From what has been said in the preceding chapters it will be seen that while some of the materials ejected from volcanic vents are, by the movements of the air and ocean, distributed over every part of the face of the globe, another, and by far the larger, part of the matter so ejected, accumulates in the immediate vicinity of the vent itself. By this accumulation of erupted materials, various structures are built up around the orifices from which the ejections take place, and the size and character of these structures vary greatly in different cases, according to the quantity and nature of the ejected materials, and the intensity of the eruptive forces by which they were thrown from the orifice. We shall proceed in the present chapter to notice the chief varieties in the forms and characters of the heaps of materials accumulated round volcanic vents.

These heaps of materials vary in size from masses no bigger than a mole-heap up to mountains like Etna, Teneriffe, and Chimborazo. The size of volcanic mountains is principally determined by the conditions of the eruptive action at the vent around which they are formed. If this action exhausts itself in a single effort, very considerable volcanic cones, like the Monte Nuovo with many similar hills in its vicinity, and the Puys of Auvergne, may be formed; but if repeated eruptions take place at longer or shorter intervals from the same vent, there appears to be scarcely any limit to the size of the structures which may, under such conditions, be formed. It is by this repeated action from the same volcanic vent going on for thousands or even millions of years, that the grandest volcanic mountains of the globe have been built up. Such volcanoes have sometimes a diameter at their base of from 30 to 100 miles, and an elevation of from 10,000 to 25,000 feet.

The form of volcanic mountains is determined in part by the nature of the materials ejected, and in part by the character of the eruptive action.

From what has been said in the preceding chapter, it will be gathered that the volcanoes built up by ejections of fragmentary materials differ in many striking particulars from those formed by the outwelling of lavas from volcanic vents. In a less degree, the volcanoes composed of the same kind of volcanic materials also vary among themselves.

When masses of scoriæ in a semi-fluid condition are thrown to only a little distance above the volcanic vent, so that they have not time to assume a perfectly solid condition before they fall round the vent, the rugged masses of lava unite to form heaps of most irregular shape. In such cases, the falling fragments being in a semi-plastic state, stick to the masses below, and do not tend to roll down the sides of the heap. Irregular heaps of such volcanic scoriæ abound on the surfaces of lava-streams, being piled up around each 'bocca' or vent which the steam-jets escaping from the lava-currents form at their surfaces. Such irregular accumulations of scoriæ were observed on the lavas of Vesuvius during the eruptions of 1822, 1855, and 1872, and have also been described in the case of many other volcanoes. In fig. 26 (p. 101) we have given representations of a group of such irregular scoria-cones which was observed by Schmidt on the Vesuvian lava of 1855. It will be seen from this drawing that there is scarcely any limit to the steepness of the sides of such scoria-heaps, in which the materials are in an imperfectly solidified condition when they reach the ground.

But in the majority of cases, the scoriæ ejected from volcanic vents are thrown to a great height, and are in a more or less perfectly solidified condition when they fell to the ground again. In such cases the fragments obey the ordinary mechanical laws of falling bodies, rolling and sliding over one another, till they acquire a slope which varies according to the size and form of the fragments. In this way the great conical mounds are formed which are known as 'cinder-cones,' or more properly as 'scoria-cones.' Scoria-cones usually vary in the slope of their sides from 35° to 40° and may differ in size from mere monticules to hills a thousand feet or more in height. Scoria-cones of this character abound in many volcanic districts, as the Auvergne, where they may be numbered by thousands. The materials forming such scoria-cones vary in size from that of a nut to masses as large as a man's head, and fragments of even larger dimensions are by no means uncommon.

When the lava in a volcanic vent is perfectly glassy, instead of being partially crystalline in structure, we find not scoriæ but pumice ejected. In such cases, as in the Lipari Islands for example, we see cones entirely built up of pumice. Such pumice-cones resemble in the angle of their slope (see fig. 41, facing p. 124), the ordinary scoria-cones, but are of a brilliant white colour, appearing as if covered with snow.

Ordinary scoriæ are usually of a black colour when first ejected, but after a short time the black oxide of iron (magnetite) which they contain, attracts the oxygen of the air and moisture, and assumes the reddish-brown colour of iron-rust. Under such circumstances the heaps of black material gradually acquire the red-brown colour which is characteristic of so many of the scoria-cones around Etna, and in the Auvergne and the Eifel. The moisture of the air, and the rain falling upon these loose cindery heaps, cause them to decompose upon their surfaces; the action is facilitated by the growth of the lower forms of vegetation, such as mosses and lichens, and thus at last a soil is produced on the surfaces of these conical piles of loose materials which may support an abundant vegetation. Cinder- or scoria-cones are not uncommonly found retaining in a most perfect manner their regular, conical form, the lips of their craters being sharp and unbroken as if the cone were formed but yesterday, while their slopes may nevertheless be covered with a rich soil supporting abundant grass and forest-trees. It may at first sight seem difficult to understand how a loose mass of scoriæ could have so long withstood the action of the rain and floods, retaining so perfectly its even slopes and sharp ridges. A little consideration will, however, convince us that it is the very loose and pervious nature of the materials of which scoria-cones are composed, which tends to their perfect preservation. The rain at once sinks into their mass, before it has time to form rivulets and streams which would wear away their surfaces and destroy the regularity of their outlines.

Scoria- and pumice-cones are frequently found to be acted upon by acid vapours to such an extent that the whole of the materials is reduced to a white pulverulent mass. In these cases the oxides of iron and the alkalis have united with the sulphuric or hydrochloric or carbonic acids, the compounds being carried away in solution by the rain-water falling on the mass; the materials left are silica, the hydrated silicate of alumina, and hydrated sulphate of lime (gypsum), all of which are of a white colour.

Cinder- or scoria-cones, and pumice-cones, are often found raised by the action of winds to a greater elevation on one side than the other, in the manner already described. One side of the cone is often seen to be more or less completely swept away by an outwelling stream of lava, and thus breached cones are formed (see fig. 40, p. 123). Not unfrequently we find a number of cones which are united more or less completely at their bases, as in Vulcanello (fig. 6, p. 43), the several vents being so near together that their ejections have mingled with one another. Cones composed entirely of fragmentary materials often show an approach to the beautifully curved slopes which we have described as being so characteristic of volcanoes, as may be seen in fig. 41, facing p. 124. In the case of scoria- and pumice-cones this curvature is probably due to the rolling downnwards and outwards of the larger fragments.

We have already pointed out that with the scoriæ there are often ejected fragments torn from the sides of the volcanic vents. Sometimes such fragments are so numerous as to make up a considerable portion of the mass of the volcanic cones. Thus in the Eifel we find hills, of by no means insignificant size, completely built up of small scoriæ and broken fragments of slate torn from the rocks through which the volcanic fissures have been opened. Occasionally we see that few or no scoriæ have been ejected, and the volcanic vents are surrounded simply by heaps of burnt slate.

The smaller fragmentary materials ejected from volcanic vents—such as lapilli and dust—rest in heaps, having a different angle of slope from those formed by scoriæ. In many cases, as we have seen, such finely-divided materials descend in the condition of mud, which flows evenly over the surface of the growing cone and consolidates in beds of very regularly stratified 'tufa' or 'tuff.'

The 'tuff-cones' thus formed differ in many important respects from the scoria-cones already described. The slope of their sides varies from 15° to 30°, and is almost always considerably less than in scoria- and pumice-cones. The tuff-cones undergo much more rapid degradation from rain and moisture than do the scoria-cones; for, though the materials of the former 'set,' as we have seen, into a substance of considerable hardness, yet this substance, being much less pervious to water than the loose scoria heaps, permits of the formation of surface-streams which furrow and wear away the sides of the cones. Sometimes the sides of the crater are found to be almost wholly removed by atmospheric denudation, and only a shallow depression is found occupying the site of the crater; such a case is represented in fig. 59. We not unfrequently find the whole slopes of such cones to be traversed by a series of radiating grooves passing from the summit to the base of the mountains, these channels being formed by water, which has collected into streams, flowing down the slopes of the mountains. The volcanic cone, under these circumstances, frequently presents the appearance of a partially opened umbrella. Owing to the impervious character of the materials composing tuff-cones, their craters are frequently found to be occupied by lakes.

Tufas have usually a white or yellowish-brown colour, and these are the colours exhibited by the cones composed of this material before they become covered by vegetation. Tufas scoriæ, and lavas usually crumble down to form a very rich soil, and many of the choicest wines are produced from grapes grown on the fertile slopes of volcanic mountains. When, however, as not unfrequently happens, the materials are finely divided and incoherent, they are so easily driven about by the winds that cultivation of any kind is rendered almost impossible. In the Islands of Stromboli and Vulcano the gardens have to be surrounded by high fences to prevent them from being overwhelmed by the ever-shifting masses of volcanic sand.

There are some cones which are composed in part of scoriæ and in part of tufa. Hence we are sometimes at a loss whether to group them with the one class of cones or the other. But in the majority of cases, scoria- and tuff-cones present the sufficiently well-marked and distinctive characters which we have described.

Lava-cones differ quite as greatly in their forms as do the cones composed of fragmentary materials, the variations being principally determined by the degree of liquidity of the lavas.

We sometimes find that outwelling masses of lava, when issuing in small quantities from a vent, accumulate in cauliflower-shaped masses, or sometimes in the form of a column, or bottle. Professor J. D. Dana describes many such fantastically-formed masses of lava as being found in Hawaii, one of which is represented in fig. 25 (p. 100).

When the lava issues from the vent in great quantities it tends to flow on all sides of it, and to build up a great conical heap above the orifice. If the lava be very liquid it flows to great distances, resting at a very slight slope. Thus we find that the volcanoes of Hawaii have been built up of successive ejections of very liquid lava, which have formed cones having a slope of only 6° to 8°, but of such enormous dimensions that the diameter of their bases is seventy miles and their height 14,000 feet.

If, on the other hand, the lava be viscid, or very imperfectly liquid in character, it tends to accumulate immediately around the vent; fresh ejections force the first extruded matter outwards, in the manner so well illustrated by Dr. Reyer's experiments, and at last a more or less steep-sided bulbous mass is formed over the vent. Such bulbous masses, composed of imperfectly fluid lavas, occur in many volcanic districts, and constitute hills of considerable size. From the tendency of matters thus extruded to choke up the vents, however, these volcanoes composed of viscid lavas cannot be expected to attain the vast dimensions reached by some of those composed of very liquid lavas. The difference in the forms of lava-cones composed of very fluid or of somewhat viscid materials is illustrated in fig. 58. When the interior of such steep-sided volcanic mountains composed of viscid materials is exposed by the action of denuding forces, the peculiar internal structure we have described is displayed by them. In the Chodi-Berg of Hungary, a great bulbous mass of andesitic rock, this endogenous structure is admirably displayed. It is also well seen in the excavation of the hill of the Grand Sarcoui, a similar mass, composed of altered trachyte, which has been erupted in the midst of a scoria-cone in the Auvergne. See fig. 44 (p. 126).

Most of the great volcanic mountains of the globe belong to the class of 'composite cones,' and are built up by alternate ejections of fluid lava and fragmentary materials. The slope of the sides in such composite cones is subject to a wide range of variation, being determined in part by the degree of liquidity of the lavas, in part by the nature of the fragmentary materials ejected, and in part by the proportions which the fragmentary and lava-ejections bear to one another.

But there is another set of causes which tends to modify the form and character of these composite, volcanic cones. As we have already pointed out, the sides of such cones are liable to be rent asunder from time to time, and the fissures so produced are injected with masses of liquid lava from below. These fissures, rent in the sides of volcanic cones, often reach the surface and eruptive action takes place, giving rise to the formation of a cone, or series of cones, upon the line of the fissure (fig. 61). Such small cones thrown up on the flanks of a great volcanic mountain are known as 'parasitic cones'; though subordinate to the great mountain mass, they may be in themselves of considerable dimensions. Among the hundreds of parasitic cones which stud the flanks of Etna, there are some which are nearly 800 feet in height.

The building up of parasitic cones upon the flanks of a volcanic mountain tends, of course, to destroy its regular conical form. This may be well seen in Etna, which, by the accumulation of materials upon its flanks, has become a remarkably 'round-shouldered' mountain. (See figs. 62 and 63.) At the same time it must be remembered that materials erupted from the central vent tend to fill up the hollows between these parasitic cones, and thus to restore to the mountain its regularly conical form.

The Island of Ischia is a good example of a great volcanic cone the flanks of which are covered with numerous small parasitic cones. While the great central volcano has evidently been long extinct, and one side of its crater-wall is completely broken down, some of the small parasitic cones around its base have been formed within the historical period—one of them as recently as the year 1302. Fig. 64 is a plan of the Island of Ischia, showing the numerous parasitic cones scattered over the slopes of the principal cone.

In one case we find that a parasitic cone, the Monte Rotaro, has itself a similar smaller cone, which is parasitic to it, at its foot; this secondary parasitic cone gives off a small lava-stream of trachyte, which has flowed down to the sea. (See fig. 65.)

Most great volcanic mountains exhibit a tendency towards a subsidence of their central portions, which may take place either during or subsequently to their period of activity. When we examine the strata upon which a volcano has been built up, but which are now exposed to our study by denuding forces, we usually find that they incline towards the centre of the eruptive activity. (See figs. 66 and 67.) Two causes may contribute to bring about this result. In the first instance, we must remark that the piling up of materials around the volcanic vent causes the subjacent strata to be subjected to a degree of pressure far is excess of that which acts upon the surrounding rocks. And secondly, it must be borne in mind that the continual removal of material from below the mountain must tend to the production of hollows, into which the overlying strata will sink. The effect of this central subsidence is to give to the flanks of volcanic cones those beautifully curved outlines which constitute so striking a feature in Vesuvius (see fig. 17, p. 87), Fusiyama (see fig. 77, No. 1, facing p. 178), and many other volcanic mountains.

There seems, at first sight, to be scarcely any limit to the dimensions which these great composite volcanic cones may attain: the lateral eruptions tending to enlarge the area of the base of the mountain, and, by the injection of the fissures, to knit together and strengthen its structure, while the central eruptions continually increase the elevation of the mass. Great, however, as is the force which is concerned in the production of our terrestrial volcanoes, it has its limits; and, at last, the piling up of materials will have gone on to such an extent, that the active forces beneath the volcano are no longer competent either to raise materials to the elevated summit of the mountain or to tear asunder its strengthened and fortified flanks. Under these circumstances, the volcanic forces, if they have not already exhausted themselves, will be compelled to find weak places in the district surrounding the volcano, at which fissures may be produced and the phenomena of eruption displayed.

Some volcanic cones exhibit evidence that during the series of eruptions by which they have gradually been built up, the centre of volcanic action has shifted to another point within the mountain. Thus Lyell has shown, in the case of Etna, that during the earlier periods in the history of the mountain the piling up of materials went on around a centre which is now situated at a distance of nearly four miles from the present focus of eruptive activity. Some of our old British volcanoes, of which the denuded wrecks exist in the Western Isles of Scotland, show similar evidence of a shifting of the axis of eruption.

One of the most conspicuous features of a volcanic cone is the great depression or crater found at its summit. In describing the internal structure of volcanic cones, we have seen how these craters are produced and acquire their inverted conical form, by the slipping and rolling back of materials towards the centre of eruptive action.

Almost all volcanic cones exhibit craters, but in those which are formed entirely by the outwelling of viscid lavas the central depression is often slight and inconspicuous, and occasionally altogether wanting. It frequently happens, however, that eruptive action has ceased at the centre of a volcano, and its summit-crater may by denudation be entirely destroyed, while new and active craters are formed upon its flanks. Stromboli furnishes us with an admirable example of this kind (see fig. 1, facing p. 10). Other volcanoes may exhibit several craters, one at the summit of the mountain and others upon its flanks. Of this we find a good example in Vulcano (fig. 6, p. 43).

When a volcano has been built up by regular and continuous eruptions from the same volcanic vent, the size of the crater remains the same, while the volcano continually grows in height and in the diameter of its base. The size of the crater will be determined by the eruptive force at the volcanic centre, the size of the mountain by the duration of the volcanic activity and the quantity of material ejected. In the earliest stage of its history, such a volcano will resemble Monte Nuovo, which has a crater reaching down almost to the base of the mountain; in the later stages of its history, such a volcano will resemble Cotopaxi (fig. 68) and Citlaltepetl (fig. 69), in which the crater, though of far greater absolute dimensions than that of Monte Nuovo, bears but a small proportion to the vast cone at the summit of which it is situated.

In the great majority of volcanoes, however, eruptive action does not go on by any means regularly and continuously, but terrible paroxysmal outbursts occur, which suddenly enlarge the dimensions of the crater to an enormous extent.

In the year 1772, there occurred a volcanic eruption in the Island of Java, which is perhaps the most violent and terrible that has happened within the historical period. A lofty volcanic cone, called Papandayang, 9,000 feet high, burst into eruption, and, in a single night, 30,000,000,000 cubic feet of materials were thrown into the atmosphere, falling upon the country around the mountain where no less than forty villages were buried. After the eruption, the volcano was found to have been reduced in height from 9,000 to 5,000 feet, and to present a vast crater in its midst, which had been formed by the ejection of the enormous mass of materials.

Many similar cases might be cited of the removal of a great part of a mountain-mass by a sudden, paroxysmal outburst. In some cases, indeed, the whole mass of a mountain has been blown away during a terrific eruption, and the site of the mountain is now occupied by a lake. This is said to have been the case with the Island of Timor, where an active volcano, which was visible from a distance of 300 miles at sea, has entirely disappeared.

The removal of the central portion of great volcanic mountains by explosive action, gives rise to the formation of those vast, circular, crater-rings of which such remarkable examples occur in many volcanic districts. These crater-rings present a wall with an outer slope agreeing with that of the volcanic cone of which they originally formed a part, but with steep inner cliffs, which exhibit good sections of the beds of tuff, ash, and lava with the intersecting dykes of which the original volcano was built up. Near Naples, one of these crater-rings, with sloping outer sides and steep inner ones, is employed to form the royal game-preserve of Astroni, the only entrance to the crater being closed by gates.

As these crater-rings are usually composed of materials more or less impervious to water, they often become the site of lakes. The beautiful circular lake of Laach, in the Rhine Provinces, with the numerous similar examples of Central Italy—Albano, Nemi, Bracciano, and Bolsena—the lakes of the Campi Phlegræi (Agnano, Avernus, &c.), and some similar lakes in the Auvergne, may be adduced as examples of crater-rings which have become the site of lakes.

One of the most beautiful of the crater-lakes in the Auvergne is Lac Paven (fig. 70), which lies at the foot of a scoria-cone, Mont Chalme, and is itself surrounded by masses of ejected materials. The crater-lake of Bagno, in Ischia (fig. 71), has had a channel cut between it and the sea, so that it serves as a natural harbour. The lake of Gustavila, in Mexico (fig. 72), is an example of a crater-lake on a much larger scale.

In many of these crater-rings the diameter of the circular space enclosed by them is often very great indeed as compared with the height of the walls.

Two of the largest crater-rings in the world are found in Central Italy, and are both occupied by lakes, the circular forms of which must strike every observer. The Lago di Bracciano, which lies to the north-west of Rome, is a circular lake six and a half miles in diameter, surrounded by hills which at their highest point rise to the height of 1,486 feet above the sea, while the surface of the waters of the lake is 640 feet above the sea-level. The Lago di Bolsena is somewhat less perfectly circular in outline than the Lago di Bracciano; it has a length from north to south of ten-and-a-quarter miles and a breadth from east to west of nine miles; the surface of the waters of this lake is 962 feet above that of the waters of the Mediterranean. The lake of Bolsena, like that of Bracciano, is surrounded by hills composed of volcanic materials; the highest points of this ring of hills rise to elevations of 684, 780, and 985 feet respectively above the waters of the lake.

In these great circular lakes of Bolsena and Bracciano, as well as in the smaller ones of Albano, Nemi, and the lakes of Frascati in the same district, the vast circular spaces enclosed by them, the gradual outer slope of the ring, and the inner precipices which bound the lake, all afford evidence of the explosive action to which they owe their origin.

But while the vast crater-rings we have mentioned are frequently found to be occupied by lakes, there are many other similar crater-rings which remain dry, either from the materials of which they are composed being of more pervious character, or from rivers having cut a channel through the walls of the crater, in this way draining off its waters.

Thus in the Campi Phlegræi, while we have the craters of Agnano and Avernus forming complete circular lakes, Astroni has only a few insignificant lakelets on its floor, and the Pianura, the Piano di Quarto, which have each a diameter of three or four miles, with many others, remain perfectly dry. In the vicinity of the great crater-lakes of Central Italy we find the crater-ring of the Vallariccia, which has evidently once been a lake but is now drained, its floor being covered with villages and vineyards.

A comparison of these vast crater-rings leads us to the conclusion that in the majority of cases, if not in every instance, they are composed almost entirely of volcanic tuff and dust. In the case of the more solidly-built composite volcanic cones, the volcanic forces, as we have seen, produce fissures in the mass, and along these fissures parasitic cones are thrown up, the tension of the mass of imprisoned vapours below the mountain being thus from time to time relieved. But in the case of a volcanic cone composed of loose fragmentary materials, such temporary relief is impossible. The cracks, as soon as they originate, will be filled up and choked by the falling in of materials from above and at their sides. In this way the eruptive action will be continually repressed, till at last the imprisoned vapours acquire such a high state of tension that the outburst, when it occurs, is of the most terrible character, and the whole central mass of the volcano is blown into the air. It may often seem surprising that the ejection of such vast masses of material from the centre of a volcanic cone does not effect more in the way of raising the height of the crater-walls. But it must be remembered that, in the case of craters of such vast area, the majority of the ejected materials must fall back again within its circumference. By repeated ejections these materials will at last be reduced to such an extreme state of comminution that they can be borne away by the winds, and spread over the country to the distance of hundreds or thousands of miles. After great volcanic outbursts enormous areas are thus found covered with fine volcanic dust to the depth of many inches or feet.

Sometimes, as in the case of the Lago di Bracciano, the eruptive forces appear to have entirely exhausted themselves in the prodigious outburst by which the great crater was produced. But in other cases, as in that of the Lago di Bolsena, the eruptive action was resumed at a later date, and small tuff-cones were thrown up upon the floor of the crater; these now rise as islands above the surface of the lake. In other cases, again, the eruptive action was resumed after the formation of the great crater-ring, with such effect that bulky volcanic cones were built up in the midst of the crater-ring which surrounds them like a vast wall; examples of this are exhibited in the extinct volcanoes of Rocca Monfina and Monte Albano. Some of the grandest volcanoes of the globe, such as Teneriffe (fig. 73), the volcanoes of Mauritius and Bourbon (figs. 74 and 75), and many others that might be cited, are thus found to be surrounded by vast crater-rings. Vesuvius itself is surrounded by the crater-ring of Somma (fig. 76).

This formation of cone within crater, often many times repeated, is very characteristic of volcanoes. The craters mark sudden and violent paroxysmal outbursts, the cones are the result of more moderate but long-continued ejection. Sometimes, as at Vesuvius in 1767 (fig. 15, p. 85), we find a nest of craters and cones which very strikingly exemplifies this kind of action.

We shall point out, hereafter, that at most volcanic centres the ejection of trachytic lavas precedes that of the basaltic lavas. Now it is these trachytic lavas which principally give rise to the formation of the light lapilli of which tuff-cones are formed. Hence it is that we so frequently find, as in the case of Vesuvius, Rocca-Monfina, and many other volcanoes, that a great crater-ring, largely composed of tuffs, encloses a cone built up of more basic lavas.

In fig. 77 we have shown by a series of outline sections the various forms assumed by volcanoes in consequence of the different kinds of eruptive action going on in them:—

1. Is an outline of Fusiyama, an almost perfect cone, with a small crater at its summit. The sides of this volcano admirably illustrate the beautiful double curves characteristic of volcanic cones.

2. Hverfjall in Iceland, a volcanic cone with a large crater, reaching almost to its base.

3. The crater-lake of Bracciano, in which the area of the crater is out of all proportion to the height of the crater-walls.

4. Rocca-Monfina, in Southern Italy, a tuff-cone of large dimensions, in the midst of which an andesitic lava-cone has been built up.

5. Teneriffe, in the Canary Islands, in which a perfect volcanic cone has been built up in the centre of an encircling crater-ring.

6. Vulcano, in the Lipari Islands, in which, by the shifting of the centre of volcanic activity along a line of fissure, a series of overlapping volcanic cones has been produced.

While speaking of the varieties of form assumed by volcanic cones and craters, we must not forget to notice the effects which are produced by denuding forces upon them. In the case of submarine volcanoes, like the celebrated island called by the English Graham Isle, by the French Isle Julie, and by the Germans the Insel Ferdinandez (fig. 78), which was thrown up off the coast of Sicily in 1831, it was evident that volcanic outbursts taking place at some depth below the level of the sea gradually piled up a cone of scoriæ with a crater in its midst. By constant accessions to its mass, this scoria-cone was eventually raised above the sea-level, but the action of the waves upon the loose materials soon destroyed the crater-walls and eventually reduced the island to a shoal. It is evident that in all cases in which eruptions take place beneath the sea-level, and the loose materials are exposed during their accumulation to the beating of the sea-waves, the form of the volcanic cone so produced will be greatly modified by the interaction of the two sets of opposed causes, the eruptive forces from below and the distributive action of the sea-waves.

Craters when once formed are often rent across, along the line of the fissure above which they are thrown up. Thus the crater of Vesuvius was in 1872 rent completely asunder on one side, so that it was possible to climb through the fissure thus produced and reach the bottom of the crater. Streams flowing down the sides of the crater, and escaping through such a rent, may in the end greatly modify the form and disguise the characters of a volcanic crater. Of this kind of action we have a striking example in the Val del Bove of Etna.

Volcanoes, as we shall point out in the sequel, are after their extinction frequently submerged beneath the waters of the ocean. The sea entering the craters, eats back their cliff-like sides and enlarges their areas. Such denuded waters are called 'calderas,' the channels into them 'barrancos.'

Sometimes the action of the waves upon a partially submerged volcano has led to the cutting back of its slopes into steep cliffs, at the same time that the crater-ring is enlarged. In such cases we have left a more or less complete rocky ring, composed of alternating lavas and fragmentary materials. Of such a ruined crater-ring, the Island of St. Paul in the South Atlantic affords an admirable example.

When the action of denudation has gone still further, all the lavas and tuffs composing the cone may be completely removed and nothing left but masses of the hard and highly-crystalline rocks which have cooled down slowly in the heart of the volcano. An example of this kind is afforded to us by St. Kilda, the remotest member of the British Archipelago.

But although the majority of volcanic craters are clearly formed by explosive action, there are some craters, like those of Kilauea in Hawaii, which probably owe their origin to quite a different set of causes. In this case the explosive action at the vent is but slight, and the crater, which is of very irregular form, appears to have originated in a fissure, which has been slowly enlarged by the liquid lavas encroaching upon and eating away its sides. Such craters as these, however, appear to be comparatively rare.

Besides the great volcanic mountains composed of lava, scoriæ, tuff and ash, there are other structures which are formed around volcanic vents even when these do not eject molten rock-masses. The water which issues in these cases either as steam or in a more or less highly heated condition frequently carries materials in suspension or solution, and these sometimes accumulate in considerable quantities around the vent.