Geology, Vol. 1 [of 3]

by Thomas C. Chamberlin; Rollin D. Salisbury

THE CAUSE OF VULCANISM.

The extraordinary facts involved in volcanic phenomena cannot well be discussed fully until the origin of the earth is considered, and the great agencies, as well as the peculiar conditions, which the earth inherited from its birth, are duly weighed, for these were, with little doubt, the true causal antecedents of vulcanism. We shall return to the subject after a sketch of the early conditions of the earth, but the views that have been entertained may be reviewed here while the phenomena are fresh in mind.

The explanation of vulcanism involves two essential elements. These are (1) the origin of the lavas, which involves a consideration of the necessary temperatures, pressures, and other conditions, and (2) the forces by which the lavas are expelled.

Nearly all current explanations of vulcanism are founded upon conditions supposed to be derived from a molten globe, and fall under two general classes: (I) those which assume that the lavas are residual portions of the original molten mass, and (II) those which assign the lavas to the local melting of rock.

I. On the Assumption that the Lavas are Original.

In this case it is not necessary to assume any special accession of heat, but merely to account for extrusion. There are two phases of this view, (1) the one postulating a general molten interior, (2) the other limiting the molten matter to local reservoirs.

Hypothesis I. Lava outflows from a molten interior.—In the early days of geology, when the earth was supposed to have a thin crust and a molten interior, it was very naturally assumed that volcanoes were but pipes leading down to the molten mass within. This view has been essentially abandoned. The independence of adjacent vents is in itself almost a fatal objection, when it is recalled that the height of recent volcanic craters ranges from nearly 20,000 feet above the sea, to 10,000 to 20,000 feet below. The view would involve the conception of lava-columns connected with a common reservoir varying possibly 30,000 to 40,000 feet in altitude, and certainly more than half that much, simultaneously. The lower outlets should as certainly be selected for the outflow of the great interior sea of fluid rock, as the lowest sag in the rim of a lake for its outflow, for no great differences in specific gravity are presumable under this hypothesis. An equally grave objection arises from tidal strain. If the earth were liquid within and merely crusted over by a shell of rock of moderate thickness, it would yield appreciably to tidal stresses, and this yielding would change the capacity of the interior so that with every distortion of the spheroid a portion of its fluid interior would be forced to the outside, and with every return to the more spheroidal form there would either be a re-flow to the interior or a shrinking of the crust. In any case a very demonstrative response to tidal influence would tell the story of interior fluidity. No such effects are observed. The tidal strains may perhaps have a slight effect in hastening a given eruption when the forces are approaching a delicate balance and an eruption is imminent, but the very triviality of this influence implies not only the absence of a general liquid interior, but also of extensive reservoirs.

Hypothesis 2. Lavas assigned to molten reservoirs.—A modification of the preceding view has been made to escape the difficulties involved in the hypothesis as stated above. It is supposed that while nearly all the subcrust solidified, numerous liquid spots were left scattered through it. This honeycombed substratum is supposed to connect the continuous outer crust with a central solid body, solidified because of pressure in spite of its high temperature. This hypothesis escapes only a portion of the objections. For instance, the lavas in Mauna Loa and Kilauea in Hawaii differ nearly 10,000 feet in height, and hence cannot well be supposed to connect with the same reservoir, but they are both on the same vast cone, which implies at least an equally large molten reservoir as its source. If there were two distinct reservoirs of the required magnitude, they must be singularly placed to supply vents so near and yet so independent. The difficulty grows greater when the whole Hawaiian chain is considered, for the points of eruption seem to have migrated from the northwesterly islands, where the volcanoes are old, to the southeastern end, where volcanic activity is now in progress.

It would be natural under this view to suppose that these residuary lakelets of liquid rock should be gradually exhausted as time goes on, and that vulcanism should be a declining phenomenon. It is not clear that this is the case. The great number of existing volcanoes in regions where great extrusions took place in earlier ages does not seem to be in harmony with the hypothesis.

II. On the Assumption that the Lavas are Secondary.

The serious difficulties that arise in interpreting volcanic lavas as remnants of an original molten mass, and the strong arguments of recent years for a very solid earth, have turned inquiry chiefly toward the second class of hypotheses, which refer the origin of lavas to the local melting of deep-seated rock. These differ widely among themselves. One group seeks for a cause of the melting in the penetration of surface air and water; another, in the relief of pressure; a third, in crushing and shearing; a fourth, in the depression of sediments into the heated interior zone; and a fifth, in the outward flow of deep-seated heat.

Hypothesis 3. Lavas assigned to the reaction of water and air penetrating to hot rocks.—As steam is one of the great factors in the explosions of volcanoes, and as water reduces the melting-point of rocks, it is a natural and simple view that water penetrating through the fissures and pores of the outer crust and coming into contact with the heated rocks below, is absorbed into them and renders them liquid, and that then, being rendered swollen and lighter by the process, they ascend and discharge quietly or explosively according to the special conditions of the case. Naturally the suggestion arises that the waters would be converted into steam long before they could reach rock hot enough to be melted, and that this steam would be forced back along its own track, as the line of least resistance, rather than force itself into the rock material and rise in the lava-column; but to this it is answered that an experiment of Daubree’s has shown that water will penetrate the capillaries of sandstone against high steam pressure and add itself to the steam within. The fact is also cited that certain substances, when highly heated, absorb gases which they give out when they cool. The absorption of hydrogen by platinum, and of oxygen by molten silver, are illustrations. It is certain that the lavas do contain large quantities of absorbed gases, and that these are partly, and in most cases largely, given out in cooling, when the cooling takes place at the surface. The presumption is that the lavas would take the gases up again on remelting under similar conditions. If the lavas of actual volcanoes had the temperatures of aqueo-igneous fusion (700°–1000° Fahr.) only, it would strengthen this view; but as temperatures of lavas often exceed 2000° Fahr., and probably sometimes reach 2500° Fahr., and perhaps 3000° Fahr., it is not easy to account for such temperatures under this hypothesis, because they would only be reached at levels far below those at which aqueo-igneous fusion might be presumed to take place. Perhaps this could be met by invoking pressure which might prevent even aqueo-igneous fusion from taking place until these temperatures were reached, but pressure brings in a grave difficulty in another line, as we shall presently see.

There is a phase of the water-penetration hypothesis which seeks to account for an accession of heat. It is affirmed that the outer rocks are oxidized, while the inner ones were not originally, or at least not completely oxidized, and that air and water from the surface, reaching the unoxidized zone, enter into combination and generate the necessary heat. This view was pardonable before the development of modern thermo-chemistry, but is now quite untenable, as may be shown by working out the reactions thermally.

All views which assign the penetration of surface air or water as a cause meet with a grave, if not insuperable, difficulty in the condition of the lower part of the earth’s crust (see [p. 218]). The fractured condition of the crust, which permits a ready penetration of water, is a very superficial phenomenon. Below the first few thousand feet the crevices and porosities of the rock are rapidly closed by the pressure of overlying rock, and all appreciable crevices and pores probably disappear at a depth of five or six miles. The effective function of fissures is, therefore, limited to the upper few miles of the crust, and even here to certain portions only. The great pressures in gas- and oil-wells show that in many quite superficial beds, even when arched, there are no fissures or pores capable of letting even gas escape effectively. The depths at which the temperatures of lavas are reached are usually estimated, from the downward increase of temperature, at 20 to 30 miles. This leaves from 14 to 24 miles of the compressed zone between the lowest assignable limit of the fissured zone, and the highest assignable zone for the origin of lavas. This thick zone of dense rock must be reckoned with in all hypotheses that involve the penetration of air and water from without, and, as well, the extrusion of lavas from within. In addition to the difficulties of the penetration of ground-water, the limitations of its heat, at penetrable depths, also bear adversely (see [p. 219]), on the descent of air and water.

Hypothesis 4. Lavas assigned to relief of pressure.—It seems to be demonstrated that pressure raises the melting-point of average rock, and hence at twenty or thirty miles’ depth there may be rocks hot enough to melt at the surface, but still solid because of high pressure. If this pressure were in some way relieved they would become liquid. Pressure may be locally relieved somewhat (1) by denudation, (2) by certain phases of faulting, (3) by anticlinal arches, and (4) by continental deformation.

(1) In most cases of denudation, cooling below probably keeps pace with loss above. At any rate, many volcanoes rise from the bottoms of the oceans where no denudation takes place, and this phase of the hypothesis is not workable there.

(2) The theory of relief by faulting finds encouragement in the fact that many volcanoes occur on fault-lines. There is no evidence, however, that this is a universal or necessary relation. Computation as to the amount of lowering of the melting-point that might arise from the faulting associated with volcanoes indicates that it is necessary to suppose that the rocks were already very close to the melting-point when the faulting took place, to make the doctrine applicable. It is to be observed that in faulting the relief of pressure on one side of the fault-line is likely to be balanced by increased pressure on the other side, and that this difference in pressure may be lost by distribution at a depth of 20 or 30 miles, where, at the nearest, this delicate relation between solidity and liquidity, on which the theory is dependent, may perhaps be reached.

(3) Immediately under an anticlinal arch there may doubtless be some relief of pressure within the limits of strength of the arch, which is not great ([p. 582]). The pressure under the arch as a whole is greater than before it was bowed up by lateral thrust, and in depth this excess becomes distributed so as to obliterate the local relief under the center of the arch, and so adds the effects of folding to the average pressure of the crust. Besides, as a matter of fact, volcanoes do not appear to be especially associated with mountain folds where arching reaches its best expression.

(4) The same general considerations bear on the assignment of liquefaction to relief of pressure in connection with the more general deformation of the earth’s body. Besides, while relief of pressure might account for liquefaction, it leaves the extrusion without an obvious cause; indeed, it would seem to furnish a condition opposed to extrusion, and if pressure were subsequently added to force the liquid out, it would tend to restore the solid condition.

Hypothesis 5. Lavas assigned to melting by crushing.—Mallet[286] and others have attributed melting to the crushing of rock. Crushing, in the ordinary sense of the term, can only take place in the zone of fracture, and that is apparently too shallow to meet the requirements of the case. Below this zone, the pressure on all sides is too great to permit any separation of fragments, and a solid mass can only change its form by what is called “solid flowage.” The rock under these conditions may be compressed, and this compression must give rise to heat, but at the same time the melting-point is raised, according to all experiments. It seems improbable that melting can be produced in this way. If great pressure could be brought to bear upon a tract of rock so as to heat it by compression, and if then the pressure were relaxed before the heat generated could be distributed by conduction, and if re-expansion did not follow, possibly melting could be effected, but this makes the process complicated and apparently inapplicable. It is scarcely possible that such a sequence of events can have affected all the tracts that are now volcanic, much less all those that have been such throughout geologic history. As noted in the preceding case, relaxation would seem to be unfavorable to expulsion. Besides, volcanoes do not seem to be confined to tracts that show signs of great crumpling and crushing, as the Alps, the Appalachians, and the closely folded ranges generally. Extrusions seem rather more common with faulted ranges where crushing is less notable and where surface tension replaces compression.

Hypothesis 6. Lavas assigned to melting by depression.—It is observed that in certain regions great thicknesses of sediments have accumulated by the slow settling of the crust below, and as these sediments obstruct the outward flow of heat while the lower beds settle nearer to the interior source of heat, it is conceived that they become heated below and, being saturated with water, take on aqueo-igneous fusion and rise as lavas, well supplied with internal gas and steam from the water and volatile constituents that were entrapped and carried down with them. The question obviously arises whether such depression is sufficient to give the temperatures the lavas show, and whether volcanic action is confined to such areas of depression and deep sedimentation. At the highest credible estimates—which are none the less to be taken with reserve—the post-Archean sediments rarely reach five or six miles in thickness at any given point, and probably never exceed ten or twelve, while twenty or thirty miles is the computed depth required for the acquisition of the temperatures the lavas actually possess. If the Archean terranes be included among the sedimentaries, the thickness may be adequate, but what then of the Archean vulcanism, which much surpasses that of later times, and the other early extrusions before the sediments were thick; and what of the moon, where there are probably no sediments at all?

Besides, it is not at all clear that the distribution of vulcanism is specially related to that of thick sediments, as it should be if this hypothesis were the true one. There are many volcanoes in the heart of the great oceans where sedimentation is now inappreciable, and probably has been in all past periods.

Hypothesis 7. Vulcanism assigned to the outflow of deep-seated heat.—If the earth grew up by slow accessions of meteoroidal or planetesimal matter, in a manner to be more fully set forth in the discussion of the origin of the earth, and if its interior heat be due chiefly to compression by its own gravity, the internal temperature would be originally distributed according to the degree of compression, and this would depend on the intensity of the internal pressure. This can be approximately computed, and is shown in the diagram on page 563, where this subject has been treated. On not improbable assumptions regarding the thermometric conductivity, the flow of heat from the deep interior to the middle zone would be greater than the loss of this zone to the superficial zone. This middle zone should, under this view, experience a rising temperature. By hypothesis, this zone is composed of various kinds of matter mixed as they happened to fall in. Hence as the temperature rises, the fusion-points of some of these constituents will be reached before those of others. More strictly, the temperatures at which some of these constituents will mutually dissolve one another will be reached, while other constituents remain undissolved, and thus a partial and distributed liquefaction will arise. The gases and volatile constituents in the mixed material would naturally enter largely into the liquefied portion. It is assumed that with a continued rise of temperature, the partial liquefactions would increase until the liquefied parts found means of uniting, and the lighter portions, embracing the gaseous contingent, were able to work their way toward the surface. As they rose by fusing or fluxing their way, the pressure upon them became less and less, and hence the temperature necessary for liquefaction gradually fell, leaving them a constantly renewed margin of temperature available for melting their way through the upper horizons. Thus it is conceived that these fusible and fluxing selections from the middle zone might thread their ways up to the zone of fracture and thence, taking advantage of fissures and fractures, reach the surface. It is conceived that such liquefaction and extrusion would carry out from the middle zone the excess of temperature received from the deeper interior, and thus regulate its temperature and forestall general liquefaction, the zone as a whole remaining always solid. The independence of volcanoes is assigned to the independence of the liquid threads that worked their way to the surface. Nothing like a reservoir or molten lake enters into the conception. The prolonged action of volcanoes is attributed to the slow feeding of the liquid threads from the locally fused middle zone. The frequent pauses in action are assigned to temporary deficiencies of supply; the renewals to the gathering of new supplies after a sufficient period of accumulation. The distribution of volcanoes in essentially all latitudes and longitudes is assigned to the general nature of the cause. The special surface distributions are assumed to be influenced, though not altogether controlled, by the favorable or unfavorable conditions for escape presented by the crustal segments of the earth. The persistence of volcanic action in time is attributed to the magnitude of the interior source, to its deep-seated location, and to the slowness of conduction of heat in the earth’s interior. The force of expulsion is found in the stress-differences in the interior, particularly the periodic tidal and other astronomic stresses (see p. 580), and in the slow pressure brought to bear on the slender threads of liquid by the creep of the adjacent rock. The violent expulsions are due to the included gases, of which steam is chief. Little efficiency is assigned to surface-waters, and that little is regarded as wholly secondary and incidental. The true volcanic gases are regarded as coming from the deep interior and as being true accessions to the atmosphere and hydrosphere. The standing of the lavas in volcanic ducts for hundreds and even thousands of years with only small outflows, as in some of the best-known volcanoes, is regarded as an exhibition of an approximate equilibrium between the hydrostatic pressure of the deep-penetrating column of lava, and the flowage-tendency of the rock-walls, the outflow being, of course, also conditioned on the slow rate of supply below, and the periodic stress-differences of the interior.

For the present these hypotheses must be left to work out their own destiny, serving in the mean time as stimulants of research. All but the last have been for some time under the consideration of geologists, and are set forth in the literature of the subject ([p. 636]).

A few special phases of the problem need further discussion, though they have been incidentally touched upon.

Modes of Reaching the Surface.

All of the views that locate the origin of the lavas deep in the earth must face the difficulties of the passage through the dense portion of the sphere below the fracture zone. Near the surface, the lavas usually take advantage of fissures or bedding-planes already existing or made by themselves. There is little evidence that they bore their way by melting, though they round out their ducts into pipes as they use them, much as streamlets on glaciers falling into crevices round out moulins. But this use of fissures and bedding-planes for passage is probably merely a matter of least resistance where the lavas are relatively cool, and their capacity for melting is low or perhaps even gone. Daly has recently urged that lavas work out reservoirs and enlarge passageways for themselves by detaching masses of rock from the roofs and sides of the spaces already occupied by them, these masses either melting and mingling with the lava, or else sinking to lower positions in the column. This process he designates stoping.[287]

In the denser and warmer zone below, the alternatives seem to be (1) melting or fluxing, or (2) mechanical penetration without fracture. As rocks “flow” in this zone by differential pressure without rupture, an included liquid mass may be forced to flow through the zone by sufficient differential pressure. If local differential pressures at the surface be neglected as probably incompetent, there only remain the stress-differences of the interior and the differences of hydrostatic pressure between the lava-column and the surrounding solid columns. The latter would not be great until a column of liquid of much depth was formed, and the former would probably not be concentrated on the liquid in such a way as to force it bodily through the solid rock. Probably fusing or fluxing its way with the aid of stress-differences is the chief resource of the lava in the initial stages. In this it may be supposed to be assisted by its gases, by its selective fusible and fluxing nature, by its very high temperature if it comes from very great depths, as held in the seventh hypothesis, and by the stress-differences which prevail in the deep interior, as shown in the last chapter. In ascending from lower to higher horizons, the lava would be constantly invading regions of lower melting-point because of less pressure. It would thus always have an excess of heat above the local melting temperature until it invaded the external, cool zone, where the regional temperature is below the melting-point of surface pressure. From that point on it must constantly lose portions of its excess of temperature by contact with cooler rocks, and probably in the process of fluxing its way in the compact zone. If this excess is insufficient to enable it to reach the zone of fracture, the ascending column is arrested and becomes merely a plutonic pipe or mass. If it suffices to reach the zone of fracture, advantage may be taken thereafter of fissures and of rupturing, and the problem of further ascent probably becomes chiefly one of hydrostatic pressure, in which the ascent of the lava-column is favored by its high temperature and its included gases. The hydrostatic contest is here between the lava-column, measured to its extreme base, and the adjacent rock-columns, measured to the same extreme depth. The result is, therefore, not necessarily dependent on the flowage of the outer rocks, but may be essentially or wholly dependent on the deep-seated flowage of the rock of the lower horizons. The ascending column may reach hydrostatic equilibrium before it reaches the surface, and may then form underground intrusions of various sorts without superficial eruption, or it may only find equilibrium by coming to the surface and pouring out a portion of its substance and discharging its gases.

Additional Considerations Relative to the Gases.

The question whether the volcanic gases are a contribution to the atmosphere and hydrosphere is so important in its bearings on the whole history of the atmosphere as to merit additional consideration here. As already noted, if the volcanic gases arise from water and absorbed air that have previously passed down through the strata, there is no real contribution to the hydrosphere and atmosphere, but merely circulation. If the gases are chiefly derived from the deep interior, they are an important accession to the atmosphere and hydrosphere.

Most views are more or less intermediate, assigning a part of the gases to the interior and a part to the exterior. No one will question that some part at least of the steam is due to the contact between the ground-waters and the hot lava, and probably no one will question that some gas comes from the interior if the lavas originate there. The vital question is, whence comes the major portion? Are the constant ebullitions of some volcanoes and the terrific explosions of others due mainly to surface-waters, or to interior gases?

It seems to be certain that in most cases the gases are diffused through the substance of the lava, and are not simply in contact with the walls of the column or with its summit. Without doubt steam is generated around the lava-column by external contact, and perhaps some explosions are due to the entrance of the rising lava upon a crevice or cavern filled with water, or to the invasion of a lake gathered in an old crater; but it still remains a question whether the importance of such explosions has not been exaggerated. Such action does not seem competent to produce inflated lavas, but merely shattered ones. Water thus “suddenly flashed into steam” could scarcely diffuse itself intimately through the lava, for the process of diffusion is exceedingly slow. But inflated lavas, pumice, scoriæ, and cinders are the typical products of explosive vulcanism. Not only in the ordinary Vesuvian type, but in the extraordinary Krakatoan type, inflated lavas are the dominant product. Prodigious quantities of this covered the sea about Krakatoa after its tremendous explosion in 1883. Judd estimates that the volume of included steam involved in the inflation of the pumice examined by him, was from three to five times that of the rock, and that the amount involved in exploding the lava into the fine dust that floated in the upper atmosphere for months, was presumably much greater.

If the sudden flashing into steam of bodies of water in external contact with hot lava be rejected as only an incidental source of explosion, it remains to be considered whether waters permeating the rock and becoming converted into steam may not be absorbed into the rising lava, become diffused through it, and ascending with it, explode at the surface. So far as access through fissures and cavities large enough to be entered by lava are concerned, it may safely be concluded that since the hydrostatic pressure of the lava must be greater than that of the water in the fissures, or else it could not rise, the lava will enter them, forcing back the water or the steam generated from it, and, having penetrated as far as accessible, will solidify as a dike, and plug up the avenue of contact between the ground-water and the portion of the lava still remaining molten. The numerous dikes that attend volcanic necks testify to the prevalence of this action. The capillary pores of the wall-rock, which cannot be thus bodily occupied by the lava, must doubtless become filled with steam, and this, following the principles of Daubree’s experiment, will force itself into contact with the lava and be absorbed by it, but whether this will be in sufficient quantity, and will become sufficiently diffused through the body of the lava-column to produce the observed effects, is an open question. The increasing testimony of deep mining is that relatively little water flows through the deeper horizons. It is urged that the water remaining in solidified lavas is very unequal in distribution, as though due to unequal access and partial diffusion. The argument seems strong, but to make it thoroughly good, it must be shown that this inequality is not due to irregularity of discharge of the gases during and after eruption, rather than to irregularity of original accession. There is, perhaps, as much ground for assigning differences in the degree of parting with the included gases, as in acquiring them. Doubtless those lavas that boiled and seethed for a long period in the caldron were more fully deprived of their gases than those that were more promptly disgorged and cooled with less convection and surface exposure.

Thermal considerations.—Probably the most important consideration relates to the heat effects. If underground-waters enter the lava-column and come forth as steam, great quantities of heat are consumed in the process. Has the lava a sufficient excess of heat to stand this? Can ebullition be maintained for the observed periods if the steam comes from ground-waters?

Many lavas probably do not carry a very large excess of heat above that necessary for liquefaction, for not a few of them contain crystals already forming, which shows that they are within the range of the temperatures of solidification of their constituents. The same conclusion is indicated by the limited fusing effects shown by the walls of dikes and sills. On the other hand, as already remarked, dikes and sills often show the effects of a rather rapid cooling from the walls. The method of flow often implies the same condition for the acidic lavas, since they usually behave as stiff, pasty masses of limited liquidity. On the other hand, the basic lavas, whose fusion-point is much lower, often flow freely and reach great distances before solidifying. The facts taken altogether imply that the average temperature of the lavas is not much above the fusing-point of the acidic lavas, while it may probably be very considerably above the fusing-point of basalt. For a rough estimate, with no pretensions to accuracy, it may be assumed that in an average case there are 500° Fahr. excess, but probably not 1000° Fahr. A computation based on even so rough an estimate as this may, by showing the order of magnitude of the thermal considerations, indicate their radical bearing. The average temperature of the ground-water of the upper two or three miles of the crust—the only portion through which water probably penetrates with sufficient freedom to be effective in this case—is probably less than 200° Fahr. The specific heat of rock appears to average somewhat less than 0.2. The temperature of the lava may be taken at 3000° Fahr. as a sufficiently high average. From these data it follows that if an amount of ground-water equal to five percent. of the volume of the lava entered the lava and was brought up to its temperature and then discharged, the temperature of the whole mass would be lowered 550° Fahr. It is therefore evident that only a small percentage of surface-water can pass through the lava consistently with its continued fluidity.

M. Fouqué estimated that the discharge of steam from a merely parasitic cone of Etna during 100 days was equal to 2,100,000 cubic meters of water. If this were ground-water, and the lava from which it issued had an excess of 500° Fahr. above the fusion-point, the formation of this steam would congeal a column 400 feet in diameter and 3000 feet deep in the time given. If this case is typical, and if Fouqué’s estimate is not greatly exaggerated or very exceptional, the view that any large portion of the steam from volcanoes comes from surface-waters seems to be incompatible with the persistence of ebullition and explosion which many of them exhibit. Stromboli has been in constant eruption as far back as the history of the region runs. It is now exploding every three to ten minutes, and yet the mass of lava seems to be small and its outflow inconsiderable. Is it possible that a current of steam, given out with this activity for so long a period, was derived from adjacent ground-waters, and has not yet solidified the lava?

The problem takes on a very different aspect if the steam, or at least some large part of it, rises from great depths and brings thence an excess of heat. It then becomes an agency for the maintenance of the liquidity of the lava, for giving it convective motion, and for promoting explosive action, so long as it continues to rise.

For these and other reasons the balance of present evidence seems to us to favor the view that most of the steam and other gases come with the lava from its original source deep in the earth.

References on vulcanism.—G. P. Scrope, Volcanoes, London, 1872. R. Mallet, on Volcanic Energy, Phil. Trans., 1873. C. Darwin, Geological Observations on Volcanic Islands, London, 1876. E. Reyer, Beitrag zur Physik der Eruptionen, Vienna, 1877; Theoretische Geologie, 1888. Fouqué, Santorin et ses Éruptions, Paris, 1879, Sartoris von Waltershausen and A. von Lasaulx, Der Aetna, Leipzig, 1880. C. E. Dutton, Geology of the High Plateaus of Utah, U. S. Geog. and Geol. Surv., 1880; The Hawaiian Volcanoes, Fourth Ann. Rept. U. S. Geol. Surv., 1883. Judd, Volcanoes, 1881; The Eruption of Krakatoa (Com. of the Roy. Soc.), 1888. J. D. Dana, Characteristics of Volcanoes, 1890. H. J. Johnston-Lavis, The South Italian Volcanoes, Naples, 1891. E. Hull, Volcanoes, Past and Present, 1892. Milne and Burton, The Volcanoes of Japan, 1892. J. P. Iddings, The Origin of Igneous Rocks, Bull. Phil. Soc., Washington, Vol. XII, 1892. A. C. Lane, Geologic Activity of the Earth’s Originally Absorbed Gases, Bull. Geol. Soc. Am., Vol. V, 1894. A. Geikie, Ancient Volcanoes of Great Britain, London, 1897. I. C. Russell, Volcanoes of North America, 1897. T. G. Bonney, Volcanoes, Their Structure and Significance, New York (and London), 1899. F. Miron, Étude des Phénomènes Volcaniques, Paris, 1903. G. C. Curtis, Secondary Phenomena of the West Indian Volcanic Eruptions of 1902, Jour. Geol., Vol. XI, No. 2, 1903. A. Heilprin, Mont Pelée and the Tragedy of Martinique, Philadelphia (and London), 1903. Robert T. Hill, Report on the Volcanic Disturbances in the West Indies, Nat’l Geog. Mag., Vol. XIII, No. 7, 1902. I. C. Russell, The Recent Volcanic Eruptions in the West Indies, Nat’l Geog. Mag., Vol. XIII, No. 7, 1902; Volcanic Eruptions on Martinique and St. Vincent, Nat’l Geog. Mag., Vol. XIII, No. 12, 1902. J. S. Diller, Volcanic Rocks of Martinique and St. Vincent, Nat’l Geog. Mag., Vol. XIII, No. 7, 1902. W. F. Hillebrand, Chemical Discussion of Analyses of Volcanic Ejecta from Martinique and St. Vincent, Nat’l Geog. Mag., Vol. XIII, No. 7, 1902. E. O. Hovey, The Eruptions of La Soufrière, St. Vincent, in May, 1902, Nat’l Geog. Mag., Vol. XIII, No. 12, 1902.