Volcanic phenomena are the outward manifestations of forces deep-seated beneath the crust of the globe; and in seeking for the causes of such phenomena we must be guided by observation of their nature and mode of action. The universality of these phenomena all over the surface of our globe, in past or present times, indicates the existence of a general cause beneath the crust. It is true that there are to be found large tracts from which volcanic rocks (except those of great geological antiquity) are absent, such as Central Russia, the Nubian Desert, and the Central States of North America; but such absence by no means implies the non-existence of the forces which give rise to volcanic action beneath those regions, but only that the forces have not been sufficiently powerful to overcome the resistance offered by the crust over those particular tracts. On the other hand, the similarity of volcanic lavas over wide regions is strong evidence that they are drawn from one continuous magma, consisting of molten matter beneath the solid exterior crust.

(a.) Lines of Volcanic Action.—It has been shown in a previous page that volcanic action of recent or Tertiary times has taken place mainly along certain lines which may be traced on the surface of a map or globe. One of these lines girdles the whole globe, while others lie in certain directions more or less coincident with lines of flexure, plication or faulting. The Isle of Sumatra offers a remarkable example of the coincidence of such lines with those of volcanic vents. Not only the great volcanic cones, but also the smaller ones, are disposed in chains which run parallel to the longitudinal axis of the island (N.W.-S.E.). The sedimentary rocks are bent and faulted in lines parallel to the main axis, and also to the chains of volcanic mountains, and the observation holds good with regard to different geological periods.[1] Another remarkable case is that of the Jordan Valley. Nowhere can the existence of a great fracture and vertical displacement of the strata be more clearly determined than along this remarkable line of depression; and it is one which is also coincident with a zone of earthquake and volcanic disturbances.

(b.) Such Lines generally lie along the Borders of the Ocean.—But even where, from some special cause, actual observation on the relations of the strata are precluded, the general configuration of the ground and the relations of the boundaries between land and sea to those of volcanic chains, evidently point in many cases to their mutual interdependence. The remarkable straightness of the coast of Western America, and of the parallel chain of the Andes, affords presumptive evidence that this line is coincident with a fracture or system of faults, along which the continent has been bodily raised out of the waters of the ocean. Of this elevation within very recent times we have abundant evidence in the existence of raised coral-reefs and oceanic shell-beds at intervals all along the coast; rising in Peru to a level of no less than 3,000 feet above the ocean, as shown by Alexander Agassiz.[2] Such elevations probably occurred at a time when the volcanoes of the Andes were much more active than at present. Considered as a whole, these great volcanic mountains may be regarded as in a dormant, or partially moribund, condition; and if the volcanic forces have to some extent lost their strength, so it would appear have those of elevation.

(c.) Areas of Volcanic Action in the British Isles.—In the case of the British Islands it may be observed that the later Tertiary volcanic districts lie along very ancient depressions, which may indicate zones of weakness in the crust. Thus the Antrim plateau, as originally constituted, lay in the lap of a range of hills formed of crystalline, or Lower Silurian, rocks; while the volcanic isles of the Inner Hebrides were enclosed between the solid range of the ArchÆan rocks of the Outer Hebrides on the one side, and the Silurian and ArchÆan ranges of the mainland on the other. And if we go back to the Carboniferous period, we find that the volcanic district of the centre of Scotland was bounded by ranges of solid strata both to the north and south, where the resistance to interior pressure from molten matter would have been greater than in the Carboniferous hollow-ground, where such molten matter has been abundantly extruded. In all these cases, the outflow of molten matter was in a direction somewhat parallel to the plications of the strata.

(d.) Special Conditions under which the Volcanic Action operates.—Assuming, then, that the molten matter, forming an interior magma or shell, is constantly exerting pressure against the inner surface of the solid crust, and can only escape where the crust is too weak (owing to faults, plications, or fissures) to resist the pressure, we have to inquire what are the special conditions under which outbursts of volcanic matter take place, and what are the general results as regards the nature of the ejecta dependent on those conditions.

(e.) Effect of the Presence or Absence of Water.—The two chief conditions determining the nature of volcanic products, considered in the mass, are the presence or absence of water. Such presence or absence does not of course affect the essential chemical composition of the ejecta, but it materially influences the form in which the matter is erupted. The agency of water in volcanic eruptions is a very interesting and important subject in connection with the history of volcanic action, and has been ably treated by Professor Prestwich.[3] At one time it was considered that water was essential to volcanic activity; and the fact that the great majority of volcanic cones are situated in the vicinity of the oceanic waters, or of inland seas, was pointed to in confirmation of this theory. But the existence in Western America and other volcanic countries of fissures of eruption along which molten lava has been extruded without explosions of steam, shows that water is not an essential factor in the production of volcanic phenomena; and, as Professor Prestwich has clearly demonstrated, it is to be regarded as an element in volcanic explosions, rather than as a prime cause of volcanic action. The main difficulty he shows to be thermo-dynamical; and calculating the rate of increase in the elastic force of steam on descending to greater and greater depths and reaching strata of higher and higher temperatures, as compared with the force of capillarity, he comes to the conclusion that water cannot penetrate to depths of more than seven or eight miles, and therefore cannot reach the seat of the eruptive forces. Professor Prestwich also points out that if the extrusion of lava were due to the elastic force of vapour of water there should be a distinct relation between the discharge of the lava and of the vapour; whereas the result of an examination of a number of well-recorded eruptions shows that the two operations are not related, and are, in fact, perfectly independent. Sometimes there has been a large discharge of lava, and little or no escape of steam; at other times there have been paroxysmal explosive eruptions with little discharge of lava. Even in the case of Vesuvius, which is close to the sea, there have been instances when the lava has welled out almost with the tranquillity of a water-spring.

(f.) Access of Surface Water to Molten Lava during Eruptions.—The existence of water during certain stages in eruptions is too frequent a phenomena to be lost sight of; but its presence may be accounted for in other ways, besides proximity to the sea or ocean. Certain volcanic mountains, such as Etna and Vesuvius, are built upon water-bearing strata, receiving their supplies from the rainfall of the surrounding country, or perhaps partly from the sea. In addition to this the ashes and scoriÆ of the mountain sides are highly porous, and rain or snow can penetrate and settle downwards around the pipe or throat through which molten lava wells up from beneath. In such cases it is easy to understand how, at the commencement of a period of activity, molten lava ascending through one or more fissures, and meeting with water-charged strata or scoriÆ, will convert the water into steam at high pressure, resulting in explosions more or less violent and prolonged, in proportion to the quantity of water and the depth to which it has penetrated. In this manner we may suppose that ashes, scoriÆ, and blocks of rock torn from the sides of the crater-throat, and hurled into the air, are piled around the vent, and accumulate into hills or mountains of conical form. After the explosion has exhausted itself, the molten lava quietly wells up and fills the crater, as in the cases of those of Auvergne and Syria, and other places. We may, therefore, adopt the general principle that in volcanic eruptions where water in large quantities is present, we shall have crater-cones built up of ashes, scoriÆ, and pumice; but where absent, the lava will be extravasated in sheets to greater or less distances without the formation of such cones; or if cones are fanned, they will be composed of solidified lava only, easily distinguishable from crater-cones of the first class.

(g.) Nature of the Interior Reservoir from which Lavas are derived.—We have now to consider the nature of the interior reservoir from which lavas are derived, and the physical conditions necessary for their eruption at the surface.

Without going back to the question of the original condition of our globe, we may safely hold the view that at a very early period of geological history it consisted of a solidified crust at a high temperature, enfolding a globe of molten matter at a still higher temperature. As time went on, and the heat radiated into space from the surface of the globe, while at the same time slowly ascending from the interior by conduction, the crust necessarily contracted, and pressing more and more on the interior molten magma, this latter was forced from time to time to break through the contracting crust along zones of weakness or fissures.

(h.) The Earth's Crust in a State of both Exterior Thrust and of Interior Tension.—As has been shown by Hopkins,[4] and more recently by Mr. Davison,[5] an exterior crust in such a condition must eventually result in being under a state of horizontal thrust towards the exterior and of tension towards the interior surface. For the exterior portion, having cooled down, and consequently contracted to its normal state, will remain rigid up to a certain point of resistance; but the interior portion still continuing to contract, owing to the conduction of the heat towards the exterior, would tend to enter upon a condition of tension, as becoming too small for the interior molten magma; and such a state of tension would tend to produce rupture of the interior part. In this manner fissures would be formed into which the molten matter would enter; and if the fissures happened to extend to the surface, owing to weakness of the crust or flexuring of the strata, or other cause, the molten matter would be extruded either in the form of dykes or volcanic vents. In this way we may account for the numerous dykes of trap by which some volcanic districts are intersected, such as those of the north of Ireland and centre of Scotland.

From the above considerations, it follows that the earth's crust must be in a condition both of pressure (or lateral thrust) towards the exterior portion, and of tension towards the interior, the former condition resulting in faulting and flexuring of the rocks, the latter in the formation of open fissures, through which lava can ascend under high pressure. These operations are of course the attempt of the natural forces to arrive at a condition of equilibrium, which is never attained because the processes are never completed; in other words, radiation and convection of heat are constantly proceeding, giving rise to new forces of thrust and tension.

It now remains for us to consider what may be the condition of the interior molten magma; and in doing so we must be guided to a large extent by considerations regarding the nature of the extruded matter at the surface.

(i.) Relative Densities of Lavas.—Now, observation shows that, as bearing on the subject under consideration, lavas occur mainly under two classes as regards their density. The most dense (or basic) are those in which silica is deficient, but iron is abundant; the least dense (or acid) are those which are rich in silica, but in which iron occurs in small quantity. This division corresponds with that proposed by Bunsen and Durocher[6] for volcanic rocks, upon the results of analyses of a large number of specimens from various districts. Rocks may be thus arranged in groups:

(1) The Basic (Heavier)—poor in silica, rich in iron; containing silica 45-58 per cent. Examples: Basalt, Dolerite, Hornblende rock, Diorite, Diabase, Gabbro, Melaphyre, and Leucite lava.

(2) The Acid (Lighter)—rich in silica, poor in iron; containing silica 62-78 per cent. Examples: Trachyte, Rhyolite, Obsidian, Domite, Felsite, Quartz-porphyry, Granite.

The Andesite group forms a connecting link between the highly acid and the basic groups, and there are many varieties of the above which it is not necessary to enumerate. Durocher supposes that the molten magmas of these various rocks are arranged in concentric shells within the solid crust in order of their respective densities, those of the lighter density, namely the acid magmas, being outside those of greater density, namely the basic; and this is a view which seems not improbable from a consideration not only of the principle itself, but of the succession of the varieties of lava in many districts. Thus we find that acid lavas have been generally extruded first, and basic afterwards—as in the cases of Western America, of Antrim, the Rhine and Central France. And if the interior of our globe had been in a condition of equilibrium from the time of the consolidation of the crust to the present, reason would induce us to conclude that the lavas would ultimately have arranged themselves in accordance with the conditions of density beneath that crust. But the state of equilibrium has been constantly disturbed. Every fresh outburst of volcanic force, and every fresh extrusion of lava, tends to disturb it, and to alter the relations of the interior viscous or molten magmas. Owing to this it happens, as we may suppose, that the order of eruption according to density is sometimes broken, and we find such rocks as granophyre (a variety of andesite) breaking through the plateau-basalts of Mull and Skye, as explained in a former chapter. Notwithstanding such variations, however, the view of Durocher may be considered as the most reasonable we can arrive at on a subject which is confessedly highly conjectural.

(j.) Conclusion as regards the Ultimate Cause of Volcanic Action.—Notwithstanding, however, the complexity of the subject, and the uncertainties which must attend an inquiry where some of the data are outside the range of our observation, sufficient evidence can be adduced to enable us to arrive at a tolerably clear view of the ultimate cause of volcanic action. So tempting a subject was sure to evoke numerous essays, some of great ingenuity, such as that of Mr. Mallet; others of great complexity, such as that of Dr. Daubeny. But more recent consideration and wider observation have tended to lead us to the conclusion that the ultimate cause is the most simple, the most powerful, and the most general which can be suggested; namely, the contraction of the crust due to secular cooling of the more deeply seated parts by conduction and radiation of heat into space. Owing to this cause, the enclosed molten matter is more or less abundantly extruded from time to time along the lines and vents of eruption, so as to accommodate itself to the ever-contracting crust. Nor can we doubt that this process has been going on from the very earliest period of the earth's history, and formerly at a greater rate than at present. When the crust was more highly heated, the radiation and conduction must have been proportionately more rapid. Owing to this cause also the contraction of the crust was accelerated. To such irresistible force we owe the wonderful flexuring, folding, and horizontal overthrusting which the rocks have undergone in some portions of the globe—such as in the Alps, the Highlands of Scotland and of Ireland, and the Alleghannies of America. It is easy to show that the acceleration of the earth's rotation must be a consequence of such contraction; but, after all, this is but one of those compensatory forces of which we see several examples in the world around us. It can also be confidently inferred that at an early period of the earth's history, when the moon was nearer to our planet than at present, the tides were far more powerful, and their effect in retarding the earth's rotation was consequently greater. During this period the acceleration due to contraction was also greater; and the two forces probably very nearly balanced each other. Both these forces (those of acceleration and retardation) have been growing weaker down to the present day, though there appears to have been a slight advantage on the side of the retarding force.[7]