formations has been swept away. It is easy, therefore, to conceive that what is gone in regions of trap may have corresponded to what is now visible in active volcanos.

It will be seen in the following chapters, that in the earth's crust there are volcanic tuffs of all ages, containing marine shells, which bear witness to eruptions at many successive geological periods. These tuffs, and the associated trappean rocks, must not be compared to lava and scoriÆ which had cooled in the open air. Their counterparts must be sought in the products of modern submarine volcanic eruptions. If it be objected that we have no opportunity of studying these last, it may be answered, that subterranean movements have caused, almost everywhere in regions of active volcanos, great changes in the relative level of land and sea, in times comparatively modern, so as to expose to view the effects of volcanic operations at the bottom of the sea.

Thus, for example, the recent examination of the igneous rocks of Sicily, especially those of the Val di Noto, has proved that all the more ordinary varieties of European trap have been there produced under the waters of the sea, at a modern period; that is to say, since the Mediterranean has been inhabited by a great proportion of the existing species of testacea.

These igneous rocks of the Val di Noto, and the more ancient trappean rocks of Scotland and other countries, differ from subaerial volcanic formations in being more compact and heavy, and in forming sometimes extensive sheets of matter intercalated between marine strata, and sometimes stratified conglomerates, of which the rounded pebbles are all trap. They differ also in the absence of regular cones and craters, and in the want of conformity of the lava to the lowest levels of existing valleys.

It is highly probable, however, that insular cones did exist in some parts of the Val di Noto: and that they were removed by the waves, in the same manner as the cone of Graham island, in the Mediterranean, was swept away in 1831, and that of NyÖe, off Iceland, in 1783.[389-A] All that would remain in such cases, after the bed of the sea has been upheaved and laid dry, would be dikes and shapeless masses of igneous rock, cutting through sheets of lava which may have spread over the level bottom of the sea, and strata of tuff, formed of materials first scattered far and wide by the winds and waves, and then deposited. Trap conglomerates also, to which the action of the waves must give rise during the denudation of such volcanic islands, will emerge from the deep whenever the bottom of the sea becomes land.

The proportion of volcanic matter which is originally submarine must always be very great, as those volcanic vents which are not entirely beneath the sea, are almost all of them in islands, or, if on continents, near the shore. This may explain why extended sheets of trap so often occur, instead of narrow threads, like lava streams. For, a multitude of causes tend, near the land, to reduce the bottom of the sea to a nearly uniform level,—the sediment of rivers,—materials transported by the waves and currents of the sea from wasting cliffs,—showers of sand and scoriÆ ejected by volcanos, and scattered by the wind and waves. When, therefore, lava is poured out on such a surface, it will spread far and wide in every direction in a liquid sheet, which may afterwards, when raised up, form the tabular capping of the land.

As to the absence of porosity in the trappean formations, the appearances are in a great degree deceptive, for all amygdaloids are, as already explained, porous rocks, into the cells of which mineral matter, such as silex, carbonate of lime, and other ingredients, have been subsequently introduced (see p. 373.); sometimes, perhaps, by secretion during the cooling and consolidation of lavas.

In the Little Cumbray, one of the Western Islands, near Arran, the amygdaloid sometimes contains elongated cavities filled with brown spar; and when the nodules have been washed out, the interior of the cavities is glazed with the vitreous varnish so characteristic of the pores of slaggy lavas. Even in some parts of this rock which are excluded from air and water, the cells are empty, and seem to have always remained in this state, and are therefore undistinguishable from some modern lavas.[390-A]

Dr. MacCulloch, after examining with great attention these and the other igneous rocks of Scotland, observes, "that it is a mere dispute about terms, to refuse to the ancient eruptions of trap the name of submarine volcanos; for they are such in every essential point, although they no longer eject fire and smoke."[390-B] The same author also considers it not improbable that some of the volcanic rocks of the same country may have been poured out in the open air.[390-C]

Although the principal component minerals of subaerial lavas are the same as those of intrusive trap, and both the columnar and globular structure are common to both, there are, nevertheless, some volcanic rocks which never occur as lava, such as greenstone, clinkstone, the more crystalline porphyries, and those traps in which quartz and mica appear as constituent parts. In short, the intrusive trap rocks, forming the intermediate step between lava and the plutonic rocks, depart in their characters from lava in proportion as they approximate to granite.

These views respecting the relations of the volcanic and trap rocks will be better understood when the reader has studied, in the 33d chapter, what is said of the plutonic formations.

FORM, STRUCTURE, AND ORIGIN OF VOLCANIC MOUNTAINS.

The origin of volcanic cones with crater-shaped summits has been alluded to in the last chapter (p. 368.), and more fully explained in the "Principles of Geology" (chaps. xxiv. to xxvii.), where Vesuvius, Etna, Santorin, and Barren Island were described. The more ancient portions of those mountains or islands, formed long before the times of history, exhibit the same external features and internal structure which belong to most of the extinct volcanos of still higher antiquity.

The island of Palma, for example, one of the Canaries, offers an excellent illustration of what, in common with many others, I regard as the ruins of a large dome-shaped mass formed by a series of eruptions proceeding from a crater at the summit, this crater having been since replaced by a larger cavity, the origin of which has afforded geologists an ample field for discussion and speculation.

Von Buch, in his excellent account of the Canaries, has given us a graphic picture of this island, which consists chiefly of a single mountain (fig. 455.). This mountain has the general form of a great truncated cone, with a huge and deep cavity in the middle, about six miles in diameter, called by the inhabitants "the Caldera," or cauldron. The range of precipices surrounding the Caldera are no less than 4000 feet in their average height; at one point, where they are highest, they are 7730 feet above the level of the sea. The external flanks of the cone incline gently in every direction towards the base of the island, and are in part cultivated; but the walls and bottom of the Caldera present on all sides rugged and uncultivated rocks, almost completely devoid of vegetation. So steep are these walls, that there is no part by which they can be descended, and the only entrance is by a great ravine, or Barranco, as it is called (see b b', map, fig. 456.), which extends from the sea to the interior of the great cavity, and by its jagged, broken, and precipitous sides, exhibits to the geologist a transverse section of the rocks of which the whole mountain is composed. By this means, we learn that the cone is made up of a great number of sloping beds, which dip outwards in every direction from the centre of the void space, or from the hollow axis of the cone. The beds consist chiefly of sheets of basalt, alternating with conglomerates; the materials of the latter being in part rounded, as if rolled by water in motion. The inclination of all the beds corresponds to that of the external slope of the island, being greatest towards the Caldera, and least steep when they are nearest the sea. There are a great number of tortuous veins, and many dikes of lava or trap, chiefly basaltic, and most of them vertical, which cut through the sloping beds laid open to view in the great gorge or Barranco. These dikes and veins are more and more abundant as we approach the Caldera, being therefore most numerous where the slope of the beds is greatest.

Assuming the cone to be a pile of volcanic materials ejected by a long succession of eruptions (a point on which all geologists are agreed), we have to account for the Caldera and the great Barranco. I conceive that the cone itself may be explained, in accordance with what we know of the ordinary growth of volcanos[392-A], by supposing most of the eruptions to have taken place from one or more central vents, at or near the summit of the cone, before it was truncated. From this culminating point, sheets of lava flowed down one after the other, and showers of ashes or ejected stones. The volcano may, in the earlier stages of its growth, have been in great part submerged, like Stromboli, in the sea; and, therefore, some of the fragments of rock cast out of its crater may not only have been rolled by torrents sweeping down the mountain's side, but have also been rounded by the waves of the sea, as we see happen on the beach near Catania, on which the modern lavas of Etna are broken up. The increased number of dykes, as we approach the axis of the cone, agrees well with the hypothesis of the eruptions having been most frequent towards the centre.

There are three known causes or modes of operation, which may have conduced towards the vast size of the Caldera. First, the summit of a conical mountain may have fallen in, as happened in the case of Capacurcu, one of the Andes, according to tradition, in the year 1462, and of many other volcanic mountains.[393-A] Sections seem wanting, to supply us with all the data required for judging fairly of the tenability of this hypothesis. It appears, however, from Captain Vidal's survey (see fig. 456.), that a hill of considerable height rises up from the bottom of the Caldera, the structure of which, if it be any where laid open, might doubtless throw much light on this subject. Secondly, an original crater may have been enlarged by a vast gaseous explosion, never followed by any subsequent eruption. A serious objection to this theory arises from our not finding that the exterior of the cone supports a mass of ruins, such as ought to cover it, had so enormous a volume of matter, partly made up of the solid contents of the dikes, been blown out into the air. In that case, an extensive bed of angular fragments of stone, and of fine dust, might be looked for, enveloping the entire exterior of the mountain up to the very rim of the Caldera, and ought nowhere to be intersected by a dike. The absence of such a formation has induced Von Buch to suppose that the missing portion of the cone was engulphed. It should, however, be remembered, that in existing volcanos, large craters, two or three miles in diameter, are sometimes formed by explosions, or by the discharge of great volumes of steam.

There is yet another cause to which the extraordinary dimensions of the Caldera may, in part at least, be owing; namely, aqueous denudation. Von Buch has observed, that the existence of a single deep ravine, like the Great Barranco, is a phenomenon common to many extinct volcanos, as well as to some active ones. Now, it will be seen by Captain Vidal's map (fig. 456. p. 391.), that the sea-cliff at Point Juan Graje, 780 feet high, now constituting the coast at the entrance of the great ravine, is continuous with an inland cliff which bounds the same ravine on its north-western side. No one will dispute that the precipice, at the base of which the waves are now beating, owes its origin to the undermining power of the sea. It is natural, therefore, to attribute the extension of the same cliff to the former action of the waves, exerted at a time when the relative level of the island and the ocean were different from what they are now. But if the waves and tides had power to remove the rocks once filling a great gorge which is 7 miles long, and, in its upper part, 2000 feet deep, can we doubt that the same power may have cleared out much of the solid mass now missing in the Great Caldera?

The theory advanced to account for the configuration of Palma, commonly called the "elevation crater theory," is this. All the alternating masses of basalt and conglomerate, intersected in the Barranco, or abruptly cut off in the escarpment or walls of the Caldera, were at first disposed in horizontal masses on the level floor of the ocean, and traversed, when in that position, by all the basaltic dikes which now cut through them. At length they were suddenly uplifted by the explosive force of elastic vapours, which raised the mass bodily, so as to tilt the beds on all sides away from the centre of elevation, causing at the same time an opening at the culminating point. Besides many other objections which may be urged against this hypothesis, it leaves unexplained the unbroken continuity of the rim of the Caldera, which is uninterrupted in all places save one[394-A], namely, that where the great gorge or Barranco occurs.

As a more natural way of explaining the phenomenon, the following series of events may be imagined. The principal vent, from which a large part of the materials of the cone were poured or thrown out, was left empty after the last escape of vapour, when the volcano became extinct. We learn from Mr. Dana's valuable work on the geology of the United States' Exploring Expedition, published in 1849, that two of the principal volcanos of the Sandwich Islands, Mounts Loa and Kea in Owyhee, are huge flattened volcanic cones, 15,000 feet high (see fig. 457.), each equalling two and a half Etnas in their dimensions.

Fig. 457.

Mount Loa, in the Sandwich Islands. (Dana)

• a. Crater at the summit.
• b. The lateral crater of Kilauea.

The dotted lines indicate a supposed column of solid rock caused by the lava consolidating after eruptions.