On the third of August, fresh floods of lava still pouring from the volcano, a new branch was sent off in a different direction; for the channel of the Skapta was now so entirely choked up, and every opening to the west and north so obstructed, that the melted matter was forced to take a new course, so that it ran in a southeast direction, and discharged itself into the bed of the river Hverfisfliot, where a scene of destruction scarcely inferior to the former was occasioned. These Icelandic lavas (like the ancient streams which are met with in Auvergne, and other provinces of Central France), are stated by Stephenson to have acc.u.mulated to a prodigious depth in narrow rocky gorges; but when they came to wide alluvial plains, they spread themselves out into broad burning lakes, sometimes from twelve to fifteen miles wide, and one hundred feet deep. When the "fiery lake"

which filled up the lower portion of the valley of the Skapta, had been augmented by new supplies, the lava flowed up the course of the river to the foot of the hills from whence the Skapta takes its rise. This affords a parallel case to one which can be shown to have happened at a remote era in the volcanic region of the Vivarais in France, where lava issued from the cone of Thueyts, and while one branch ran down, another more powerful stream flowed up the channel of the river Ardeche.

The sides of the valley of the Skapta present superb ranges of basaltic columns of older lava, resembling those which are laid open in the valleys descending from Mont Dor, in Auvergne, where more modern lava-currents, on a scale very inferior in magnitude to those of Iceland, have also usurped the beds of the existing rivers. The eruption of Skaptar Jokul did not entirely cease till the end of two years; and when Mr. Paulson visited the tract eleven years afterwards, in 1794, he found columns of smoke still rising from parts of the lava, and several rents filled with hot water.[585]

Although the population of Iceland was very much scattered, and did not exceed fifty thousand, no less than twenty villages were destroyed, besides those inundated by water; and more than nine thousand human beings perished, together with an immense number of cattle, partly by the depredations of the lava, partly by the noxious vapors which impregnated the air, and, in part, by the famine caused by showers of ashes throughout the island, and the desertion of the coasts by the fish.

_Immense volume of the lava._--But the extraordinary volume of melted matter produced in this eruption deserves the particular attention of the geologist. Of the two branches, which flowed in nearly opposite directions, the greatest was fifty, and the lesser forty miles in length. The extreme breadth which the Skapta branch attained in the low countries was from twelve to fifteen miles, that of the other about seven. The ordinary height of both currents was one hundred feet, but in narrow defiles it sometimes amounted to six hundred. Professor Bischoff has calculated that the ma.s.s of lava brought up from the subterranean regions by this single eruption "surpa.s.sed in magnitude the bulk of Mont Blanc."[586] But a more distinct idea will be formed of the dimensions of the two streams, if we consider how striking a feature they would now form in the geology of England, had they been poured out on the bottom of the sea after the deposition and before the elevation of our secondary and tertiary rocks. The same causes which have excavated valleys through parts of our marine strata, once continuous, might have acted with equal force on the igneous rocks, leaving, at the same time, a sufficient portion undestroyed to enable us to discover their former extent. Let us, then, imagine the termination of the Skapta branch of lava to rest on the escarpment of the inferior and middle oolite, where it commands the vale of Gloucester. The great platform might be one hundred feet thick, and from ten to fifteen miles broad, exceeding any which can be found in Central France. We may also suppose great tabular ma.s.ses to occur at intervals, capping the summit of the Cotswold Hills between Gloucester and Oxford, by Northleach, Burford, and other towns.

The wide valley of the Oxford clay would then occasion an interruption for many miles; but the same rocks might recur on the summit of c.u.mnor and Shotover Hills, and all the other oolitic eminences of that district. On the chalk of Berkshire, extensive plateaus, six or seven miles wide, would again be formed; and lastly, crowning the highest sands of Highgate and Hampstead, we might behold some remnants of the current five or six hundred feet in thickness, causing those hills to rival, or even to surpa.s.s, in height, Salisbury Craigs and Arthur's Seat.

The distance between the extreme points here indicated would not exceed ninety miles in a direct line; and we might then add, at the distance of nearly two hundred miles from London, along the coast of Dorsetshire and Devonshire, for example, a great ma.s.s of igneous rocks, to represent those of contemporary origin, which were produced beneath the level of the sea, where the island of Nyoe rose up.

_Volume of ancient and modern flows of lava compared._--Yet, gigantic as must appear the scale of these modern volcanic operations, we must be content to regard them as perfectly insignificant in comparison to currents of the primeval ages, if we embrace the theoretical views of many geologists, which were not inaccurately expressed by the late Professor Alexander Brongniart, when he declared that "aux epoques geognostiques anciennes, tous les phenomenes geologiques se pa.s.soient dans des dimensions _centuples_ de celles qu'ils presentent aujourd'hui."[587] Had Skaptar Jokul, therefore, been a volcano of the olden time, it would have poured forth lavas at a single eruption a hundred times more voluminous than those which were witnessed by the present generation in 1783. But it may, on the contrary, be affirmed that, among the older formations, no igneous rock of such colossal magnitude has yet been met with; nay, it would be most difficult to point out a ma.s.s of ancient date (distinctly referable to a single eruption) which would even rival in volume the matter poured out from Skaptar Jokul in 1783.

_Eruption of Jorullo in 1759._--As another example of the stupendous scale of modern volcanic eruptions, I may mention that of Jorullo in Mexico, in 1759. The great region to which this mountain belongs has already been described. The plain of Malpais forms part of an elevated platform, between two and three thousand feet above the level of the sea, and is bounded by hills composed of basalt, trachyte, and volcanic tuff, clearly indicating that the country had previously, though probably at a remote period, been the theatre of igneous action. From the era of the discovery of the New World to the middle of the last century, the district had remained undisturbed, and the s.p.a.ce, now the site of the volcano, which is thirty-six leagues distant from the nearest sea, was occupied by fertile fields of sugar-cane and indigo, and watered by the two brooks Cuitimba and San Pedro. In the month of June, 1759, hollow sounds of an alarming nature were heard, and earthquakes succeeded each other for two months, until, at the end of September, flames issued from the ground, and fragments of burning rocks were thrown to prodigious heights. Six volcanic cones, composed of scoriae and fragmentary lava, were formed on the line of a chasm which ran in the direction from N. N. E. to S. S. W. The least of these cones was 300 feet in height; and Jorullo, the central volcano, was elevated 1600 feet above the level of the plain. It sent forth great streams of basaltic lava, containing included fragments of granitic rocks, and its ejections did not cease till the month of February, 1760.[588]

[Ill.u.s.tration: Fig. 57.

_a_, Summit of Jorullo. _b_, _c_, Inclined plane sloping at an angle of 6 from the base of the cones.]

Humboldt visited the country more than forty years after this occurrence, and was informed by the Indians, that when they returned, long after the catastrophe, to the plain, they found the ground uninhabitable from the excessive heat. When he himself visited the place, there appeared, around the base of the cones, and spreading from them, as from a centre, over an extent of four square miles, a ma.s.s of matter of a convex form, about 550 feet high at its junction with the cones, and gradually sloping from them in all directions towards the plain. This ma.s.s was still in a heated state, the temperature in the fissures being on the decrease from year to year, but in 1780 it was still sufficient to light a cigar at the depth of a few inches. On this slightly convex protuberance, the slope of which must form an angle of about 6 with the horizon, were thousands of flattish conical mounds, from six to nine feet high, which, as well as large fissures traversing the plain, acted as fumeroles, giving out clouds of sulphurous acid and hot aqueous vapor. The two small rivers before mentioned disappeared during the eruption, losing themselves below the eastern extremity of the plain, and reappearing as hot springs at its western limit.

_Cause of the convexity of the plain of Malpais._--Humboldt attributed the convexity of the plain to inflation from below; supposing the ground, for four square miles in extent, to have risen up in the shape of a bladder to the elevation of 550 feet above the plain in the highest part. But Mr. Scrope has suggested that the phenomena may be accounted for far more naturally, by supposing that lava flowing simultaneously from the different orifices, and princ.i.p.ally from Jorullo, united into a sort of pool or lake. As they were poured forth on a surface previously flat, they would, if their liquidity was not very great, remain thickest and deepest near their source, and diminish in bulk from thence towards the limits of the s.p.a.ce which they covered. Fresh supplies were probably emitted successively during the course of an eruption which lasted more than half a year; and some of these, resting on those first emitted, might only spread to a small distance from the foot of the cone, where they would necessarily acc.u.mulate to a great height. The average slope of the great dome-shaped volcanoes of the Sandwich Islands, formed almost exclusively of lava, with scarce any scoriae, is between 6 30'

and 7 46', so that the inclination of the convex ma.s.s around Jorullo, if we adopt Mr. Scrope's explanation (see fig. 57), is quite in accordance with the known laws which govern the flow of lava.

The showers, also, of loose and pulverulent matter from the six craters, and princ.i.p.ally from Jorullo, would be composed of heavier and more bulky particles near the cones, and would raise the ground at their base, where, mixing with rain, they might have given rise to the stratum of black clay, which is described as covering the lava. The small conical mounds (called "hornitos," or little ovens) may resemble those five or six small hillocks which existed in 1823 on the Vesuvian lava, and sent forth columns of vapor, having been produced by the disengagement of elastic fluids heaping up small dome-shaped ma.s.ses of lava. The fissures mentioned by Humboldt as of frequent occurrence, are such as might naturally accompany the consolidation of a thick bed of lava, contracting as it congeals; and the disappearance of rivers is the usual result of the occupation of the lower part of a valley or plain by lava, of which there are many beautiful examples in the old lava-currents of Auvergne. The heat of the "hornitos" is stated to have diminished from the first; and Mr. Bullock, who visited the spot many years after Humboldt, found the temperature of the hot spring very low,--a fact which seems clearly to indicate the gradual congelation of a subjacent bed of lava, which from its immense thickness may have been enabled to retain its heat for half a century. The reader may be reminded, that when we thus suppose the lava near the volcano to have been, together with the ejected ashes, more than five hundred feet in depth, we merely a.s.sign a thickness which the current of Skaptar Jokul attained in some places in 1783.

_Hollow sound of the plain when struck._--Another argument adduced in support of the theory of inflation from below, was, the hollow sound made by the steps of a horse upon the plain; which, however, proves nothing more than that the materials of which the convex ma.s.s is composed are light and porous. The sound called "rimbombo" by the Italians is very commonly returned by _made ground_ when struck sharply; and has been observed not only on the sides of Vesuvius and other volcanic cones where there is a cavity below, but in such regions as the Campagna di Roma, composed in a great measure of tuff and porous volcanic rocks. The reverberation, however, may perhaps be a.s.sisted by grottoes and caverns, for these may be as numerous in the lavas of Jorullo as in many of those of Etna; but their existence would lend no countenance to the hypothesis of a great arched cavity, four square miles in extent, and in the centre 550 feet high.[589]

_No recent eruptions of Jorullo._--In a former edition I stated that I had been informed by Captain Vetch, that in 1819 a tower at Guadalaxara was thrown down by an earthquake, and that ashes, supposed to have come from Jorullo, fell at the same time at Guanaxuato, a town situated 140 English miles from the volcano. But Mr. Burkhardt, a German director of mines, who examined Jorullo in 1827, ascertained that there had been no eruption there since Humboldt's visit in 1803. He went to the bottom of the crater, and observed a slight evolution of sulphurous acid vapors, but the "hornitos" had entirely ceased to send forth steam. During the twenty-four years intervening between his visit and that of Humboldt, vegetation had made great progress on the flanks of the new hills; the rich soil of the surrounding country was once more covered with luxuriant crops of sugar-cane and indigo, and there was an abundant growth of natural underwood on all the uncultivated tracts.[590]

_Galongoon, Java_, 1822.--The mountain of Galongoon (or Galung Gung) was in 1822 covered by a dense forest, and situated in a fruitful and thickly-peopled part of Java. There was a circular hollow at its summit, but no tradition existed of any former eruption. In July, 1822, the waters of the river Kunir, one of those which flowed from its flanks, became for a time hot and turbid. On the 8th of October following a loud explosion was heard, the earth shook, and immense columns of hot water and boiling mud, mixed with burning brimstone, ashes, and lapilli, of the size of nuts, were projected from the mountain like a waterspout, with such prodigious violence that large quant.i.ties fell beyond the river Tandoi, which is forty miles distant. Every valley within the range of this eruption became filled with a burning torrent, and the rivers, swollen with hot water and mud, overflowed their banks, and carried away great numbers of the people, who were endeavoring to escape, and the bodies of cattle, wild beasts, and birds. A s.p.a.ce of twenty-four miles between the mountain and the river Tandoi was covered to such a depth with bluish mud that people were buried in their houses, and not a trace of the numerous villages and plantations throughout that extent was visible. Within this s.p.a.ce the bodies of those who perished were buried in mud and concealed, but near the limits of the volcanic action they were exposed, and strewed over the ground in great numbers, partly boiled and partly burnt.

It was remarked, that the boiling mud and cinders were projected with such violence from the mountain, that while many remote villages were utterly destroyed and buried, others much nearer the volcano were scarcely injured.

The first eruption lasted nearly five hours, and on the following days the rain fell in torrents, and the rivers, densely charged with mud, deluged the country far and wide. At the end of four days (October 12th) a second eruption occurred more violent than the first, in which hot water and mud were again vomited, and great blocks of basalt were thrown to the distance of seven miles from the volcano. There was at the same time a violent earthquake, and in one account it is stated that the face of the mountain was utterly changed, its summits broken down, and one side, which had been covered with trees, became an enormous gulf in the form of a semicircle. This cavity was about midway between the summit and the plain, and surrounded by steep rocks, said to be newly heaped up during the eruption. New hills and valleys are said to have been formed, and the rivers Banjarang and Wulan changed their course, and in one night (October 12th) 2000 persons were killed.

The first intimation which the inhabitants of Bandong received of this calamity on the 8th of October, was the news that the river Wulna was bearing down into the sea the dead bodies of men, and the carca.s.ses of stags, rhinoceroses, tigers, and other animals. The Dutch painter Payen determined to travel from thence to the volcano, and he found that the quant.i.ty of the ashes diminished as he approached the base of the mountain. He alludes to the altered form of the mountain after the 12th, but does not describe the new semicircular gulf on its side.

The official accounts state that 114 villages were destroyed, and above 4000 persons killed.[591]

_Submarine volcanoes._--Although we have every reason to believe that volcanic eruptions as well as earthquakes are common in the bed of the sea, it was not to be expected that many opportunities would occur to scientific observers of witnessing the phenomena. The crews of vessels have sometimes reported that they have seen in different places sulphurous smoke, flame, jets of water, and steam, rising up from the sea, or they have observed the waters greatly discolored, and in a state of violent agitation as if boiling. New shoals have also been encountered, or a reef of rocks just emerging above the surface, where previously there was always supposed to have been deep water. On some few occasions the gradual formation of an island by a submarine eruption has been observed, as that of Sabrina, in the year 1811, off St.

Michael's in the Azores. The throwing up of ashes in that case, and the formation of a cone about three hundred feet in height, with a crater in the centre, closely resembled the phenomena usually accompanying a volcanic eruption on land. Sabrina was soon washed away by the waves.

Previous eruptions in the same part of the sea were recorded to have happened in 1691 and 1720. The rise of Nyoe, also, a small island off the coast of Iceland, in 1783, has already been alluded to; and another volcanic isle was produced by an eruption near Reikiavig, on the same coast, in June, 1830.[592]

_Graham Island_[593], 1831.--We have still more recent and minute information respecting the appearance, in 1831, of a new volcanic island in the Mediterranean, between the S. W. coast of Sicily and that projecting part of the African coast where ancient Carthage stood. The site of the island was not any part of the great shoal, or bank, called "Nerita," as was first a.s.serted, but a spot where Captain W. H. Smyth had found, in his survey a few years before, a depth of more than one hundred fathoms water.[594]

[Ill.u.s.tration: Fig. 58. Form of the cliffs of Graham Island, as seen from S. S. E., distant one mile, 7th August, 1831.[596]]

[Ill.u.s.tration: Fig. 59. View of the interior of Graham Island, 29th Sept., 1831.]

[Ill.u.s.tration: Fig. 60. Graham Island, 29th Sept., 1831.[597]]

The position of the island (lat. 37 8' 30" N., long. 12 42' 15" E.) was about thirty miles S. W. of Sciacca, in Sicily, and thirty-three miles N. E. of Pantellaria.[595] On the 28th of June, about a fortnight before the eruption was visible, Sir Pulteney Malcolm, in pa.s.sing over the spot in his ship, felt the shocks of an earthquake, as if he had struck on a sand-bank; and the same shocks were felt on the west coast of Sicily, in a direction from S. W. to N. E. About the 10th of July, John Corrao, the captain of a Sicilian vessel, reported that, as he pa.s.sed near the place, he saw a column of water like a water-spout, sixty feet high, and 800 yards in circ.u.mference, rising from the sea, and soon afterwards a dense steam in its place, which ascended to the height of 1800 feet. The same Corrao, on his return from Girgenti, on the 18th of July, found a small island, twelve feet high with a crater in its centre, ejecting volcanic matter, and immense columns of vapor; the sea around being covered with floating cinders and dead fish. The scoriae were of a chocolate color, and the water which boiled in the circular basin was of a dingy red. The eruption continued with great violence to the end of the same month; at which time the island was visited by several persons, and among others by Capt. Swinburne, R. N., and M. Hoffmann, the Prussian geologist. It was then from fifty to ninety feet in height, and three-quarters of a mile in circ.u.mference. By the 4th of August it became, according to some accounts, above 200 feet high, and three miles in circ.u.mference; after which it began to diminish in size by the action of the waves, and it was only two miles round on the 25th of August; and on the 3d of September, when it was carefully examined by Captain Wodehouse, only three-fifths of a mile in circ.u.mference; its greatest height being then 107 feet. At this time the crater was about 780 feet in circ.u.mference. On the 29th of September, when it was visited by Mons. C. Prevost, its circ.u.mference was reduced to about 700 yards. It was composed entirely of incoherent ejected matter, scoriae, pumice, and lapilli, forming regular strata, some of which are described as having been parallel to the steep inward slope of the crater, while the rest were inclined outwards, like those of Vesuvius.[598] When the arrangement of the ejected materials has been determined by their falling continually on two steep slopes, that of the external cone and that of the crater, which is always a hollow inverted cone, a transverse section would probably resemble that given in the annexed figure (61). But when I visited Vesuvius, in 1828, I saw no beds of scoriae inclined towards the axis of the cone. (See fig. 45, p. 381.) Such may have once existed; but the explosions or subsidences, or whatever causes produced the great crater of 1822, had possibly destroyed them.

[Ill.u.s.tration: Fig. 61.]

Few of the pieces of stone thrown out from Graham Island exceeded a foot in diameter. Some fragments of dolomitic limestone were intermixed; but these were the only non-volcanic substances. During the month of August, there occurred on the S. W. side of the new island a violent ebullition and agitation of the sea, accompanied by the constant ascension of a column of dense white steam, indicating the existence of a second vent at no great depth from the surface. Towards the close of October, no vestige of the crater remained, and the island was nearly levelled with the surface of the ocean, with the exception, at one point, of a small monticule of sand and scoriae. It was reported that, at the commencement of the year following (1832), there was a depth of 150 feet where the island had been: but this account was quite erroneous; for in the early part of that year Captain Swinburne found a shoal and discolored water there, and towards the end of 1833 a dangerous reef existed of an oval figure, about three-fifths of a mile in extent. In the centre was a black rock, of the diameter of about twenty-six fathoms, from nine to eleven feet under water; and round this rock are banks of black volcanic stones and loose sand. At the distance of sixty fathoms from this central ma.s.s, the depth increased rapidly. There was also a second shoal at the distance of 450 feet S. W. of the great reef, with fifteen feet water over it, also composed of rock, surrounded by deep sea. We can scarcely doubt that the rock in the middle of the larger reef is solid lava, which rose up in the princ.i.p.al crater, and that the second shoal marks the site of the submarine eruption observed in August, 1831, to the S. W. of the island.

From the whole of the facts above detailed, it appears that a hill eight hundred feet or more in height was formed by a submarine volcanic vent, of which the upper part (only about two hundred feet high) emerged above the waters, so as to form an island. This cone must have been equal in size to one of the largest of the lateral volcanoes on the flanks of Etna, and about half the height of the mountain Jorullo in Mexico, which was formed in the course of nine months, in 1759. In the centre of the new volcano a large cavity was kept open by gaseous discharges, which threw out scoriae; and fluid lava probably rose up in this cavity. It is not uncommon for small subsidiary craters to open near the summit of a cone, and one of these may have been formed in the case of Graham Island; a vent, perhaps, connected with the main channel of discharge which gave pa.s.sage in that direction to elastic fluids, scoriae, and melted lava. It does not appear that, either from this duct, or from the princ.i.p.al vent, there was any overflowing of lava; but melted rock may have flowed from the flanks or base of the cone (a common occurrence on land), and may have spread in a broad sheet over the bottom of the sea.

[Ill.u.s.tration: Fig. 62.

Supposed section of Graham Island. (C. Maclaren.[599])]

The dotted lines in the annexed figure are an imaginary restoration of the upper part of the cone, now removed by the waves: the strong lines represent the part of the volcano which is still under water: in the centre is a great column, or dike, of solid lava, two hundred feet in diameter, supposed to fill the s.p.a.ce by which the gaseous fluids rose; and on each side of the dike is a stratified ma.s.s of scoriae and fragmentary lava. The solid nucleus of the reef, where the black rock is now found, withstands the movements of the sea; while the surrounding loose tuffs are cut away to a somewhat lower level. In this manner the lava, which was the lowest part of the island, or, to speak more correctly, which scarcely ever rose above the level of the sea when the island existed, has now become the highest point in the reef.

No appearances observed, either during the eruption or since the island disappeared, gave the least support to the opinion promulgated by some writers, that part of the ancient bed of the sea had been lifted up bodily.

The solid products, says Dr. John Davy, whether they consisted of sand, light cinders, or vesicular lava, differed more in form than in composition. The lava contained augite; and the specific gravity was 207 and 270. When the light spongy cinder, which floated on the sea, was reduced to fine powder by trituration, and the greater part of the entangled air got rid of, it was found to be of the specific gravity 264; and that of some of the sand which fell in the eruption was 275;[600] so that the materials equalled ordinary granites in weight and solidity. The only gas evolved in any considerable quant.i.ty was carbonic acid.[601]

_Submarine eruptions in mid-Atlantic._--In the Nautical Magazine for 1835, p. 642, and for 1838, p. 361, and in the Comptes Rendus, April, 1838, accounts are given of a series of volcanic phenomena, earthquakes, troubled water, floating scoriae and columns of smoke, which have been observed at intervals since the middle of the last century, in a s.p.a.ce of open sea between longitudes 20 and 22 west, about half a degree south of the equator. These facts, says Mr. Darwin, seem to show, that an island or an archipelago is in process of formation in the middle of the Atlantic; a line joining St. Helena and Ascension would, if prolonged, intersect this slowly nascent focus of volcanic action.[602]

Should land be eventually formed here, it will not be the first that has been produced by igneous action in this ocean since it was inhabited by the existing species of testacea. At Porto Praya in St. Jago, one of the Azores, a horizontal, calcareous stratum occurs, containing sh.e.l.ls of recent marine species, covered by a great sheet of basalt eighty feet thick.[603] It would be difficult to estimate too highly the commercial and political importance which a group of islands might acquire, if in the next two or three thousand years they should rise in mid-ocean between St. Helena and Ascension.

CANARY ISLANDS.

_Eruption in Lancerote, 1730 to 1736._--The effects of an eruption which happened in Lancerote, one of the Canary Islands, between the years 1730 and 1736, were very remarkable; and a detailed description has been published by Von Buch, who had an opportunity, when he visited that island in 1815, of comparing the accounts transmitted to us of the event, with the present state and geological appearances of the country.[604] On the 1st of September, 1730, the earth split open on a sudden two leagues from Yaira. In one night a considerable hill of ejected matter was thrown up; and, a few days later, another vent opened, and gave out a lava-stream, which overran Chinanfaya and other villages. It flowed first rapidly, like water, but became afterwards heavy and slow, like honey. On the 7th of September an immense rock was protruded from the bottom of the lava with a noise like thunder, and the stream was forced to change its course from N. to N. W., so that St.

Catalina and other villages were overflowed.

Whether this ma.s.s was protruded by an earthquake, or was a ma.s.s of ancient lava, blown up like that before mentioned in 1783 in Iceland, is not explained.

On the 11th of September more lava flowed out, and covered the village of Maso entirely, and for the s.p.a.ce of eight days precipitated itself with a horrible roar into the sea. Dead fish floated on the waters in indescribable mult.i.tudes, or were thrown dying on the sh.o.r.e. After a brief interval of repose, three new openings broke forth immediately from the site of the consumed St. Catalina, and sent out an enormous quant.i.ty of lapilli, sand, and ashes. On the 28th of October the cattle throughout the whole country dropped lifeless to the ground, suffocated by putrid vapors, which condensed and fell down in drops. On the 1st of December a lava-stream reached the sea, and formed an island, round which dead fish were strewed.

_Number of cones thrown up._--It is unnecessary here to give the details of the overwhelming of other places by fiery torrents, or of a storm which was equally new and terrifying to the inhabitants, as they had never known one in their country before. On the 10th of January, 1731, a high hill was thrown up, which, on the same day, precipitated itself back again into its own crater; fiery brooks of lava flowed from it to the sea. On the 3d of February a new cone arose. Others were thrown up in March, and poured forth lava-streams. Numerous other volcanic cones were subsequently formed in succession, till at last their number amounted to about thirty. In June, 1731, during a renewal of the eruptions, all the banks and sh.o.r.es in the western part of the island were covered with dying fish, of different species, some of which had never before been seen. Smoke and flame arose from the sea, with loud detonations. These dreadful commotions lasted without interruption for _five successive years_, so that a great emigration of the inhabitants became necessary.

_Their linear direction._--As to the height of the new cones, Von Buch was a.s.sured that the formerly great and flourishing St. Catalina lay buried under hills 400 feet in height; and he observes that the most elevated cone of the series rose 600 feet above its base, and 1378 feet above the sea, and that several others were nearly as high. The new vents were all arranged _in one line_, about two geographical miles long, and in a direction nearly east and west. If we admit the probability of Von Buch's conjecture, that these vents opened along the line of a cleft, it seems necessary to suppose that this subterranean fissure was only prolonged upwards to the surface by degrees, and that the rent was narrow at first, as is usually the case with fissures caused by earthquakes. Lava and elastic fluids might escape from some point on the rent where there was least resistance, till, the first aperture becoming obstructed by ejections and the consolidation of lava, other orifices burst open in succession along the line of the original fissure. Von Buch found that each crater was lowest on that side on which lava had issued; but some craters were not breached, and were without any lava streams. In one of these were open fissures, out of which hot vapors rose, which in 1815 raised the thermometer to 145 Fahrenheit, and was probably at the boiling point lower down. The exhalations seemed to consist of aqueous vapor; yet they could not be pure steam, for the crevices were incrusted on either side by siliceous sinter (an opal-like hydrate of silica of a white color), which extended almost to the middle. This important fact attests the length of time during which chemical processes continue after eruptions, and how open fissures may be filled up laterally by mineral matter, sublimed from volcanic exhalations. The lavas of this eruption covered nearly a third of the whole island, often forming on slightly inclined planes great horizontal sheets several square leagues in area, resembling very much the basaltic platforms of Auvergne.

_Pretended distinction between ancient and modern lavas._--One of the new lavas was observed to contain ma.s.ses of olivine of an olive-green color, resembling those which occur in one of the lavas of the Vivarais.

Von Buch supposes the great crystals of olivine to have been derived from a previously existing basalt melted up by the new volcanoes; but we have scarcely sufficient data to bear out such a conjecture. The older rocks of the island consist, in a great measure, of that kind of basaltic lava called dolerite, sometimes columnar, and partly of common basalt and amygdaloid. Some recent lavas a.s.sumed, on entering the sea, a prismatic form, and so much resembled the older lavas of the Canaries, that the only geological distinction which Von Buch appears to have been able to draw between them was, that they did not alternate with conglomerates, like the ancient basalts. Some modern writers have endeavored to discover, in the abundance of these conglomerates, a proof of the dissimilarity of the volcanic action in ancient and modern times; but this character is more probably attributable to the difference between submarine operations and those on the land. All the blocks and imperfectly rounded fragments of lava, transported during the intervals of eruption, by rivers and torrents, into the adjoining sea, or torn by the continued action of the waves from cliffs which are undermined, must acc.u.mulate in stratified breccias and conglomerates, and be covered again and again by other lavas. This is now taking place on the sh.o.r.es of Sicily, between Catania and Trezza, where the sea breaks down and covers the sh.o.r.e with blocks and pebbles of the modern lavas of Etna; and on parts of the coast of Ischia, where numerous currents of trachyte are in like manner undermined in lofty precipices. So often, then, as an island is raised in a volcanic archipelago by earthquakes from the deep, the fundamental and (relatively to all above) the oldest lava will often be distinguishable from those formed by subsequent eruptions on dry land, by their alternation with beds of sandstone and fragmentary rocks.

The supposed want of ident.i.ty, then, between the volcanic phenomena of different epochs resolves itself partly at least into the marked difference between the operations simultaneously in progress, above and below the waters. Such, indeed, is the source, as was before stated in the First Book (Chap. V.), of many of our strongest theoretical prejudices in geology. No sooner do we study and endeavor to explain submarine appearances, than we feel, to use a common expression, out of our element; and unwilling to concede that our extreme ignorance of processes now continually going on can be the cause of our perplexity, we take refuge in a "pre-existent order of nature."

_Recent formation of oolitic travertin in Lancerote._--Throughout a considerable part of Lancerote, the old lavas are covered by a thin stratum of limestone, from an inch to two feet in thickness. It is of a hard stalact.i.tic nature, sometimes oolitic, like the Jura limestone, and contains fragments of lava and terrestrial sh.e.l.ls, chiefly helices and spiral bulimi. It sometimes rises to the height of 800 feet above the level of the sea. Von Buch imagines that this remarkable superstratum has been produced by the furious northwest storms, which in winter drive the spray of the sea in clouds over the whole island; from whence calcareous particles may be deposited stalact.i.tically. Mr. Darwin informs me that he found a limestone in St. Helena, the harder parts of which correspond precisely to the stone of Lancerote. He attributes the origin of this rock in St. Helena not to the spray of the sea, but to drifting by violent winds of the finer particles of sh.e.l.ls from the sea-beach. Some parts of this drift are subsequently dissolved by atmospheric moisture, and redeposited, so as to convert calcareous sand into oolite.

_Recent eruption in Lancerote._--From the year 1736 to 1815, when Von Buch visited Lancerote, there had been no eruption; but, in August, 1824, a crater opened near the port of Rescif, and formed by its ejections, in the s.p.a.ce of twenty-four hours, a considerable hill.

Violent earthquakes preceded and accompanied this eruption.[605]

_Teneriffe._--The Peak of Teneriffe is about 12,000 feet high, and stands, says Von Buch, like a tower encircled by its fosse and bastion.

The bastion consists, like the semicircular escarpment of Somma turned towards Vesuvius, of precipitous cliffs, composed of trachyte, basalt, coa.r.s.e conglomerates, and tuffs, traversed by volcanic dikes, mostly vertical, and of basalt. These cliffs vary in height from 1000 to 1800 feet, and are supposed by Von Buch to have been heaved up into their present position by a force exerted from below, in accordance with the theory proposed by the same author for the origin of the cones of Vesuvius and Etna. According to the observations of M. Deville in 1839[606], the trachytes are often granitoid in their aspect, and contain instead of gla.s.sy felspar the allied mineral called oligoclase, which had been previously considered as characteristic of more ancient igneous rocks. The same traveller supposes, although he found no limestone or trace of fossils in any of the rocks of Teneriffe, that the alternating trachytes and trachytic conglomerates originated beneath the sea. If this opinion be correct, and it is at least very probable, geologists may still speculate on two modes in which the ma.s.s of the island acquired its present form and elevation above the sea. 1st, The advocates of Von Buch's crater-of-elevation hypothesis may imagine that a succession of horizontally superimposed beds were upheaved by a sudden movement, and tilted so as to dip in all directions outwards from the centre of a new dome-shaped eminence, in the middle of which a large opening or bowl-shaped cavity was produced. 2dly, Or according to the theory which to me appears preferable, a submarine hill in the form of a flattened dome may have gradually acc.u.mulated, partly below the waters and partly above by the continued outpourings of sheets of lava and the ejection of ashes from a central orifice. In this case the dikes would represent the fissures, which were filled during successive eruptions, and the original inclination of the beds may have been increased by the distension and upheaval of the ma.s.s during reiterated convulsions, acting most forcibly at or near the channel of discharge, which would become partially sealed up with lava from time to time, and then be burst open again during eruptions. At length the whole island may have been raised bodily out of the sea by a gradual upward movement.