Modern views respecting the cause of earthquakes—Earthquakes due to faulting—To explosions of steam—To volcanic evisceration—To chemical degradation—Attractive influence of the heavenly bodies—The effect of oceanic tides—Variation in atmospheric pressure—Fluctuation in temperature—Winds and earthquakes—Rain and earthquakes—Conclusion.

As the results of modern inquiries respecting the cause of earthquakes, we see many investigators chiefly attributing these phenomena to special causes. A few attribute them to several causes. It seems to us that they might be attributed to very many causes which often act in a complex manner. The primary causes are telluric heats, solar heat, and variations in gravitating influences. These may be the principal, and sometimes the immediate, cause of an earthquake. The secondary causes are those dependent upon the primary causes, such as expansions and contractions of the earth’s crust, variations in temperature, barometrical pressure, rain, wind, the attractive influences of the sun and moon in producing tides in the ocean or the earth’s crust, variations in the distribution of stress upon the earth’s surface caused by processes of degradation, the alterations in the position of isogeothermal surfaces, &c.

The part which may be played by these various causes in the production of oscillations, pulsations, and tremors will be referred to.

Earthquakes consequent on faulting.—In the chapter on Earth Oscillations, the causes producing the phenomena of elevation and depression are briefly indicated.

By the variations in stress accompanying elevations and depressions, cracks are produced. Inasmuch as compression would crush the rocks constituting the earth’s crust, we must conclude with Captain Dutton that these cracks are formed by tension. By elevation, the upper rigid crust of the earth is stretched, and fissures are produced. The sudden formation of these fissures or faults gives rise to earthquakes, and perhaps also to volcanic vents. That earthquake and volcanic regions are situated on areas where there is evidence of rapid elevation is strikingly illustrated round the shores of the Pacific.

Lasaulx considered that the earthquake of Herzogenrath was more or less intimately connected with the great mountain fissure—the Feldbiss—which crosses the coal region of the Wurm.[126] The sudden elevation or sinking of large areas at the time of an earthquake may be a consequence of these dislocations.

It has already been pointed out that the earthquake region of Japan is the one where we have evidence of recent and rapid elevation. That certain earthquakes of this region may possibly be the result of faulting we have the evidence of our senses and of our instruments. The sudden blows and jolts which are sometimes felt are indicative of the sliding of one mass of rock across another.

Should the ground be simply torn asunder, this tearing would give rise to a series of waves of distortion, vibrating in directions parallel to the plane of the fissure. Supposing this motion to be propagated to a number of surrounding stations, it would be recorded at each of these as having the same direction. To those situated on a line forming a continuation of the strike of the fissure, the vibrations would advance so to speak end on, whilst to those stations lying in a line perpendicular to the strike of the fissure, the motion would advance broadside on.

Motions like these latter have been recorded in Tokio, where earthquakes which from time observations were known to have come from the faulted and rising region to the south have been registered as a series of east and west motions, or vibrations transverse to this line of propagation.

It must, however, be here mentioned that the registration of only transverse motion may possibly be due to the extinction of normal motion, although this is not generally regarded as probable.

It would therefore appear that certain earthquakes and faults are closely related phenomena, the former being an immediate effect of the latter. Faults are due to earth oscillations, and to a variety of causes producing disturbances in the equilibrium of the earth’s crust; the principal cause of all these phenomena being alterations in the distribution of heat, and the attractive force of gravity.

Earthquakes consequent on the explosion of steam.—Humboldt regarded volcanoes and earthquakes as the results of a common cause, which he formulated as ‘the reaction of the fiery interior of the earth upon its rigid crust.’ Certain investigators, who have endeavoured to reduce Humboldt’s explanation to definite limits, have suggested that earthquakes may be due to sudden outbursts of steam beneath the crust of the earth, and its final escape through cracks and fissures.

Admitting that steam may accumulate by separating out from the cooling interior of our globe, its sudden explosion might be brought about by its own expansive force, or by the movements in the bubbling mass from which it originated.

Others, however, rather than regard the steam as being a primeval constituent of the earth’s interior, imagine it arises from the gradual percolation of water from the surface of the earth down to volcanic foci, into which it is admitted against opposing pressures, by virtue of capillary action.

Mallet, in his account of the Neapolitan earthquake, shows that the whole of the observed phenomena can be accounted for by the admission of steam into a fissure, which by the expansive force exerted on its walls was rent open. Just as at the Geysers we hear the thud and feel the trembling produced by the sudden evolution and condensation of steam, so may steam by its sudden evolution and condensation in the ground beneath us give rise to a series of shocks of varying intensity, accompanied by intermediate vibratory motions—that is to say, a motion which, as judged of by our feelings, is not unlike many earthquakes. Often it may happen that the result of the explosion may be the production of a fault, or at least a fissure; and thus in the resulting movements we may have a variety of vibrations, some being those of compression and distortion, produced by the blow of the explosion, and others being those of distortion alone, produced by the shearing action which may have taken place by the opening of the fault. Sometimes one set of these vibrations may be prominent, and sometimes the other. Thus, when we say that an earthquake has shown evidence by the nature of its vibrations that it was produced by a fault, this by no means precludes the possibility that an explosion of steam may also have been connected with the production of the disturbance. Mallet threw out the suggestion that the opening of fissures beneath the ocean might admit water to volcanic foci. During the time that the water was in the spheroidal state, the preliminary tremors, so common to many earthquakes, would be produced. These would be followed by the explosion, or series of explosions, constituting the shock or shocks of the earthquakes.

The chief reasons for believing that the earthquakes of North-Eastern Japan are partly due to explosive efforts are:—

1. That the greater number of disturbances, perhaps ninety per cent., originate beneath the sea, where we may imagine that the ground, under the superincumbent hydrostatic pressure, is continuously being saturated with moisture.

2. Many of the diagrams show that the prominent vibrations, of which there are usually from one to three, in a given disturbance have the same character as those produced by an explosive like dynamite, the greatest and probably the most rapid motions being inwards towards the origin.

It may here be remarked that a very large proportion of the destructive earthquakes of the world have originated beneath the sea, as has often been testified by the succeeding sea waves. Also, it must be observed, that earthquake countries, like volcanic countries, are chiefly those which have a coast line sloping at a steep angle beneath the sea—that is to say, earthquakes are frequent along coasts bordered by deep water.

The earthquakes which occur at volcanic foci constitute another class of disturbances which may be accredited to the explosive efforts of steam.

Earthquakes due to volcanic evisceration.—By the ejection of ashes and lava from volcanic vents, there is an extensive evisceration of the neighbouring ground. When we look at a volcano like Fujiyama, 13,000 feet in height, and at least fifty miles in circumference, and remember that the mass of cinders and slag of which it is composed came from beneath the area on which it rests, the point to be wondered at is, that earthquakes, consequent on the collapse of subterranean hollows, are not more frequent than they are. At the time of a single eruption of a volcano, the quantity of lava ejected amounts to many thousand millions of cubic feet. In 1783 the quantity of lava ejected from Skaptas Joknee, in Iceland, was estimated as surpassing ‘in magnitude the bulk of Mont Blanc.’[127] Admitting that hollow spaces are the results of these eruptions, and that in consequence of this evisceration the ground is rendered unstable, the instability being increased by the additional load placed above the eviscerated area, it would seem that from time to time earthquakes are inevitable.

Facts, however, teach us that volcanoes act as safety valves, and that, as a rule, at or shortly after an eruption, earthquakes cease. The relationship of earthquakes to volcanic eruptions would therefore indicate, notwithstanding the arguments put forward to show that an area loaded by a volcano has in consequence of the evisceration and the load a quaquaversal dip, that evisceration does not take place beneath volcanoes as is usually supposed, and we may conclude that it is but few earthquakes which have an origin due to these causes.

Earthquakes and evisceration by chemical degradation.—A powerful agent, which tends to the formation of subterranean hollows, is chemical degradation. The effects of this have been often measured by quantitative analysis of the solid materials which are daily carried away by many of our springs. In limestone districts this is very great. Prof. Ramsay estimates that the mineral matter discharged annually by the hot springs of Bath is equivalent in bulk to a column 140 feet in height and 9 feet in diameter. At San Filippo, in Tuscany, the solid matter discharged from the springs has formed a hill a mile and a quarter long, a third of a mile broad, and 250 feet in thickness.[128] Many other examples of subterranean chemical degradation will be found in text-books of geology.

By this chemical action large cavernous hollows are produced. Beneath a volcano it is probable that liquid material immediately takes the place of that which is ejected, and that hollows are not formed as in the case of chemical degradation. If a cavern becomes too large, it eventually collapses.

Of the falling in of large excavations we have examples in large mines. As a consequence, not only is a trembling produced, but also a noise, which is so like that produced by certain earthquakes that the South American miners have but one word, ‘bramido,’ to express both.[129]

Boussingault, who was an advocate for the theory that many earthquakes are produced by the sinking of the ground, calls attention to the fact that we have evidences of the subsidence of great mountains, like the Andes, the districts around which are so well known for their earthquakes. Capac Urcu is one of these mountains which legends tell us has decreased in height.

The variation in the height of mountains is a subject which deserves attention. That mountains may possibly be hollow, we have the remarkable results attained by Captain Herschel, who found that the attractive force of gravity in the neighbourhood of the Himalayas was not so great as it ought to have been had these mountains been solid. The Rev. O. Fisher gives another explanation of this phenomenon. Palmieri considers that the terrible earthquake which devastated Casamicciola (1881) was due to the hot springs having gradually eaten out cavernous spaces beneath the town. The extremely local character of this shock was certainly favourable to such a view.

The earthquake which, in 1840, caused Mount Cernans, in the Jura, to fall, is also attributed to the solvent action of waters in undermining its foundations. This undermining action was in great measure probably due to a large spring, which, twenty-five years previously, had disappeared, and which subsequently may possibly have been slowly disintegrating the foundations of the mountain. Earthquakes of this order would be principally confined to districts where there are rocks which are more or less soluble, as, for instance, rock salt, gypsum, and limestone.

Earthquakes and the attractive influences of the heavenly bodies.—The most important attractions exercised upon our planet are those due to the sun and moon. To these influences we owe the tides in our ocean, and possibly elastic tides in the earth’s crust. Some theorists would also insist upon liquid tides in the fluid interior of our earth. The nature of the earth’s interior is, however, a question on which there is a diversity of opinion.

One doctrine, which, until recent years, received much support, was that the interior of the earth was a reservoir of molten matter contained within a thin crust. Hopkins showed that the least possible thickness of such a crust must be from 800 or 1,000 miles, otherwise the motions of precession and nutation would be subject to interference.

M. Delauney objected to the views of Hopkins, on the supposition that the fluid interior of the earth had a certain viscosity.

Sir William Thomson arrives at the conclusion that the earth on the whole must be more rigid than a continuous solid globe of glass. Mr. George H. Darwin’s investigations on the bodily tides of viscous or semi-elastic spheroids tend to strengthen the arguments of Sir William Thomson.

Some philosophers hold the view that the central portion of the earth, although intensely hot, is solid by pressure, whilst the outer crust is solid by cooling. Between the two there is a shell of liquid or viscous molten matter.

Another argument is, that although the interior of the globe may be solid, it is only retained in that condition by an immense pressure, on the relief of which it is liquefied—it is potentially liquid.

As these views, and the arguments for and against them, are to be found in all modern text-books of geology, we will at once proceed to consider the effect of solar and lunar attractive influences in producing earthquakes upon a globe which is either solid, partially solid, or which has an interior wholly liquid.

Effect of the attractive influences of the sun and moon.—In 1854 M. F. Zantedeschi put forward the view—that it is probable there is a continual tendency of the earth to protuberance in the direction of the radii vectores of the two luminaries which attract it. In consequence of these protuberances, pendulums ought at one time to swing more slowly than at others. Zantedeschi remarks that the periods of earthquakes appear to confirm such a view, insomuch as they occur more often at the syzygies, or epoch of the spring tides, than at neap tides—an observation found in the works of Georges Baglivi (1703) and Joseph Toaldo (1770).[130]

Prof. Perrey, of Dijon, who did so much for seismology, held the view that the preponderance in the number of earthquakes felt at particular seasons was possibly due to the attractive influence of the sun and moon producing a tide in the fluid interior of the earth, which, acting on the solid crust, produced fractures.

Rudolf Falb, whose writings have of late years attracted considerable attention, brings forward views which may be regarded as amplifications of those suggested by Perrey.

According to Falb, the inner portion of the earth must be regarded as fluid. In the crust above this fluid reservoir are cracks and channels, into which, by the attraction of the moon and sun, the fluid is drawn. On entering these cracks cooling takes place, together with explosions of gas and subterranean volcanic disturbances. The attractions producing the internal tides required by Falb are chiefly dependent upon the following factors:—

1. The nearness and distance of the sun from the earth (January 1 and July 1).

2. The position of the moon with regard to the earth, which in every twenty-seven days is once near and once distant.

3. The phases of the moon—whether full or new moon (syzygies), or whether first or last quarter (quadratures).

4. The equinoxes, the position of the sun in the equator, and the relative position of the earth.

5. The position of the moon relative to the equator.

6. The concurrence of the ‘centrifugal force’ of the earth with the last quarter of the moon.

7. The entrance of the moon on the ecliptic—the so-called nodes.

Assuming that earthquakes are wholly consequent on these attractions, it at once becomes possible to predict their occurrence. This Falb does, and when his predictions have been fulfilled he has certainly gained notoriety.

He commenced by the predictions of great storms. In 1873 he predicted the destructive earthquake of Belluno, which earned for himself a eulogistic poem, which he has republished in his ‘Gedanken und Studien Über Vulkanismus.’ After this, in 1874, he predicted the eruption of Etna. He also explained why, in b.c. 4000, there should have been a great flood, and for a.d. 6400 he predicts a repetition of such an occurrence.

When we approach the question of the extent to which the attraction of the sun and moon may influence the production of earthquakes, a question which we have to answer is, whether it is likely that the attractive power of the moon is so great that it could draw up the crust the earth beyond its elastic limits. We know what it can do with water. It can lift up a hemispherical shell 8,000 miles in diameter about two or three feet higher at its crown than it lifts the earth. Even supposing the solid crust to be lifted 100 times the apparent rise of the tide, is it likely that a hemispherical arch 8,000 miles in diameter when it is raised 200 feet at its crown could by any possibility suffer fracture? If an arch is 12,000 miles in length, all that we here ask is, whether the materials which compose the arch are sufficiently elastic to allow themselves to be so far stretched that the crown may be raised 200 feet. The result which we should arrive at is apparently so obvious that actual calculation seems hardly necessary. If we regard the earth as being solid, the question resolves itself into the inquiry as to whether a column of rock, which is equal in length to the diameter of the earth, or about 8,000 miles, can be elongated 200 feet without a fracture. This is equivalent to asking whether a piece of rock one yard in length can be stretched one seventy thousandth of a foot. Considering that this is a quantity which is scarcely appreciable under the most powerful of our microscopes, we must also regard this as a question which it is hardly necessary to enter into calculations about before giving it an answer. To vary the method of treating such a question, may we not ask what is the utmost limit to which it would be possible to raise up or stretch the crust of the earth without danger of a fracture? Thus, for instance, to what extent might a column of rock be elongated without danger of its being broken? From what we know of the tenacity of materials like brick and their moduli of elasticity, it would seem possible to stretch a bar of rock 8,000 miles in length for approximately half a mile before expecting it to break. As to whether there is a wave, the height of which is equal to half this quantity, running round our earth as successive portions of its surface pass beneath the attracting influences of the sun and moon, is a phenomenon which, if it exists, would probably long ago have met with a practical demonstration.

The deformation which a solid globe or spherical shell would experience under the attractive influences of the sun and moon has been investigated by LamÉ, Thomson, Darwin, and other physicists and mathematicians.

A conclusion that we are led to as one result of these valuable investigations is, that if the interior of the earth be fluid, and covered with a thin shell, then enormous elastic tides must be produced. A consequent phenomenon, dependent on the existence of these tides, would be a marked regularity in the occurrence of earthquakes. As this marked regularity does not exist, we must conclude that earthquakes are not due to the attractive influences of the sun and moon acting upon the thin crust of the earth covering a fluid interior. The periodicity of earthquakes corroborates the conclusions of Sir William Thomson, who remarks that if the earth were not extremely rigid the enormous elastic tides which must result would be sufficient to lift the waters of the ocean up and down so that the oceanic tide would be obliterated.

Assuming that the earth has the rigidity assigned to it by mathematical and physical investigators, we nevertheless have travelling round our earth, following the attractions of the moon and sun, a tidal stress. This stress, imposed upon an area in a critical state, may cause it to give way, and thus be the origin of an earthquake. Earthquakes ought therefore to be more numerous when these stresses are the greatest.

The periods of maximum stress or greatest pull exerted by the moon and sun will occur when these bodies are nearest to our planet—that is, in perigee and perihelion, and again when they are acting in conjunction or at the syzygies. That earthquakes are slightly more numerous at these particular periods than at others is a strong reason for believing that the attractions of the moon and sun enter into the list of causes producing these phenomena.

Had there been a strongly marked distinction in the number of earthquakes occurring at these particular seasons as compared with others, we might have attributed earthquakes to the existence of elastic tides of a sensible magnitude. As the facts stand, it appears that the maximum pulls exerted by the moon and sun are only sufficient to cause a slight preponderance in the number of earthquakes felt at particular seasons, and therefore that these pulls only result in earthquakes when the distorting effort has been exerted on an area which, by volcanic evisceration, the pressure of included gases, and other causes, is on the verge of yielding.

Earthquakes and the tides.—If we assume that earthquakes are in many cases due to the overloading of an area and its consequent fracture, such loading may occur by the rising of the tide. A belief that the earthquakes of Japan were attributable to the tides may be found in the diary of Richard Cocks under the date November 7, 1618, who remarks:—

‘And, as we retorned, about ten aclock, hapned a greate earthquake, which caused many people to run out of their howses. And about the lyke hower the night following hapned an other, this countrey being much subject to them. And that which is comunely markd, they allwais hapen at a hie water (or full sea); so it is thought it chauseth per reason is much wind blowen into hollow caves under ground at a loe water, and the sea flowing in after, and stoping the passage out, causeth these earthquakes, to fynd passage or vent for the wind shut up.’[131]

Although we may not acquiesce in Cock’s views respecting the imprisoned wind, it would seem that a comparison of the occurrence of earthquakes and the state of the tide would be a legitimate research. Inasmuch as the stresses which are brought to bear upon an area by the rising of the tide are so very much greater than those due to barometrical changes, it is not unlikely that a marked connection would be found. But it must be remembered that because researches, so far as they have gone, tend to show that earth movements are more frequent when an area is relieved of a load, it is not unlikely that the greatest number of earthquakes may be found to occur at low water. Prof. W. S. Chaplin attempted to make this investigation in Japan, but not being able to obtain the necessary information respecting the tides, was compelled to relinquish this interesting work.

Every foot of rise in a tide is equivalent to a load being placed on the area over which the tide takes place of sixty-two pounds to the square foot. This load is not evenly distributed, but stops abruptly at a coast line. Lastly, it may be observed that many coast lines are not simultaneously subjected to stresses consequent upon this load. Japan, for instance, may be regarded as an arch placed horizontally. The area near the crown of this arch is loaded by the tidal wave crossing the Pacific before the areas near the abutment, and farther there is a horizontal pressure at the crown which, if Japan were like a raft, would tend, as the tide advanced, to straighten its bow-like form, but as the wave passed its abutments to increase its curvature.

Prof. G. Darwin has calculated the amount of rise and fall of a shore line due to tidal loads (see p. 336, ‘Earth Pulsations’). The result of these calculations apparently indicates that these loads may have a considerable influence upon the stability of an area in a more or less critical condition.

Mr. J. Carruthers suggests that tidal action may hold a general but indirect relationship to volcanic and seismic action by the retardation it causes on the earth’s rotation. By this retardation the polar axis tends to lengthen, and tensile stresses are induced, resulting in fracture. The fluid interior of the earth, being no longer restrained, would move polewards, and, leaving equatorial portions unsupported, this would gradually collapse. The primary fractures would be north and south, while the secondary fractures would be east and west.[132]

That the rise of the tide is accompanied by a greater percolation of water to volcanic foci, which, in consequence, assume a greater state of activity, is a theory which was advanced many years ago. To determine how far tides may directly be connected with earthquakes, the necessary records have yet to be examined.

Variations in atmospheric pressure.—When we consider the immense load which, by a sudden rise of the barometer, is placed upon the area over which this rise takes place, it is not difficult to imagine that this rise may occasionally be the final cause which makes the crust of the earth to give way. A barometric rise of an inch is equivalent to a load of about seventy-two pounds being put upon every square foot of area over which this rise takes place. On the other hand, a fall in the barometric column indicates that a load has been removed, and whatever elastic effort may be exerted by subterranean forces in endeavouring to escape, being met by less resistance, they may burst these bonds, and an earthquake will result. For reasons such as these the final cause of earthquakes has often been attributed to variations in atmospheric pressure. In Japan there are practically as many earthquakes with a high barometer as with a low one.

The extent to which barometric fluctuations have acted as final causes in the production of earthquakes may be judged of by a comparison of the times of barometric variation and the times at which earthquakes have occurred.

Three important laws of barometric variation are the following:—

1. In the world generally the average barometric pressure is highest in winter. (Exceptions occur near Iceland and in the North Pacific.)

2. The summer and winter monthly mean barometer differs least near the equator and over the great oceans. They differ most over the great continents and generally with increasing latitude.

3. The greatest number of barometrical fluctuations usually take place in winter.

Inasmuch as there are generally more earthquakes in winter than in summer, the first of these laws would indicate that this might be due to the greater load which acts upon the crust of the earth at that season. The second law would indicate that the distinction between the winter and summer earthquakes ought to be most marked in high latitudes, which, if we refer to the table on p. 257, we observe to be borne out by the results of observation. The countries where there are as many earthquakes in winter as in summer are chiefly those in low latitudes. The number of these countries from which we have records are, however, few.

Facts opposed to the idea that earthquakes may be caused by an increase of barometric pressure are the results of observations like those of Schmidt and Rossi, which show that earthquakes chiefly occur with a low barometer.

Assuming that these latter observations will be found by future investigators to be generally true, we must conclude that the relief of atmospheric pressure has an influence upon the occurrence of earthquakes. Such a conclusion would partially accord with the third barometrical law, or the fact that there are more occasions on which we get a low barometer during the winter months.

Other writers who have examined this question are Volger, Kluge, AndrÈs, and Poly. The latter investigator sought a connection between earthquakes and revolving storms, in the centres of which there is usually an abnormal decrease of atmospheric pressure. If an area over which such a sudden change in pressure took place was in a critical state, it is not difficult to see that storms such as Poly refers to might sometimes be accompanied by earthquakes.

Fluctuations in temperature.—Inasmuch as fluctuations in temperature are governed by the sun, it may at once be said that there is a connection between earthquakes and readings of the thermometer. Certainly earthquakes occur mostly during the cold months or in winter. Similarly, as changes in temperature are so closely connected with barometric fluctuations, and these are said to have a direct influence upon the yielding of the earth’s crust, seismic phenomena are indirectly linked to fluctuations in temperature. A rise in temperature is usually accompanied by a fall in the barometer, and this in turn may be a condition favourable for the occurrence of an earthquake.

If we regard solar heat as an agent causing expansions or contractions in the earth’s crust, then fluctuations in temperature become an immediate cause of earthquakes. The probability, however, is that solar heat has little or no connection with the final cause producing earthquakes, although at the same time coincidences between the occurrence of earthquakes and unusual fluctuations in temperature may from time to time be observed.

Winds and earthquakes.—Although it may be admitted that high winds exert enormous pressures upon mountain ranges, and might occasionally give rise to stresses causing rocky masses in unstable equilibrium to give way, the coincidences which have been established between the occurrence of storms and earthquakes can usually only be regarded as occurrences which have synchronised by chance.

Storms are usually accompanied with a barometric depression, and the relation of diminutions in atmospheric pressure to earthquakes has been discussed.

Rain and earthquakes.—It has already been shown that earthquakes have occasionally been found to coincide with rain and rainy seasons. Whether the saturation of the ground with moisture or the percolation of the same to volcanic foci may be a direct effect producing earthquakes it is difficult to say. The probability, however, is that, rain being dependent on phenomena like changes in temperature, barometric fluctuations, and winds, we must regard it and the earthquakes which happen to coincide with these precipitations of moisture as congruent effects of more general causes.

Conclusion.—Although it would be an easy matter to discuss the relationship of earthquakes and other phenomena, we must conclude that the primary cause of earthquakes is endogenous to our earth, and that exogenous phenomena, like the attraction of the sun and moon and barometric fluctuations, play but a small part in the actual production of these phenomena, their greatest effect being to cause a slight preponderance in the number of earthquakes at particular seasons. They may, therefore, sometimes be regarded as final causes. The majority of earthquakes are due to explosive efforts at volcanic foci. The greater number of these explosions take place beneath the sea, and are probably due to the admission of water through fissures to the heated rocks beneath. A smaller number of earthquakes originate at actual volcanoes. Some earthquakes are produced by the sudden fracture of rocky strata or the production of faults. This may be attributable to stresses brought about by elevatory pressure. Lastly, we have earthquakes due to the collapse of underground excavations.

---