fragments torn from the sides of the crevice through which they escape. These ejected matters, on falling, accumulate round the opening into a circular bank, which, by the continuance of the process, becomes a truncated cone, with an internal funnel. This is the common form of a volcanic cone, thrown up by the explosions of a single eruption. If lava flows from the same orifice, after the formation of the cone, it breaks down the side; if before, the cone is often raised upon the hardened surface of the lava-current, which flows underneath, in a sort of canal, without damaging the bank above. Should subsequent eruptions take place on the same point, the hillock becomes more complicated in its structure, but the conical form is still preserved with sufficient regularity, the ejected matters mantling round the outside of the hill, and the lava, which pours over the lips of the crater, or forces its way through crevices in the sides of the cone, hardening into massive ribs or coatings, by which its bulk is at the same time increased, and a durable skeleton supplied. After repeated eruptions from the same opening, the simple coue becomes in this way enlarged into the volcanic mountain.

Vague and incorrect ideas are often attached to what is called the crater of a volcano. Some have erroneously supposed that every volcano must at all times have a crater-confounding it with the vent of the erupted matter, which is often no more than a narrow crevice, and, being filled up by the products of the eruption, is not easily to be discovered afterwards. A crater is the cup-shaped hollow left by the repeated explosions of elastic fluids, which usually, but not always, accompany the emission of lava from a crevice, and often occur without any overflow of lava. The crater of a simple cone, formed of fragmentary matter alone, is, as we have seen, a hollow inverted cone, circumscribed by the talus of debris heaped up round the vent. But, in volcanic mountains, after explosions of paroxysmal violence, the whole solid centre of the mountain is often blown into the air, and its contents scattered over the outer slopes, or worn to powder by repeated ejections, and carried by winds to vast distances. The crater left by such an eruption is a deep and often wide cavity, bordered by abrupt rocky precipices, in which sections are exposed of the successively-accumulated beds that form the substance of the mountain. Such a crater is wholly different in appearance from the smoothsided and regularly-sloping funnel of a simple cone. The former deserve the

distinguishing appellation of craters of paroxysmal explosion. Nor are they broken through volcanic mountains alone, but not unfrequently through granite or stratified rocks, which may be seen surrounding them in deep escarpments, and supporting the fragments of those rocks and scoriæ thrown out. The width of a crater seems to depend on the bulk of the volumes of vapour discharged at once, and does not always correspond with the quan tity of matter ejected, or the duration of the eruption. After the formation of a crater of great size, in the manner we have described, succeeding eruptions, from the same central vent, only throw up secondary cones and lava streams at the bottom of this gulph, which, accumulating on one another, by degrees fill it up entirely. At this time the volcanic mountain may exhibit no crater at all; and this is by no means an unfrequent condition of extinct or dormant volcanos. But the weight and coherence of these accumulations over the mouth of the volcano seem, by repressing, to increase its latent energy; and it often again bursts forth in a paroxysm of explosions, which blow off the whole summit of the mountain, and leave a fresh central cavity, of proportionate dimensions, sometimes several miles in diameter. Almost every volcanic mountain, habitually eruptive, is thus undergoing a succession of destructions and repairs, and none could better illustrate this law than Vesuvius during the past century. Those who will take the trouble to consult Hamilton's plates and relations, will trace the process we have described several times repeated, up to the publication of his work. The last phenomenon, described by him, was the paroxysmal eruption of 1794, which gutted the cone, and left a vast crater, three miles in circumference. This cavity was gradually filled by the falling in of its sides, and the subsequent minor eruptions from that time to 1822, when a high convexity had replaced the hollow on the summit of the cone. In October of that year, an eruption occurred, accompanied by explosions of great violence, which lasted twenty days, and once more hollowed out the cone, leaving a crater a mile in diameter, and two thousand feet deep. Since that time, fresh eruptions have been going on from the bottom of the crater: a secondary cone is thrown up there, and produces lava and scoriæ, which already have half filled the great crater.

The cliff-range of Somma, which half encircles the upper cone of Vesuvius, is, without doubt, the remaining segment of

the walls of the vast crater produced by the explosions of 79 A. D., which entombed Herculaneum and Pompeii beneath the fragments of the shattered mountain.

From what we have said, it will appear bow incorrect is the popular notion, that, in every eruption, the crater of a volcano is filled to the brim with lava, which pours thence over the outer slope. The violent explosions of a single eruption occasionally blow nearly the whole mountain into the air, leaving only its skirts as a low truncated cone, surrounding a basin, several miles in diameter, After such a paroxysm, hundreds of eruptions may take place within this vast crater before it is filled, and a new mountain reared in place of the old one. We may mention here that we are very sceptical as to the accounts received, from popular report, of the sinking in of volcanic mountains during eruptions. We know the ordinary course to be, that they are blown outwards, and their fragments scattered on all sides by the violence of the aeriform explosions, which sometimes continue for weeks, and reduce the wreck of the mountain to an impalpable powder, which the winds bear off to enormous distances. Nor do we recollect any relation of the disappearance of a mountain, and the substitution of a cavity, perhaps a lake (as the Peak of Timor, destroyed in 1637; Papandayang, in Java, 1772), without the accompaniment of tremendous discharges of fragmentary matter, which is described as covering the whole face of the country around, to a distance sometimes of hundreds of miles: from which circumstances we conclude, that the bulk of the mountain was broken up and scattered to the winds by repeated explosions, not that it fell in though it is natural that the inhabitants, finding on their return a deep cavity in place of a mountain, should imagine it the effect of subsidence rather than explosion. In fact, all the phenomena of volcanos tend to show their origin in a mass of matter, confined at an intense temperature, and struggling to escape; and, therefore, make it very improbable that any vast subterranean caverns can exist, into which the mountain could be precipitated. That the cliffs, surrounding a deep crater, occasionally fall inwards during earthquakes, so as to soften their declivity, and truncate the mountain at a lower pomt, is very true, and this has probably given rise to some of the stories as to the engulphing of mountains. The appearances of the volcano of Kirauea, in Owyhee, described by Mr. Ellis, are very peculiar, but afford no countenance to the idea of subterranean cavities. It

seems that some vast and ancient crater of this mountain has been nearly filled with a sort of bath, or pool, of liquid Java, on the surface of which a crust forms, but as fast as fresh lava wells up from below, the crust is broken through by minor eruptions. As this mass of lava rose in the crater, the weight of its increasing column has, at intervals, burst a lateral crevice in the side of the mountain, through which the reservoir of lava has been tapped of its excess, and cirenlar subsidences been successively formed in the crust above-the broken edges of which form a series of terraced ledges, at different heights, surrounding the present hollow. This is a remarkable, but very intelligible, variation of the volcanic phenomena, perfectly in harmony with their known laws of opera

tion.

Immense volumes of aqueous vapours are evolved from a crater during eruptions, and for a long time after the discharge of lava and scoriæ has ceased. They are condensed in the cold atmosphere surrounding the volcanic peak, and heavy rains are often caused, even in countries where, under other circumstances, rain is unknown. Falling on a surface which the eruption has thickly coated with fine ashes and loose fragments of all sizes, the rains sweep them along in a flood of mud and stones, which often does far more mischief than the ignited lava or earthquakes, and deposit at the foot of the mountain massive beds of conglomerate. If snow covers the cone, still more extensive deluges are produced through its sudden melting by contact with the red-hot lava. Etna, as might be expected, presents many traces of such floods; but it is in Iceland that they are exhibited on the most powerful scale. Conglomerates of immense extent and thickness have been spread in this manner within a late period, over the plains at the base of Hecla. On Etna itself a thick bed of solid ice has lately been found under an ancient current of lava. It is very conceivable that a coating of sand and scoriæ, the best possible non-conductors of heat, may enable snow to bear a stream of red-hot lava over it without being melted. It is probable, that in Iceland the circumstance has been often repeated, and we may expect to find glaciers alternating there with beds of lava and volcanic conglomerate.

One continuous eruption will frequently throw up a number of simple cones. Every considerable eruption is described as commencing with the splitting of the solid ground, and the production of a crevice prolonged sometimes many miles. The explosions, as well as the lava

streams, then break out from one, or from several points on this great crack. Thus, in the eruption of Etna in 1811, seven cones were successively thrown up in a line from the summit nearly to the foot of the mountain. In 1536, twelve mouths opened one below the other, and threw out lava and scoriæ. In 1669, the whole flank of the mountain was split open, a wide fissure showing itself, twelve miles in length, from the top halfway to the base. This crevice is figured in the old engravings of Etna, and is reported to have emitted a vivid light, showing it to be filled to some height with incandescent Java. Two cones were formed upon it. These circumstances are not confined to the flanks of a volcanic mountain, but take place equally when the eruption breaks through horizontal strata. In 1730, the Island of Lancerote, one of the Canaries, was split by longitudinal fissures running the whole length of the island, from which so much matter was discharged during five successive years, as formed thirty cones, some of them six hundred feet high, and overwhelmed with a flood of lava nearly the entire island. The eruption of Jorullo, in 1759, threw up six cones upon one line in the middle of a flat plain. That of Skapta Jokul, in 1783, was accompanied by the outburst of three copious sources of lava in the plain, stretching from the foot of that mountain, about eight miles apart; while a fourth, on a continuation of the same line, but beneath the sea, created a new island, at a distance of thirty miles from the coast. The lava produced by the three iuland vents alone covered a space of one thousand square miles, with a thick mass of solid rock. It is probable that many of the volcanic cones of Auvergne and the Velay, some hundreds of which are arranged in a linear chain, were the product of continuous eruptions. Such lengthened subterranean fissures do not always show themselves on the surface, the loose earth sinking into, and concealing them; and hence partial subsidences are usually observed along the line of volcanic orifices. Nor are they in general opened at once throughout their whole length, but prolonged by degrees, the first orifices becoming obstructed by the ejections and the consolidation of lava, so as to cause others to be burst in succession along the line of the original cleft. Analogy leads us to conclude, that the linear arrangement of the principal vents in a volcanic train or system, even when they stretch across half the globe, is owing to the same general cause as that of the secondary apertures, the creation, namely, of a fissure through the crust of the globe. The law already noticed, that the neigh

bouring volcanos of a train or group are found in activity by turns, the one serving for a time as a vent for the energy of the whole district, is as true on the small as on the large scale, and is shown from a great body of concurrent facts, to have prevailed in ages preceding any historical records of eruptions, as well as since.

Mr. Lyell very properly draws attention to the enormous quantity of new rock produced at once upon the surface of the globe by single eruptions. That of Skapta Jokul, for instance, already mentioned, discharged two streams of lava in opposite directions, one of forty, the other fifty miles in length, and averaging eleven miles in breadth, and perhaps fifty feet in thickness. The fragmentary matter ejected at the same time, and carried down the slopes of the volcano by deluges of rain, must have been of proportionate magnis tude. This example alone invalidates the assumption that the igneous forces have been impaired and enfeebled in latter times. It would be most difficult to point out a mass of igneous origin of ancient date, distinctly referrible to a single eruption, which would rival in volume the matter poured out by Skapta Jokul in 1783.

Next in order, M. Lyell discusses the changes effected by earthquakes. These are principally alterations in the superficial levels, and the production of crevices in solid strata. Unfortunately the relations of earthquakes are usually confined to the damage sustained by towns or villages, and little notice is taken of phenomena interesting only to the naturalist. Moreover, the extent of alterations in level can hardly be ascertained at all, except along the shore of the sea, which supplies a stationary base from whence to measure the change. Mr. Lyell has, however, collected a sufficient number of wellauthenticated facts, to prove that both subsidence and elevation, on a very extensive scale, occasionally accompany earthquakes. The most remarkable, perhaps, is the well-known elevation, in 1821, of the whole coast of Chili, through a space of above one hundred miles, to a height of from three to four feet along the shore, and, according to all appearance, much more at some distance inland: Older terraces of shingle and shell range along the same coast to a height of fifty feet, showing the land to have been raised that mach above the sea by preceding shocks at no very distant date. The earthquake of the Caraccas in 1812 is described as terrific. The surface undulated like a boiling liquid, producing all the effects of sea-sickness. Enormous rocks were detached from the mountain, one of which,

Silla, lost three hundred feet of its height. The year before, the valley of the Missis sippi was similarly convulsed. The inhabitants relate that the earth rose in great waves; and when they reached a certain fearful height, the surface burst, and volumes of water, sand, and coal, the materials of the soil, were discharged to the height of a hundred feet and more. The chasms were all parallel, and in a direction from S. W. to N.E. (the direction of the Alleghany chain which borders the basin of the Mississippi), and many of the inhabitants saved themselves from being swallowed up by felling tall trees, laying them at right angles to the direction of the crevices, and stationing themselves upon them.

The sea shares in the agitation of the solid earth. Ships feel every shock as if they had struck on a shoal, and loose articles lying on their decks are often thrown several feet into the air, showing the violence of the upward movement communicated to the water. The sea often deserts the coast, and returns immediately in a terrific wave (that of Lisbon and the coast of Spain in 1755 was fifty feet high), which sweeps over the shore, and must leave lasting traces of its devastating power. It is probably caused by the sudden upheaving of a portion of the bed of the sea, the first effect of which would be to raise a body of water over the elevated part, its momentum carrying it much above the level it would afterwards assume, and causing a draught or receding of the water from the neighbouring coasts, immediately followed by the return of the displaced water, which will be also impelled by its momentum much further and higher on the coast than its former level. The undulatory shocks of the earthquake of 1755 travelled over sca and land at the rate of twenty miles in a minute, as appears from the interval between the time when the first shock was felt in Lisbon, and that of its occurrence at distant places in the West Indies, Scotland, Norway, Switzerland, Italy, and North Africa. The earthquake felt at Conception in 1750 uplifted the bed of the sea to the height of twenty-four feet at the least, and it seems probable that the adjoining coast shared in the elevation, for an enormous bed of shells, of the same species as those now living in the bay, is seen raised above high-water mark along the beach. These shells, as well as others which cover the adjoining hills of mica-schist, to the height even of fifteen hundred feet, have been identified with some taken at the same time in a living state from the bay. There is, therefore, every reason to conclude that the

whole extent of this coast, so often visited by severe carthquakes, has suffered a very great amount of elevation within an exceedingly recent period.

Mr. Lyell discusses at length the much controverted question of the apparent changes of level in the neighbourhood of Pozzuoli, since the Roman era, and brings forward an overwhelming mass of evidence in proof of the fact that this part of the Campanian coast was lowered at least twenty feet some time between the third and the sixteenth century, and re-elevated about as much again at the epoch of the eruption which produced the Monte Nuovo. The circumstances which demonstrate this are so clearly legible, that it would never perhaps have been disputed but for the natural repugnance to admit so remarkable a local coincidence of depression and elevation to nearly the same extent, as well as the strong prejudices existing in regard to the immobility of the land, by which we have probably been blinded to the force of many other similar facts. But it is time the geologist, at least, should overcome those first and natural impressions which induced the poets of old to select the rock as the emblem of stability, the sea of mutability. Paradoxical as it may appear, truth compels us to reverse the opinion; and, with respect to periods of long duration, to attribute invariability of level to the ocean, fluctuation and inconstancy to the land.

With regard to the exciting cause of earthquakes and eruptions, our author expresses no decided opinion: he admits, however, that the phenomena prove the existence of vast bodies of intensely heated rock, probably in a liquefied state, like lava, beneath the solid crust of the earth, and also that there is a continual transmission of heat from thence to the surface, more or less regular or interrupted, according to the obstacles it meets with, or creates, to its own development. Now, it does appear to us that this undeniable evolution of heat from the interior of the globe towards its surface is alone fully sufficient to account for all the phenomena of earthquakes and volcanos, which seem to follow necessarily from its action by the simple laws of mechanic and hydrostatic forces. It is evidently only by the formation of habitual vents or chimneys for the free passage of hot vapour, that the internal heat can be discharged through the imperfectly conducting superficial strata, in sufficient abundance to obviate the more violent outbursts or expansions of the matter confined immediately below them at an increasing temperature. But the circumstances which allow of a permanently eruptive vent, as Stromboli,