SB ID 
 
CALIFORWJA 
 
 EAHTH 
 
 SCIENCE* 
 UBRARY 
 
THE CAUSE OF EARTHQUAKES, MOUNTAIN FORMATION 
 
 AND KINDRED PHENOMENA CONNECTED WITH 
 
 THE PHYSICS OF THE EARTH 
 
 BY T. J. J. SEE, A.M., Lx.M., Sc.M. (Missou.), A.M., Pn.D. (BEROL.) 
 
 Reprinted from 
 
 PROCEEDINGS OF THE AMERICAN PHILOSOPHICAL SOCIETY 
 VOL. XLV. 1907 
 
; 
 
THE CAUSE OF EARTHQUAKES, MOUNTAIN FORMA- 
 TION AND KINDRED PHENOMENA CONNECTED 
 WITH THE PHYSICS OF THE EARTH. 
 
 BY T. J. J. SEE, A.M., Lx.M., Sc.M. (Missou.), A.M., PH.D. (BEROL.), 
 
 PROFESSOR OF MATHEMATICS, U. S. NAVY, IN CHARGE OF THE NAVAL 
 OBSERVATORY, MARE ISLAND, CALIFORNIA. 
 
 Reprinted from Proceedings American Philosophical Society, Vol. xlv,^ 1906. 
 
THE CAUSE OF EARTHQUAKES, MOUNTAIN FORMA- 
 TION AND KINDRED PHENOMENA CONNECTED 
 WITH THE PHYSICS OF THE EARTH. 
 
 BY T. J. J. SEE, A.M., Lx.M., Sc.M. (Missou.), A.M., PH.D. (BEROL.), 
 
 PROFESSOR OF MATHEMATICS, U. S. NAVY, IN CHARGE OF THE NAVAL 
 OBSERVATORY, MARE ISLAND, CALIFORNIA. 
 
 (Read October 19, 1906.} 
 
 I. GENERAL CONSIDERATIONS ON THE CAUSE OF EARTHQUAKES. 
 i. Introduction. 
 
 The great San Francisco earthquake of April 18, 1906, presented 
 certain remarkable characteristics which immediately became a sub- 
 ject of investigation on the part of men of science resident in this 
 part of the United States. One very striking feature of this earth- 
 quake was the conspicuous rotatory motion of the earth particle; 
 and another was the long duration of the disturbance. The rotatory 
 motion appeared so remarkable and so difficult to reconcile with 
 theories very generally held by geologists and seismologists that it 
 seemed worth while to make a somewhat comprehensive survey of 
 the general subject of earthquakes, in the hope of reaching a better 
 understanding of the cause of these phenomena. And as the details 
 of this particular earthquake will be fully treated by others, 1 the 
 result of the present inquiry 2 into the physical cause of earthquakes 
 
 1 The Committee of Investigation appointed by the Governor of Cali- 
 fornia: Professors A. C. Lawson, George Davidson, A. O. Leuschner, G. K. 
 Gilbert, W. W. Campbell, H. F. Ried, J. C. Branner, Chas. Burkhalter. 
 Investigations are being made also by Professor Omori of the Imperial 
 University of Tokio, Messrs. Otto Von Geldern, Luther Wagoner, and Mr. 
 Hoehl of the American Society of Civil Engineers, and perhaps by others. 
 
 2 Rear Admiral H. H. Rousseau, U. S. Navy, Chief of the Bureau of 
 Yards and Docks, has read this paper throughout, and made a number of 
 suggestions which proved valuable. The independent judgment of an experi- 
 enced engineer was felt to be no inconsiderable advantage in weighing some 
 of the difficult questions here treated, and my most cordial acknowledgements 
 are due to Rear Admiral Rousseau for his great kindness. 
 
 274 
 
I9 o6.] SEE THE CAUSE OF EARTHQUAKES. 275 
 
 and other related phenomena may not be without interest to investi- 
 gators of the physics of the earth. 
 
 Almost exactly four months after the earthquake of April 18, 
 namely, August 16, 1906, another, much more terrible, laid waste 
 Valparaiso and the surrounding cities of Chili, producing scenes of 
 desolation which are rare even in South America. The scientific 
 need and the humanitarian demand for an investigation of the cause 
 of these disturbances could, therefore, hardly be greater than it is 
 at the present time. But if it be said that the researches of science 
 are powerless to stay the hand of the destroyer, and only the laws 
 of these terrible phenomena can be discovered, yet even the intelli- 
 gent appreciation of natural laws may greatly mitigate the extent 
 of the disaster and suffering which follow ; and on both humane and 
 scientific grounds, the prospects of extending the domain of useful 
 knowledge furnish a high inspiration for earnest endeavor to pene- 
 trate the mystery of these hidden forces of nature, which so long 
 have baffled the skill of philosophers. 
 
 \Earthquakes and volcanoes were among the earliest physical 
 phenomena to receive the attention of the ancients, and they have 
 always occupied a prominent place in natural philosophy. Although 
 the importance of the subject was derived originally from the ter- 
 rible disasters which these mysterious agencies of unknown forces 
 occasionally inflict upon large portions of mankind, in more recent 
 times earthquakes have been studied also as about the only available 
 means of throwing light upon the physics of the globe. No arti- 
 ficial forces at the command of the experimenter are great enough 
 to produce vibrations of the earth's crust or to transmit them through 
 the body of the planet when once established in the surface layers. 
 
 But notwithstanding all the labor and research which has been 
 bestowed upon the subject, it can hardly be said that we yet have 
 any satisfactory theory of the cause of these phenomena. This is 
 the more regrettable, because, on the one hand, it places it beyond 
 the power of science to predict earthquakes, or even to foretell 
 the regions of their occurrence, which might afford some measure 
 of security to life and property; while, on the other, it leaves 
 many men of science without adequate hope that the true cause of 
 these phenomena will ever be discovered, and at the same time so 
 
276 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 completely bewildered by a multitude of unsatisfactory theories that 
 the progress of discovery itself is seriously embarrassed. 
 
 There will naturally be those who doubt the existence of one 
 common and universal cause of earthquake and volcanic phenomena. 
 Nevertheless, difficult as the subject is, we believe that such a 
 cause exists, and that it is capable of demonstration, if not with 
 mathematical rigor at least with such high degree of probability 1 
 as to render the resulting theory practically useful, and we ask 
 nothing of the reader except a careful examination of the facts 
 as interpreted in the light of the cause assigned in this paper. 
 If such a view, associating the varied phenomena of earthquakes 
 and volcanoes, with mountain formation and the development of 
 great sea waves, under one common cause, renders them more intel- 
 ligible, and enables us to see the relations of all the observed phe- 
 nomena in a clearer and simpler light, there will be presumptive 
 evidence of the truth of the proposed theory ; and the probability of 
 its correctness will increase with the harmony existing among all 
 the known facts, and the effectiveness with which contradictions 
 of other theories may be established. The final test of the theory 
 will depend upon its usefulness in the advancement of discovery, 
 so as to harmonize the whole body of earthquake and volcanic 
 phenomena, including those associated with the origin and structure 
 of mountains, the observations of geodesy, and of great sea waves, 
 in their mutual relations, and in respect to the undisturbed parts 
 of our globe. If the theory shall meet this test satisfactorily, we may 
 feel confident that it assigns the true cause of the phenomena, and 
 within certain limits the resulting laws of nature may be used to 
 foretell events which will contribute to the repose and safety of 
 mankind, and to the progress and usefulness of discovery in this 
 interesting branch of natural philosophy. 
 
 j 2.. The dynamical cause of earthquakes and volcanoes probably 
 
 ^he unequivocal proof of the elevation of the coast at Yakutat Bay, 
 Alaska, Sept. 10-15, 1899, seems to remove the last trace of uncertainty re- 
 garding the chief function of earthquakes, and makes the demonstration as 
 rigorous as that of any theorem in geometry. See the important memoir of 
 Tarr and Martin, Bulletin of the Geological Society of America, vol. 17, May, 
 1906. Professor Georee Davidson, President of the Seismological Society 
 of America, kindly called my attention to this classic work after the present 
 investigation was finished. Note added December 12, 1906. 
 
Igo6 . SEE -THE CAUSE OF EARTHQUAKES. 277 
 
 depends upon the explosive pozver of steam formed within or just 
 beneath the heated rocks of the earth's crust chiefly by the leakage 
 of the ocean beds. 
 
 Some of the most complicated phenomena in nature depend 
 upon the simplest and most obvious of causes, but there are several 
 reasons why the true cause often proves very difficult to discover. 
 On the one hand our mental operations are not infrequently 
 thwarted by conflicting prejudices and contradictory theories, so 
 that attention is diverted from the real questions ; and, on the other, 
 our clearness of vision and power of intuition are blinded by the 
 very closeness and familiarity of the true cause, which is least sus- 
 pected. Success in interpreting nature depends upon a combination 
 of the proper elements of thought into one simple connected view 
 which deals not with details but with the general tendencies. In the 
 case of earthquakes and volcanoes this general view has been very 
 difficult to obtain ; and with the growth of elaborate scientific inves- 
 tigation and classification of earthquakes the difficulty has increased 
 rather than diminished. For attention has been given to the 
 attainment of high accuracy in the measurement of tremors by 
 seismographs and other apparatus, and investigators have been oc- 
 cupied with the registration and discussion of the details of phe- 
 nomena rather than with the general underlying causes. 
 
 We shall hereafter examine the porosity of matter and the prob- 
 lem of the penetration of water into the rocks of the earth's crust, 
 both from the experimental and historical standpoints, but let us 
 first consider the probable state of the internal heat of the earth. 
 In Astron. Nach., No. 4053, the writer has shown that when we 
 consider the force of gravity alone, and suppose a body to be made 
 up of gas reduced to the state of single atoms, over one-half of the 
 primordial supply of heat is stored up within the condensing mass, 
 while still in the gaseous stage ; and in a later paper on the rigidity 
 of the heavenly bodies (A. N., 4104), it is shown that circulation 
 and radiation become retarded and greatly restricted with increasing 
 density, so that in the later stages of the development of a mass like 
 the earth, much more than one-half of the heat generated is retained 
 within the mass for raising the temperature. It is shown that all 
 the heat of our earth depending on gravitation would raise the 
 
278 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 temperature of an equal mass of water 9954 C. ; and as decidedly 
 more than half of it is still retained in the globe, we may conclude 
 that the internal heat of the earth "is ample to raise the whole globe 
 to a temperature of something like 20,000 or 25,000 C., accord- 
 ing to the average specific heat of the earth's matter. If radium and 
 other related elements exist within the earth in appreciable quan- 
 tities, the amount of heat stored up, as Sir G. H. Darwin and 
 others 1 have remarked, may be vastly greater yet. Now it is recog- 
 nized that the crust or cooled layer on the outside of our planet is 
 extremely thin, and we know that the temperature increases down- 
 ward at an average rate of something like i C. for each 30 metres 
 of descent. This accords also with Lord Kelvin's calculations on 
 the cooling of a molten globe, carried out in conformity with 
 Fourier's Analytical Theory of the propagation of heat in solid 
 bodies. 2 
 
 From this we may infer, as geologists have long since remarked, 
 that, even without the penetration of steam, molten rock would be 
 encountered at a depth of decidedly less than 30 kilometres. As the 
 percolation of hot water and steam appreciably lowers the melting 
 point of silicious and perhaps other rocks (the lavas are mainly 
 silicates), and itself develops at the very low temperature of only 
 100 C. under atmospheric pressure, we may infer that it would 
 form in the earth at a depth much smaller than 30 kms. At no more 
 than 10 or 15 kms. under the ocean beds large quantities of it might 
 be produced and give rise to imprisoned forces of tremendous power. 
 Besides it would rapidly absorb and spread in the hotter layers of 
 rock beneath, just as in the case of gases absorbed in hot steel, cited 
 by Tait and quoted in 5. That this absorption actually takes place 
 is proved by the vast clouds of steam given off by melted lava after 
 it pours from a volcano, such as Vesuvius. 
 
 We are thus confronted with the following situation : 
 \ The internal temperature of the earth is extremely high, with 
 
 1 Presidential address to the British Association for the Advancement of 
 Science, Capetown, 1905; also a very recent paper presented to the Royal 
 Society, April 5, 1906, by the Hon. R. J. Strutt, F.R.S., reported in Nature 
 of May 17, 1906. 
 
 2 "The Secular Cooling of the Earth," Appendix D, Thomson and Tait's 
 "Nat. Phil." 
 
I 9 o6j 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 FIG. i. 
 
 279 
 
 Mt. Pelee. The burning cloud of December 16, 1902, seen from the sea. 
 (From the Belgian Astronomical Society, tenth year. Plate V.) 
 
 heated rocks quite near the surface, while the crust is fractured 
 and leaky everywhere, and especially where the depth of the sea is 
 greatest. The sea covers three-fourths of the earth's surface, and 
 earthquakes are found to be most violent -where the sea is deepest, v 
 
 PROC. AMER. PHIL. SOC., XLV. <8 4 R, PRINTED FEBRUARY I 9 , I 9 O 7 . 
 
280 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 , 
 
 and volcanoes most numerous on the adjacent shores.^ Could then 
 anything be more probable than to suppose that both of these great 
 natural phenomena depend simply and wholly upon the explosive 
 power of steam which has developed in the heated rock of the earth's 
 crust ? 
 
 The mere statement of the facts seems almost enough to con- 
 vince one of the truth of this theory. But in view of the wide 
 differences of opinion heretofore prevailing we shall examine it in 
 detail, and we believe it will be possible to show that no contradiction 
 can be established, and that it probably is the correct explanation 
 of the mysterious forces which have so long baffled investigators 
 and wrought such havoc in numerous places throughout the world. 
 
 It would seem that the obvious fact of the leaky character of 
 the sea bottom, covering three-fourth's of the earth, with great in- 
 ternal heat everywhere so close beneath and volcanoes not only 
 abundant on the shores adjacent to the deepest seas, but pouring 
 forth vast quantities of steam when in eruption long ago suggested 
 and apparently ought to have convinced investigators of the validity 
 of this natural and simple explanation. But it appears to have been 
 generally rejected, owing to several circumstances which did not 
 enable investigators to obtain the proper point of view. On the 
 one hand there were traditional theories of volcanoes and their 
 relations to a supposed liquid or molten globe; and on the other 
 little or no adequate knowledge of the enormous number and great 
 violence of submarine earthquakes, which have recently been shown, 
 mainly through the important researches of Professor Milne, to be 
 the most powerful in the deepest oceans. 
 
 While volcanoes and earthquakes have been associated from the 
 time of Aristotle and Pliny, and we think justly so, and some'mutual 
 connection could hardly be denied ; yet even after this relation was 
 especially affirmed by great original investigators like Humboldt 
 and Charles Darwin, it has unfortunately become customary of late 
 years to class earthquakes as volcanic and tectonic or structural. 
 Instead of viewing volcanoes as outlets of pent-up-steam, which 
 blows out if possible the molten rock in which it develops a clear 
 indication of every great eruption an effort was made to explain 
 earthquakes as volcanic, with only partial success, whereas both 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 281 
 
 phenomena depend upon the common cause of steam pressure 
 formed deep in the earth's crust, principally by the leakage of waters 
 from the sea. This highly explosive agency is developed so abun- 
 dantly in the infinitely thin crust between the underlying molten 
 globe, and the overlying oceans, the outcome of a fire beneath and 
 of water above, as in a boiler, that one should not wonder at ter- 
 rible explosive or eruptive phenomena appearing upon our planet. 
 Considering the vast extent of the oceans it would be strange indeed 
 if something like volcanoes and earthquakes were not inseparably, 
 associated with the very nature of the terrestrial spheroid. 
 
 If we consider with attention the various causes which might 
 be assigned to explain earthquakes and volcanoes, taking into account 
 their recognized geographical distribution and relation, the relative 
 situation of the inner globe of fire and the overlying layer of water 
 separated from it only by the thin and leaky bottom of the sea, and 
 remembering that both phenomena are augmented to the maximum 
 in regions characterized by high mountains near the deepest oceans, 
 as on the west coast of South and Central America, the Aleutian 
 and Kurile Islands, Japan, Sumatra, Java and other islands of the 
 East Indies, bordering on the deep waters of the Indian Ocean, New 
 Zealand, and the Lesser Antilles in the West Indies, Iceland, Italy, 
 Greece, etc., we shall find the probabilities that steam pressure 
 developing in the earth's crust is the true and common cause both 
 of earthquakes and volcanoes, are as infinity to one against any other 
 conceivable cause, or all other causes combined. The widely ex- 
 tended relationship here pointed out is so intimate and everywhere 
 so confirmatory of the theory that we cannot suppose it to be due 
 to chance. 
 
 3. Views of Professor Milne and his methods of analysis. 
 
 It has been justly remarked by many seismologists that the 
 greatest belt of earthquakes surrounds the Pacific Ocean. Now each 
 part of this great " fire girdle of volcanoes " with innumerable 
 earthquake disturbances has been studied with care by one or more 
 investigators. Without going into the detailed methods of record- 
 ing and charting developed by Professor Milne, Professor Ewing, 
 Dr. Davidson, Major De Montessus de Ballore, Dr. Agamennone, 
 Dr. Cancani, Dr. Vicentini, Grablowitz, Omori, Koto, Nagaoka, and 
 
282 SEE THE CAUSE OF EARTHQUAKES. [October , 9 , 
 
 others, which are of. great value for the close study of particular 
 regions, we may call attention to the conclusions of Milne, and 
 Montessus de Ballore regarding the slope of the seashores as im- 
 portant factors in the development of earthquakes. 
 
 As a result of careful study of Japanese earthquakes covering 
 eight years, Professor Milne found that "the central portions of 
 Japan where there are a considerable number of active volcanoes 
 are singularly free from earthquakes. The greater number of dis- 
 turbances originate along the eastern coast of the Empire and many 
 of them have a submarine origin." " Lines 120 geographical miles 
 in extent in running in an easterly or southeasterly direction from 
 the highlands of Japan into the Pacific Ocean, like similar lines 
 drawn from the Andes westwards into the same ocean, have a slope 
 of I in 20, or I in 30, and in both of these districts earthquakes are 
 frequent. On the contrary, along the faces of flexures which are 
 comparatively gentle, being less than half of these amounts, which 
 may be seen along the borders of most of the continents and islands 
 of the world, earthquakes are comparatively rare. The inference is 
 that where there is the greatest bending it is there that sudden 
 yielding is most frequent." 1 
 
 It seems advisable to quote more at length the full line of thought 
 laid down by Professor Milne in his classic work on " Seismology " 
 (London, 1898). On page 31 Professor Milne says: 
 
 \ "A very much more serious objection to the volcanic origin of the 
 majority of earthquakes is the fact that these disturbances are common in 
 the Himalaya, Switzerland, and other non-volcanic regions. The destructive 
 earthquake in 1891 in Mino and Owari occurred in a region of metamor- 
 phic and stratified rocks. Again, an analysis of some ten thousand earth- 
 quake observations of Japan shows that there have been but comparatively 
 few which had their origin near to the volcanoes in the country. The 
 greater number of this series originated beneath the ocean or along the 
 seaboard, and as they radiated inland they became more and more feeble, 
 until, on reaching the backbone of the country, which is drilled by numerous 
 volcanic vents, they were almost imperceptible. Beyond this central range 
 of mountains, earthquakes are only rarely experienced, and what is true of 
 Japan seems to be generally true for the coasts of North and South America/'^ 
 
 " Throughout the world we find that seismic energy is most marked 
 along the steeper flexures in the earth's crust, in localities where there is 
 evidence of secular movement, and in mountains which are geologically new 
 
 1 Cf. Scismological Journal of Japan, 1895, p. xv ; and Dutton, " Earth- 
 quakes in the Light of the New Seismology," chapter III. 
 
I9 o 5 ] SEE THE CAUSE OF EARTHQUAKES. 283 
 
 and where we have no reason for supposing that brady-seismic movements 
 have yet ceased. 
 
 " As examples of the flexures to which reference is here made, we may 
 take sections running at right angles to the coast lines of the various conti- 
 nents. The unit of distance over which such slopes have been measured is 
 taken at two degrees, or one hundred and twenty geographical miles. 
 
 "The following are a few of such slopes:' 
 
 West Coast, South America, near Aconcagua i in 20.2 "1 
 
 The Kurils from Urup I in 22.1 Seismic 
 
 in 30.4 | districts, 
 in 23.5 
 
 in 
 
 J 
 
 91 ] 
 158 
 
 in 158 I Non-seismic 
 
 73 r districts, 
 in 243 J 
 
 in 
 
 Japan, east coast of Nipon 
 Sandwich Islands, northwards 
 
 Australia generally 
 Scotland from Ben Nevis 
 South Norway 
 South America, eastwards 
 
 " The conclusion derived from this is that if we find slopes of consid- 
 erable length extending downwards beneath the ocean steeper than I in 35, 
 at such places submarine earthquakes, and their accompanying landslips 
 may be expected. On the summit of these slopes, whether they terminate 
 in a plateau or as a range of mountains, volcanic action is frequent, while 
 the earthquakes originate on the lower portions of the face and base of 
 these declivities. Districts where earthquakes, often followed by submarine 
 disturbances, are most frequent are regions like the northeast portion of 
 Japan and the South American coast between Valparaiso and Iquique. 
 Here we have a double folding. The sea bed, as it approaches the shore 
 line, instead of rising gradually, sinks downward to form a trough parallel 
 to the coast, after which it rises to culminate in mountain ranges. The 
 South American trough which lies within fifty or sixty miles off the coast, 
 like the Tuscarora deep off Japan, attains depths of over four thousand 
 fathoms, and the bottoms of these double folds are well known origins of 
 earthquakes and sea wavesjj 
 
 Professor Milne then goes on to show that where secular move- 
 ments are active, " the forces which have brought these mighty folds 
 (mountains) into existence have not yet ceased to act." The most 
 important question of all, however, is what are these forces ? He says 
 they appear where " mountain formation is geologically of recent 
 origin," and adds i 1 
 
 "The conclusion to which such observations lead is that wherever we 
 find in progress those secular movements which result in the building up of 
 countries or mountain ranges, there we should expect also to find a pro- 
 nounced seismic activity. Thus, while admitting a few small earthquakes 
 to be volcanic in their origin, we recognize the majority of these disturb- 
 ances as the result of the sudden fracturing of the rocky crust under the 
 
 1 " Seismology," by John Milne, F.R.S., p. 33. 
 
284 SEE THE CAUSE OF EARTHQUAKES. [October x 9 , 
 
 influence of bending. The after-shocks which so frequently follow large 
 earthquakes announce that the disturbed strata are gradually accommodating 
 themselves to their new position."l 
 
 Professor Milne's statement that " the greater number of this 
 series (10,000 Japanese earthquakes) originated beneath the ocean 
 or along the seaboard, and as they radiated inward they became more 
 and more feeble, until, on reaching the backbone of the country, 
 which is drilled by numerous volcanic vents, they were almost 
 imperceptible " seems to point directly to the cause set forth in this 
 paper. If earthquakes depend upon the explosive power of steam, 
 they ought not to be numerous near the volcanoes (unless these 
 vents get stopped up), but they ought to be very numerous under 
 the sea in the deep trough just east of Japan, which he says is found 
 to be true by laborious and extensive observations covering a vast 
 number of these phenomena. In order to leave no doubt as to the 
 significance of these results we shall consider also the other lines of 
 thought which he has worked out with so much care. 
 
 4. Inadequacy of the tectonic theory based on slipping and 
 bending, and dislocational and fault movements. 
 
 At present we shall not touch upon all the questions discussed 
 by Professor Milne, but we may remark that slopes of I in 20 given 
 above probably are not great enough to produce the least slipping, 
 or fracturing or bending of rocks. The most effectual way to 
 convince ourselves of the truth of this view is by an appeal to the 
 cones of actual volcanoes. Take Mount Cotopaxi, for example. It 
 is one of the tallest active volcanoes in the world, and the most 
 regularly built of all the large volcanoes. The slope is 30, the 
 angle of the apex being 120. A slope of 30 corresponds to I in 
 1.732; and thus Professor Milne's ratio of I in 20 is less than one- 
 tenth that required to produce stability; and it has escaped his 
 notice that slopes steeper than I in 35 are not such that the steepness 
 could give rise to submarine earthquakes and their accompanying 
 landslips. If the cones of volcanoes like Cotopaxi do not slip, when 
 they are more than ten times steeper than the steepest sea slopes, 
 and over twenty times that mentioned as unstable by Professor 
 Milne, why should slips occur under the sea? Obviously the steep- 
 ness, though no doubt considerable in certain places, is not the cause 
 
19 o6.] SEE THE CAUSE OF EARTHQUAKES. 285 
 
 of earthquakes. Rains, snows and glaciers on Mount Cotopaxi 
 ought to produce slipping of rocks, if anywhere, because the angle 
 is steep and the material loose and unsettled. We are not aware that 
 the slipping of any volcanic cone or other similar mountain has ever 
 been observed to produce a real earthquake ; and if slipping were the 
 order of nature, we should expect some enormous slips with corre- 
 sponding tremors due to this cause near Cotopaxi, Aconcagua, and 
 other great volcanoes (especially when these mountains are shaken 
 at the times of eruption), which are not observed. 
 
 We seem compelled therefore to abandon the theory of slipping 
 and bending of rocks 1 except as producing all the time infinitesimal 
 tremors called microseisms, which very likely depend to a con- 
 siderable extent on this cause. Glaciers are known to be fluid 
 masses, and they move accordingly, though very slowly. It has 
 been shown by the writer (in Nature, 1902) that a rock such as 
 marble undergoes secular bending, and is therefore fluid; and we 
 take it that all large rock masses are very similar in their behavior, 
 though their viscosity may be and generally is greater than that of 
 marble; and hence if movement of mountain masses or other large 
 rocks take place, it would seem that, wherever possible, it should 
 be by a very gradual yielding. The cases in which very large 
 masses of rock, like the sides of a mountain, acquire such unstable 
 positions as to fall, do not seem to be very numerous. 2 Accordingly, 
 it is difficult to believe that this cause is very effective in producing 
 earthquakes ; for such shocks as might result from it would be rare, 
 small and unimportant. And moreover they could never occur 
 where the average slope is anything like so small as I in 20. Be- 
 sides the arguments here outlined there is another hardly less effec- 
 tive which we shall merely mention, namely : That the forces which 
 may shake an entire continent and send waves of compression and 
 
 1 This hypothesis was originally proposed by Boussingault, from observa- 
 tions made on earthquakes noticed in the Andes remote from known vol- 
 canoes, and has at length developed into the tectonic theory now widely 
 held by seismologists and geologists. 
 
 2 The movement by sliding of one or two mountains in the Alps is 
 recorded within the historical period. Among the Andes the most noted 
 change is the collapse of the crater of Carihuairazo, adjacent to Chimborazo, 
 during a violent earthquake on the night of 19-20 of June, 1698. Before 
 this disaster Carihuairazo is said to have been taller than Chimborazo. 
 
286 SEE THE CAUSE OF EARTHQUAKES. [October 19 
 
 distortion through all the rocks between the two oceans, and disturb 
 the whole earth, are not produced by so small a cause as the slipping 
 and bending of ledges of rock. 
 
 Humboldt and Charles Darwin long ago associated earthquakes 
 with secular elevations and depressions, and it is noticeable that 
 Professor Milne likewise thinks these disturbances occur with in- 
 creased frequency in regions where such changes are still in progress. 
 
 Montessus de Ballore concluded from his elaborate study of sta- 
 tistical data that in adjacent seismic regions, instability of the earth 
 is increased by differences of topographic relief; and that the 
 unstable regions are associated with the greatest lines of corrugation 
 of the earth's crust. Like Professor Milne, he observes that rapidly 
 deepening shores which slope gently, especially if they are the con- 
 tinuations of flat or moderately falling coast plains, are stable. His 
 results are illustrated by steep regions of the seashore in South 
 America, Japan, and other parts of the world, and by other regions 
 where the slope into the sea is more gradual. 
 
 These views and others of similar tenor by several investigators 
 have led some geologists and seismologists to conclude that many 
 of the earthquakes noticed along shores which are steep are due to 
 the sliding of unstable deposits of sediment settling on the rock 
 slopes. But if we recall, as above, the smallness of these slopes, even 
 where the descent is most rapid it never exceeds that of our 
 mountains upon the land, and is seldom as steep, and observe that 
 the surrounding sea water is quiescent and would both greatly buoy up 
 and resist the motion of any supposed sliding deposit, so that it is 
 doubtful if appreciable sliding really takes place, and certain that 
 if it does occur the effect in disturbing the earth would be very 
 slight, we shall find it difficult to believe that the theory is well 
 founded. It appears that such a deposit, resisted by the surrounding 
 water, would slide with extreme slowness, and settle gently without 
 any appreciable jar, and consequently no earthquake of importance 
 could be produced in this way. 
 
I9o6 .| SEE THE CAUSE OF EARTHQUAKES. 287 
 
 II. ON THE POROSITY OF MATTER AND ON THE LEAKAGE OF THE 
 
 OCEAN BOTTOMS. 
 
 5. On the porosity and penetrability of matter under the enor- 
 mous fluid pressure operating in the deepest oceans, and the under- 
 lying crust of the earth. 
 
 Somewhat extensive researches on the internal pressures, con- 
 stitution, and rigidities of the sun and planets, carried out during 
 the past two years and published in the Astronomische Nachrichten, 
 have led the writer to the conviction that many of the laws of matter 
 depending on molecular forces, such as impenetrability and solidity, 
 are quite inapplicable to the conditions prevailing in the interior of 
 the earth and other bodies of our solar system; that under the 
 immense pressures there operating, whatever be the temperatures, 
 but especially under the high temperatures known to prevail in the 
 interior of these masses, the hardest natural bodies would yield 
 like sponges, and admit of the most perfect interpenetrability of 
 all the elements. The conclusion was reached from the study of 
 forces of somewhat impressive magnitude that all matter is enor- 
 mously porous, and quite leaky under forces much smaller even than 
 those operating in the interior of the earth; so that solidity and 
 impenetrability, long held to be among the most universal properties 
 of matter, far from being absolute, appeared to be very relative 
 properties, appropriate to very small, but wholly inappropriate to 
 large, forces, and sometimes set aside by the direct evidence of our 
 senses in common laboratory experiments. 
 
 There doubtless are many experiments which would enable us to 
 appreciate the significance of these general principles in specific 
 cases, but it will suffice to recall one close at hand, and directly 
 connected with the question under discussion. In the series of 
 soundings of the depths of the sea carried out some years ago by 
 certain officers of the United States navy occupied with hydrographic 
 and ocean surveys it was found that hollow glass balls with walls 
 several centimetres thick, when subjected to increasing pressure at 
 various depths, came up more and more completely filled with water, 
 in proportion as the depth increased, though no fracture of the 
 glass had occurred, and no holes in it could be discovered by ex- 
 amination of the surface under the highest microscopic power. 
 
288 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 After a careful inquiry by many experienced physicists the conclu- 
 sion was reached that the water had been forced slowly but bodily 
 through the thick walls of the glass under a pressure of less than 
 1,000 atmospheres, in an interval of less than an hour's time. 
 
 In the year 1661 a well-known 'experiment was made by the 
 Florentine academicians who forced water through the solid walls 
 of a sealed hollow sphere of gold, and other metals, by changing the 
 shape of the sphere under mechanical applications of pressure, so 
 as to diminish the volume. The present case of the porosity of glass 
 was thus verified from the opposite point of view, by the steady 
 application of external fluid pressure, on the spherical surfaces of 
 glass balls sent down in modern soundings of the ocean depths. 
 
 The great porosity of all matter has of course long been recog- 
 nized by physicists, but we are so accustomed to dealing with small 
 forces and the resulting doctrine of the impenetrability of matter 
 that it is doubtful whether our appreciation of this fact has yet 
 passed beyond the academic stage. In his well-known " Properties 
 of Matter," fourth edition, p. 87, Tait says : 
 
 " The porosity of wood, necessary for the circulation of sap, is beauti- 
 fully shown by the fact that, from microscopic examination of a thin slice 
 of fossil tree, a botanist can tell at once the species to which it belonged. 
 The greater part of the material of the wood has disappeared for it may 
 be millions of years, but its microscopic structure has been preserved by the 
 infiltration of silicious or calcareous materials which, hardening in the pores, 
 have thus preserved a perfect copy of the original. The rapid passage of 
 gases through unglazed pottery, iron and (hot) steel, etc., shows the porosity 
 of these bodies in a very remarkable manner. So does the strange absorp- 
 tion of hydrogen by a mass of palladium. The porosity of steel has recently 
 been shown in a most remarkable manner by Amagat, who forced mercury 
 through a thickness of more than three inches under a presure of at least 
 four thousand atmospheres. The metal was quite impervious to glycerine 
 under the same pressure." 
 
 At the time this passage was written, some twenty years ago, 
 Tait remarked that decisive proof of the porosity of vitreous bodies, 
 such as glass, had not yet been obtained, but added " that they 
 form almost a solitary class of exceptions to an otherwise general 
 rule seems highly improbable." He then proceeded to show that all 
 bodies whatsoever must necessarily be porous and leaky when sub- 
 jected to great fluid pressure, and he pointed out that the penetra- 
 bility depended greatly on the character of the fluid, thus indicating 
 the great influence of molecular and atomic forces. 
 
19 o6] SEE THE CAUSE OF EARTHQUAKES. 289 
 
 To make a practical application of these principles, what shall 
 we now say with respect to the ocean bottoms? In deep places 
 the pressure of the sea water upon them is very great, sufficient 
 to force the water through walls of solid glass several centimeters 
 thick in a short time, and the bed itself in general no tighter than 
 that of a pond in a common field. Obviously, most of these bot- 
 toms will leak, and leak at a rapid rate under the enormous pres- 
 sure operating in the greatest depths of the sea. The bed of 
 the ocean will not leak with equal rapidity in all places, but almost 
 universal leakage will certainly develop; and the water will be 
 driven down into the earth at various rates depending upon the 
 fluid pressure and temperature and the physical character of the 
 sea bottom. Where the rock is volcanic, and badly fractured, or 
 sandy, the leakage will be most rapid, and where the bed is made of 
 fine clay or unbroken granite, the leakage will be much more 
 gradual. It will also depend directly on the depth of the sea, being a 
 maximum where the ocean is deepest, and generally quite insig- 
 nificant in shallow water. The amount of water leaking through 
 any square meter of the sea bottom will be given by an expression 
 of the form 
 
 w=P.p.f(t).4>(T), 
 
 where P is the fluid pressure in the bed of the sea, and thus directly 
 proportional to the depth ; p the average porosity of the ocean bot- 
 tom, and thus depending on the kind of ooze, dust, sediment and 
 rocks underlying the sea and their state of compression; and f(t) 
 is some function of the time, depending on the average rate of 
 leakage through the successive strata; and (f>(T) is a function 
 of the temperature, and thus -increasing with the descent into the 
 rocks of the earth's crust. As water is almost incompressible for 
 small or moderate forces, its escape downward would depend upon 
 the continued descent of that which first entered the bed of the 
 ocean, the rate of which would be diminished under the increasing 
 pressure and density encountered in the lower strata, but on the 
 other hand increased by the rising temperature which makes the 
 rocks more penetrable and also augments their power of absorption. 
 Various values of these quantities, P, p, /(/), <1>(T), would give the 
 
290 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 several rates of leakage for the corresponding areas of the bottom 
 of the sea. In general it is obvious that the leakage will be most 
 rapid where the sea bottom is fractured or porous, the underlying 
 temperature high, and the depth very great. A rapid rate of leakage 
 would imply that large quantities of water quickly come in contact 
 with the heated rock and develop correspondingly great steam pres- 
 sure in the crust which underlies that part of the ocean. Tait's 
 remark about the rapid passage of gases through hot steel ob- 
 viously applies to the absorption and diffusion of steam in hot 
 rock; for this is found by experiment to be quite general for many 
 of the metals. And in the case of lava as it pours from a volcano, 
 it is observed that the molten rock emits vast quantities of vapor, 
 of which, according to Sir Archibald Geikie, 999 parts in 1,000 is 
 steam. This fact in itself is extremely impressive; for it indicates 
 that the remaining thousandth part of the gases emitted, including 
 vapors of sulphur, hydrogen sulphide, hydrochloric and carbonic 
 acid, are derived from the rocks of the earth's crust under the action 
 of steam and the high temperature. We may therefore consider that 
 steam is the only original vapor operating in the crust of the earth. 
 
 6. Daubree's experiments on the effects of capillarity. 
 
 After this paper was fully outlined and some references were 
 being verified, the author had the good fortune to notice the fol- 
 lowing significant statement in Sir Archibald Geikie's admirable 
 " Text Book of Geology," fourth edition, 1903, p. 354: 
 
 "An obvious objection to this explanation is the difficulty of conceiving 
 that water should descend at all against the expansive force within. But 
 Daubree's experiments have shown that, owing to capillarity, water may 
 permeate rocks against a high counter pressure of steam on the further side, 
 and that so long as the water is supplied, whether by minute fissures or 
 through pores of the rocks, it may, under pressure of its own superincumbent 
 column, make its way to highly heated regions. Experience in deep mines 
 rather goes to show that the permeation of water through the pores of the 
 rocks gets feebler as we descend." 
 
 In his " Physics of the Earth's Crust," second edition, p. 144, Rev. 
 O. Fisher also makes some interesting remarks on Daubree's ex- 
 periments, which are included in his " Rapport sur les progres de 
 la Geologic experimentale,-" Paris, 1867. After describing Daubree's 
 experiment, Rev. Fisher remarks : 
 
r 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 291 
 
 M. Daubree conceives that if the layer of rock were of great thickness, 
 and a very high temperature maintained in the cavity, a correspondingly high 
 steam pressure would result, which would be sufficient to raise lava in the 
 vent of a volcano, and to produce earthquakes; while the force so obtained 
 might after expenditure be again and again renewed. 
 
 "This theory requires the occurrence of cavities at great depth (' sup- 
 posons une cavite separee des eaux de la surface') communicating with the 
 volcanic vents. But the only argument in favor of cavities existing seems 
 to be that the requisite mechanical force is obtainable by means of them; 
 but it seems a priori impossible that there should be such cavities." 
 
 These passages are of interest in connection with Part VIII. 
 of this paper, where it is shown that such cavities or partial cavities 
 develop from the expulsion of lava from under the bed of the 
 sea, and the resulting subsidence of the bottom causes the great 
 sea waves which so frequently follow violent earthquakes. 
 
 7. Historical development of the theory of the penetration of 
 sea water. 
 
 Although these passages were found too late to have influenced 
 the theory developed in this paper, they are cited here for conveni- 
 ence, and to show some of the historical aspects of the problem of 
 the penetration of sea water. It was much discussed also in Hum- 
 boldt's time, as we learn from his remarks in the Cosmos: 
 
 " The geographical distribution of the volcanoes which have been in a 
 state of activity during historical time, the great number of insular and lit- 
 toral volcanic mountains, and the occasional, although ephemeral, eruptions 
 in the bottom of the sea, early led to the belief that volcanic activity was 
 connected with the neighborhood of the sea, and was dependent upon it for 
 its continuance." 
 
 "For many hundred years," says Justinian, or rather Trogus Pompeius, 
 whom we follow, " Etna and the Eolian islands have been burning, and 
 how could this have continued so long, if the fire had not been fed by 
 the neighboring sea?" In order to explain the necessity of the vicinity of 
 the sea, recourse has been had even in modern times, to the hypothesis 
 of the penetration of sea-water into the foci of volcanic agency, that is to 
 say, into deep-seated terrestrial strata. When I collect together all the facts 
 that may be derived from my own observations and the laborious researches 
 of others, it appears to me that everything in this involved investigation 
 depends upon the questions whether the great quantity of aqueous vapours, 
 which are unquestionably exhaled from volcanoes even when in a state of 
 rest, be derived from sea-water impregnated with salt, or rather, perhaps, 
 with fresh meteoric water; or whether the expansive vapours (which at a 
 depth of nearly 94,000 feet is equal to 2,800 atmospheres) would be able at 
 
292 SEE THE CAUSE OF EARTHQUAKES. [October 1 8, 
 
 different depths to counterbalance the hydrostatic pressure of the sea, and 
 thus afford them under certain conditions a free access to the focus." * 
 
 Again : 2 
 
 " The great number of volcanoes on the islands and on the shores of 
 continents must have early led to the investigation by geologists of the 
 causes of this phenomenon. I have already, in another place (Cosmos, Vol. 
 I, p. 242), mentioned the confused theory of Trogus Pompeius under Augus- 
 tus, who supposed that the sea-water excited the volcanic fire. Chemical 
 and mechanical reasons for this supposed effect of the sea have been adduced 
 to the latest times. The old hypothesis of the sea-water penetrating into 
 the volcanic focus seemed to acquire a firmer foundation at the time of 
 the discovery of the metals of the earth by Davy, but the great discoverer 
 himself soon abandoned the theory to which even Gay-Lussac inclined, in 
 spite of the rare occurrence or total absence of hydrogen gas. Mechanical, 
 or rather dynamical causes, whether sought for in the contraction of the 
 upper crust of the earth and the rising of continents, or in the locally dimin- 
 ished thickness of the inflexible portion of the earth's crust, might, in my 
 opinion, offer a greater appearance of probability. It is not difficult to 
 imagine that at the margins of the up-heaving continents which now form 
 the more or less precipitous littoral boundary visible over the surface of the 
 sea, fissures have been produced by the simultaneous sinking of the adjoin- 
 ing bottom of the sea, through which the communication with the molten 
 interior is promoted. On the ridge of the elevations, far from that area of 
 depression in the oceanic basin, the same occasion for the existence of such 
 vents does not exist. Volcanoes follow the present sea-shores in single, 
 sometimes double, and sometimes even triple parallel rows. These are con- 
 nected by short chains of mountains, raised on transverse fissures, and form- 
 ing mountain-nodes. The range nearest to the shore is frequently (but by 
 no means always) the most active, while the more distant, those more in 
 the interior of the country, appear to be extinct or approaching extinction. 
 It is sometimes thought that, in a particular direction in one and the same 
 range of volcanoes, an increase or diminution in the frequency of the erup- 
 tions may be perceived, but the phenomena of renewed activity after long 
 intervals of rest render this perception very uncertain." 
 
 8. Views of Lucretius on the penetration of sea water into 
 Mtna. 
 
 We have quote the above passage because of Humboldt's saga- 
 cious remarks, some of which deal with the theory of the penetra- 
 tion of sea water as held by the ancients. He mentions Trogus 
 Pompeius under Augustus as the author of the theory, but it is 
 remarkable that the same views were held by the poet Lucretius 
 more than half a century before. 
 
 1 Cosmos, Vol. I, p. 242. Bohn's translation. 
 
 2 Cosmos, Vol. V, pp. 431-2. All the citations of Humboldt's works are 
 from the Bohn translations. 
 
i 9 o6.] SEE THE CAUSE OF EARTHQUAKES. 293 
 
 In " De Rerum Natura," Lib. VI, 680 et seq., we read, accord- 
 ing to Munro's translation : 
 
 " And now at last I will explain in what ways yon flame roused to fury 
 in a moment blazes forth from the huge furnaces of ^Etna. And first the 
 nature of the whole mountain is hollow underneath, underpropped through- 
 out with caverns of basaltic rocks. Furthermore, in all caves are wind and 
 air; for wind is produced when the air has been stirred and put in motion. 
 When this air has been thoroughly heated and raging about has imparted 
 its heat to all the rocks round, wherever it comes in contact with them, 
 and to the earth, and has struck out from them fire burning with swift 
 flames, it rises up and then forces itself out on high, straight through the 
 gorges ; and so carries its heat far and scatters far its ashes and rolls on 
 smoke of a thick pitchy blackness and flings out at the same time stones 
 of prodigious weight; leaving no doubt that this is the stormy force of 
 air. Again the sea to a great extent breaks its waves and sucks back its 
 surf at the roots of that mountain. Caverns reach from this sea as far as 
 the deep gorges of the mountain below. Through these you must admit 
 (that air mixed up in water passes; and) the nature of the case compels 
 (this air to enter in from that) open sea and pass right within and then 
 go out in blasts and so lift up flame and throw out stones and raise clouds 
 of sand; for on the summit are craters, as they name them in their own 
 language ; what we call gorges and mouths." 
 
 In one important part of this passage, the text is corrupt and 
 the context, therefore, supplied; yet there is absolutely no doubt, 
 from preceding passages stating that the sea penetrates the land, 
 that Lucretius held that the mountain is hollow, the water filters 
 through the crevices and cracks in the rocks, until it comes into con- 
 tact with the subterranean fires which convert it into vapors that 
 give rise to the explosive violence witnessed in the eruptions of 
 
 We shall see hereafter that Aristotle describes a volcanic eruption 
 as due to the urging blast of pent-up vapor, but it does not seem that 
 he gave any satisfactory explanation of how the vapor developed 
 within the earth's crust. 
 
 9. Lucretius' views on earthquakes, 
 
 " Now mark and learn what the law of earthquakes is. And first of 
 all take for granted that the earth below us as well as above is filled in 
 all parts with windy caverns and bears within its bosom many lakes and 
 many chasms, cliffs and craggy rocks; and you must suppose that many 
 rivers hidden beneath the crust of the earth roll on with violence waves 
 and submerged stones; for the very nature of the case requires it to be 
 throughout like to itself. With such things then attached and placed below, 
 the earth quakes above from the shock of great falling masses, when under- 
 
294 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 neath time has undermined vast caverns; whole mountains indeed fall in, 
 and in an instant from the mighty shock tremblings spread themselves far 
 and wide from that centre. And with good cause, since buildings beside a 
 road tremble throughout when shaken by a waggon of not such very great 
 weight; and they rock no less, where any sharp pebble on the road jolts 
 up the iron tires of the wheels on both sides. Sometimes, too, when an 
 enormous mass of soil through age rolls down from the land into great 
 and extensive pools of water, the earth rocks and sways with the undula- 
 tion of the water just as a vessel at times cannot rest, until the liquid 
 within has ceased to sway about in unsteady undulations. . . . 
 
 " The same great quaking likewise arises from this cause, when on a 
 sudden the wind and some enormous force of air gathering either from 
 without or within the earth have flung themselves into the hollows of the 
 earth, and there chafe at first with much uproar among the great caverns 
 and are carried on with a whirling motion, and when their force afterwards 
 stirred and lashed into fury bursts abroad and at the same moment cleaves 
 the deep earth and opens up a great yawning chasm. This fell out in 
 Syrian Sidon and took place at JEg'mm in the Peloponnese, two towns 
 which an outbreak of wind of this sort and! the ensuing earthquake threw 
 down. And many walled places besides fell down by great commotions on 
 land and many towns sank down engulphed in the sea together with their 
 burghers. And if they do not break out, still the impetuous fury of the 
 air and the fierce violence of the wind spread over the numerous passages 
 of the earth like a shivering-fit and thereby cause a trembling" (" De Rerum 
 Natura," Lib. VI, Munro's translation). 
 
 III. THE GEOGRAPHICAL DISTRIBUTION OF VOLCANOES AND THEIR 
 RELATION TO EARTHQUAKE PHENOMENA. 
 
 10. Four fundamental facts to be explained by a theory of 
 volcanoes. 
 
 A satisfactory theory of the cause of volcanic action must account 
 for the following phenomena : 
 
 1. The distribution of some 400 active volcanoes about the mar- 
 gins of the sea, and the numerous eruptions which take place in the 
 sea or on islands, while none at all occur inland at distances exceed- 
 ing about 100 miles from the ocean or equivalent large bodies of 
 water. 
 
 2. The fact that 999 in 1,000 parts of the vapors emitted by 
 volcanoes is steam, as if produced by the leakage of the oceans, 
 near which the volcanic vents always are situated. 
 
 3. Volcanoes are particular mountains, and all mountains follow 
 the seashore as if formed in some way by the action of the sea upon 
 the adjacent land. 
 
I 9 o6] 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 295 
 
 
 r ^VT( '.'; 
 i- ? f r.ff- < f 
 
 i i *Y i >\K 
 
 1'ROC. AMER. PHIL. SOC. , XI.V. 184 S, PRINTED FEBRUARY 19, 
 
296 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 4. The close geographical relationship existing between vol- 
 canoes and earthquakes throughout the world, and the part played 
 by earthquakes in mountain formation, and the eruption of volcanoes. 
 
 These four fundamental facts seem to admit of easy and natural 
 explanation on the hypothesis that the penetration of sea water 
 develops steam just under the crust of the earth, and the result is 
 the upheaval of mountains and the eruption of volcanoes^ 
 
 ii. Professor Milne's researches on the distribution of earth- 
 quakes. 
 
 H)ne of the most remarkable results of recent research is the 
 discovery of numerous regions greatly affected by submarine earth- 
 quakes, so that it is now known that these phenomena occur not 
 only on land, but more especially under the sea. As we shall treat 
 of this remarkable result hereafter, we shall at present confine our 
 attention to the relations of earthquakes and volcanoes as observed 
 upon the continents. It has long been recognized that both groups 
 of phenomena occur in a series of belts, which follow the same gen- 
 eral regions of the world, along certain so-called lines of weakness 
 in the earth's crust. 1 
 
 In a recent review of earthquakes published in the British As- 
 sociation Report for 1902, Professor John Milne has outlined twelve 
 principal seismic regions, some of them of great extent. These 
 several belts include the wide boundaries of the Pacific Ocean, the 
 Antilles and Caribbean Sea region, and the great belt beginning at 
 the Azores, and extending through the Mediterranean to the Hima- 
 layas and India. This last great belt is the only one in which the 
 sea does not predominate over the land, and even here, the sea is 
 paramount over a large part of the area included, while the rest 
 includes or lies adjacent to the highest mountain range in the world^ 
 As Major Button has remarked, it may be doubted whether all of this 
 last region should be included in one area, except, perhaps, as an 
 outline to aid the memory ; but at all events, the Azores and southern 
 
 1 Cf. Professor Milne's work on " Earthquakes," edition 1903, which 
 includes an excellent map of the world giving the distribution of both earth- 
 quakes and volcanoes. As -earthquakes in the interior of the oceans until 
 recently were seldom recorded, unless of great violence, the earthquakes 
 charted on the map are chiefly those observed on the land, so that the centres 
 of the oceans appear unduly vacant. 
 
I 9 o6] 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 297 
 
 Europe, made up throughout of broken mountainous regions ex- 
 tending into the Mediterranean, with the Black Sea and Caspian on 
 the east, are not essentially different from the earthquake regions 
 surrounding the Pacific Ocean. 
 
 In his recent Bakerian Lecture at the Royal Society, March 12, 
 1906, Professor Milne explains his latest classification of seismic 
 regions as follows : 
 
 " Regions which lie on the western suboceanic frontier of the American 
 and the eastern frontier of the Asiatic continents, and regions which lie on 
 a band passing from the West Indies through the Mediterranean to the 
 Himalayas. 
 
 " In addition to these there are two minor regions, one following the 
 eastern suboceanic frontier of the African continent, which I have called 
 the Malagassy region, and an Antarctic region which lies to the southwest 
 of New Zealand. 
 
 " The following table gives the number of large earthquakes or mass 
 displacements which have occurred in the subdivisions of these regions since 
 1899- 
 
 
 
 1899 
 
 1900 
 
 1901 
 
 1902 
 
 1903 
 ii 
 
 6 
 
 i 
 3 
 
 2 
 
 22 
 
 3 
 [902, 
 
 ind s 
 
 1904 
 
 Total. 
 
 Region of the Pacific 
 Ocean. 
 
 Western Atlantic and 
 Eurasian regions. 
 
 I. East Indian Archi- 
 pelago 
 
 ii 
 19 
 H 
 
 6 
 
 9 
 6 
 
 13 
 
 4 
 9 
 B 
 her, 
 ance 
 
 17 
 5 
 ii 
 
 4 
 
 o 
 
 7 
 6 
 
 2 
 
 4 
 etwee 
 1903 
 s wet 
 
 13 
 5 
 i 
 
 4 
 
 2 
 
 3 
 3 
 
 8 
 
 4 
 :n MJ 
 
 ,751 
 e rec 
 
 14 
 
 9 
 i 
 
 8 
 
 6 
 o 
 
 22 
 
 I 
 
 irch, 
 arge 
 orded 
 
 9 
 14 
 
 
 
 
 o 
 o 
 I 
 
 4 
 
 
 
 and 
 mall 
 
 75 
 59 
 3o 
 28 
 
 16 
 
 25 
 25 
 
 62 
 
 21 
 
 S"ovem- 
 disturb- 
 
 2. The coast of Japan... 
 3 Alaskan coast . .. 
 
 4. Central America 
 5. West of South Amer- 
 ica 
 
 6. Antillian region. ... 
 
 7. Azores 
 
 8. Alpine, Balkan, Cau- 
 casian, Himalayan 
 region 
 
 9. Malagassy district 
 10. Antarctic district 
 
 
 .Totals... 
 
 Oi 
 
 56 
 
 4r1 
 
 64. 1 *8 
 
 20 
 
 341 
 
 " Many of the disturbances included in this table are known to have been 
 followed by hundreds and even thousands of after-shocks. The most active 
 district is at present that of the East Indies, which might well be consid- 
 ered as an eastern prolongation of the Himalayan region. The scene of 
 this activity it may be noticed, is at the junction of two lines of rock folding, 
 which meet almost at right angles. Whether the Antillean and Central 
 American region should be separated is open to question. If we unite their 
 registers as belonging to two comparatively near and parallel earth ridges, 
 the movements of one influencing those 'of the other, we have a region of 
 hypogenic activity approximate to that of the Japan seas. 
 
298 SEE THE CAUSE OF EARTHQUAKES. [October 18, 
 
 " Generally it would appear that these regions of instability are to be 
 found along the margins of continents or tablelands, which rise suddenly 
 to considerable heights above oceanic or other plains. 
 
 " At the present time we may, therefore, say that megaseismic disturb- 
 ances do not occur anywhere, but only in districts with similar contours. 
 Are we dealing with primitive troughs and ridges which are simply altering 
 their dimensions under the continued influence of secular contraction, or do 
 these reliefs of seismic strain represent isostatic adjustments which denuda- 
 tion and sedimentation demand ? " 
 
 Professor Milne then discusses other possible causes, such as 
 the effects of ocean currents and the seasons, including * meteoro- 
 logical causes, such as accumulations of ice and snow at the poles, 
 and finally the motion of the pole in the body of the earth ; and he 
 says that in about thirteen years between 1892 and 1904, he " finds 
 records for at least 750 world-shaking earthquakes," which affords 
 one an impressive idea of the extent of his researches, and of the 
 importance of the subject. 
 
 \ In general the "geographical distribution of volcanoes is closely 
 similar to that of the earthquakes, but the latter are the more general 
 and widely extended phenomena, while the former are more special. 
 It is remarkable that the volcanoes break out in the centers of the 
 earthquake belts. This relation can not be accidental, but points to 
 a common cause underlying both phenomena^] 
 
 Besides the active volcanoes near the seashore, and on islands, 
 many of which were heaved up originally by submarine eruptions, 
 nearly every country has a long list of extinct volcanoes. The 
 islands in which volcanic eruptions have ceased, may also be viewed 
 as extinct volcanoes in the sea. In this respect the southern and 
 central Pacific Ocean is particularly rich in extinct volcanoes, and 
 there also, a great many submarine earthquakes are supposed to 
 occur. But the greatest breeding ground for world-shaking earth- 
 quakes, as Professor Milne says, are the deep troughs along the 
 continents, near which many volcanoes usually are burning. As 
 volcanic regions, we may mention, especially, the west coast of 
 South and Central America, the Aleutian and Kurile Islands, Japan, 
 the Philippines, Sumatra, Java, and adjacent islands of the East 
 Indies, New Zealand, the region of Erebus and Terror in the 
 Antarctic, and Iceland, the Caribbean Sea, with the Azores and 
 Canaries, the region of the Mediterranean and Central Asia, west of 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 299 
 
 the Himalayas. In such of these regions as fall far within the con- 
 tinents, the volcanoes have in all cases died out for lack of water, 
 but the earthquakes still exist as a survival of former conditions. 
 This is true, for example, in the regions of Central Asia, which no 
 longer has any active volcanoes, though some were active there not 
 very long ago geologically, and thus exhibit still a fresh and even 
 sulphurous appearance. 
 
 12. The outbreak of new volcanoes within the historical period. 
 
 It may be mentioned here that within historical times the follow- 
 ing new volcanoes have broken forth : 
 
 1. Monte Nuovo, eight miles from Naples,. September 28, 1538. 
 
 2. Jorullo, in Mexico, September 29, 1759. 
 
 3. Izalco, San Salvador, February 23, 1770, 5,000 feet high, 
 thrown up on what was formerly a cattle farm. 
 
 4. Las Pilas, on the Plains of Leon, Nicaragua, April n, 1850, a 
 small volcano. 
 
 5. Ilopango, Nicaragua, January 20, 1880, a small volcano, 
 thrown up in a lake 600 feet deep. 
 
 6. Fusiyama, Japan, 12,365 feet high, which tradition says was 
 thrown up in a single night, about 300 B. C. 
 
 7. Tarewera, New Zealand, January 10, 1886, a mountain with a 
 flat top, which previously had given no volcanic indications. 
 
 There are perhaps other volcanoes, some of them mentioned by 
 Strabo, which have broken forth on land ; and a good many more 
 which have been upheaved in the sea. 
 
 Many old volcanes long extinct have burst forth into renewed 
 activity, generally with terrible violence. Any volcano may become 
 extinct or dormant, and then again break forth. In his work, on 
 " Volcanoes," Professor Bonney often speaks of a mountain as 
 having lost its crater ; and mentions in this class Ixtaccihuatl and 
 Chimborazo ; but although it is probable, it is not certain that either 
 of these has ever been active. 1 Yet according to the view developed 
 
 1 In his two ascents of Chimborazo, Whymper found lava and other 
 volcanic indications, which had escaped the notice of Humboldt and Bous- 
 singault, who did not regard the mountain as volcanic. The adjacent moun- 
 tain of Carihuairazo, said to have been higher than Chimborazo before the 
 crater collapsed, June 19-20, 1698, might have ejected the volcanic products 
 noticed by Whymper on Chimborazo, though it seems improbable. 
 
300 SEE THE CAUSE OF EARTHQUAKES. |p c tob r , 9 , 
 
 in this paper any mountain may become a volcano, on short notice, 
 if the internal violence is sufficient to break open an outlet for the 
 vapors which always slumber beneath. We shall see, hereafter, that 
 the mountains are all filled with volcanic materials, and an explosion 
 is all that is required to set them going, and this is usually effected 
 by the throes of an earthquake. All new volcanoes, and old ones 
 when they burst forth into renewed activity, do so with violent 
 earthquake shocks. The shocks of an earthquake almost always have 
 some effect on a burning volcano, and in earthquakes remote from 
 erupting centers, the breaking out of a volcano causes the shocks to 
 cease, as was long ago noticed by Strabo. 
 
 Since this intimate connection has been observed again and 
 again, and the volcanic and earthquake belts are generally similar, 
 though not strictly identical, throughout the world, there is a very 
 strong indication that both depend upon a common cause, and that 
 cause is nothing else than ordinary steam. It is worth while to 
 notice that as Central America is a narrow country, with fairly deep 
 seas on both sides, it is exactly where we should expect volcanic 
 forces to have great sway, and observation shows that this is true 
 for earthquakes as w r ell as volcanoes. The recurrence of frightful 
 earthquakes in that region, and the upheaval of three new volcanoes 
 within historical times speaks for itself, and shows that all the 
 mountains are not yet finished ; and that some of the land in Central 
 America is being elevated by forces depending on the influence of 
 the sea, whether volcanic or seismic. As the result of his observa- 
 tions Darwin held that volcanoes break out in rising areas, most 
 likely because an outlet is easily established when the outer layers 
 are cracked open to a great depth. 
 
 13. The relation of earthquakes to volcanoes. . 
 
 In his interesting work on " Earthquakes in the Light of the New 
 Seismology," p. 43, Major Dutton follows Professor Milne in his 
 classifications, and remarks : 
 
 " Though it is possible to indicate regions which present both volcanoes 
 and earthquakes, there is no proof of interdependence between seismicity 
 and vulcanicity in general. While there are earthquakes which are certainly 
 of volcanic origin, the one phenomenon does not necessarily imply the other." 
 
 Professor Milne, Omori, Dutton, and others have recently at- 
 tempted to disprove the relationship of earthquakes and volcanoes 
 
i9o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 301 
 
 exhibited by the shores of the Pacific Ocean, and their" mode of at- 
 tack has been to show that the volcanoes around the Pacific are not 
 a continuous " girdle of fire," but are bunched here and there, with 
 large spaces between; and that while the earthquakes are also dis- 
 tributed with some irregularity, there is no visible connection be- 
 tween them and the volcanoes. 
 
 But if steam forming in the earth's crust from sea water leaking 
 down is the common cause of both earthquakes and volcanoes, should 
 there really be any immediate connection between the two classes 
 of phenomena? Would not volcanoes develop chiefly where the 
 force of the steam was sufficiently powerful and suddenly exerted 
 to break through the crust or mountains, and therefore chiefly in 
 the mountains along the seashore, where the crust is greatly frac- 
 tured, and enables the violent explosions of steam to blow open an 
 outlet by raising a mountain which would burst into a volcano? 
 It is along such shores also that the leakage would be greatest and 
 most volcanoes should exist, provided the crust becomes badly 
 fractured. 
 
 If, on the other hand, the crust is not much broken and ex- 
 plosions of steam cannot break through, would there not result a 
 great many earthquakes of the class now called tectonic because not 
 visibly connected with volcanoes and supposed to be due to slipping 
 of rocks or faults? When the crust is wholly unbroken, it would 
 naturally be very difficult, even for deep-seated forces of enormous 
 magnitude, to raise up a mountain that would become a volcano, 
 because all the overlying strata would have to be violently broken in 
 such a way as to radiate from a point like a star, and ordinarily the 
 strain of the imprisoned steam is much more easily released by an 
 earthquake which merely shakes up the crust in such a way that a 
 neighboring fault moves and the internal pressure is relieved and 
 equalized by scattering, without breaking through all the overlying 
 strata at one time. 
 
 This indeed appears to be the process of nature, and if we 
 consider it in relation to volcanoes and earthquakes we shall per- 
 ceive, in accordance with observation, that the former should be the 
 more special, the latter the more general phenomena. Also both 
 phenomena should occur under the sea, and along the shores of the 
 
302 SEE-THE CAUSE OF EARTHQUAKES. [Octob.r 19. 
 
 deepest oceans ; but volcanoes would develop chiefly in certain 
 regions where the rocks are already broken in the uplift of moun- 
 tains and therefore easily burst open, whereas earthquakes might 
 occur in any locality where the leakage of the sea developed suffi- 
 cient steam. It is undeniable that this is in accordance with obser- 
 vation on our actual earth ; and it shows that while both volcanoes 
 and earthquakes should surround the Pacific Ocean, earthquakes 
 are much more widely and uniformly distributed than volcanoes. 
 Also in those regions near actual volcanoes, where the imprisoned 
 steam has a vent, violent earthquakes should not occur; but if the 
 activity of the volcano ceases, the danger of earthquakes would be 
 increased. 
 
 Humboldt remarks that this opinion was widely spread among 
 the people of the Andes, and it would be difficult to deny that this 
 result of their long experience was well founded, though confessedly 
 they did not know upon what principle the dreaded explosions 
 depended. 
 
 If the leakage from the sea has moderate uniformity with respect 
 to the time, it is clear that the cessation of the smoke of a volcano 
 is really one of nature's danger signals, since the pressure within 
 the subterranean reservoirs of the mountain and adjacent regions 
 may then increase to such a degree as to become extremely dan- 
 gerous. Neither Krakatoa nor Pelee had been active for long 
 periods before the fearful explosions of 1883 and 1902. Krakatoa 
 had been practically dormant for two hundred years, and while 
 Pelee had experienced an eruption in 1851, it was small, and no 
 Important explosion had occurred since I7&2. 1 In the case of Ve- 
 suvius the general experience is the same the longer the eruptions 
 are delayed the more violent they become. For it appears that in 
 79 A. D. no eruption had occurred for about six centuries, and 
 Pliny's description of that outbreak shows that it was more violent 
 than any that has occurred since. The volcano of Conseguina in 
 Central America illustrates the same principle by the long repose 
 preceding the frightful eruption of 1835, which spread devastation 
 far and wide, and in many ways resembled the terrible outbreak of 
 
 1 Cf. Heilprin, " Mont Pelee and the Tragedy of Martinique," pp. 61-187, 
 188. 
 
a9 o6 ] SEE THE CAUSE OF EARTHQUAKES. 303 
 
 Krakatoa. 1 It has so often been observed that the earthquakes 
 ceased on the eruption of a neighboring volcano, that one cannot 
 doubt that direct relief was afforded by the eruption. 
 
 Viewing the relation of earthquakes and volcanoes in this light, 
 we can easily understand why many of the so-called tectonic earth- 
 quakes are doubly severe much more violent than those closely 
 connected with volcanoes because where no volcanic outlet has 
 been available, the explosive strain increases to an enormous extent 
 before it can obtain any relief whatever ; and when the yielding does 
 occur the shock is one of appalling violence, and does great damage 
 causing the slipping of rocks, faults and subsidences, and is felt 
 over a very large area, because the explosive strain has become deep- 
 seated and intense. 
 
 The great depth at which many of the so-called tectonic earth- 
 quakes have been proved to occur, is at once an argument against 
 the dislocational or faulting theory, and a convincing proof that 
 shocks of this type are due to the explosive power of superheated 
 steam. These shocks are obviously too deep-seated to be accounted 
 for by subsidences, and moreover the resulting vibrations are too 
 complex to be due to mere slipping of a ledge of rock, as will be 
 more fully explained hereafter. It may be shown that no possible 
 subsidence of rock faults could produce a conspicuously rotatory 
 earthquake like that which destroyed San Francisco. 
 
 IV. THE GENERAL CAUSE OF THE FORMATION OF MOUNTAINS AND 
 THEIR GEOGRAPHICAL DISTRIBUTION. 
 
 14. On the formation of mountains and Cordilleras, as illus- 
 trated by the Andes. 
 
 If we consider the deep trough running for a great distance 
 parallel to the coast line of the western shore of South America 2 
 and recall that other deep troughs of the same kind exist parallel 
 
 1 In chapter XVI of his valuable work on " Earthquakes," edition of 
 1903, Professor .Milne cites several other eruptions of fearful violence accom- 
 panied in each case by terrible earthquakes. 
 
 2 The trough is not of uniform depth throughout its course, but the 
 depression is always conspicuous, so that everywhere the earth's crust is 
 arched downward. 
 
304 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 [October 19, 
 
 to the Aleutian and Kurile Islands, the east shore of Japan, the 
 west shore of Sumatra and Java ; also near New Zealand and vari- 
 ous islands in the deepest oceans, as Guam, the Bahamas and other 
 West Indian Islands, we shall perceive that this arrangement is 
 not by chance. The South American trough always appears parallel 
 to the great mountain ranges of the Andes, and the fact that it is 
 
 FIG. 3. 
 
 Outline Map of South America, showing the great Ocean Trough parallel 
 
 to the Andes. 
 
 of about the same volume as the matter included in the Cordilleras 
 appeared to be a suspicious circumstance. The relation above cited 
 for Japan, Java, and other islands is similar, but in the case of 
 oval islands the adjacent depression may not be a trough, but rather 
 a hole of somewhat oval or elliptical figure. 
 
 Some years ago the writer noticed these remarkable sinks while 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 305 
 
 examining a relief cast of the Atlantic Ocean exhibited in the office 
 of the United States Coast and Geodetic Survey at Washington, and 
 remarked that it seemed as if the volumes of the upraised islands 
 were not very much larger than those of the adjacent depressions. 
 Why should depressions exist so near these elevations above the 
 sea, and, in the case of mountain ranges, so nearly parallel to them 
 for long distances? Is there not obviously a direct connection be- 
 tween the elevated land and the unusual depression in the adjacent 
 sea bottom ? 
 
 To understand just what this connection is, we may recall that 
 after the great eruption of Mt. Pelee in 1902, it was found by actual 
 measurement that a considerable portion of the adjacent sea bottom 
 
 FIG. 4. 
 
 Vertical Section perpendicular to the Andes and Andean trough, drawn to 
 natural scale, and showing the mode of operation of the trough in the 
 formation of mountain and cordilleras. 
 
 had sunk down hundreds of fathoms. 1 It is impossible to be- 
 lieve that this settling of the bottom of the sea could be due to the 
 
 1 This statement is perhaps somewhat too positive, for Dr. O .H. Titt- 
 man, superintendent of the U. S. Coast Survey, informs me that reliable 
 determinations of depth before the eruption of Pelee seems to have been 
 insufficient to decide the question satisfactorily. The cables were broken, 
 and the subject investigated by the French Commission (Lacroix, Alf, La 
 Montagne Pelee et ses eruptions, Paris, 1904), which includes M. Rollet de 
 1'Isle's investigation of the reported changes of depth in the vicinity of Mar- 
 tinique. The French commission was inclined to ascribe the disturbances to 
 submarine volcanic action, rather than to subsidences. This, however, is not 
 a matter of great importance ; for in his work on " Seismology," p. 35-36, 
 Professor Milne mentions several well established cases of subsidences in 
 the Mediterranean and in the Pacific Ocean off the Esmeralda River in 
 , Ecuador. He points out that " disturbances originating beneath the sea, 
 which are much more numerous than those originating beneath the land, 
 likewise emanate from a region of strain. Mr. W. G. Forster, who has 
 paid so much attention to the earthquakes of the Mediterranean, tells us 
 that they have been accompanied by great subsidences of the sea bottom." 
 
306 SEE THE CAUSE OF EARTHQUAKES. [October i 9 
 
 mere shaking of the earthquakes accompanying that eruption; and 
 we must, therefore, suppose that after matter had been expelled by the 
 dreadful explosions which destroyed St. Pierre and devastated Mar- 
 tinique, or in earlier eruptions, a subsidence near the roots of the 
 mountain actually took place. Again, in 1835, Captain Fitzroy and 
 Charles Darwin observed that after the violent earthquake which 
 destroyed Conception, the Chilian coast line in that region had 
 been elevated from three to five feet for several hundred miles. Not 
 only the coast but also the whole country back to the Andes was 
 raised. This could only be explained by the injection or forcing in 
 of a corresponding bulk of lava under the land ; and this lava could 
 come from nowhere except from under the bed of the great trough 
 in the adjacent sea. The ultimate effect would be to cause the trough 
 of the ocean to deepen correspondingly. And, moreover, such peri- 
 odic injections from under the sea trough would not only push along 
 the ejected stream of lava, step by step, until the end of the column 
 reached the mountains, but the forces thus arising would supply 
 the " lateral thrusts " which are said to be much needed for the 
 explanation of the upheavals, the tipping of the strata, the inclina- 
 tions and sometimes reversed positions of the rocks, and other geo- 
 logical phenomena observed in mountains like the Andes. Hereto- 
 fore the abundant phenomena of this kind noticed in all high moun- 
 tains have not been satisfactorily explained. A correct theory must 
 account for the inclinations seen in the mountains as well as the 
 rising of the coast, and such submarine earthquakes as Darwin 
 observed to precede the uplift of the beach at Conception. The 
 present theory seems to be capable of meeting this severe test, and 
 it requires us to make no assumption except that molten lava may 
 be forced from under the bed of the trough, and pushed along its 
 course beneath the crust by the throes of successive earthquakes. 
 It is recorded that a great sea wave followed the Chilian earth- 
 quake of 1835, and such waves are very frequent along the Chilian 
 and Peruvian coasts. They almost always follow an earthquake, and 
 begin by a recession of the sea from the shore, which then returns 
 as a great wave, carrying everything before it. Some have supposed 
 the sea bottom to subside, thus withdrawing the water toward the 
 sink, till it flows in on all sides to fill up the depression, and then 
 
I90 6.] SEE THE CAUSE OF EARTHQUAKES. 307 
 
 piles up and returns as a great wave which continues to oscillate 
 furiously, sometimes for days after the earthquake. We shall con- 
 sider these waves more fully hereafter, and at present it is sufficient 
 to remark that this explanation is satisfactory for the kind of waves 
 usually observed along the west coast of South America. 
 
 We may then suppose that in such earthquakes a very large 
 mass of lava is forced from under the sea, which then settles below 
 its former level, and the great wave follows. If the lava is forced 
 toward the land, the coast or mountains are upraised ; if towards the 
 ocean, a ridge may be upheaved there, or possibly a submarine vol- 
 cano of large extent. In either case the trough of the sea bottom 
 parallel to the coast eventually becomes less stable, and, at certain 
 intervals, settles little by little, when the consistency of underlying 
 
 FIG. 5. 
 
 lava has been thinned by successive ejections ; and thus with the 
 settling stability is again restored. 
 
 As the trough is arched downwards towards thfe exploding lava 
 the steam pressure from beneath cannot force it upward ; and the 
 strain is necessarily relieved by motion of the lava towards the 
 Andes or the ocean usually towards the mountains till the trough 
 gets broad and deep and the mountains very far away and so high 
 that the movement of the column offers unprecedentedly great re- 
 sistance, when the release will at length become easier towards the 
 ocean by the forcing up of ridges or volcanoes along the other 
 margin of the trough. Ridges with peaks in them will usually re- 
 sult, and this is the beginning of the new Andes or Cordilleras, which 
 are destined to rise slowly from the sea, leaving a deep valley 
 towards the ancient shore, to be drained and filled in by erosion. 
 
 Thus we explain some of the remarkable parallel ridges of 
 
308 SEE- THE CAUSE OF EARTHQUAKES. [October i 9 | 
 
 the Cordilleras. And it is natural that in this upheaval of the 
 crust, release of strain due to subterranean steam pressure should 
 occasionally come by the throwing up of cross ridges, sometimes 
 enclosing undrained areas, and thus lakes like Titicaca are formed. 
 Proceeding upon this simple and natural principle, we may easily 
 explain all the chief characteristics of the Andes. It is impossible 
 to doubt that these mountains have been formed by the very forces 
 which we see still at work there. The upheavals have been step by 
 step, and earthquakes forcing up the mountains have at the same 
 time caused the ejection, by the pushing along of a column or rather 
 a layer, of the necessary matter from under the sea, thus sinking the 
 bottom into a permanent trough, while the subsidences accompany- 
 ing some of the earthquakes have produced enormous sea waves. 
 Countless thousands and perhaps millions of these earthquakes ancl 
 sea waves have occurred throughout past geological ages. 
 
 The trough parallel to the coast and the upheavals and sea 
 waves now observed are a survival to show us just how the moun- 
 tains have been formed, and where the next mountain range will 
 form in the sea. We can predict the formation of the new Andes as 
 confidently as we can an eclipse, though it will be a much longer time 
 before the new mountains develop ; for we recognize the cause to be 
 a true one, and see just how it works by a kind of self-regulating 
 automatic process. The working of the cause has been observed 
 near Mt. Pelee, and the operation of the same process along the 
 South American coast is proved by the observations of Charles Dar- 
 win and by the great sea waves frequently observed within historical 
 times. 
 
 15. Investigation of the significance of the observed lay of 
 mountain chains by means of the theory of probability. 
 
 In their new work on " Geology," Vol. I (p. 543), Chamberlin 
 and Salisbury remark that the relationship between the direction of 
 folded ranges of mountains and the adjacent seacoast is " a coinci- 
 dence that is only in part causal.'' There are, doubtless, many ways 
 in which this problem could be treated by the methods employed in 
 theory of probability. Without claiming to exhaust the various lines 
 along which the discussion might be developed, we believe the fol- 
 lowing method rests on equitable considerations. 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 309 
 
 Imagine a mountain chain like the Andes which runs near the 
 shore cut up into pieces each long enough to reach the sea. If 
 there is no physical cause why the chain should be parallel to the 
 seashore, some of the pieces might be expected to lie at all angles 
 with respect to the shore line of the coast from o to 90, and thus 
 the chain's most probable form is that of a zig-zag line made up of 
 
 FIG. 6. 
 
 short pieces lying at all angles. Take the intervals of angular dis- 
 tribution of the pieces of the chain at 9 ; then, excluding the exis- 
 tence of physical causes, any angle from the first interval, o to 9, 
 to the last of these subdivisions, 81 to 90, must be held to be 
 equally probable. At any place there are ten divisions of the quad- 
 rant, all equally available for the mountain chain to follow. The 
 
310 SEE THE CAUSE OF EARTHQUAKES. [October , 9 , 
 
 probability that the direction of a piece of the chain will fall in any 
 one of these divisions of the right angle is one-tenth ; and the prob- 
 ability that all the n pieces throughout the whole chain will fall in 
 the same angular division is (i/io)". When n is a large number, or 
 the chain is long and close to the sea, this probability becomes prac- 
 tically zero. In such a range as the Andes it is observed that all the 
 pieces of the chain do fall in the first division of the angle, between 
 o and 9; and, hence, in general, the probability is (io) n : i that 
 the observed lay of the chain so exactly parallel to the shore of the 
 sea is not the result of mere chance, but depends directly on some 
 physical cause which has made the chain essentially straight as well 
 as laid out the general course parallel to the sea coast. 
 
 If the chain had the short bends in it here assumed to be pos- 
 sible, the total length would be greater than that of the existing 
 chain, and n would be correspondingly increased. The data used, 
 therefore, make n a minimum, and P = i/(io) n a maximum for the 
 given chain everywhere so closely following the seashore. Thus 
 the calculated value of P is too large rather than too small, with the 
 existing lay of the chain. 
 
 Another way of reaching analogous results is to consider what 
 the deviation from strict parallelism is in so long a chain, and the 
 probability that the coincidence would be so exact throughout, when 
 all angles between i" and 324,000" (the equivalent of 90) are 
 equally probable. If no physical cause is involved depending on 
 the sea, there is no reason why the chain should not run at any angle 
 across the shore line. Now, the Andes are made up on the average 
 of at least two parallel chains, and, the probability of this double 
 parallel trend throughout would be only 
 
 324,000 324,000 114,976,000,000 
 
 In any case, we see that for the double chain, the chances are hun- 
 dreds of billions to one against strict parallelism to the seashore. 
 In the first method of treatment each part of the chain is con- 
 sidered, without regard to the rest ; in the second method, the chain 
 is viewed more as a whole, and the individual parts neglected. In 
 respect to the first method, it might be claimed that as mountains 
 
19 o6.] SEE THE CAUSE OF EARTHQUAKES. 311 
 
 arise from foldings of the crust, and as the earth's crust is thick, a 
 zig-zag form with many bends in it would be improbable, because a 
 crack started in the rocks anywhere would be likely to run in. a 
 straight line. There may be some justice in this criticism, but it 
 seems quite fully compensated for by the fact that the chain actually 
 bends wherever the coast line alters it's course. Thus in practice 
 the chain is shown to be capable of flexure wherever the shore line 
 changes its direction, and there does not seem anything improbable 
 in many short bends, unless the trend has some connection with the 
 coast. Whether the chain could fairly be conceived as bending at 
 such short intervals is a question we need not enter into, for we may 
 observe that short bends actually appear in certain chains, and thus 
 the hypothesis is not contrary to nature under certain conditions. In 
 the present case we have made no assumption as to causes, except 
 that the lay of the chain is independent of the seacoast, and hence 
 the hypothesis postulates nothing improbable. 
 
 To reduce the first method to numbers, we may observe that the 
 length of the chain of the Andes is 4,400 miles and the average 
 distance from the sea about 66 miles. Thus 
 
 4,400 i 
 
 n= - = 66^ or P= - = i : (decillion) 2 
 
 If other parallel ranges be included, this divisor would be much 
 increased, perhaps nearly squared. On the other hand, the method 
 of viewing the chain as a whole makes the divisor a quantity of 
 the order of one hundred billions. The truth must, I think, lie 
 somewhere between these extremes. If we were to take account also 
 of the mountains parallel to the Atlantic coast, it would certainly be 
 moderate to conclude that the parallelism to the seashore noticed 
 in the whole of South America, so far as it depends on chance, would 
 be less than i : (decillion) 2 . 
 
 Thus it is clear that the probability of a physical cause connect- 
 ing the mountains with the parallel shore of the sea is probably 
 more than a decillion decillions to unity, and certainly more than 
 one hundred billions to unity. 
 
 The parallelism noticed in North America along the Pacific and 
 Atlantic coasts, is not less pronounced, nor is there less extent of 
 
 PROC. AMER. PHIL. SOC., XLV. 184 T, PRINTED FEBRUARY 2O, 1907. 
 
312 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 [October 19, 
 
 mountains involved. On the contrary, the ranges of North America 
 exceed in length those of South America. The Rocky Mountains 
 are indeed farther from the coast than the Andes, but the shore line 
 has receded since they were formed; while the Sierra Nevada and 
 Coast Ranges are nearer the present shore. When one considers 
 all these circumstances, including the greater length of the chains, 
 
 FIG. 7. 
 
 and the greater number of parallel ranges, notwithstanding their 
 distance from the seacoast, it will be found that in North America 
 P' is certainly not larger than P as found for South America. 
 Thus we may put them approximately equal to each other, and write 
 
 P. P'=-r- r ' 7 ^= l ' (decillion) 4 
 (io) 66 (io) b6 t 
 
 In Africa the mountain ranges are not high, but they run quite 
 
19o6 .] SEE THE CAUSE OF EARTHQUAKES. 313 
 
 parallel to the shore. We may, therefore, without appreciable 
 error put 
 
 P>=- 
 
 (io) 66 
 
 and 
 
 p. P'. P" = i: (decillion) 6 
 
 In the other three continents, Europe, Asia, Australia, the lay 
 
 FIG. 8. 
 
 of the mountains is such as to justify us in taking P'" . P iv . P v 
 P. P'. P". And, therefore, for the whole world, we may safely take 
 
 P. P f . P". P'". P iv . P v =i: (decillion) 12 
 
 This number is so infmitesimally small, or the divisor is so 
 fabulously large, that it becomes an absolute certainty that the paral- 
 lelism of the mountains to the seashore depends on a true physical 
 cause, and that cause can be nothing but the action of the sea itself. 
 Since the mountains always wall in the land, it follows that they 
 are erected by the sea, through injection of the coast by lava ex- 
 pelled from under the ocean bed. Accordingly, it follows that the 
 crust is bent parallel to the seashore by a true physical cause. 
 
314 
 
 SEE-THE CAUSE OF EARTHQUAKES. 
 
 [October 19, 
 
 If instead of putting P = i : (decillion) 2 we had used P = 
 i: (100 billion), the final result would have been P. P'. P" . P'". 
 P*. PV. = I : (100 billion) 6 . 
 
 Even this divisor is so infinitely large that the fraction totally 
 disappears, and it becomes an absolute certainty that the lay of the 
 mountains depends on the action of the sea as a physical cause ; and 
 that action can be nothing else than the injection of the coast with 
 lava. By this process the mountains were upheaved. The lay of the 
 
 FIG. 9. 
 
 
 mountains parallel to the sea is, therefore, no " coincidence that is 
 only in part causal," but the direct outcome of a general law of 
 nature. How universal this law is may be inferred from the unerr- 
 ing precision with which it is exemplified in a region such as that 
 about the Bay of San Francisco. Here the mountains line the shores 
 on all sides, and change their course in such a way as to enclose the 
 bay with walls so close by and well fitting as to leave no doubt 
 about the origin of the surrounding mountains. 
 
 A deceptive way of viewing such an argument as the foregoing 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 315 
 
 is to claim that the sea, because of its fluid nature, flows into the 
 depressions in the earth's crust, which are due to subsidence or 
 collapse; and when these depressions are filled up they are bound 
 to be surrounded by higher regions of hills or mountains, due to 
 wrinkles in the crust. This method of reasoning ignores the exact 
 
 FIG. 10. 
 
 parallelism to the seashore, which held for every mountain chain in 
 the world at the time of its formation, and still holds for nearly all 
 the principal ranges, though in a few cases the lapse of ages has 
 modified the direction of the shore of the adjacent sea. 
 
 The result here established is, therefore, a fundamental law of 
 nature, and it gives the key to the leading phenomena of the earth's 
 surface. 
 
316 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 1 6. Hozv a sea valley develops and gives rise to parallel moun- 
 tain chains, as illustrated in the San Joaquin, between the Sierra 
 Nevada and Coast Range, in California. 
 
 We have already seen that mountains are raised by the injection 
 of the coast by steam-saturated lava exploding beneath the earth's 
 crust. To understand the entire working of this process under good 
 conditions, we might study the different sea valleys now existing in 
 various parts of the world, or take one which shows the characteris- 
 tic features of the process. If we could find one in which the process 
 is complete, but of recent date geologically, its present form would, 
 no doubt, enable us to make out the transformation which is under- 
 gone at different stages. The Sierra Nevada mountains, with the 
 adjacent San Joaquin valley, appear to be an ideal case to illustrate 
 the process in question. 
 
 A study of the Sierras in California shows that the western 
 slopes of these gigantic mountains are very gradual less than 
 one in fifty and braced by many spurs, with deep intervening 
 canons, of which Yosemite is the most famous. The eastern slopes 
 of the Sierras are about ten times steeper than the western, and the 
 jutting spurs are largely wanting. This shows that the injecting 
 forces which raised these mountains came almost entirely from the 
 west ; and they continued so long that they finally gave the range an 
 unsymmetrical form 1 gently sloping and deeply corrugated with 
 canons on the west, and steep and precipitous on the east. 
 
 The form thus taken by the Sierras would indicate that at a 
 late stage in their history the San Joaquin valley took the form 
 shown in the accompanying figure, long and sloping on the east and 
 
 1 The more gradual slope of a mountain range toward the sea is due to 
 two causes more or less distinct: (i) The vertical upheaval of the chain, 
 with the successive horizontal thrusts which push it little by little from the 
 sea, thus naturally making the farther slope the steeper; (2) the subse- 
 quent elevation of the shore, by injections under the crust, which tips the 
 range still further over, by raising the base of an incline that is already 
 gradual. These two causes combined will be found to explain the principal 
 inequalities in the slopes of mountain ranges as observed in different parts 
 of the world. When the range is being first upheaved, the injections give 
 a nearly vertical uplift; as the range gets older the injections come more 
 and more from the seaward side; when the range itself is finished, the 
 injections only tip its base upward, and make the seaward slope more and 
 more gradual. 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 317 
 
 steep and precipitous on the west. And when this form was once 
 attained the upheaval of the Coast Range became inevitable. The 
 present position of the San Joaquin river near the western side of 
 the valley still shows this form of the valley, and enables us to see 
 the precise process of transformation. 
 
 It is, perhaps, doubtful whether the elevation of these mountains 
 has yet ceased; the earthquakes in California and the low level of 
 the valleys would seem to show that the whole state is still rising, 
 
 FIG. ii. 
 
 a. Mountain formation just beginning. 
 
 Mountain formation in the middle stages. 
 
 c. Mountain formation in the later stages. 
 
 d. New range rising from the sea. 
 
 and I am told that this is also shown by beaches and shells at many 
 places along the coast. 
 
 The same principles which are here applied to the Californian 
 mountains can be applied to other mountain ranges throughout 
 the world. 
 
 17. Significance of the asymmetry of a mountain chain with 
 respect to its principal axis. 
 
 It is generally noticed that most mountains are unsymmetrical on 
 the two sides ; one side has a gradual slope, while the other is steep 
 and precipitous. Moreover, the spurs jutting out from the range 
 are unequally divided between the two sides, the larger number of 
 
318 SEE THE CAUSE OF EARTHQUAKES. [October*,, 
 
 spurs being on the side where the slope is gradual,- which is always 
 turned toward the sea. Let us now examine the meaning of this 
 arrangement. In the new work on Geology by Chamberlin and 
 Salisbury (Vol. I, p. 542-3) we read: 
 
 " Mountain- forming Movements. Along certain tracts, usually near the 
 borders of the continents, and at certain times, usually separated by long 
 intervals, the crust was folded into gigantic wrinkles, and these constitute 
 the chief type of mountains, though not the only type. The characteristic 
 force in this folding was lateral thrust. The strata were not only arched, 
 but often closely folded, and sometimes intensely crumpled. In extreme 
 cases, like the Alps, the folds flared out above, giving overturn dips and 
 reverse strata, as illustrated in the chapter on " Structural Geology," pp. 
 501-511. In these cases there was an upward as well as a horizontal move- 
 ment, for the folds themselves were lifted; but the more horizontal thrust 
 so much preponderated, and was so much the more remarkable, that the 
 upward movement was overshadowed. It is well to note, however, that 
 these mountain ranges are crumpled outward and not inward, as might 
 be expected if they resulted simply from the shrinkage of the under side 
 of a thin shell. The folds are sometimes nearly upright and symmetrical, 
 and sometimes inclined and asymmetrical, as illustrated in the chapter referred 
 to. Where the folds lean, the inference has been drawn that the active 
 thrust came from the side of the gentler slope, the folds being pushed over 
 toward the resisting side, and this seems to be commonly true." 
 
 Thus we have good authority for the statement that the lateral 
 thrust came from the side of the gentler slope. It is well known 
 that the gentler slope is turned towards the sea, and thus we realize 
 that the forces which pushed the mountains horizontally was directed 
 from the adjacent ocean. 
 
 In his great work on " Face of the Earth," * Professor Suess 
 records the following facts, all bearing on the above view : 
 
 1. In Vol. I, p. 452, Suess shows that the tangential movement 
 in the Himalayas and Burmese Mountains on the opposite sides of 
 the Bramaputra are in opposite directions, each being directed from 
 the center of the river exactly what the present theory requires, 
 and not explainable on any other hypothesis. 
 
 2. In Vol. II, p. 34, Suess shows that the mountain folds in the 
 eastern part of the United States " have been produced by a tan- 
 gential movement directed from the existing Atlantic Ocean toward 
 the mainland." On page 139 he shows that the same principle holds 
 for South America. 
 
 1 Oxford Translation by Dr. Hertha Sollas. 
 
I9o6 | SEE THE CAUSE OF EARTHQUAKES. 31i) 
 
 3. In Vol. II, p. 121, Suess shows that the " prevailing tangential 
 movement in the Alps and Pyrenees " is toward the north. Thus 
 for the Alps, Lombardy is the most active ancient sea trough ; and 
 in the case of the Pyrenees, Suess shows that most of Spain was then 
 in the bed of the sea, and has since arisen (pp. 123-128). 
 
 4. That the Sierra Nevada folds were produced by tangential 
 movement from the side of the Pacific is generally recognized by 
 geologists. Dana and others long ago have remarked the same 
 thing about the Andes (Suess, Vol. I, p. 539), and recently, Cham- 
 berlin has written me that most of the mountains give evidence of 
 having been upheaved by lateral thrusts from the direction of the sea. 
 
 In view of these facts is not the significance of the asymmetry of 
 mountain chains perfectly plain? The unsymmetrical build of the 
 chains shows that the forces by which they were formed were di- 
 rected from the sea. And, as these forces could not have been 
 subaerial, they must have been subterranean, working just under 
 the earth's crust, and identical with those observed in earthquakes, 
 when lava is expelled from the sea and pushed under the adjacent 
 coast, along which the mountains always run so exactly parallel. 
 
 By no possibility could any supposed contraction of the earth 
 have given rise to these features in mountain structure, since in 
 that case it would be impossible for the shape of the ranges and 
 tilting of the strata, pointing to lateral thrust, to be always directed 
 from the sea. The arrangement and structure of mountains thus 
 contradicts the contraction theory of the globe, and shows that any 
 supposed effect of contraction was insensible. 
 
 1 8. Criticism of the contraction theory of the formation of a 
 range such as the Alps. 
 
 In addition to the considerations advanced by Fisher, as ex- 
 plained elsewhere in this paper, to show that mountains formed by 
 wrinkles in the crust would in no case exceed a very small height, 
 and should, moreover, be distributed with some uniformity over 
 the globe, we may here consider some objections to the view that 
 a chain such as the Alps has been produced by shrinkage. 
 
 To get Elie de Beaumont's theory clearly before us, we quote it 
 as given by Lyell, " Principles of Geology," I2th ed., Vol. I, p. 119: 
 
 " The origin of these chains depends not on partial volcanic action or a 
 
320 SEE THE CAUSE OF EARTHQUAKES. [October*,, 
 
 reiteration of ordinary earthquakes, but on the secular refrigeration of the 
 entire planet. For the whole globe, with the exception of a thin envelope, 
 much thinner in proportion than the shell to an egg, is a fused mass, kept 
 fluid by heat, but constantly cooling and contracting its dimensions. The 
 external crust does not gradually collapse and accommodate itself century 
 after century to the shrunken nucleus, subsiding as often as there is a 
 slight failure of support, but it is sustained throughout whole geological 
 periods, so as to become partially separated from the nucleus until at last 
 it gives way suddenly, cracking and falling in along determinate lines of 
 fracture. During such a crisis the rocks are subjected to great lateral pres- 
 sure, the unyielding ones are crushed, and the pliant strata bent, and are 
 forced to pack themselves more closely into a smaller space, having no 
 longer the same room to spread themselves out horizontally. At the same 
 time, a large portion of the mass is squeezed upwards, because it is in the 
 upward direction only that the excess in size of the envelope, as compared 
 to the contracted nucleus can find relief' This excess produces one or more 
 of those folds or wrinkles in the earth's crust which we call mountain-chains." 
 
 It is unnecessary to dwell on the violence of the hypothesis that 
 the nucleus has shrunk away from the crust, as here outlined. So 
 far as one can see, no such result is possible, nor is the shrinkage 
 ever appreciable. 
 
 But we may here remark that a radial shrinkage of one mile 
 will give a shrinkage in the semi-circumference amounting to 3.14 
 miles. Whether correctly or not, it has been estimated by geologists 
 that the amount of folding in the Alps exceeds one-third of the 
 whole space now occupied by these mountains, the shortening of the 
 original crust being placed by Heim at 74 miles. Analogous results 
 have been reached by Claypole and others regarding the mountain 
 ranges of America. 
 
 It should be observed that to afford a slack for folds amounting 
 to 74 miles, a radial shrinkage of about 12 miles is required, even 
 when all of the tangential movement is carried round to one point. 
 To carry all the tangential movement round to one point would 
 imply one of two things : ( i ) That the crust is loose from the 
 globe it surrounds, and thus can be carried around to one point 
 from a whole semi-circumference, which seems altogether improb- 
 able; (2) that if the crust is thus shrunk up without being carried 
 around loose from the globe, the matter underlying it must be 
 condensed to about three-halves its former density. The cone of 
 matter underlying the Alps, and extending to the center of the 
 earth would thus have a density 50 per cent, greater than that 
 
i 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 321 
 
 occurring under neighboring areas. This latter result is impossible, 
 since observations show that the matter under mountains not only 
 is not denser than the average, but actually lighter by an appreciable 
 quantity. We know, therefore, that the wrinkling has not con- 
 densed the matter underlying the mountains. 
 
 On the other hand, it seems equally incredible that the shrink- 
 age of the whole globe should be brought forward to one point, as 
 if the crust were loose from the globe it covers. No part of the 
 theory of mountains formed by wrinkles of the crust is to be seri- 
 ously entertained. Besides the difficulties just mentioned, the 
 postulated radial shrinkage of 12 miles is too great. It may well be 
 doubted whether a shrinkage of one mile in the radius has taken 
 place since the continents began to emerge from the oceans. 
 
 Professor Suess (Vol. II, p. 552) says: 
 
 "As a result of tangential thrusts, the sediment of this (Mediterranean) 
 Sea were folded together and driven upwards as a great mountain range, 
 and the Alps have, therefore, been described as a compressed sea." 
 
 We must, therefore, seek the explanation of the formation of the 
 Alps in some process by which this folding can have taken place in 
 the sea, or along its borders, and thus we reach the theory outlined 
 in this paper. 
 
 19. Why ive abandon the contraction theory of mountain for- 
 mation. 
 
 While the considerations here adduced for the origin of moun- 
 tains seem conclusive, it may not be wholly without interest to point 
 out some difficulties which are not satisfactorily met by the con- 
 traction theory, which is the only one now in general use. It is 
 usually stated that mountains result from a crumpling of the 
 earth's crust, and that the crumpling takes place along the principal 
 lines of weakness. This theory fails to explain the origin of isolated 
 peaks or associated groups of peaks which sometimes rise like cones 
 or groups of cones, often more or less intersecting, in the midst of 
 comparatively regular plains. A theory w r ith this serious defect is 
 highly unsatisfactory. 
 
 If then the contraction theory fails to explain isolated peaks and 
 groups, which are sometimes pushed up in comparatively level 
 plains, and fails to explain the conspicuous parallelism to the sea- 
 
322 SEE THE CAUSE OF EARTHQUAKES. [October 19 
 
 shores, while the account given of cross ranges and parallel ranges 
 standing in isolation is unsatisfactory, it must be admitted that the 
 theory itself is not well founded. 
 
 The assumption that since the continents began to rise from the 
 oceans the earth has shrunk enough to produce great wrinkles in 
 the crust comparable with our high mountains is undeniably a violent 
 hypothesis. For in the writer's paper on the rigidity of the heavenly 
 bodies (A. N. t 4104), it is shown that no circulation of currents 
 within the earth has been possible since the globe became encrusted. 
 If there are no currents within, the propagation of heat outward 
 could take place only by conduction; and from Fourier's analytical 
 theory of heat we know that the loss of heat would be extremely 
 slow, and confined almost wholly to a shallow layer near the surface. 
 Indeed it seems probable that the shrinkage of the entire globe is 
 barely comparable to the secular contraction of the cooling crust 
 alone. The approximate accuracy of this view is confirmed by the 
 fact that the crust has not cracked open by pulling apart, as it 
 would do if the crust shrank much more rapidly than the globe as 
 a whole. The fact that the interior of the globe lost very little 
 heat, while the crust cooled all the time, would lead one to think that 
 so far from wrinkling by contraction, the crust ought to have 
 cracked open by the shrinkage of the shell over a nearly unyielding 
 nucleus; but no doubt the process was too slow and the rocks too 
 plastic to give rise to actual rupture of the earth's crust. 
 
 Moreover, if the globe shrank, it is inconceivable that this shrink- 
 age could fail to be fairly uniform in the different equal areas of the 
 surface, and thus we should expect the resulting wrinkles to be 
 distributed over the globe with . moderate equality and uniformity. 
 Instead of this, we find the mountains, heretofore assumed to be 
 wrinkles, bunched into congested systems, and almost always paral- 
 lel to the seashore, and larger in proportion to the depth of the 
 adjacent ocean. Is it therefore at all credible that the mountains 
 have really been formed by the shrinkage of the earth ? Would it be 
 going too far to say that the whole theory of secular contraction 
 as applied to our encrusted planet is a misconception dating from 
 a time when currents were supposed to circulate freely throughout 
 a liquid globe ? 
 
19o6 .] SEE THE CAUSE OF EARTHQUAKES. 323 
 
 After a careful consideration of the whole question, including 
 all the forces at work, one may well doubt whether the radius of the 
 earth -has shrunk a mile since the continents began to emerge from 
 the oceans. The crust could easily accommodate itself to such a 
 small shrinkage as one part in 4,000, without producing any wrink- 
 ling whatever. 
 
 This subject of planetary wrinkles has been treated mathe- 
 matically by Professor Sir G. H. Darwin, in his researches on the 
 "Tides of a Viscous Spheroid" (Phil. Trans. Roy. Soc., Part n, 
 1879, P- 588)- And while his results are recognized to be correct, 
 on the hypothesis, they seem to me to be inapplicable to the remote 
 history of the earth, because I believe the shrinkage to have been 
 nearly insensible, and certainly much less effective than has been gen- 
 erally supposed. 
 
 In his " Physics of the Earth's Crust," page 118, Rev. O. Fisher 
 has discussed Darwin's theory thus : 
 
 " It may be replied to this theory, that the formation of the existing 
 continents cannot be looked at apart from their geological history, and that 
 they are evidently dependent on, and as it were, gathered round, the great 
 mountain ranges in which they culminate. Although these ranges primarily 
 originated long ago in very early geological times, their present loftiness is 
 due to quite late movements ; and, if these had not subsequently occurred, 
 they would before now have probably have been razed to the sea-level and 
 have disappeared, so that, whatever cause it was which wrinkled the conti- 
 nents, seems to have continued active to times comparatively, if not quite, 
 recent; and the moon is too far off now. The occurrence of great changes 
 of level at no very distant geological period are manifest from such instances 
 as that related by the elder Darwin." 
 
 It has always been extremely difficult to show how contraction 
 could produce elevation of ranges, and the mechanical explanations 
 which have been put forward are admittedly unsatisfactory. 1 The 
 spurs which so often jut out from the main ranges do not look like 
 wrinkles in the crust ; for they are too numerous and terminate too 
 suddenly at their extreme ends. The isolated parallel ranges so 
 often met with in Nevada and Southern California also terminate 
 too suddenly to be explained by shrinkage, or by lines of weakness. 
 
 In his work on the " Physics of the Earth's Crust," second 
 edition, the Rev. O. Fisher has discussed with much care the inade- 
 
 1 Fisher's " Physics of the Earth's Crust," second edition, p. 123. 
 
324 SEE THE CAUSE OF EARTHQUAKES. | October 19,. 
 
 quacy of the contraction theory to account for the elevations of 
 mountains actually observed upon the earth. In the case of a solid 
 globe, he finds (p. 122) that the average height of the elevations 
 would be only six and one-third feet. By the theory of probability 
 we easily see that this would absolutely prevent individual elevations, 
 even when exceptionally favored by circumstances, from attaining 
 any considerable height. In the appendix to the second edition, page 
 58, he examines the mean elevations which will result from the 
 hypothesis of a liquid substratum, and finds that when the radial 
 contraction is 12 miles, "the mean height of all the elevations, due 
 to the corresponding corrugation of the matter above the level owing 
 to secular cooling, is about 44 feet." " It appears," he adds, " there- 
 fore as the result of the investigations in this chapter, that the 
 hypothesis of a liquid substratum does not afford such an increased 
 amount of compression as to render it possible to attribute the 
 elevation of mountains to contraction through cooling in that case,, 
 any more than in the case of solidity." 
 
 When one recalls that our actual mountains are many hundreds' 
 and even thousands of times higher than the mean elevations re- 
 sulting from the contraction theory, it is readily seen how utterly 
 devoid of foundation that theory really is. Considerations adduced 
 in the writer's paper on the rigidity of the heavenly bodies show that 
 the radial contraction of twelve miles used by the Rev. O. Fisher 
 is probably at least twelve times too large; so that the highest ad- 
 missible mean elevations due to shrinkage would be only a very 
 few feet. 
 
 It need scarcely be added that the Rev. O. Fisher must be 
 given the chief credit for showing by long and patient research 
 the inadequacy of this time-honored theory, which was originally 
 suggested by Elie de Beaumont in 1829, and was no doubt a direct 
 outgrowth of Laplace's nebular hypothesis. 
 
 We thus seem compelled to abandon the contraction theory en- 
 tirely, and to explain both peaks and ranges with their striking 
 parallelism to the coast by upheavals occurring near the sea, due to 
 the explosive power of steam, which has heaved up the mountains 
 from beneath. The mountains apparently show this mode of forma- 
 tion, and it explains with equal satisfaction cordilleras and ranges,. 
 
I 9 o6.j 
 
 SEE THE CAUSE OF EARTHQUAKES. 325 
 
 whether of continued or isolated character, with their numerous 
 jutting spurs and cross ranges, and isolated peaks, which are well- 
 nigh unintelligible on any other hypothesis. And lastly it shows 
 that all mountains are alike inside, whether they burst open and 
 become volcanoes or remain intact. 
 
 A theory presenting so many desirable points should have a 
 strong claim to acceptance. 
 
 In this connection one geological term in extensive use might 
 perhaps be explained. We refer to the phrase Line of Weakness of 
 the earth's crust, which was employed by Leopold von Buch to 
 explain the arrangement of volcanoes along the seashore. It .forms 
 wherever the sea stands some time, especially if the sea is deep, 
 because the explosive paroxysms of steam work under the edge of 
 the sea, but not under the land, and, therefore, " lateral thrusts " 
 from the sea begin, while they cease on the land; the result is an 
 injection of the coast line from the direction of the sea, and moun- 
 tains and volcanoes are upraised, according to the intensity and 
 especially the difference of these forces, from the sea and land, and 
 their duration. 
 
 It is not without significance that the height of the mountains 
 are in general proportional to the depth of the adjacent sea, because 
 the forces of injection depend upon the depth, and the elevations 
 produced are proportional to the intensity of these forces. When 
 the sea recedes, however, the extent of the land gained is propor- 
 tional to the shallowness of the water, and hence arise the large flat 
 plains in many countries. This explains the arrangement of moun- 
 tains and volcanoes along the sea coast, which has, therefore, been 
 called a line of weakness in the earth's crust. As a matter of fact, 
 any line will prove to be weak where the sea stands for a long time, 
 for mountains and volcanoes will be upheaved there. Thus I con- 
 ceive that there is originally no such thing as a line of weakness in 
 the crust, and we may with advantage dispense with that unfortu- 
 nate term. This seems the more advisable, since the earth behaves 
 as a solid, and local weakness developed in the formation of moun- 
 tains has little effect at a distance, except in volcanic regions or 
 ocean troughs, which act together sometimes throughout their whole 
 extent. 
 
326 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 It is shown in the paper on the rigidity of the heavenly bodies 
 that the strength of a planet like the earth is not appreciably de- 
 pendent upon the crust, but arises primarily from the great pressure 
 acting throughout the body, which itself in turn depends upon the 
 mass and density of the globe. The earth's crust, therefore, has 
 little importance in the theory of the eartji, except in our treatment 
 of surface phenomena. 
 
 20. Is there a creeping movement of the fluid substratum be- 
 neath the crust? 
 
 The existence of such a powerful seismic zone around the 
 Pacific Ocean, which is surrounded by unfinished mountains and 
 a * fire girdle of volcanoes ' leads one to inquire whether there may 
 not be throughout this vast ocean, as well as in smaller seas, a 
 tendency for the explosive stresses to find relief at the margins, 
 by a slow creeping movement of the particles of the substratum 
 towards the periphery, where the chief relief is afforded. For 
 those stresses arising under the crust in the middle of such an 
 ocean, some relief would be afforded by the rising and sink- 
 ing of certain oceanic islands ; but a greater relief would be 
 afforded around the periphery of the sea, where the great moun- 
 tain chains are in process of formation. As the crust under the sea 
 is incessantly strained by the heaving of subterranean forces, some 
 parts rising and others sinking, a slow creeping movement of the 
 fluid substratum towards the periphery seems not only possible, but 
 perhaps probable. Such a final movement would be the result of 
 the countless earthquakes which disturb the sea bottom, and in any 
 given earthquake the motion would be extremely slight. The 
 creeping fluid would tend towards the avenues of escape in islands 
 and on the margins of the sea, as well as towards areas still sub- 
 merged but rising; and thus we recognize forces which under cer- 
 tain conditions may both elevate and depress islands in the sea ; but 
 in the long run the sinking tendency will predominate where there 
 is water, and the rising tendency where there is land. 
 
 All along the west coast of South America Charles Darwin 
 found conspicuous evidence of elevation within recent geological 
 times ; and at Valparaiso the amount was no less than 1 ,300 feet. 
 The periodic subsidences indicated in certain places by beds of ma- 
 
i9o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 327 
 
 rine fossils of past geological ages, could, I think, be explained by 
 tendencies developed under the crust, according to which the fluid 
 substratum is alternately thickened and thinned, owing to the con- 
 currence or non-concurrence of the subterranean forces. When 
 they work towards a point the result is elevation, and when they tend 
 to diverge from a point the result is depression, and the elevation 
 is transferred to neighboring areas. This is a modern view of the 
 periodic movement of the earth's crust so clearly foreseen by Strabo 
 nearly 2,000 years ago. 
 
 According to this view the sea bottoms may oscillate, but on the 
 whole tend to subside, not on account of the shrinkage of the globe, 
 but by virtue of the gradual working out of the underlying fluid 
 substratum, which in the long run pushes up the land. 
 
 In his " Principles of Geology," I2th edition, Vol. II, page 155, 
 Lyell discusses with characteristic fairness the historical cases of 
 elevation of coasts noticed in different parts of the world. We shall 
 content ourselves with citing a very few cases of this type :. 
 
 1. Islands in the sea innumerable, both volcanic and non-volcanic 
 (apparently, though all are raised by volcanic forces). 
 
 2. The southwestern end of the Island of Crete, which even 
 Professor Suess admits to have experienced undeniable secular 
 elevation within the historical period. 
 
 3. The region about Pozzuoli and the Bay of Naples. This is 
 shown by the famous temple of Jupiter Serapis, and by the eleva- 
 tion of the coast actually witnessed at the time of the eruption of 
 Monte Nuovo in 1538. This raising of the land was confirmed on 
 a larger scale for the whole Bay of Naples during the Vesuvian 
 eruption of April, 1906 (cf. Quarterly Journal of the Geological 
 Society, No. 247, 1906), by Professor Lorenzo, who found the ele- 
 vation of the land at Pozzuoli to be six inches, and at Portici one 
 foot. 
 
 4. The foundations of both ^Etna and Vesuvius were laid in 
 the sea. 
 
 5. Professor Suess cites the most ample evidence of raised 
 beaches and other sea marks high above the present strand in almost 
 all parts of the world. As these heights are very unequal, they can- 
 not be explained by a simple sinking of the sea level, but there must 
 
 PROC. AMER. PHIL. SOC., XLV. 184 U, PRINTED FEBRUARY 2O, 1907. 
 
328 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 have been undeniable oscillations of the land, similar to, but on a 
 larger scale, than those observed during the historical period. 
 
 6. Conclusive proofs of the upheavals of the Chilian coast during 
 the earthquakes of 1822, 1835, etc., have been given by Lyell, and 
 will not be repeated here. We content ourselves, therefore, with the 
 following account of the Valparaiso earthquake of August 16, 1906, 
 which speaks for itself: 
 
 THE GREAT EARTHQUAKE AT VALPARAISO, AUGUST 16, 1906. 
 
 A special copyrighted cablegram to the San Francisco Examiner of 
 August 23, dated Valparaiso, Chili, August 22, says : 
 
 " The recent seismic disturbances in this region have thrown up several 
 new islands in Valparaiso Bay. These islands are of various dimensions, 
 some being very extensive while others appear to be mere cone-like rocks 
 jutting above the waters. It is reported that islands have appeared at dif- 
 ferent points along the coast of Chili. . . . 
 
 " The wrenching given the earth's surface is still showing more and more 
 day by day. In sections of the harbor and on the coast, the shore line has 
 been materially changed. Promontories have slid bodily into the sea and 
 in other places strips of coast line have been completely submerged. The 
 theory is that there has been a great uplift of the Andes so as to change 
 almost entirely the contour of the hilly region of the republic. Landslides 
 are everywhere in evidence. Mountain sides have been stripped away and 
 chasms in the hills filled up. 
 
 " Persons who have arrived here on horseback from points along the 
 coast say that they witnessed nothing but devastation. Whole villages were 
 wiped out." * 
 
 1 Since this paper was finished, Professor H. D. Curtis, in charge of the 
 D. O. Mills Expedition of the Lick Observatory, at Santiago, Chili, has 
 written an interesting letter to Professor Kroeck of the Pacific University at 
 San Jose, which is published in the San Francisco Argonaut of November 3. 
 Professor Curtis says: 
 
 " A Commission has been appointed to study the shock and its causes. I 
 published a statement that the primary cause was doubtless the same as at 
 San Francisco, the slipping or sliding of one stratum past another, due to 
 the well-known geological fact that the Coast of Chili is very slowly rising. 
 I learn since that the Bay of Valparaiso is now ten feet shallower. So I 
 think the displacement in this shock will prove to be mainly vertical. It 
 may be that the centre of disturbance was under the sea, as Valparaiso suf- 
 fered much more than Santiago." 
 
 Professor George Davidson, President of the Seismological Society of 
 America, informs me that during the great earthquake at Yakutat Bay, near 
 Mt. St. Elias, Alaska, September 3-20, 1899, the land at the head of Yakutat 
 Bay was raised 47% feet. In the Bulletin of the Geological Society of 
 America, vol. 17, May, 1906, will be found a careful investigation by Tarr 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 329 
 
 21. Avicenna's views on mountain formation. 
 
 Lyell .justly remarks that it is surprising to find among the 
 extant fragments of Avicenna, Arabian physician and astronomer 
 of the tenth century, a treatise on the " Formation and Classification 
 of Minerals," characterized by considerable merit, the second chap- 
 ter of which is " On the Cause of Mountains." Mountains, accord- 
 ing to Avicenna, are formed, some by essential, others by accidental 
 causes. And in illustration of the essential causes, he cites " a 
 violent earthquake, by which land is elevated, and becomes a moun- 
 tain." In regard to the accidental causes he mentions excavation 
 by water or erosion, which produces cavities, such that adjoining 
 land is made to stand out and form eminences. 
 
 The theory of mountain formation adopted in this paper was 
 therefore foreshadowed by Avicenna in the tenth century of our era. 
 It is extremely remarkable that so simple an explanation should 
 have been allowed to slumber for so many centuries, while artificial 
 and highly unsatisfactory hypotheses were in use. 
 
 V. EXPLANATION OF THE ELEVATION OF PARTICULAR MOUNTAIN 
 RANGES AND PLATEAUS. 
 
 22. On the uplifting of the Andes. 
 
 We have already seen that the Andes have been uplifted by the 
 injection of lava beneath the crust in the earthquakes incident to 
 the heaving of the Andean Valley in the adjacent sea. This has 
 been the chief cause of the original uplift of these great mountains, 
 and the resulting explanation suffices to account for all the principal 
 phenomena. Thus we explain the gentle slope of the mountains 
 
 and Martin, who show that the coast was elevated for more than a hundred 
 miles, though slight depressions also occurred in a few places. Elevations of 7 
 to 20 feet were common, and so little change had occurred in 1905 that Tarr 
 and Martin were able to illustrate their memoir by photographs of the most 
 convincing character. The barnacles and other marine animals were still 
 adhering to the rocks, and there could be no possible doubt about the fact 
 of the elevation. The depression of some areas was made equally clear by 
 the encroachment of the salt water upon forests, which were thus killed. Two 
 of the shocks at this great earthquake (September 10-15) were particularly 
 terrible, the motions recorded in Tokio, 3,300 miles away, being ^ and y% 
 inch respectively. Note added December 3, 1906. 
 
330 SEE THE CAUSE OF EARTHQUAKES. [October^, 
 
 on the west and the extreme steepness of the descent on the east; 1 
 and we also account for the more numerous jutting spurs on the 
 west. On the whole side spurs are much less conspicuous on the 
 east, where descent is more rapid. These characteristics are prob- 
 ably the leading features of the Andes, but there are others deserv- 
 ing of attention, among which we may mention the following : 
 
 1. Very great volcanic violence throughout the whole range, 
 and in the peaks of the eastern as well as of the western cordillera. 
 
 2. Enormous vertical uplifts, or fault movements, often amount- 
 ing to thousands of feet, occurring throughout the cordillera, but be- 
 coming especially predominant on the eastern side. 
 
 3. The vertical uplifting of enormous plateaus such as those of 
 Quito, Caxamarca, Cuzco and Titicaca, the latter being 12,500 feet 
 above the sea. 
 
 The heaving of the Andean Valley in the sea seems to be the 
 principal cause of the original elevation of the mountains, but it 
 appears probable that after the mountains were raised to great 
 height another secondary cause contributed to the forces operative 
 in producing the present enormous elevation. This secondary force 
 was nothing else than the soaking tropical rains constantly drench- 
 ing the eastern slope of the mountains. As the earth's crust was 
 already broken and faulted, the leakage of the water downward 
 would be facilitated, while the ceaseless character of the rainfall 
 would make the eastern slope of the Andes to all essential purposes 
 an inland sea. Effectively, therefore, this great range of mountains 
 is built upon a narrow strip of land with seas on both sides like the 
 mountains in Central America, and hence, the violence of th*e vol- 
 canoes and earthquakes becomes more easily intelligible. 
 
 To make this theory more specific we may recall that all the 
 principal peaks about Quito have been volcanic, and three or four 
 volcanoes are still terribly active there now. Just east of Quito, at 
 the head of the Amazon Valley, are the most terrible rainfalls on 
 
 1 Professor Solon I. Bailey, of Harvard Observatory, who crossed the 
 Andes twice, once near Sorata, Bolivia, and again at the Aricoma Pass, and 
 traversed the eastern ran<ye a third time through the river Urubamba, writes 
 me that his impression is that the eastern slope is two or three times steeper 
 than the western slope. Few observers have had better opportunities of 
 judging of the general structure of the Central r\ndes than Professor Bailey. 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 331 
 
 the earth ; and this region is so near to Cotopaxi, Sangay and other 
 active vents as to leave little doubt that it contributes to the activity 
 of the volcanoes. In Whymper's account of his " Travels among 
 the Great Andes of the Equator," p. 240, he cites an interesting pas- 
 sage from the " Royal Commentaries of Peru," p. 632, which runs 
 thus : 
 
 " By reason of the continual Rains, and moisture of the Earth, their 
 woolen Clothes and linen being always wet, became rotten, and dropped from 
 their Bodies, so that from the highest to the lowest every Man was naked, 
 and had no other covering than some few Leaves. ... So great, and so 
 insupportable were the Miseries which Gonzalo Pigarro and his Companions 
 endured for want of Food, that the four thousand Indians which attended 
 him in this Discovery, perished with Famine. . . . Likewise of the three 
 hundred and forty Spaniards which entred on this Discovery, two hun- 
 dred and ten dyed, besides the fifty which were carried away by Orellana. 
 . . . Their Swords they carried without Scabbards, all covered with rust, 
 and they walked barefoot, and their Visages were become so black, dry and 
 withered, that they scarce knew one the other; in which condition they came 
 at length to the Frontiers of Quitu, where they kissed the Ground, and re- 
 turned Thanks to Almighty God, who had delivered them out of so many 
 and so imminent dangers." 
 
 In his account of the ascent of Sara-urcu, Whymper says 
 (p. 241): 
 
 . . . The scouts came back with -bad reports. The animals, they said, 
 could go no farther; there was an end to paths and trails, except occasional 
 wild-beast tracks ; there was nothing whatever to eat, and everything must 
 be carried; there was no place to camp upon, the whole country was a 
 dismal swamp ; and everlasting rain was falling ; so much so that, although 
 they supposed they had been near to Sara-urcu, they were quite unable to be 
 sure. ..." 
 
 Pages 241242 : 
 
 ". . . This (food) arrived late, and delayed us so much that we could 
 not reach the next camping-place by nightfall, and had to stop in a swamp, 
 on a spot where, if you stood still, you sank up to the knees in slime. This 
 place was just on the divide, nearly 13,000 feet above the sea, and during the 
 greater part of the eleven hours' night sleet or rain fell, rendering it well-nigh 
 impossible to keep up a fire out of the sodden materials. For me the men 
 constructed a sort of floating bed, cutting down reeds, and crossing and 
 recrossing them, piling them up until they no longer sank in the slime. For 
 themselves they made smaller platforms of a similar description, and sat on 
 their heels during the whole night, trying to keep up a fire. ..." 
 
 Page 242 : 
 
 "... The land was entirely marshy, even where the slopes were con- 
 siderable; and upon it there was growing a reedy grass to the height of 
 
332 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 eight to ten feet, in such dense mass as to be nearly impenetrable. The 
 machetas were found inadequate. It would have taken several weeks' labour 
 of our whole party to have cleared a track over a single mile. The only 
 way of getting through was by continually parting the reeds with the hands 
 (as if swimming), and as they were exceedingly stiff, they sprang back 
 directly we let go, and shut us out of each other's sight. The edges of the 
 leaves cut like razors, and in a short time our hands were streaming with 
 blood, for we were compelled to grasp the stems to prevent ourselves from 
 sinking into the boggy soil. On this day we crossed the divide, and the 
 streams now flowed toward the Atlantic. The whole country was like a satu- 
 rated sponge. ..." 
 
 Page 243 : 
 
 "... everything burnable was dripping with moisture, and the sur- 
 rounding land was so wet that water oozed or even squirted out in jets when 
 it was trodden upon. ..." 
 
 Page 245 : 
 
 "... Rain continued without intermission. No one at Cayambe had 
 spoken about these incessant rains. From the aspect of the country (so 
 different from any other part of Ecuador), from the saturation of the hills, 
 the innumerable small pools, streamlets and springs, I am convinced they 
 .are nearly perpetual. ..." 
 
 In an excellent account of " A New Peruvian Route to the Plain 
 of the Amazon," published in the National Geographic Magazine 
 for August, 1906, Professor Solon I. Bailey, of Harvard College 
 Observatory, describes the route through the Aricoma Pass at an 
 altitude of 16,500 feet, and continues (p. 439) : 
 
 " . . . On reaching the eastern crest of these mountains, if the view 
 is clear, one seems to be standing on the edge of the world. The eye, in- 
 deed, can reach but little of the vast panorama, but just at one's feet the 
 earth drops away into apparently endless and almost bottomless valleys. We 
 may call them valleys, but this does not express the idea; they are gorges, 
 deep ravines in whose gloomy depths rage the torrents which fall from the 
 snowy summits of the Andes down toward the plain. We might hunt the 
 world over for a better example of the power of running water. The 
 whole country is on edge. There all the moisture from the wet air, borne 
 by the trade winds across Brazil from the distant Atlantic, is wrung by the 
 mountain barrier and falls in almost continual rain. 
 
 " Near the summit of the pass only the lowest and scantiest forms of 
 vegetable life are seen. In a single day, however, even by the slow march 
 of weary mules, in many places literally stepping ' downstairs ' from stone 
 to stone, we drop 7,000 feet. Here the forest begins, first in stunted growths, 
 and then, a little lower down, in all the wild luxuriance of the tropics, where 
 moisture never fails. The lower eastern foot-hills of the Andes are more 
 heavily watered and more densely overgrown than the great plain farther 
 
I9 o6.] SEE THE CAUSE OF EARTHQUAKES. 333 
 
 down. Here is a land drenched in rain and reeking with mists, where the 
 bright sun is a surprise and a joy in spite of his heat. In these dense forests, 
 with their twisting vines and hanging lianas, a man without a path can 
 force his way with difficulty a mile a day. ..." 
 
 From these accounts, and from the greater steepness of the 
 eastern side of the Andes, I think it clear that the perpetual rain 
 which falls there sinks down as if the country were overlaid by a 
 deep sea, and thus in effect the Andes are on a narrow strip between 
 two oceans. Hence the terrific effects of the volcanic forces, which 
 have not only upraised the peaks and chains, but also plateaus like 
 that of Titacaca. While the western ocean furnished the forces 
 for the original uplift of the chain, the reeking tropical rains must 
 have augmented these forces in the later stages of the Andean de- 
 velopment, and this amply accounts for the activity of the volcanoes 
 in the eastern range. These* volcanoes might, it is true, be accounted 
 for by the leakage of the ocean, yet it seems probable that the 
 enormous surface rainfall can hardly fail to increase the volcanic 
 violence where the range is already formed, and the rocks broken 
 and tilted to permit of a maximum seepage of the ceaseless tropical 
 rains which constantly soak the eastern side of the mountains. 
 
 23. On the process involved in the elevation of the Alps. 
 
 In the "Face of the Earth," Vol. II, p. 121, Professor Suess 
 shows that " the prevailing tangential movement in the Alps and 
 Pyrenees " is towards the north. Thus we see that the plains of 
 Lombardy and the Valley of the Po constituted the sea valley which 
 was most active in exerting the northern tangential thrusts in fold- 
 ing the Swiss Alps. Professor Suess (Vol. I, p. 274) remarks on 
 the similarity of the Swabian-Franconian sunken area north of the 
 Alps, and the depression of the Adriatic to the south. An exami- 
 nation of almost any good map will convince anyone that the 
 Adriatic once covered the whole of the valley of the Po. The 
 maps given by Reclus in his large work " La Terre," p. 184, show 
 the jutting spurs radiating from the plains of Lombardy into the 
 surrounding mountains on all sides. This result, therefore, is a very 
 happy confirmation of the theory. It is also satisfactory to find 
 that the Pennine range of the Alps, nearest the valley of the Po, 
 including Mt. Blanc, is the highest of these great mountains. The 
 highest range of the Himalayas, including Mt. Everest, also stands 
 
334 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 nearest the valley of the Ganges; and the process of mountain- 
 formation seems to be such that when several successive ranges are 
 formed and completed that nearest the sea is the highest, probably 
 because the forces there became most nearly vertical. 
 
 In Vol. II, p. 552, Professor Suess says : " As a result of tangen- 
 tial thrusts the sediment of this (Mediterranean) Sea was folded to- 
 gether and driven upwards as a great mountain range, and the Alps 
 have, therefore, been described as a compressed sea." 
 
 From these considerations, it is obvious that the Alps were 
 injected from several sides, but especially from the side of the 
 valley of the Po. In this way, the successive ranges of the Alps 
 were formed, probably beginning near the north, and working 
 southward; and thus we see that the valleys of Switzerland are the 
 results of this successive wrinkling of the crust. Many lakes were 
 formed in the Alps, and. these contributed their part to the final 
 shaping of the contours of the country. As the country was sur- 
 rounded by seas and traversed by many valleys and lakes, all of 
 which gave the water access to the bowels of the earth where breaks 
 of the rocks were once started, the movements finally became very 
 complex, and hence the great difficulty of unraveling the tangled 
 skein of Alpine development. In no other way than this could such 
 a system of mountains have arisen. The average height of the 
 Alpine region is about 4,000 feet, but instead of a level tableland, 
 it is a mass of broken chains and valleys, showing great horizontal 
 crumpling, and also conspicuous and uneven vertical uplifts. The 
 Alps, therefore, afford one of the best illustrations of the theory. 
 For details of the various valleys of the Alps, and the great faults 
 which mark these sunken areas, one may consult Suess, Vol. I, p. 
 200, et seq. 
 
 24. On the origin of the Himalayas and of the Plateau of Tibet. 
 
 If we study the general character of the Himalayas by means 
 of the excellent map given in the article, " India," Encyclopedia 
 Britannica, ninth edition, we shall find that the most conspicuous 
 feature of this great chain is the prominence of the jutting spurs on 
 the south, facing the ocean. Perhaps a few remarks ought to be 
 made about the process by which these spurs originate. 
 
 We have seen that it was by the injection of lava from the sea 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 335 
 
 s 
 
 p 
 X) 
 
 valleys now occupied by the Ganges and Bramaputra that these 
 mighty mountains were upheaved. In this process there would 
 naturally be certain paths or outlets under the crust, along which 
 the lava would escape most easily ; and these outlets would depend 
 upon the upheaval of the overlying strata. A break in the strata 
 perpendicular to the direction of the chain would develop into a 
 
336 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 spur, 1 because the upheaval would then become easy, and more and 
 more lava would be forced into the outlet thus afforded. By means 
 of the earthquakes accompanying the expulsion of lava from the 
 sea valley the crust is broken at various points, and thus the spurs 
 are gradually developed. It is noticeable in all the great chains 
 that these spurs are directed towards the sea. In the Himalayas, 
 for example, they are on the convex side, while in the Sierra Ne- 
 vadas they are on the concave side of the chain ; which shows that 
 the spurs are not due to any process of shrinkage of the earth. How- 
 ever the chains may curve, the spurs will be found towards the 'sea 
 valleys from which the expulsions of lava have taken place. This 
 arrangement of the spurs shows the process of mountain formation 
 very clearly. 
 
 If we examine the map of Tibet given in the Encyclopedia 
 Britannica, ninth edition, we shall find these spurs mainly on the 
 outside of this great tableland. It was therefore injected from the 
 seas on the north as well as on the south; for the Kuen Lun and 
 Altin Tagh mountains on the north show the spurs almost as dis- 
 tinctly as the Himalayas, on the south ; and on the north of Tibet 
 the spurs point towards the Arctic ocean. Tibet was thus upheaved 
 by forces injecting this great tableland on all sides. It is thus a 
 very elevated plateau, enclosed by terribly high mountains. Before 
 the upheaval had attained such great height, no doubt the enormous 
 rainfall produced by these mountains, and hence sinking down in 
 the faults thus opened to the bowels of the earth, contributed greatly 
 to the uplifting forces due to the surrounding seas. In this way one 
 may account easily and naturally for the gradual upheaval of the 
 highest plateau in the world. 
 
 The forces depending on the valleys of the Ganges and Brama- 
 putra are still active. And the bones of elephants and rhinoceroses 
 now found 15,000 feet above the sea, an elevation at which these 
 animals could not possibly have lived (cf. article " Himalayas," 
 Ency. Brit., by Gen. Strachey), show that the vertical uplift of Tibet 
 took place in comparatively recent geological times. When this 
 tableland had an elevation of a mile or less, it was no doubt inhabited 
 
 1 In 7 we have cited Humboldt's remarks about volcanoes in parallel 
 ranges being connected by cross ranges, forming mountain-nodes. These 
 are similar to spurs, and hence we see why eruptions occur at such points. 
 
SEE THE CAUSE OF EARTHQUAKES. 
 
 337 
 
 by -numerous large quadrupeds ; but after it attained an altitude of 
 two miles they had naturally deserted it ; and now, at an altitude of 
 nearly three miles we find only the bones to show that it was once 
 habitable by such animals as now flourish in the low plains of India. 
 
338 SEE THE CAUSE OF EARTHQUAKES. [October , 9t 
 
 The terrific violence of the earthquakes which are still felt in northern 
 India gives us an idea of the amazing power of the subterranean 
 forces by which this great uplift was produced. In elevating the 
 highest mountains of the world, it is not remarkable that they also 
 raised the highest plateau. 
 
 In his work on the " Face of the Earth," Vol. I, p. 452, Professor 
 Suess remarks with surprise that the tangential movement shown 
 by the mountains has opposite direction on the two sides of the 
 Bramaputra ; this confirms the present theory in the most conclusive 
 manner. 
 
 25. On the elevation of plateaus. 
 
 It will be seen that in this paper we abandon contraction as 
 the principal modifying cause and consider only the light matter 
 proved to be injected under mountains and narrow plateaus by the 
 power of steam, which gives all these masses a substratum of honey- 
 combed material identical with the lighter lava and denser pumice. 
 That this material now lies under these regions of greatest elevation 
 seems certain, and in place of the contraction theory, we may sub- 
 stitute that of steam expansion and solidification, upon which moun- 
 tain building depends. This is, of course, effected in accordance 
 with Henry's law of gaseous absorption, and the still more general 
 law that matter adapts itself to the pressure to which it is subjected. 
 
 It will be seen that all the mathematical reasoning is the same 
 whether we suppose greater contraction under the oceans, or an 
 actual heaving up of the land areas by the injection of light vol- 
 canic material more or less full of bubbles, which decreases the 
 average specific gravity of the land and mountains. 
 
 Pendulum observations made at many points on land and sea also 
 give the same indication as respects the earth's arrangement of 
 density ; and if the present view be sound the interpretation of these 
 observations will now become more obvious. 
 
 If we apply the foregoing theory to a narrow plateau, like that 
 of Mexico, which is a uniform tableland, with mountains on both 
 sides, it will become evident that the mountains were formed when 
 they were near the sea ; and that when the shore line had remained 
 fixed there for a long time it again receded, after the elevation of 
 the mountains and the tableland, both of which probably were 
 
1906.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 339 
 
 injected with light volcanic materials. Hence the substratum of 
 such a plateau is composed of light materials similar to porous lava 
 and pumice. Only a few mountains, as Orizaba, Popocatapetl, 
 Colima, Jorullo, experienced such violent forces as to come to 
 eruption, but the materials under all of them are the same. The 
 detritus of ashes which the volcanoes eject is nothing but pumice 
 ground to dust by friction and explosive violence. It is obvious 
 that the pumice, ashes and other volcanic materials are not made at 
 the time of ejection, but are always in store, and simply happen to 
 be thrown out of any mountain which experiences eruption. We 
 may in all probability conclude that the plateaus of Tibet, Quito, 
 Caxamarca, Cuzco, Titicaca, and other elevated regions are under- 
 laid by light materials like that thrown from volcanoes. This is 
 certainly true of the great ridges of the Himalayas, Alps, Andes 
 and other mountains which form cordilleras. In fact, the ridges 
 of cordilleras always rest upon a substratum of light volcanic 
 materials, and if they could be exploded from within they would 
 dispense ashes, pumice and scoriae as abundantly as any volcano. 
 The quantity of this light material forced up under the ridges of 
 mountains depends upon the elevation, and breadth, and is thus enor- 
 mous in our highest ranges like the Andes and Himalayas. In this 
 way we may explain many of the anomalies of geodesy, and pen- 
 dulum observations, without any other hypothesis. Is not the ac- 
 cordance of observation with so simple a theory the best proof that 
 the result represents a general law of nature ? 
 
 26. On the theory that large segments of the lithosphere act as 
 units and squeeze those segments which lie between. 
 
 In the new work on " Geology " by Chamberlin and Salisbury, 
 which represents the trend of current geological thought, the authors 
 adopt a subdivision of the earth into large segments which are sup- 
 posed to act as units. The contraction theory is the basis of the 
 reasoning, and part of the discussion is as follows : 
 
 " The downward movements are unquestionably the primary ones, and 
 the horizontal ones are secondary and incidental. The fundamental feature 
 is doubtless central condensation actuated by gravity, and the master move- 
 ments are the sinking of the ocean-basins. The great periodic movements 
 that made mountains and plateaus, and changed the capacity of the ocean- 
 basins, probably started with the sinking of part or all of the ocean-bottoms. 
 
340 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 In the greater periodic movements, probably all the basins participated more 
 or less, but some seem to have been more active than others. For example, 
 in the last great mountain-making period, the Pacific basin seems to have 
 been more active than the Atlantic, while in the similar great event at the 
 close of the Paleozoic, the opposite seems to have been true. The squeezing 
 up of the continents doubtless took place simultaneously with the settling of 
 the basins. The true conception is perhaps that the ocean-basins and con- 
 tinental platforms are but the surface forms of great segments of the litho- 
 sphere, all of which crowd towards the center, the stronger and heavier seg- 
 ments taking precedence and squeezing the weaker and lighter ones between 
 them. The area of the more depressed or master segments is almost exactly 
 twice that of the protruding or squeezed ones. This estimate includes in the 
 latter about 10,000,000 square miles now covered with shallow water. The 
 volume of the hydrosphere is a little too great for the true basins, and it 
 runs over, covering the borders of the continents. The amount of the over- 
 flow fluctuates from time to time, and may be neglected in a study of the 
 movements and deformation of the lithosphere." 
 
 Among the major group of squeezed segments we find: (i) 
 Eurasia, (2) Africa, (3) North America, (4) South America; and 
 the minor group includes Australia, Antarctica, the East Indian 
 platform, and Greenland. The depressed or master segments are the 
 oceans: (i) The Pacific, (2) the Indian, (3) the North Atlantic, 
 (4) the South Atlantic ; and a minor group of smaller seas, as the 
 Arctic, the Mediterranean, the Caribbean, and the chain of deep 
 pits between the Philippines and the platform of Borneo. 
 
 The authors discuss the crowding of these segments towards the 
 center and the crumpling which follows along their edges, accom- 
 panied by fracture and slipping. 
 
 "If these segments be regarded as the great integers of body-movement, 
 two-thirds of them taking precedence in sinking and the other third in suffer- 
 ing distortion, it is easy to pass to the conception of subsegments, moving 
 somewhat differently from the main segments, so as to aid in their adjust- 
 ment to one another, and thus to the conception of plateaus and deeps. It 
 is easy also to pass to the conception of mutual crowding and crumpling at the 
 edges of these segments, accompanied by fracture and slipping. These con- 
 ceptions perhaps represent the true relations between the massive move- 
 ments of the abysmal and continental segments, as well as the less massive 
 plateau-forming movements and the mountain-forming distortions. The 
 mountains and plateaus are probably the incidental results of the great 
 abysmal and continental readjustments. 
 
 " The great movements are probably to be attributed to stresses that 
 gradually accumulated until they overcame the rigidity of the thick massive 
 segments involved, and forced a readjustment. In accumulating these 
 stresses, some local yielding on weak lines and at special points was an 
 
1906.] SEE THE CAUSE OF EARTHQUAKES. 341 
 
 inevitable incident in distributing more equably the accumulating stresses. 
 So, also the first great readjustments probably left many local strains and 
 unequal stresses which gradually eased themselves by warpings, minor fault- 
 ings, etc., so that some minor movements were a natural sequence of the 
 great movements." 
 
 From these citations one gets a good idea of the explanations 
 heretofore considered most plausible. 
 
 Let us now examine this reasoning a little more closely. Is it 
 mechanically conceivable that large segments of the lithosphere should 
 thus act together as units? Would not such action imply that the 
 the earth is cut deep down into the lithosphere and free to move 
 along the boundaries of these severed segments? If the lithosphere 
 were thus cut up into pieces, and all the segments could be regarded 
 as solid, with abundant lubricating oil between them, they might 
 by their mutual gravity crowd towards the center, and possibly the 
 small ones would be squeezed between the larger. But when we 
 recall that the earth is not cut up in this way, with oil between the 
 pieces, but is one unbroken mass, it is clear that the greatest resis- 
 tance would arise to relative motion of the parts. The friction of 
 one part against the other would be so amazingly great that no 
 motion would take place at the supposed joint. Even if motion oc- 
 curred it is impossible to see how the crust could be crumpled without 
 producing a corresponding condensation of the matter underlying 
 the folded area, the density of which would thus be greatly increased, 
 in some cases by 50 per cent. Now geodetic observations show that 
 the density of the matter under the mountains not only is not greater 
 than the average, but actually less; and thus we see that the folds 
 of the crust could not be produced in this way. For in the first 
 place, motion could not take place on account of friction; and in 
 the second place, if it took place, the resulting condensation of the 
 underlying matter would become sensible to geodetic measurement, 
 which is contrary to observation. 
 
 Is it not therefore impossible to entertain the doctrine of large seg- 
 ments of the lithosphere moving together and squeezing others 
 between them? Should not this whole theory, along with the hy- 
 pothesis of contraction in general, be entirely given up? 
 
342 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 VI. CONFIRMATION OF THE FOREGOING THEORY OF MOUNTAIN 
 BUILDING BY GEODESY. 
 
 27. General considerations on the attraction of mountains. 
 
 If the theory above outlined be admissible it will follow that 
 all mountains are filled with light material like porous lava, which 
 was originally injected as hot lava under great steam pressure, full 
 of bubbles, and is thus similar to dense pumice. When the lava 
 solidified, the bubbles dried up, and left a honeycombed structure 
 of great strength, but comparatively small weight. The deep in- 
 terior of all large mountains should be filled with material of this 
 kind, which gives strength, but has low density, and the effect of its 
 injection is to reduce the average density of the mountain below 
 that of an equal mass of similar material taken from a plain. 
 
 It is remarkable that although some long tunnels through moun- 
 tain chains have been bored by human effort, none of them has gone 
 more than a mile or so deep, and thus all are too shallow to give us 
 any experimental knowledge of the materials in the depths of the 
 mountains. No large mountain has been eroded to great depth, 
 and hence our only knowledge of the interior of mountains is de- 
 rived from volcanic action, which blows out the inner portions of 
 some of them. The mountains which happen to break out as vol- 
 canoes, either because the subterranean steam pressure is suddenly 
 applied and abnormally great, or because they are weak in some 
 point of their construction, are obviously not different from ordi- 
 nary chains and peaks, until after they break forth. 
 
 Eruptions are always accompanied by violent earthquakes, and 
 a mountain does not take fire and form ashes and cinders by burn- 
 ing, but blows out the volcanic materials already stored up in vast 
 quantities. Thus we naturally infer that all mountains are essen- 
 tially alike, but we are never able to see the inner contents except 
 of the few of them which have become volcanoes. And if we are 
 able to explore the interior of all mountains, it is only by studying 
 the materials expelled from volcanoes by the explosive power of 
 steam. It is doubtful if even the oldest mountains show erosion a 
 mile deep, and hence we cannot penetrate the earth's crust to any 
 depth except by an analysis of the materials which come out of the 
 mountains after they are blown open in eruption. It is scarcely 
 
i 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 343 
 
 necessary to add that new ashes, cinders, scoriae, pumice and other 
 volcanic materials are not as a whole red hot, and therefore the bulk 
 of them is not formed at the time of the first volcanic outbreak. 
 They have been resting quietly under the mountain for immeasurable 
 ages, and one cannot doubt that they exist in all mountains. 
 
 It is unfortunate that a custom has arisen of speaking of vol- 
 canic action as essentially superficial. While it cannot be said to 
 have any fixed depth, and in many cases may not be extremely deep, 
 it is safe to say that, as a general rule, it is by no means very 
 shallow. On the average the depth is at least of the same order as 
 the height of the mountains, plus the depth of the adjacent sea. 
 There is every reason to suppose the forces which pushed up the 
 Andes and keeps some of them in active eruption arises from about 
 the same depth as the most violent earthquakes which visit that 
 region. Only a few of the peaks of the Andes have become vol- 
 canoes, but some of them, as Aconcagua, in Chili, Gualateiri and 
 Sahama, of the Sorata Range in Bolivia, and Cotopaxi and its 
 associates near Quito in Ecuador are all very high ; and the ejection 
 of materials at that height requires deep-seated and most tremen- 
 dous forces. Yet obviously only a small part of the energy origi- 
 nating under the base of a volcano has been spent in ejecting lava, 
 rocks, ashes and cinders; a very large part of the energy exerted 
 from below has been spent in raising the sides of the mountain before 
 eruption broke out at the top. 
 
 We may then consider all mountains to be made up internally of 
 light material like that blown out of volcanoes. Chimborazo, for 
 example, may never have had an eruption, and yet it is impossible 
 to believe that its constitution is essentially different from that of 
 the neighboring mountains of Cotopaxi, Pinchinchi, Antisana, 
 Tunguragua, Sangay, and Cayambe, all of which are volcanoes. 
 
 Now it is a very remarkable fact of observation that geodetic 
 researches in many countries, and pretty much ever since the ex- 
 periments of Bouguer and La Condamaine on Chimborazo in the 
 year 1738, have tended to show that the matter under the roots of a 
 great mountain is lighter than the average matter of the adjacent 
 plain. In their experiments on the deviation of the plumb line 
 near Chimborazo the existence of great cavities in this colossal 
 
 PROC. AMER. PHIL. SOC., XLV. 184 V, PRINTED FEBRUARY 23, 1907. 
 
344 SEE THE CAUSE OF EARTHQUAKES. [October , 9 , 
 
 trachytic mountain was suspected by Bouguer and La Condamaine, 
 though, as Humboldt remarks, the evidence was unsatisfactory 
 (Cosmos, Vol. V, p. 30). Yet the foregoing theory enables us to 
 recognize that the similar widely extended indications of modern 
 geodesy can hardly fail to be real; and according to the principle 
 of continuity the law thus found for certain mountains must be held 
 to be general for all the great peaks and chains of our globe. In 
 fact the satisfactory explanation of this long-standing geodetic prob- 
 lem would seem to be one of the strongest arguments which could 
 be adduced in favor of the truth of the present theory of mountain 
 constitution and formation. For the facts of geodesy are based upon 
 exact observation and experiment, and unbiased by any theory ; and 
 yet they become intelligible only when interpreted by means of a 
 general theory like the present. The support thus furnished by a 
 wide body of geodetical facts would seem to place the theory of 
 mountain constitution on a basis where it may become a useful 
 adjunct to the principles of geodetic measurement. 
 
 Probably no measurements involving long distances upon the 
 earth have so high an order of accuracy as those of geodesy. Is it 
 too much to hope that the principle of mountain constitution here 
 outlined may become the means of still greater refinement, in the 
 triangulation of the earth's surface ? 
 
 In his well-known work on the " Figure of the Earth," x Pratt 
 announces a proposition : 
 
 " To deduce from the previous calculations some probable conclusions re- 
 garding the constitution of the earth's crust. The first thing to be observed 
 in the results given in the last paragraph is the very small amount of the 
 resulting deflection at the two extremities of the Indian Arc Punnae close to 
 Cape Comorin, and Kaliana the nearest station to the Himalayan Mountains; 
 whereas the effect of the Ocean and Mountains has been shown to be very 
 large. This shows that the effect of variations of density in the crust must 
 be very great, in order to bring about this near compensation. In fact the 
 density of the crust beneath the mountains must be less than that below the 
 plains, and still less than that below the ocean-bed." 
 
 Again : 2 
 
 " The circumstance already noticed, that at seven coast stations out of 
 thirteen the deflection is toward the sea, seems to bear testimony to the truth 
 of the theory, that the crust below the ocean must have undergone greater 
 
 " Figure of the Earth," 3d edition, p. 134. 
 2 Ibid., p. 136. 
 
I9 o6.] v SEE THE CAUSE OF EARTHQUAKES. 345 
 
 contraction than other parts. The deflection toward the land at the other 
 six coast-stations can of course easily be understood without at all calling in 
 question the theory. The proximity of the land may easily be conceived 
 sufficient to counteract any effect of the more distant parts of the ocean. It 
 is the fact of even some of the deflections being toward the sea, that bears 
 testimony to the theory, while the others offer no argument to the contrary." 
 "The least, then, that can be gathered from the deflections of these 
 coast-stations is, that they present no obstacle to the theory so remakably 
 suggested by the facts brought to light in India, viz., that mountain-regions 
 and oceans on a large scale have been produced by the contraction of the 
 materials, as the surface of the earth has passed from a fluid state to a con- 
 dition of solidity the amount of contraction beneath the mountain-region 
 having been less than that beneath the ordinary surface, and still less than 
 that beneath the ocean-bed, by which process the hollows have been produced 
 into which the ocean has flowed. In fact the testimony of these coast-sta- 
 tions is rather in favour of the theory, as they seem to indicate, by excess of 
 attraction towards the sea, that the contraction of the crust beneath the 
 ocean has gone on increasing in some instances still further since the crust 
 became too thick to be influenced by the principles of floatation, and that 
 an additional flow of water into the increasing hollow has increased the 
 amount of attraction upon stations on its shores." 1 
 
 We have quoted this paragraph to recall how geodesists have 
 heretofore attempted to explain the comparative lightness of the 
 matter beneath the mountains, and the greater density of that 
 beneath the plains and especially beneath the sea. 
 
 Pratt elsewhere says that the hidden cause of the deflection of 
 the plumb line between the mountain masses on the north and the 
 Indian Ocean on the south " may lie below, in the variations of the 
 density of the earth's crust." 2 He then examines various arrange- 
 ments of density, and finally calculates by three hypotheses that 
 the sea level in Great Britain stands at a mean height of 1,567 
 feet higher than it would if the ocean hemisphere with pole in New 
 Zealand became land, other things remaining equal. 
 
 28. Pratt' 's theorem on the equilibrium of the earth between the 
 land and water hemispheres? 
 
 "There is no doubt that the solid parts of the earth's crust, beneath the 
 Pacific Ocean must be denser than in the corresponding parts on the op- 
 posite side, otherwise the ocean would flow away to the other parts of the 
 earth. The following reasoning will explain this. Suppose the earth to be 
 a sphere. Through any point on it suppose a surface drawn separating a 
 
 1 " Figure of the Earth," 3d edition, p. 137. 
 
 2 Ibid., p. 148. 
 
 3 Ibid., pp. 159-160. 
 
346 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 thin portion on the right hand and through the same point a similar surface 
 separating a like portion on the left. The sphere consists, then, of three 
 parts, the middle portion being of a symmetrical form and attracting the point 
 in the direction of the radius, and the two slender slices attracting it equally 
 to the right and left of that radius. If one of these slices became fluid and of 
 less density than the other, its attraction would be overcome by that of the 
 other, and the fluid would be drawn away to the other parts of the sphere. 
 
 " It does not follow that the whole of the fluid would be drawn over. The 
 above process would go in till the surface of the fluid at the circumference of 
 the slice had become so inclined as to be at right angles to the direction of 
 the resultant attraction of the whole mass, solid and fluid. If, however, a 
 narrow channel were cut through this circumference (which would otherwise 
 act as an embankment) the whole of the water would be drawn off. 
 
 " Now in the case of the earth there is a channel opening a passage from 
 the New Zealand hemisphere into the opposite one, viz., the North -and South 
 Atlantic, and yet the ocean remains in that hemisphere. There must, there- 
 fore, be some excess of matter in the solid parts of the earth between the 
 Pacific Ocean and the earth's centre which retains the water in its place. 
 This effect may be produced in an infinite variety of ways; and therefore, 
 without data, it is useless to speculate regarding the arrangement of matter 
 which actually exists in the solid parts below." 
 
 Pratt then discusses the original fluidity of the earth, and says 
 finally : 
 
 "The conclusion at which we have arrived in Art. 132, that the parts 
 of the crust below the more elevated regions are of less density, and the 
 parts beneath the depressed regions in the oceans are of greater density than 
 the average portions of the surface, seems to bear additional testimony to the 
 fluid theory. For it shows, that notwithstanding the varied surface, seen 
 at present in mountains and oceans, the amount of matter in a vertical 
 prism drawn down at various places to any given spheroidal stratum is the 
 same although its length varies from place to place as the earth's contour 
 varies. No better explanation of this phenomenon can be conceived than 
 that which the fluid hypothesis furnishes; viz., that these prisms though now 
 of different lengths, were, when the crust was fluid, of the same length ; and 
 as their lengths are now various simply from the fact that the surface in 
 solidifying has contracted unequally, of course the amount of matter which 
 they contain is the same in all of tljem." 1 
 
 The views here expressed are very generally recognized by 
 geodesists, and have been treated in different ways by Clarke, 
 Tisserand, Helmert, and other leading authorities. 
 
 29. Results obtained by the United States Coast and Geodetic 
 Survey from recent researches on the average depth of isostatic 
 compensation. 
 
 1 " Figure of the Earth," 3d edition, p. 162. 
 
I9o6 .j SEE THE CAUSE OF EARTHQUAKES. 347 
 
 With the approval of Dr. O. H. Tittman, superintendent, Mr. 
 J. F. Hayford has recently presented to the Washington Academy of 
 Sciences and published in the Proceedings for May 18, a valuable 
 summary of the results deduced from the geodetic operations in 
 the United States. He has discussed with care the long series of 
 deflections of the vertical as determined by Clarke's Standard 
 Spheroid of 1866, which has been found to fit best for the area 
 covered by the United States, and determined the depths at which 
 isostatic compensation takes place. By five solutions of the prob- 
 lem he finds the residuals least for a depth of 71 miles. Some of the 
 conclusions are stated thus : 
 
 " The evidence shows clearly and decisively that the assumption of com- 
 plete isostatic compensation within the depth of 71 miles is a comparatively 
 close approximation to the truth, that the assumption of extreme rigidity is 
 far from the truth that the United States is not maintained in its position 
 above the sea level by the rigidity of the earth, but is, in the main, buoyed up, 
 floated, upon underlying material of deficient density. 
 
 "The conclusions just stated were based upon the 507 residuals con- 
 sidered as one group. The residuals have been examined in separate groups 
 of 25, each group covering a small region. Not a single group of 25 con- 
 tradicts the conclusion just stated. 
 
 "It is certain that for the United States and adjacent regions including 
 oceans, the isostatic compensation is more than two-thirds complete perhaps 
 much more. 
 
 " The departure from perfect compensation may be, in some regions, in 
 the direction of overcompensation rather than undercompensation but in 
 either case the departure from perfect compensation is less than one-third. 
 
 " In terms of stresses, it is safe to say that these geodetic observations 
 prove that the actual stresses in and about the United States have been so 
 reduced by isostatic adjustment that they are less than one-tenth as great 
 as they would be if the continent were maintained in its elevated position, 
 and the ocean floor maintained in its depressed position, by the rigidity of 
 the earth." 
 
 Mr. Hayford assumed that the compensation is uniformly dis- 
 tributed with respect to the depth, but he remarks that this is only a 
 convenient working hypothesis. Pendulum observations combined 
 with deflection observations, he thinks, may furnish the means of 
 detecting the distribution of compensation. Assuming that the 
 compensation all occurs within a stratum 10 miles thick, the bot- 
 tom of the stratum of isostatic compensation comes out 37 miles. 
 Mr. Hayford, however, prefers the depth of 71 mijes, though he 
 adds that it rests upon an insecure foundation. He concurs with 
 
348 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 the view now generally held that compensation is always going on, 
 through the changes of levels due to erosion and other causes. 
 Mr. Hayford's paper is extremely important in showing that the 
 depth of compensation is not great, and the principal inequalities 
 of density are essentially limited to the earth's crust, complete aver- 
 age compensation for the whole country occurring within about 
 three or four times that depth. The result, therefore, accords well 
 with the views developed in the present investigation. In his dis- 
 cussion of Hayford's paper, Major Button, who was present, re- 
 marked that " the heavy masses of sediment which are formed upon 
 the bottom of the sea can, I conceive, only be elevated by a positive 
 uplifting force." 
 
 VII. THE ORIGIN OF ISLANDS AND THE SECULAR MOVEMENT OF 
 
 THE STRAND. 
 
 30. Submarine earthquakes and volcanoes. 
 
 The eruption of volcanoes from the sea was noted by the Greeks, 
 as we learn from Aristotle (Meteor., ii, 8, 17-19), quoted by Hum- 
 boldt in Cosmos, Vol. I, p. 240, Bohn's translation: 
 
 "The heaving of the earth does not cease till the wind (dvquof) which 
 occasions the shocks has made its escape into the crust of the earth. It is 
 not long ago since this actually happened at Heraclea in Pontus, and a 
 similar event formerly occurred at Hiera, one of the Aeolian Islands. A por- 
 tion of the earth swelled up, and with loud noise rose into the form of a hill, 
 till the mighty urging blast (Trvev^a) found an outlet, and ejected sparks and 
 ashes which covered the neighborhood of Lapari, and even extended to sev- 
 eral Italian cities." 
 
 Strabo (lib. i, p. 59) describes the naming eruption observed at 
 Methone in the year 282 B. C, near the town in the Bay of Her- 
 mione, saying a fiery mountain arose seven stadia in height, inac- 
 cessible by day on account of its heat and sulphurous flames, but 
 emitting an agreeable odor at night. It was so hot that the sea 
 boiled for a distance of five stadia, and was turbid and filled with 
 detached masses of rock for full twenty stadia. This eruption is 
 also mentioned by Ovid 1 (Metam., xv). 
 
 " Near Troezen is a tumulus, steep and devoid of trees, once a plain, 
 now a mountain. The vapours enclosed in dark caverns in vain seek a pas- 
 sage by which they may escape. The heaving earth, inflated by the force 
 of the compressed vapours, expands like a bladder filled with air, or like a 
 goat skin. The ground has remained thus inflated and the high projecting 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 349 
 
 Another island eruption on Thera occurred forty-five years 
 later, and the t ancients were perhaps familiar with additional sub- 
 marine outbreaks on the Italian coast ; otherwise probably few if any 
 more submarine volcanoes were noted till modern times. 1 
 
 A few submarine volcanoes, either new or of recent origin, were 
 noted by the early navigators, but the phenomenon of volcanic 
 upheavals at sea did not attract much attention till the time of 
 Humboldt, who made a careful study of all such outbursts. In the 
 Cosmos Humboldt mentions five submarine volcanoes which ap- 
 peared in the first half of the nineteenth century. Probably it 
 would be difficult to give a complete list of all the submarine vol- 
 canoes which have ever appeared, as the records are widely scat- 
 tered, and to be of value would have to be critically examined in 
 each case. It is also clear that more such volcanoes would now be 
 noticed than in former centuries, because the travel and explora- 
 tion of the sea is more extensive than in former ages ; yet even now 
 great areas of the ocean are quite seldom visited by ships, so that 
 the world knows little or nothing of what is going on in a large part 
 of the globe. If a submarine volcano of short duration should be 
 upheaved, it might again disappear and leave no record of its ex- 
 istence. Numerous islands are also raised in the sea without visible 
 volcanic outlets. The average number of submarine volcanoes 
 upheaved in a century, if careful watch could be kept on the oceans 
 throughout the world, might prove to be something like one every 
 two years, or 50 in a century. This, however, is merely the num- 
 ber of peaks which would raise their heads above the water. As 
 most of the oceans are much too deep for them to show above the 
 sea level, it is clear that the number which remain covered by the 
 sea is very much greater than those which rise into view, in a ratio 
 of perhaps something like 20 to I. Thus in the whole earth there 
 may be, and probably are, 1,000 submarine volcanoes erupted in a 
 century, or an average of ten in a year. In this way we may ex- 
 plain some of the great sea waves so often encountered by navi- 
 gators, and frequently noted by the tide gauges in civilized coun- 
 
 eminence has been solidified by time into a naked rock" (Humboldt's, 
 Cosmos, Vol. I, p. 239). 
 
 1 The list island chances in the Mediterranean recorded by Pliny in his 
 Natural History is quite impressive (cf. Lib. ii.). 
 
350 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 tries. Not all volcanoes thus upheaved would be sufficiently large 
 and energetic in their action to be widely or distinctly felt, because 
 some would be small and slow in their action and hence produce 
 no sensible sea waves. Yet when we consider how much of the sea 
 passes unobserved, and how many sea waves actually are reported, 
 an average of ten submarine volcanoes per year does not seem too 
 high an estimate. Those which are small and in deep water will 
 escape notice entirely; and in.midocean none except the powerful 
 eruptions could make themselves felt, because there is no means of 
 detecting a disturbance of sea level, unless it happens to be large 
 enough to disturb a passing ship. 
 
 In his well-known article on " Geology " in the Encyclopedia 
 Britannica, ninth edition, Sir Archibald Geikie says : 
 
 "At the Hawaii Islands, on 25th February, 1877, masses of pumice, 
 during a submarine volcanic explosion, were ejected to the surface, one of 
 which struck the bottom of a boat with considerable violence and then floated. 
 At the same time when we reflect to what a considerable extent the bottom 
 of the great ocean basins is dotted over with volcanic cones, rising often 
 solitary from profound depths, we can understand how large a proportion of 
 the actual eruptions may take place under the sea. The foundations of these 
 volcanic islands doubtless consist of submarine lavas and fragmentary ma- 
 terials, which, in each case, continued to accumulate to a height of two or 
 three miles, until the pile reached the surface of the water and the phenomena 
 became subaerial. The immense abundance and wide diffusion of volcanic 
 detritus over the bottom of the Pacific and Atlantic Oceans, even at dis- 
 tances remote from land, as has been made known by the voyage of the 
 ' Challenger/ may indicate the prevalence and persistence of submarine 
 volcanic action, though at the same time, it must be admitted that an ex- 
 tensive diffusion of volcanic debris from the islands is effected by winds 
 and ocean-currents." 
 
 ^ evidence therefore that the sea bottom is very leaky and in 
 a constant state of eruption in many places seems conclusive. In 
 other places very few eruptions occur, because the underlying rocks 
 are less leaky and the sea is too shallow to exert much pressure. 
 The number of earthquakes under the sea must be much greater 
 than those on land, area for area ; but here again few are observed, 
 and still fewer traced to their centers, because the data are in- 
 sufficient. 
 
 It is fully recognized, however, that many submarine earth- 
 quakes occur off the east coast of Japan and the Philippines, in the 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 351 
 
 Aleutian Islands, off the coast of Peru and Chili, near St. Paul's 
 Island in the Atlantic, the Antilles and the East Indies. And 
 while the investigation of these phenomena is still in its infancy, 
 we may be sure they are of wide and universal distribution, espe- 
 cially in the deepest oceans, where the shores are steep and broken, 
 as in volcanic regions. The southern Pacific Ocean, where so 
 many volcanic islands exist, must be a prolific but unexplored 
 center of such disturbances. 
 
 31. Views of Charles Darwin on the distribution of volcanic 
 islands. 
 
 " During my investigations on coral reefs, I had occasion to consult the 
 works of many voyagers, and I was invariably struck with the fact that with 
 rare exceptions, the innumerable islands scattered throughout the Pacific, 
 Indian, and Atlantic Oceans, were composed either of volcanic or of 
 modern coral-rocks. It would be tedious to give a long catalogue of all 
 the volcanic islands, but the exceptions I have found are easily enumerated: 
 In the Atlantic we have St. Paul's Rock, described in this volume, and the 
 Falkland Islands, composed of quartz and clay slate ; but these latter 
 islands are of considerable size, and lie not very far from the South Amer- 
 ican coast: in the Indian Ocean, the Seychelles (situated in a line prolonged 
 from Madagascar) consists of granite and quartz : in the Pacific Ocean, New 
 Caledonia, an island of large size, belongs (as far as is known) to the 
 primary class. New Zealand, which contains much volcanic rock and some 
 active volcanoes, from its size cannot be classed with the small islands, 
 which we are now considering. The presence of a small quantity of non- 
 volcanic rock as of clay slate on three of the Azores, or of the tertiary lime- 
 stone of Maderia, or of clay-slate at Chatham Island in the Pacific, or of 
 lignite at Kerguelen land, ought not to exclude such islands or achipelagoes, 
 if formed chiefly of erupted matter, from the volcanic class. 
 
 "The composition of the numerous islands scattered through the great 
 oceans with such rare exceptions volcanic, is evidently an extension of that 
 law, and the effect of the same causes, whether chemical or mechanical, from 
 which it results, that a vast majority of the valcanoes now in action stand 
 either as islands in the sea, or near its shores. This fact of the oceanic is- 
 lands being so generally volcanic is also interesting in relation to the nature 
 of the mountain-chains on our continents, which are comparatively seldom 
 volcanic; and yet we are led to suppose that where our continents now 
 stand an ocean once extended. Do volcanic eruptions, we may ask, reach 
 the surface more readily through fissures formed during the first stages of the 
 conversion of the bed of the ocean into a tract of land ? " * 
 
 In connection with the above views of Darwin it should be noted 
 that not all of the islands which are upheaved in the sea could be 
 1 " Geological Observations on Volcanic Islands," chapter VI. 
 
352 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 expected to pour forth molten rock so as to give above the water 
 a volcanic aspect. In some cases the islands would be raised with- 
 out eruption breaking out at all, while in others the eruptive out- 
 breaks would occur beneath the sea, and the surface appearance of 
 such islands would be non-volcanic. 
 
 32. Dr. Rudolph's views on submarine earthquakes and erup- 
 tions. 
 
 In the first volumes of " Beitrage zur Geophysik " (especially 
 Vols. I and II) Dr. E. Rudolph has several elaborate papers dealing 
 with the subject of submarine earthquakes and eruptions in an 
 exhaustive manner. The large catalogues of such phenomena col- 
 lected in this careful inquiry are certainly impressive, and the 
 results of Dr. Rudolph's investigations may be said to confirm the 
 view here taken that the sea bottom is in a constant state of vol- 
 canic disturbance. These disturbances are in fact so frequent that 
 a similar view is held by many experienced navigators as the result 
 of their own observations. What explanation can be given for such 
 phenomena except the penetration of the sea water into the earth's 
 crust? Volcanic outbreaks never occur except along the shores 
 or under the sea. The significance of these disturbances of the sea 
 bottom and their connection with the formation of islands seems 
 therefore obvious. In the " Face of the Earth," Vol. I, p. 61, 
 Professor Suess remarks on the calm attitude of the scholar in the 
 midst of dangerous natural phenomena, and gives the following 
 interesting account of Apollonius of Tyana: 
 
 " In the year 62 or 65 A. D. Apollonius of Tyana was on the island of 
 Crete. He was on that coast of the island which is washed by the Libyan 
 Sea, on a promontory in the neighborhood of Phastus, and was engaged in 
 conversation with a number of men who had come to do honor to the 
 sanctuary on the promontory, when suddenly an earthquake took place. The 
 roar of the thunder, says Philostratus, did not proceed from the clouds, but 
 came from the depths of the sea, and the sea retired at least seven stadia, 
 so that the crowd were afraid that in its retreat it would carry the temple 
 with it, and wash them all away. Apollonius however said: 'Be com- 
 forted; the sea has brought forth new land! A few days later they heard 
 that a new island had risen between Thera and Crete." 
 
 This account not only illustrates the character of a philosopher, 
 but also shows that even Apollonius recognized that the sea brought 
 forth new land. 
 
I9 o6.] SEE THE CAUSE OF EARTHQUAKES. 353 
 
 33. On the upbuilding of the smaller areas by gradual upheaval. 
 
 We have already studied the process of injection from under the 
 bed of the sea and have seen that the mountains have been upheaved 
 in this way. 
 
 This same reasoning is immediately applicable to narrow islands 
 like Japan and Java, with water on both sides. Not only have their 
 mountains been thus upraised, and the volcanoes formed in the back- 
 bones of these islands, but the whole tablelands of the islands have 
 been built up in the same way, by the gradual injection of porous 
 lava from under the bed of the sea. If the sea was originally deeper 
 on one side, it has given the backbone of the islands a somewhat 
 unsymmetrical form, the stronger injection and development coming 
 from the deeper sea, but the uplift sometimes makes a wider extent 
 of land towards the shallower water. 
 
 This is clearly shown in the formation of numerous peninsulas, 
 as Athos, Longos, Cassandra, Pelion, Attica, Corinth, Argolis, and 
 the three peninsulas of the Peloponnesus in Greece ; in Italy as a 
 whole, and especially Calabria and Sicily, the latter being a tri- 
 angular island with mountains facing the Tyrrhenian, Ionian and 
 Mediterranean Seas; also in Scandinavia and Scotland; in Kamt- 
 chatka, Corea and the Malay Peninsula of Asia; and many other 
 places. The same principle is beautifully shown in such islands as 
 Cyprus, Crete, Euboea and nearly all the islands of the yEgean Sea ; 
 in numerous islands along the coast of Dalmatia ; in the ^Eolian 
 Islands, Sardinia, Corsica, Elbe, Ischia and the island of Capri; 
 Minorca, Majorca, Pine Islands, Isle of Wight, and numerous others 
 around Scotland and Ireland. Good illustrations in America are 
 found in the West Indies and the Catalina Islands off the Cali- 
 fornia coast ; and in Asia, it is shown beautifully by the form of 
 Saghallien, the islands of Japan, Formosa, Sumatra, Java, and 
 numerous islands in the East Indies; also in New Zealand; and in 
 fact almost universally throughout the world. 
 
 In the case of Italy, the Apennines are nearly in the center of the 
 country, but slightly nearer the Adriatic, which was the deeper sea, 
 and did most to elevate the peninsula. The same arrangement is 
 well illustrated in the island of Sicily, where the highest mountains 
 face the deepest seas. All of these and many other obvious illus- 
 
354 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 trations of the principle may be readily found, and must convince 
 even the most sceptical of its general validity. 
 
 If we apply this principle to Guam, the Philippines, the Aleutian 
 Islands, Puerto Rico and other islands of the Antilles, we shall be 
 able to explain how the mountains of these countries have been 
 raised, and the islands themselves developed by matter injected from 
 under the troughs near them. 1 This seems to remove all doubt as 
 to the process involved in the formation of islands, and the smaller 
 land areas, and we may consider the plains and slopes of the larger 
 land areas bordering on the sea. 
 
 If the" injections are powerful, as in South America, the up- 
 heaval of matter by the power of steam brings the subsidence of the 
 adjacent bottom and the accompanying sea waves. We need not 
 assume, and we do not assume, any original weakness or line of 
 weakness of the earth's crust, but simply that the ocean bed leaks, 
 and the steam pressure tends to heave it up, and that the process 
 works at various depths from ten to twenty miles, and is continued 
 at irregular intervals over long periods. 
 
 This principle of ridges enables us to understand why islands 
 rise from the sea in chains in such cases they are really submerged 
 and perhaps immature mountains. The Aleutian Islands are now in 
 this state, and many others may be pointed out. When a ridge is 
 once started, the chances greatly favor another parallel to the first, 
 because of the way the sea bottom is depressed nearby, in forcing 
 up the islands from beneath. 
 
 If we examine relief maps of the world, 2 we shall see by the 
 mountains parallel to the shores exactly where the ocean used to be 
 and how it gradually withdrew. Thus, the successive ridges of the 
 
 1 As throwing light upon some remarkable processes of stability in nature, 
 it is interesting to notice that when a volcanic cone has been upheaved in 
 the sea, and by successive uplifts raised above the water to become an island, 
 the subsidence that eventually takes place, after much matter has been ex- 
 pelled from beneath the crust to make the island, does not appreciably en- 
 danger the island itself, which is powerfully braced by the conical form of 
 its base. When a portion of the sea bottom sinks near by, to partially fill up 
 the cavities in the crust, the settlement of the bed of the ocean scarcely dis- 
 turbs the island, on account of the way in which the crust is upraised about 
 its base, and braces it on all sides. 
 
 'Those of Rand, McNally & Co., given in the Encyclopedia Americana, 
 have been found extremely useful. 
 
19o6t] SEE THE CAUSE OF EARTHQUAKES. 355 
 
 Andes were formed, beginning on the east, and working westward ; 
 as the movement progressed the chains got higher and higher, 
 requiring increased volcanic forces ; and in the last range, therefore, 
 occur the most and the highest volcanoes. This principle seems to 
 hold not only in the Andes, but generally throughout the world. 
 
 34. On the formation of islands, and on the significance of the 
 symmetrical disposition of their mountains. 
 
 We have already explained the formation of mountains and have 
 seen that they are not due to any shrinkage of the globe with result- 
 ing collapse of the crust, but to the injection of lava from the sea, 
 which uplifts the crust into a parallel ridge. The subject of island 
 formation, already alluded to in 33, has not been much discussed, 
 and it is impossible to say that there is any recognized theory on the 
 subject. When the islands are small they are justly considered 
 mountains in the sea. Some islands are volcanic and are thus vol- 
 canoes pushed up in the sea like those occasionally raised on the 
 land. That volcanic islands are similar to ordinary land volcanoes 
 seems clear enough ; but how about the larger islands ? How did 
 they originate? To answer this question satisfactorily, we may 
 remark that when we look at an ocean like the Pacific and notice 
 how it is dotted all over with peaks projecting above the water, and 
 remember that a still greater number do not reach the surface, we 
 recognize that such peaks were no part of the original constitution 
 of the globe, but have been developed in the sea in the course of 
 immeasurable ages. 
 
 Now, as to the larger islands, many of them have conspicuous 
 mountain chains, and the study of the lay of these chains is very 
 instructive. If mountains are wrinkles in the earth's crust, due to 
 the shrinkage of the globe, there is certainly no reason why they 
 should be symmetrically placed on islands. On this theory, the 
 mountains might cross the islands at any angle, or even miss the 
 islands entirely. On actual examination, what do we find to be the 
 fact? When we look at any good map we see the mountains run- 
 ning through the islands with the utmost symmetry, making in all 
 cases a veritable backbone or central axis for the land on both sides 
 In no cases do the mountains run diagonally or crosswise. To 
 appreciate the significance of this arrangement, look at the map of 
 
356 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 such islands as Cyprus, Crete, and others in the Mediterranean, 
 Saghallien, Formosa, Sumatra, Java, and others in the East Indies. 
 The regular symmetry of the mountains makes them in all cases the 
 backbones of the islands, and this can only mean that they were 
 formed in their present position by injections from the ocean on 
 both sides ; in no other way could such a symmetrical arrangement 
 arise. The universality of this law shows that the mountains depend 
 on the sea, and not at all on the shrinkage of the globe. The same 
 symmetrical arrangement is shown in numerous peninsulas through- 
 out the world. 
 
 35- Why all upheavals do not produce volcanoes. 
 
 It is not necessary for the subterranean forces to break through 
 and form volcanoes the movement often becomes so deep-seated 
 or feeble with the raising of the mountains that outbreaks do not 
 occur. What takes place in this respect depends on the depth and 
 recession of the sea and the suddenness and violence with which the 
 upheaving pressure is exerted. Thus, along the east coast of South 
 America no large volcanoes were formed, and such small ones as 
 may have once existed have now lost all trace of a volcanic aspect, 
 because the sea was shallow and kept retreating. On the northeast 
 of South America, however, the Lesser Antilles are in deeper water, 
 and when they rise to full growth may form a somewhat imposing 
 chain of mountains, exhibiting volcanic violence depending on the 
 depth of the sea. 
 
 In the case of the Alps the development was arrested by the rise 
 of Italy from the Mediterranean, which stopped the sinking of the 
 deep trough which has since become Lombardy. This was formerly 
 the Alpine trough, and it is now so filled up by erosion that the 
 Adriatic is the nearest sea, the recognized recession of which con- 
 firms the law. 
 
 In the case of the Himalayas also the development was arrested 
 by the rise of the vast plain of India. And while the resulting moun- 
 tain range became high, volcanic force was at length enfeebled by 
 the shallowness of the troughs where the Indus and the Ganges now 
 flow, and thus it is supposed that no active volcanoes broke forth 
 on the tops of these mighty mountains. If the adjacent water had 
 remained deep, as off the west coast of South America, the Hima- 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 357 
 
 lavas would doubtless have been broken through by the resulting 
 violence of the volcanic forces. As it was the shallow water, 
 eventually supervening, gave the power for heaving the mountains 
 little by little, and when they attained great height became so feeble 
 or deep-seated that it left them unbroken by volcanic violence. Thus 
 we see why the Alps and the Himalayas, in the main, failed to form 
 volcanoes, and why Africa and Australia are also devoid of these 
 vents (many small ones may have existed and have since lost all 
 trace of this appearance), which chiefly develop near the deep sea, 
 where the sudden exertion of these forces break through the moun- 
 tain tops. This happens in some cases where the mountains are not 
 very high, either because the seat of the explosion is shallow or the 
 fractures such as to offer but little resistance from greater depth. 
 
 In order to break through high mountains, the force has to be 
 extremely powerful, and this is not likely to be the case where the 
 adjacent sea is shallow, as was true south of the Alps and the 
 Himalayas. 
 
 Earthquakes in these regions, however, still continue, and have 
 always been abundant, but they are deep-seated, owing largely to 
 the filling in of the Alpine and Himalayan troughs, and lead to no 
 eruptions, and hence have been called tectonic. They are clearly a 
 survival, due to the same forces which upheaved the mountains, but 
 the sea having so far receded they cannot blow open any cones at 
 this late date. In fact the centers of disturbances usually are some- 
 what remote from the mountains at present and diffused over such 
 an area in the ancient trough that their power for rupture is slight. 
 
 36. Gradual secular desiccation of the oceans indicated by the 
 lowering of the strand lines throughout the world. 
 
 The raising and lowering of the land by subterranean forces 
 which have effected the withdrawal and encroachment of the sea over 
 the land was first advocated by the Greek geographer Strabo, who 
 adopted the theory of Archimedes that the figure of the ocean sur- 
 face is that of a sphere (cf. Suess, " Face of the Earth," Vol. II, 
 p. 2). This theory has been much developed in modern times and 
 explains numerous movements of the strand line. But according 
 to the elaborate study of this question made by Professor Suess, the 
 elevation and subsidence theory even in the oscillatory form adopted 
 
358 SEE THE CAUSE OF EARTHQUAKES. [October i 9 , 
 
 by Lyell, is inadequate to account for the general lowering of the 
 strand line noticed throughout the world. After his careful study 
 of this question, Professor Suess is unable to adopt any satisfactory 
 explanation of the observed phenomenon, and the whole problem 
 remains an enigma to the investigator of the physics of the earth. 
 An impartial examination of all the arguments leads me to believe 
 that the conceptions of Strabo, as developed into the oscillatory theory 
 advocated by Lyell, is adequate to account for many more of the 
 phenomena of the earth's surface than Professor Suess concedes; 
 yet it seems not only possible, but even probable, that an additional 
 cause is at work which must contribute somewhat to the general 
 lowering of the strand line throughout the world. This additional 
 cause is nothing else than the secular desiccation of the oceans, 
 brought about mainly by the sinking of the waters into the crust of 
 the earth, where the resulting steam becomes the motive power in 
 earthquakes. 1 A certain amount of water is also taken up as water 
 of crystallization in the development of crystalline rocks, which 
 extend somewhat deeper from age to age. 
 
 The view of earthquakes here adopted makes sea water the prin- 
 cipal disturbing cause. Of the water which thus sinks below the 
 crust, an appreciable part, but by no means all of it is expelled by 
 volcanic action ; and thus there is a steady accumulation of water 
 deep down in the earth's crust. If this view be well founded, there 
 is thus a secular desiccation of the waters of the globe, depending 
 on the penetration of water to the deeper parts of the crust, and also 
 on the absorption of water by rocks in crystallization and otherwise. 
 How rapidly this process goes on cannot be accurately known at 
 present; but on the basis of data observed within the historical 
 period, we may reasonably hold that the fall of the strand line due 
 to this cause is on the average less than a meter in 2,000 years, and 
 most likely of the order of one-tenth of this amount. 
 
 The desiccation here postulated seems to be a vera causa, and 
 this process, in connection with the oscillatory movement of the 
 land recognized by Lyell and other geologists as necessary for the 
 
 1 The sinking of the sea bottom when lava is expelled from under it also 
 increases the capacity of the ocean basins, and thus slightly lowers the strand 
 line. 
 
xgo6] SEE THE CAUSE OF EARTHQUAKES. 359 
 
 explanation of such phenomena as the coal measures, gives us a 
 simple and natural way of explaining leading phenomena of the 
 strand. We have elsewhere explained why the movements of faults 
 must be attributed primarily to the formation of steam-saturated 
 lava at the depths whence earthquakes originate. 
 
 The steam which induces this movement naturally remains hidden 
 in the earth, and as the process goes on uninterruptedly from one 
 geological age to another, there is a gradual secular desiccation of 
 the waters of the sea, and a correspondingly slow lowering of the 
 strand line. 
 
 To determine the average rate of this lowering of the strand we 
 would have to take a figure somewhat smaller than that found in the 
 different countries since a given geological epoch, and even then, 
 the result would be partly vitiated by the effects of secular elevation 
 of the land. In any case, the movement depending on secular desic- 
 cation of the oceans is extremely slow. 
 
 This view that there is a secular desiccation of the waters of the 
 sea is not new, but was entertained in different forms by Benoist de 
 Maillet (1692), Celsius (1743), Von Hoff (1822), Goethe and 
 others (cf. Suess, " Face of the Earth," Vol. II, Ch. I). 
 
 Professor Suess' exhaustive discussion of the movements of the 
 strand line will be found chiefly in chapters XII-XIV of Volume 
 II of the " Face of the Earth." He considers the lowering of the 
 sea level to the extent of hundreds of meters within recent geo- 
 logical time to be proved. The cause here suggested gives the only 
 explanation of the phenomenon which seems at all probable or 
 consistent with known facts. 
 
 37. On the gentle movements of the land. 
 
 In many parts of the world, the rocks are comparatively un- 
 broken, and leakage is very slow and gradual. This may corre- 
 spond to the beds of shallow seas or to the land when level and 
 unbroken by mountains. In all such regions the water which may 
 seep down would give rise to a very evenly diffused subterranean 
 steam pressure and the chances are that any movement which might 
 take place would prove to be very slow and gradual. The strain 
 being nearly equalized at all points, there would be no heaving re- 
 quired to adjust the nearly even balance of the forces, and conse- 
 
 PROC. AMER. PHIL. SOC., XLV. 184 W, PRINTED FEBRUARY 2J, 1907. 
 
360 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 quently earthquakes of sensible strength would seldom or never 
 occur. In this way we may explain the comparative immunity of 
 many districts from earthquakes. 
 
 Yet, when an even pressure thus arises, it may produce a steady 
 elevation of the land. Bending and warping of the strata may also 
 occur where the stresses are steadily applied, and under the cir- 
 cumstances the rocks would probably change their figure slowly 
 without snapping; when the earthquakes are more violent the rocks, 
 are broken into smaller pieces and fault movements increase. 
 
 38. Oscillations of the strand shown in such phenomena as the 
 coal measures and fossil beds. 
 
 No special effort is made in this paper to explain all the phe- 
 nomena of the earth's crust which offer difficulty to the geologist; 
 and hence we have been chiefly concerned with phenomena of eleva- 
 tion. But ever since the^time of Aristotle and Strabo it has been 
 justly remarked by sagacious observers that there has been not 
 only elevation of the land, but under certain conditions also subsi- 
 dence. Lyell and many other writers have discussed these oscillatory 
 movements which are well exhibited by the successive layers seen 
 in many of the coal measures. Some of these layers may be ex- 
 plained by the effects of damming up and drifting of vegetation to 
 places where it did not grow ; but even when allowance is made for 
 these causes, there still seems ample evidence of an oscillatory move- 
 ment of the land in many places. This is also well shown in fossil 
 beds, where sea shells often alternate with brackish water species. 
 
 The expansion and contraction of the limits of the sea over large 
 areas of the low-lying shore is a frequent phenomenon, and a slow 
 oscillation of the strand seems the only rational explanation of it yet 
 offered. Such an oscillation is most easily explained by a sub- 
 stratum of fluid beneath the earth's crust, such as we show to exist. 
 Under just what conditions the land sinks is not clear. The insta- 
 bility may result from a number of causes of which the most probable 
 would seem to be: upward movement of neighboring regions, thus 
 weakening the support of the region in question and perhaps putting 
 additional load upon it, while the underlying fluid layer slowly 
 yields, thus causing subsidence. If during an earthquake a neigh- 
 boring area should be started upward, the strain would naturally 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 361 
 
 equalize itself in that direction, and the movement might continue 
 for a long time until subterranean conditions changed. The earth's 
 crust is complex, of unequal thickness in different parts, and broken 
 into unequal blocks by various faults, which are continually adjust- 
 ing themselves to the strains arising in the underlying substratum 
 supporting them. That some areas should go up while others go 
 down is therefore not at all remarkable ; and most of the oscillations 
 of the land will be found to depend upon causes of this kind. In 
 my opinion the seat of the forces will be found to lie mainly in the 
 fluid substratum beneath. Considering the great number of blocks 
 into which the earth's crust has been shown to be broken by faults 
 such as Professor Suess has so fully discussed for the regions of 
 the Alps and the Tyrol, the oscillations of level with the changes of 
 the strand in salt and fresh water regions seem easily accounted for. 
 In his work on " Meteorics " (lib. I, cap. 12) Aristotle justly 
 remarks : 
 
 " The distribution of land and sea in particular regions does not endure 
 throughout all time, but it becomes sea in those parts where it was land, 
 and again it becomes land where it was sea. . . . 
 
 " And the sea also continually deserts some lands and invades others. 
 The same tracts, therefore, of the Earth are not, some always sea, and others 
 always continents, but everything changes in the course of time." 1 
 
 39. Strabo's vieivs on the elevation and depression of the land. 
 
 In his " Principles of Geology " (pp. 24-25, I2th edition) Lyell 
 quotes the views of Strabo regarding the elevation and depression 
 of the land as follows : 
 
 " It is not," says Strabo, " because the lands covered by seas were orig- 
 inally at different altitudes, that the waters have risen, or subsided, or receded 
 from some parts and inundated others. But the reason is, that the same 
 land is sometimes raised up and sometimes depressed, and the sea also is 
 simultaneously raised and depressed, so that it either: overflows or returns 
 into its own place again. We must, therefore, ascribe the cause to the 
 ground, either to that ground which is under the sea, or to that which be- 
 comes flooded by it, but rather to that which lies beneath the sea, for this 
 is more movable and, on account of its humidity, can be altered with great 
 celerity. It is proper to derive our explanations from things which are 
 obvious, and in some measure of daily occurrence, such as deluges, earth- 
 quakes and volcanic eruptions, and sudden swellings of the land beneath the 
 , sea ; for the last raise up the sea also ; and when the land subsides again, they 
 occasion the sea to be let down. And it is not merely the small, but the 
 
 1 Cf. Lyell's " Principles of Geology}" I2th edition, Vol. I, pp. 21-22. 
 
362 SEE -THE CAUSE OF EARTHQUAKES. [October 19, 
 
 large islands also, and not merely the islands, but the continents which can 
 be lifted up together with the sea ; and both large and small tracts may sub- 
 side, . for habitations and cities, like Bure, Bizona, and many others, have 
 been engulfed by earthquakes." 
 
 VIII. ON THE GREAT SEA WAVES WHICH FREQUENTLY ACCOMPANY 
 VIOLENT EARTHQUAKES. 
 
 40. Great sea zvaves caused chiefly by the subsidence of the sea 
 bottom after the lava has been expelled by the throes of earthquakes. 
 
 We have already shown how the sea bottom may subside over a 
 large area after a violent earthquake which has forced out from be- 
 neath the bed a large amount of steam-saturated lava. This lava 
 is usually forced up under adjacent mountain ranges, mountain 
 peaks and volcanoes, or under the intervening coast plains. If 
 relief is afforded by forcing the column of molten matter along so 
 that the remotest part raises in the mountains, which are at the top 
 of the arched portion of the neighboring region of crust, we should 
 have no means of discovering the resulting slight elevation of the 
 peaks. For the heights of mountains are always uncertain by several 
 feet, because extremely exact levels are difficult to establish so far 
 above the sea. We are thus unable to say whether the various 
 peaks and ranges are rising in height, or to tell what fluctuation of 
 altitude they may undergo from time to time, by the heaving of 
 earthquakes. In the case of plains near the sea it is easier to detect 
 changes of level, but in many places characterized by violent earth- 
 quakes unfortunately no such observations are taken. Tidal obser- 
 vatories are best adapted to keeping record of any changes of level 
 that may occur, because it is only by analyzing the tides carefully 
 that we can detect changes of the sea level. 
 
 As already remarked, the sudden upheaval noticed in the sea- 
 coast of Chili after certain earthquakes was especially remarked by 
 Charles Darwin and Capt. Fitzroy, who experienced the severe 
 earthquake at Conception, February 20, 1835, and noticed the result- 
 ing elevations of the coast line. 
 
 Let us consider briefly the upheavals of the Chilian coast wit- 
 nessed by Darwin and Fitzroy. The only reasonable explanation 
 of this fact is that lava had been forced under the coast, and it is 
 obvious that the matter must have been expelled from under the 
 
i9o6.] SEE THE CAUSE OF EARTHQUAKES. 363 
 
 bed of the sea. We do not know whether the lava was injected as 
 a thin layer or the whole body of the material under the coast slightly 
 pushed back to afford relief of the strain under the sea, but a dis- 
 placement of the latter kind seems the more probable. A sea wave 
 of considerable magnitude 1 was noticed on that occasion, but it was 
 not so large as sometimes develops, and the sea bottom may have 
 subsided only very slightly. But on August 13, 1868, the whole 
 South American coast from Valdivia in Chili to Guayaquil in Ecua-' 
 dor was violently shaken by a terrible earthquake, with its highest 
 intensity near Arica. A few minutes after the earthquake the ob- 
 servers were surprised and alarmed to notice the sea slowly receding 
 from the land, and very soon vessels which had been anchored in 
 seven fathoms of water were left high and dry, with no means <5f 
 escape. In a short time their surprise was converted into terror at 
 the sight of a mighty ocean wave fifty or sixty feet high returning 
 with terrible velocity and carrying everything before it. The vessels 
 stranded on the beach at Arica, including the U. S. S. Wateree, 
 were swept up by the gigantic wave and carried nearly a. half mile 
 inland and again left stranded higher than before. 2 The wave rolled 
 back and after a short interval again swept the shore ; and the 
 furious oscillation of the water thus started continued for a day 
 or two before the sea finally quieted down. This great sea wave was 
 propagated over the Pacific and observed almost all over the world. 
 In 1877, May 9, another great earthquake visited the same 
 region and was followed by a wave of even greater magnitude, of 
 exactly the same type, the water first slowly receding from the land, 
 and then returning as a gigantic wave carrying everything before it. 
 This wave of 1877 is known as the Iquique wave. At Arica the 
 hulk of the stranded U. S. S. Wateree was again picked up and 
 carried still furthe'r inland, which would indicate that at Arica the 
 height of this wave surpassed that of 1868. The sea continued to 
 oscillate in periods of something like an hour and did not subside 
 for a couple of days. 
 
 1 In his valuable work on " Earthquakes," chapter IX, Professor Milne 
 gives a catalogue of sea waves. Different waves present different phenomena, 
 and we here treat only of the best established types. 
 
 2 Button, " Earthquakes in the Light of the New Seismology," p. 281. 
 
364 SEE THE CAUSE OF EARTHQUAKES. [October i 9 , 
 
 These two are good types of those great sea waves in which the 
 water recedes from the land in the first few minutes following the 
 earthquake. 
 
 Major Button and others have suggested that the sea bottom 
 sinks, and the explanation we have given of how this takes place 
 seems satisfactory and free from objections. It is clear that if by 
 the throes of the earthquake a large body of lava is forced from 
 under the bed of the Andean trough, the bed might thereupon settle 
 from twenty to fifty feet over a large area. 1 The great inrush of 
 the water following this subsidence would withdraw it from the land, 
 and as soon as the rushing currents met in the center of the trough, 
 they would raise the water into a high ridge, and its subsidence would 
 give the first great wave which rolled in upon the devastated shore. 
 With the first depression of the water over the ridge, another in- 
 rush would take place, again withdrawing the sea from the shore, 
 and another great wave would follow like the first, but of slightly 
 feebler intensity. And so the oscillations of the sea would con- 
 tinue for a day or two, till they became reduced by friction to in- 
 sensible magnitude. This explanation accords with all the known 
 facts, and the recognized laws of fluid motion. Assuming it to be 
 ^correct, the result is of interest as showing the effect of friction in 
 destroying the motion of the sea, which has often been discussed in 
 connection with the problem of the tides. In this case, the length of 
 the wave of 1868 has been calculated to be about 100 geographical 
 miles ; and as the depth of the sea is between four and five miles, we 
 see that the wave length involved is from twenty to twenty-five times 
 the depth of the sea. 
 
 At the close of this section we shall give another possible ex- 
 planation of waves of this kind, which begins with a recession of 
 the water from the shore ; but meanwhile we shall notice waves of 
 a different class, sometimes encountered, which begin by a sudden 
 rising of the water near shore. 
 
 The great wave which overwhelmed Simoda, Japan, December 
 29, 1854, may be taken as a type of those which are characterized 
 by the sudden inrush of a great wave without any previous recession 
 
 subsidence might be much greater if the area affected was propor- 
 tionally diminished. 
 
tgob.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 365 
 
 of the sea near the shore. At 9:15 a. m. Vice Admiral Putiatin on 
 board the Russian frigate Diana noted very powerful shocks of an 
 earthquake, and a little before ten o'clock a huge wave was seen 
 coming which quickly overflowed the city. Major Button remarks 
 that the time here involved, and the known speed of propagation of 
 such a wave, indicates that it originated over 100 miles from the 
 shore, in the Tuscarora trough, where the depth attained is over 
 4,000 fathoms. 
 
 If it originated so far away, it might have resulted from the 
 sinking of the sea bottom, as in the Arica and Iquique waves, the 
 withdrawal of the sea when the waters rush into the sink becoming 
 so nearly insensible at the great distance of Simoda as to escape 
 notice in the bay. But a more probable explanation is that the sea 
 bottom just east of the Tuscarora deep was heaved up into a ridge 
 with elevation of 20, 30, or 50 feet. This would produce the 
 great wave which so suddenly appeared to overwhelm the city. 
 In this case there would be no preliminary recession of the water 
 whatever, and the wave would come without the least warning, as 
 appears to have been the case. 
 
 In case the wave originated by a subsidence of the bottom of 
 that trough, a slight withdrawal of the water from shore should 
 have been noticed even at that great distance; but if the upheaval 
 occurred beyond the trough, the greater inrush of water from that 
 side may have obscured the slight recession which otherwise might be 
 expected at Simoda. The upheaval is conceivable in the manner 
 we have described, either with or without subsidence of the bed, 
 and ordinarily the disturbance might be on either side of the 
 trough ; but in this case the time shows that the uplift probably was 
 beyond the trough. 
 
 41. Another explanation of sea waves on the hypothesis of 
 submarine eruptions. 
 
 The only other rational way of explaining these great sea waves 
 is by means of the uprush following the explosion of a submarine 
 volcano. We consider his explanation much less satisfactory than 
 that already given, but it is undeniable that in certain cases it might 
 account for both classes of sea waves, especially where the water is 
 deep and we can suppose the volcano to be upheaved near shore and 
 of large size. 
 
366 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 Let the accompanying figure illustrate a section perpendicular 
 to the shore of the sea and land, on a coast like that of Chili or Peru, 
 where this phenomenon has so often been observed in a typical way, 
 and the water is deep. 
 
 Suppose a volcano to be formed at some distance, say 50 or 100 
 kilometers from the shore. If a great wave is produced we may 
 assume the volcano to be raised to a height of 1,000 or 2,000 meters, 
 or even higher ; but as the sea is seven or eight kilometers deep, this 
 will only reach one-seventh or, at most, one-fourth of the way to the 
 surface. In the throes of the earthquake the volcano is raised, and 
 the water forced up immediately over the eruptive center; and the 
 steam, stones, lava, dust and ashes are driven upward towards the 
 surface, as in the explosion of a land volcano. The explosion is 
 
 FIG. 14. 
 
 Submarine volcano. 
 
 resisted by the great depth of the water, which is hurled upward in 
 a violent current from the orifice. The steam condenses to water 
 by the low temperature of the ocean, and the other gases are ab- 
 sorbed, whether coming from the bed of the sea or formed in the 
 water by the intense heat of the red hot lava. The current of steam 
 and flying stones, lava, sand and ashes, by beating against the over- 
 lying stratum of water, forces such rapid upward movement that the 
 level above is forced bodily upward, it may be several hundred 
 meters. But the fluid medium is continuous and presses in on all 
 sides, and is therefore drawn upward on all sides about the base of 
 the cone to supply the uprush of water. The currents thus forced 
 with enormous violence are shown in the figure. The drawing 
 upward of the water about the base of the cone causes the inflow 
 of water from the bed of the sea towards the base to maintain the 
 upward movement, following the upheaval and explosion of the 
 volcano. Thus the water near the shore some distance away is 
 sucked down in the general lowering of the level and the sea is 
 observed to slowly recede at the shore. All this is done in a short 
 
19o6 .] SEE-THE CAUSE OF EARTHQUAKES. 367 
 
 time, say 30 minutes, and then follows the sea wave which has taken 
 form owing to the forcing up of the water over the volcano, thus 
 forming the crest, and its withdrawal from the shore, forming the 
 trough. 
 
 The water heaved up gradually settles and the wave approaches 
 the shore often with a velocity of something like eight kms. per 
 minute and thus sweeps everything before it. It then oscillates 
 back and forth with period of some 60 minutes, and for a day or two 
 the sea may continue to be agitated with appalling violence, and the 
 wave propagated to the remotest parts of the earth. 
 
 If an observer were to witness such an earthquake in a region 
 where the shore was steep and the sea of uniform depth, and should 
 note the time of the sea wave and the direction of the normal to 
 the wave front as it first returns, he would have a very approximate 
 means of locating the situation of the new submarine volcano. It 
 would lie on the normal to the circular w r ave front, and at a distance 
 corresponding to the time of arrival in a sea of the given depth. 
 
 The interval r required for the oscillation of the wave being 
 known, the theory of the wave motion could be worked out by the 
 general formula 1 for a wave of any length A and any depth of the 
 water k, 
 
 ^ -^T~ (I) 
 
 *- I 
 
 Perhaps it could be found with sufficient approximation by the more 
 special forms, in which the velocity becomes, when the wave is long 
 compared to depth of the water: 
 
 V= Vgk, or V= (k + E) (2k + E) ; (2)' 
 
 E being the height of the crest of the wave above the normal level 
 of the water. 
 
 In this way we could find not only the distance of the eruption 
 from the observer, but the direction, so as to fix its place with con- 
 
 1 Airy, " Tides and Waves," Art. 169. 
 
 2 Report of Committee of the Royal Society on the Krakatoa Eruption, 
 p. 94. 
 
368 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 siderable accuracy. If the depth is not uniform, but the topography 
 of the ocean basin known, the observer would still be able to locate 
 the hypothetical submarine volcano with considerable precision. 1 
 
 Within historical times several submarine volcanoes have been 
 observed to rise above the sea about South America and elsewhere, 
 in places where the depth was small. From the circumstances that 
 the sea is generally very deep off the Chilian and Peruvian coast, 
 where the most violent earthquakes occur, one would expect but 
 very few of these volcanoes to reach the surface. Yet the large 
 number of sea waves following violent earthquakes may afford us 
 some idea of the activity of submarine earthquakes and perhaps 
 volcanoes in that part of the world. It is probable that not less 
 than one hundred such earthquakes with sea waves occur along the 
 South American coast in a century, and of these not less than ten 
 have done great damage. 
 
 If the supposed eruption is some distance out at sea, the effect 
 on shore would be small, because the level is not so much changed, 
 owing to the great body of the intervening water. Also, when a 
 violent earthquake occurs and but slight recession of the water is 
 noted, followed by a wave after considerable interval, the indications 
 would point to a great eruption at considerable distance. On the other 
 hand, if the recession of the water is quick and the wave returns 
 after a short interval, the eruption should be comparatively near the 
 shore. Thus by a study of the waves observed the place and char- 
 acter of the eruption may be approximately determined. 
 
 In some cases the sea is said to be bodily upheaved, and rises 
 with the utmost suddenness. In such cases the volcano may be 
 very near shore, or the sea bottom may be upheaved in the form of 
 a ridge or cone without submarine eruption. Since the earthquakes 
 under the sea are very numerous, there is a great probability that 
 all these movements of the sea should be observed occasionally. 
 
 Major Button adopts the view that the sea bottom sinks when 
 the sea withdraws from the shore, and I also consider this the 
 most probable cause in the great majority of cases. It is the more 
 logical to accept this view since we now see how a sinking of the 
 
 1 If such an eruption occurred the surface of the sea would be likely to 
 show evidence of it, by ejected pumice, ashes, and other volcanic debris. 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 369 
 
 bottom can take place by natural process. So long as there was no 
 means of explaining the subsidence of the bottom as a part of a 
 general process in nature, the acceptance of such a violent hypothesis 
 presented great difficulty. 
 
 To make entirely clear how collapse of the bottom may occur 
 after the expulsion of the steam-saturated lava from under the bed 
 of the Andean trough, we may observe that the release of the in- 
 tense pent-up pressure must tend to produce a sudden and somewhat 
 violent cooling in the stratum from which the lava is expelled. Its 
 support of the overlying bed of the ocean trough is thus largely 
 withdrawn, and sinking may easily follow. If this does not happen 
 in every case (and we have no reason to think it so frequent an 
 occurrence), it would probably follow at certain intervals, when the 
 successive expulsions of material have reduced the underlying 
 stratum of lava to a state of small density, in which the medium is 
 filled very largely with bubbles of steam and therefore rapidly cooled 
 when the pressure is released by an ejection of lava. This gives, I 
 think, a simple conception of a self-adjusting system, such as is so 
 often found in nature, by which the continuous process of expulsion 
 of lava may go on, and the level of the sea bottom be adjusted auto- 
 matically. But whether this is the exact process or some improve- 
 ment may be suggested when our knowledge is more extended, it is 
 clear that some such automatic mechanism is at work, and that it has 
 operated in similar troughs all over the world throughout geological 
 history. 
 
 As the water did not withdraw and later return as a great wave 
 during the recent San Francisco and Valparaiso earthquakes, we 
 know that the sea bottom did not sink in the case of either of these 
 great disturbances. The expulsion of the lava, however, must leave 
 the sea bottom less stable and increase the probability of its sinking 
 when the next severe earthquakes occur at these places. As the 
 San Francisco earthquake was much less severe than that at Val- 
 paraiso, the probability of the sea bottom sinking off the Californian 
 coast is much less than off the coast of Chili ; yet such a subsidence 
 with the accompanying seismic sea wave is sure to come sooner or 
 later in all places subject to heavy earthquakes. In 1812 the whole 
 of southern California was severely shaken by earthquakes; on De- 
 
370 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 cember 8, the water withdrew from the shore at Santa Barbara, and 
 after a short interval returned as a great wave, which overflowed the 
 coast and did much damage to life and property. All cities situated 
 on such coasts should be prepared for the emergency of a seismic sea 
 wave that is, there should be a place of refuge for the people, and 
 houses thus exposed to the inundations of the sea should not contain 
 treasures and historical articles which cannot be replaced. A mu- 
 seum, for example, should never be built in such an exposed 
 situation. 
 
 Perhaps a few words may be added in regard to the handling 
 of ships, so many of which have been lost in the past, owing to the 
 mystery surrounding these great waves. In South America the 
 people have learned by bitter experience that when an earthquake 
 occurs the first thing to do is to watch the sea. If it begins to 
 withdraw from the shore they at once flee to high ground, for they 
 know that the wave will follow. This same principle is eminently 
 applicable to commanders of ships in the harbor. On the first indi- 
 cation of the retreat of the water from the shore the ships should 
 be headed with all possible speed for the open sea. For if the ships 
 remain in the harbor they may soon be stranded and unable to move, 
 and sure to be carried inland when the wave returns ; whereas if an 
 effort is made to get out to sea, the ships may ride over the wave 
 without difficulty and suffer no damage whatever. This rule is 
 easily applied to all steam ships, whether belonging to the navy or 
 merchant marine. Sailing vessels, being less under control than 
 steamships, might be unable to escape in some cases; yet, if the 
 state of the wind gave them the requisite motive power, even they 
 might make the open sea. A wave does not come immediately after 
 an earthquake, but something like half an hour or an hour after- 
 wards, and this usually gives time for escape. After the sea bottom 
 subsides, the water must flow from the shore into the depression, 
 and then when the water piles up, it must again flow back to land 
 to produce the wave ; and if the ships are properly handled in this 
 interval, most of them will escape undamaged. 
 
I9o6 .j SEE THE CAUSE OF EARTHQUAKES. 371 
 
 IX. CONCLUDED THEORY OF VOLCANOES. 
 
 42. Other theories of volcanic action. 
 
 The four fundamental facts mentioned in 10 have been fully 
 considered, and we have found that the hypothesis of the penetra- 
 tion of sea water into the crust of the earth affords a natural and 
 satisfactory explanation of all volcanic phenomena. Such hypotheses 
 as the following : ( I ) Lava flowing out of a molten interior, is 
 contradicted by the rise and fall of the columns of lava in volcanoes, 
 as if forced up by the elastic pressure of steam, which also escapes 
 in eruptions; (2) molten reservoirs, contradicted by the same phe- 
 nomena; and moreover neither (i) nor (2) enables us to account 
 for the observed distribution of volcanoes; (3) melting by relief of 
 pressure, and (4) melting by crushing, encounter the same diffi- 
 culties, and others besides. None of these four hypotheses can be 
 seriously considered. 
 
 There remains Major Button's recent suggestion that radium 
 is the exciting cause. But the researches of the Hon. R. J. Strutt 
 have shown that all the principal rocks of the earth's crust, especi- 
 ally granite, contain large quantities of radium, and since these 
 rocks underlie all the continents, we should expect abundant active 
 volcanoes everywhere inland if radium were the exciting cause, 
 whereas in fact they appear in the depths of the sea or along the 
 shores of the oceans. The cause of volcanic action is thus narrowed 
 down to the penetration of water into the heated rocks of the earth's 
 crust, and all other hypotheses may be unhesitatingly rejected. 
 
 We shall now adduce some further considerations bearing on 
 the aqueo-igneous theory^ with a view of throwing additional light 
 upon particular phenomena. 
 
 43. Certain objections to the theory of the penetration of sea 
 water. 
 
 The beginning of this theory may be traced back to Lucretius, 
 and perhaps to Aristotle, 1 and hence we shall first answer two ob- 
 jections which have been urged against it. 
 
 First, it is held that the temperatures of the lavas are too high, 
 2,000 to 3,000, whereas one would expect the temperature to be 
 
 1 In more recent times it has been treated by Sir J. Prestwich, in a paper 
 " On the agency of water in volcanic eruptions," Proc. Roy Soc., April 16, 1885. 
 
372 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 no higher than from 700 to 1,000. A sufficient answer to this 
 objection is the pressure within the earth's crust, which removes the 
 depth of fusion to a lower layer where the temperature is higher ; 
 and, moreover, in all eruptions the temperature is no doubt greatly 
 raised by the violent churning the lava receives before reaching the 
 orifice of the volcano, and by superheated steam escaping through 
 it. The lava would naturally flow from that depth at which motion 
 under the earth's crust is easiest, and the temperature observed is 
 therefore about what should be expected. 
 
 Second, it is held that the fractured portion of the earth's crust 
 which would permit a ready penetration of water is confined to a 
 layer not more than five or six miles deep, the great pressure lower 
 down operating to close all crevices ; and it is therefore claimed that 
 water going down fifteen or twenty miles would penetrate the re- 
 maining ten to fourteen miles of unbroken rock with great difficulty. 
 In answer to this objection it may be said that the depth to which 
 rocks are fractured is not certainly known ; but whatever it may be, 
 Daubree's experiments show that the force of capillarity may cause 
 the water and steam to keep on descending till the vapor reaches a 
 temperature where it is rapidly absorbed and diffused among the 
 rock, just as gases are in hot steel ; and when the vapor becomes 
 superheated its explosive violence is greatly increased, and hence 
 this also would tend to raise the observed temperature of the lava, 
 because it is chiefly the hotter lava, still further heated in ejection, 
 which would be forced out of volcanoes. 
 
 These objections, therefore, present no serious difficulty to the 
 theory that volcanic action depends on nothing but the penetration 
 of sea water. 
 
 It is sometimes said that earthquakes accompanying volcanic 
 eruptions are shallow, and it has therefore been inferred that the 
 lava comes from no great depth. Perhaps a more correct view 
 would be to hold that the throat of the volcano does not be- 
 come closed to a great depth, by partial cooling of rock since the 
 last eruption, and the shocks naturally proceed from this point of 
 resistance rather than from the source of the steam and lava rising 
 beneath the volcano, which may be much deeper, and yet give no 
 sensible indication of their movement till the resistance becomes con- 
 
19o6 .j SEE THE CAUSE OF EARTHQUAKES. 373 
 
 siderable. Besides the honeycombed pumice which underlies vol- 
 canoes, it may be supposed that they frequently contain a certain 
 number of passages out of which the pumice has been blown, some 
 of which may become real caverns when the lava subsides after an 
 eruption. When a new eruption begins these old passages might 
 offer little resistance till the lava column came within a short dis- 
 tance of the surface, and hence the shallowness of the shocks wit- 
 nessed in eruptions. The shallowness of these shocks does not 
 prove the superficial character of the lava erupted; on the contrary 
 the earthquake shocks felt over the whole region around every 
 active volcano shows that the subterranean disturbances arise in a 
 layer which acts as fluid just beneath the crust. The forces de- 
 veloping in this layer find their relief in the eruption of steam and 
 lava from the volcano. 
 
 44. The origin of volcanic ashes due principally to the break- 
 ing and grinding up of pumice. 
 
 In his useful work on the " Volcanoes of North America," the 
 late Professor Russell, of the University of Michigan, makes the 
 following explanation of the origin of volcanic ashes (pp. 7576) : 
 
 " Sheets of Volcanic Sand and Dust. In the case of volcanic eruptions 
 of the explosive type, the steam occluded in the lava expands as external pres- 
 sure is relieved ; this expansion is frequently so violent that the rock is dis- 
 integrated and the fragments projected high in the air. Besides this primary 
 mode of reducing the lava to fragments, and much of it to the condition of 
 dust, the larger fragments as they are shot upwards with a velocity in some 
 instances even greater than the initial velocity of shells fired from modern 
 rifle-cannon, strike against one another and against falling fragments, and 
 are shattered, thus tending to increase the quantity of fine dust-like particles 
 produced. While much fine material originates thus, and is carried away 
 by the wind, many of the fragments that escape comminution fall into the 
 the crater from which they were thrown and are again violently ejected, 
 thus multiplying the chances of their being reduced to powder. An erup- 
 tion of the explosive type thus tends to form much fine dust, which is car- 
 ried high into the air by the upward rushing steam and falls most abundantly 
 near the place of discharge. Should a strong wind be blowing, the dust is 
 carried to leeward of the volcano, and on reaching the earth forms a sheet, 
 which, owing to the winnowing action of the wind, is composed of finer and 
 finer fragments, the greater the distance from the volcano.". 
 
 Professor Russell appreciated more fully than many geologists 
 the necessity of explaining the enormous clouds of dust which 
 arise from volcanoes, but it is difficult to escape the impression 
 
374 SEE THE CAUSE OF EARTHQUAKES. [October^, 
 
 that his explanation is somewhat labored. It would be more natural 
 to say that a volcanic mountain is underlaid and filled with pumice 
 and when the explosions become violent some of this porous ma- 
 terial is ground up and blown out as dust. Much of it may already 
 be in the form of powder from former earthquake shocks when the 
 mountain was packed with pumice, and simply requires to be blown 
 out; and hence the vast clouds which obscure the sun and darken 
 the earth for hundreds of miles ! If the rock broken up by the 
 explosion were solid, as supposed by Russell, it would be less easy 
 to account for the enormous outpourings of ashes, observed in 
 such volcanoes as Hecla, yEtna, Vesuvius and Conseguina. Solid 
 rocks would produce lapillse and sand rather than fine ashes, which 
 result from the breaking up of pumice with very thin bubbles. 
 
 When lava is forced up in the throat of a volcano some of it 
 may run out, relieving the pressure which raised it, and the rest 
 then sinks back into some of the passages which lead to the throat 
 of the volcano. There may be, and in general probably are, several 
 of these passages, unequally opened at different times, and the lava 
 is forced up from some of them. After the lava is poured out and 
 subsides, other passages formerly closed may be opened and eject 
 vast quantities of volcanic ashes without encountering any molten 
 rock whatever. It would be a great mistake to suppose that all 
 ashes which pour from a volcano are forced through a layer of 
 liquid lava before ejection. If the ashes were forced through a 
 layer of liquid they would be red hot when cast out, and such heat 
 would give the particles a ruddy glow. As a general rule, such a 
 glow is not observed, and hence the theory that the ashes are ejected 
 through a layer of liquid is untenable. The outpouring of lava is 
 only a part of the operations of a volcano; the ejection of vapor, 
 ashes and pumice being, perhaps, even more important. Steam is 
 the one force which has to be relieved, and the other substances 
 ejected are incidental thereto. 
 
 In his well-known on "Geology" (fourth edition, 1903), p. 
 173, Sir Archibald Geikie says: 
 
 "... The finest dust is in a state of extremely minute subdivision. 
 When examined under the microscope, it is sometimes found to consist not 
 only of minute crystals and microlites, but of volcanic glass, which may be 
 observed adhering to the microlites or crystals round which it flowed when 
 
I9o6<] SEE THE CAUSE OF EARTHQUAKES. 375 
 
 still part of the fluid lava. The presence of minutely cellular fragments is 
 characteristic of most volcanic fragmental rocks, and this structure may 
 commonly be observed in the microscopic fragments and filaments of ghss. 
 A characteristic feature of these minute fragments is the frequent occur- 
 rence among them of semi-circular or elliptical ('hour-glass') shapes, which 
 evidently represent the sides of vesicles or pores that enclosed vapour or 
 gas in the molten rock, and were disrupted and blown out during volcanic 
 explosions." 
 
 45. On the supposed absence of volcanoes in the Alps and 
 Himalayas, and on the former existence of these vents in the interior 
 of continents. 
 
 It is frequently remarked that volcanoes do not appear in the 
 Alps and Himalayas, and the inference has been drawn that no 
 volcanoes originated in the formation of these great mountain 
 ranges. But it is well known that at some time in the past geological 
 ages volcanoes existed in almost every part of every country, and 
 mountain chains like the Alps and Himalayas are no exceptions to 
 the general rule. Professor Suess (" Face of the Earth," Vol. I, 
 pp. 201-274) mentions some volcanoes formerly active in the Alps, 
 and undoubtedly similar vents once existed in the Himalayas. In 
 the course of time nearly all surface trace of eruptions is lost 
 where the glaciation, denudation and sedimentation are active, as in 
 the Himalayas. When we consider how imperfect our knowledge 
 of those mountains is, not only because they are high, but also 
 inaccessible to exploration, the failure up to this time to find extinct 
 volcanoes or their products is not remarkable. Craters are soon 
 covered by ice and worn down by the grinding action of glaciers, 
 while their ashes and lavas are equally covered and lost from view. 
 
 Major Dutton justly remarks that the regions which have been 
 exempt from volcanoes are small in comparison with those which 
 have had them; and he observes that going back to early Tertiary 
 times we find them occurring where they have long been extinct. 
 
 " The grandest volcanic field in the world was central and southern India 
 in Cretaceous times, when there was not a volcano in all Europe, and ex- 
 tremely few in North and South America. In the Jura-Trias, the Appa- 
 lachian region, from Labrador to the Gulf of Mexico, bristled with them, 
 and vast plateaux of lava were outpoured. In Paleozoic, they abounded in 
 the region of the Great Lakes, in Missouri, in Arkansas, and in eastern Texas. 
 There is hardly a county or bailiwick on the whole mundane sphere which 
 has not had its volcanic cycle at some time or other, and there are many 
 
 PROC. AMER. PHIL. SOC., XLV. 184 X, PRINTED FEBRUARY 25, 1907 . 
 
376 SEE THE CAUSE OF EARTHQUAKES. [October 19 
 
 which have had two, three, four or five cycles. Volcanoes are not local 
 phenomena, nor yet are they strictly universal. But they come very near 
 being universal. I think that Charles Darwin's observation that they are 
 associated with regions of elevation is very generally sustained, even in the 
 depths of the ocean." 
 
 These remarks of Major Button, communicated to the writer in 
 a private letter, are of unusual interest. At the present time we 
 have no active volcano in the world more than 100 miles from the 
 sea or equivalent large body of water. It has been remarked that 
 Mt. Demavend is about 320 miles from the Mediterranean, but only 
 about 50 miles from the Caspian Sea, which is a deep body of salt 
 water. In the same way Jorullo in Mexico is about 80 miles from 
 the Pacific coast, but it has never been active since the first out- 
 break in 1759. Some of the volcanoes in the eastern range of the 
 Andes of Bolivia may be over 100 miles from the sea, but they are 
 much nearer Lake Titicaca and the terrible tropical rains which 
 constantly drench the eastern slopes of the Andes. 
 
 Thus all active volcanoes, to the number of about 400, are very 
 near the sea or equivalent large lakes. In the interior of continents 
 they die out for lack of adequate water supply. Unfortunately, 
 we do not know the contours of the sea in past geological ages very 
 accurately, but from Major Button's remarks, quoted above, it seems 
 probable that volcanoes have always developed in the neighborhood 
 of the sea and died out when the water receded to a considerable 
 distance from them. This is well illustrated by the extinct volcanoes 
 now found in the western part of the United States. Thus volcanoes 
 of former ages seem to follow the same law as those now active. 
 
 The present distribution of volcanoes proves conclusively that 
 they depend upon the sea. The erupting force is shown to be steam 
 by the great preponderance (999 parts in 1,000) which that vapor 
 has over all others. 
 
 The progressive extinction of volcanoes in the interior of 
 continents is, therefore, clearly intelligible. Not only do the vol- 
 canoes die out for lack of motive power to keep open the orifices, 
 but the earthquakes also famish in the same way, though to a lesser 
 degree, because their explosive force is distributed over a wider 
 area and does not require to be so concentrated. Unless a volcano 
 keeps moderately active it becomes permanently closed by lava 
 
I 9 o6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 377 
 
 hardening in its throat ; yet earthquakes which have no surface out- 
 let may successfully maintain a languid existence on a small supply 
 of water, and hence they may continue to be felt in a region long 
 after all volcanoes have been extinguished. 
 
 46. Explanation of immense outflows of lava such as are seen 
 in the plateau of Deccan and in Oregon and Utah. 
 
 For a number of years it has been a subject of remark among 
 geologists that the largest lava flows are not the output of volcanoes, 
 but of immense fissures which opened in the earth's crust and per- 
 mitted the welling forth of vast quantities of molten rock. Sir 
 Archibald Geikie (cf. Suess, Vol. I, p. 145) emphasized this view 
 as long ago as 1880. In recent years this theory of the origin of the 
 immense deposits of sheet lava seen in the region of the Columbia 
 River in Oregon and in Utah, as well as in the great tableland of 
 Deccan in India, has been very generally adopted. In all such cases 
 fissures no doubt opened and poured forth the molten rock through- 
 out their length. The subsidence of considerable areas of the earth's 
 crust may have contributed to this outflow, and different degrees 
 of liquidity are invoked to explain the observed phenomena. Reyer 
 suggests (cf. Geikie's "Geology," Vol. I, p. 301) that the degree 
 of saturation with gases and vapors may have influenced the form 
 of eruption, volcanic discharges resulting when the impregnation 
 was strong enough to cause eruption, and tranquil outpourings whea 
 the rock is but feebly saturated with explosive gases. 
 
 Major Dutton thinks differences of temperature as well as 
 chemical differences may have been more important in giving the 
 great lava flows their peculiar aspects. In regard to these outflows 
 in general, I believe that the crust cracked open on account of 'the 
 relative movement of neighboring portions. There are many ways 
 in which this could occur. If there were any appreciable tangential 
 pressure between two portions of the crust, the outpouring of lava 
 would be less easy ; but since we abandon the contraction theory and 
 deny that the mountains are due to the shrinkage of the crust, there 
 is on this hypothesis no pressure between the two portions of the 
 crust except that due to their weight when resting side by side. If, 
 therefore, the subterranean movements under two neighboring parts 
 should be such as to force a fault apart, there would be nothing to 
 prevent the lava from rising and pouring forth. 
 
378 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 In the valley of the Ganges the earthquake thrusts were towards 
 the Himalayas, and at some period in the upheaval of these moun- 
 tains the crust may have been cracked and also so pulled apart as 
 to allow the great lava flows observed in the plateau of Deccan. 
 Much of this may have occurred when the peninsula of India was 
 still under the sea. No doubt the great lava flows of Oregon and 
 Utah were similar in character, but the details of the process must 
 be left to future investigation. 
 
 47. On the vapors exhaled by volcanoes. 
 
 We have seen that, according to Geikie, 999 parts in 1,000 of the 
 vapors emitted by volcanoes is composed of steam, and have con- 
 cluded that this is the only original vapor operating under the earth's 
 crust. For the one-thousandth part of other gases such as sulphur- 
 ous oxide, carbon dioxide, hydrochloric acid, et cetera, may well 
 be derived from the heated rocks when saturated with steam, with- 
 out any other original gas existing in the earth. If any other gas 
 except steam were really active in the earth, either under the vol- 
 yanoes or in remote regions which are disturbed by earthquakes, we 
 may be very sure that at some of the volcanoes some of this gas 
 would escape and we should be able to recognize it. As such escape 
 does not occur we may be quite sure that steam is the only original 
 vapor operating within the earth. 
 
 The question of the dissociation of water vapor by the intense 
 heat of subterranean lava is also worthy of remark. It has been 
 found by analyses of the vapors escaping from Thera, where the 
 outlets are generally submarine, that the water gases are dissociated. 
 Immediately over the focus of eruption free hydrogen formed 30 
 per cent, of the gases emitted. In general the free hydrogen is fully 
 twice as abundant as the free oxygen, so that the mixture on coming 
 in contact with a burning body at once ignites with a sharp ex- 
 plosion. A considerable quantity of free nitrogen is also present, 
 and traces of other gases. Fouque, who has given most attention 
 to this subject, infers that the water- vapor of volcanic vents may 
 exist in a state of dissociation in the magma just beneath the earth's 
 crust. The free nitrogen is supposed to be derived from the air 
 absorbed in the water which percolates downward. 
 
 The question of the existence of true volcanic flames was first 
 
X9o6 .] SEE THE CAUSE OF EARTHQUAKES. 379 
 
 settled by the observations of Fouque, who showed that true flames 
 may arise when the free gaseous mixtures are ignited by red hot 
 stones from the volcano or by the strokes of lightning which play 
 so actively about the orifice of a volcano in eruption. No doubt the 
 dreadful tongues of fire so often seen to radiate from erupting vol- 
 canoes are to be explained very largely by the ignition of free gases 
 by thunderbolts produced by the condensation of clouds of aqueous 
 vapor. 
 
 48. Strabo speaks of the volcanoes as safety valves. 
 
 In his " Principles of Geology," 12 edition, Vol. I, p. 25, Lyell 
 remarks that the gifted Amasean geographer, Strabo, alluding to the 
 tradition that Sicily had been separated by a convulsion from Italy, 
 adds (Lib. vl, p. 396, edit. Almelov. Amst. 1707) that in his time 
 the land near the sea in those parts was rarely shaken by earth- 
 quakes, since open orifices exist whereby fire and burning matter 
 and water escape ; but formerly, when the volcanoes of ^Etna, the 
 Lapari Islands, Ischia and others were closed up, the imprisoned 
 fire and wind might have produced far more vehement movements. 
 " The doctrine, therefore," continues Lyell, " that volcanoes are 
 safety valves, and that the subterranean convulsions are probably 
 most violent when first the volcanic energy shifts itself to a new 
 quarter, is not modern." 
 
 X. CONCLUDED THEORY OF EARTHQUAKES. 
 
 49. All important earthquakes due to the action of explosive 
 forces within or just under the earth's crust. 
 
 Major Dutton seems to have had an inkling of the process here 
 involved when he wrote the following (" Earthquakes," p. 49-5) : 
 
 Fit remains now to refer to the possibility that many quakes whose 
 origin is unknown, or extremely doubtful, may, after all, be volcanic. This 
 must be fully admitted, and indeed, it is in many cases highly probable. 
 Evidences that volcanic action has taken place in the depths of the earth with- 
 out visible, permanent results on the surface abound in ancient rock ex- 
 posures. Formations of great geological age, once deeply buried and 
 brought to daylight by secular denudation, show that lavas have penetrated 
 surrounding rock-masses in many astonishing ways. Sometimes they have 
 intruded between strata, lifting or floating up the overlying beds without any 
 indication of escaping to the surface. Sometinies the lava breaks across a 
 series of strata and finds its way into the partings between higher beds. Or 
 
380 SEE -THE CAUSE OF EARTHQUAKES. 
 
 [October 19, 
 
 it forces its way into a fissure to form a dyke which may never reach the 
 surface. In one place a long arm or sheet of lava has in a most surprising 
 and inexplicable manner thrust itself into the enveloping rock-mass, and in 
 the older of metamorphic rocks these offshoots or apophyses cross each other 
 in great numbers and form a tangled network of intrusive dykes. In other 
 places the intruded lava formed immense lenticular masses (laccolites), which 
 have domed up the overlying strata into mountain masses. These intrusions, 
 almost infinitely varied in form and condition, are often, in fact usually, in- 
 explicable as mechanical problems, but their reality is vouched for by the 
 evidence of our senses. What concerns us here is the great energy which 
 they suggest and their adequacy to generate in the rocks those sudden, elastic 
 displacements which are the real initiatory impulses of an earthquake. They 
 assure us that a great deal of volcanic action has transpired in past ages far 
 underground, w r hich makes no either sign at the surface than those vibrations 
 which we call an earthquake." <$> ^ ^^ 
 
 The blowing out of a huge obelisk of granite some 300 meters 
 long and 120 meters in diameter but too large to get through 
 the orifice of Mount Pelee, and which therefore hung in the mouth of 
 that volcano, while the cone itself was split on all sides by the fearful 
 force of the explosion which had ejected the plug of granite from 
 the roots of the volcano, affords an excellent and familiar example 
 of what may be done by volcanic forces. This obelisk is illustrated 
 in the National Geographic Magazine for August, .1906, by photo- 
 graphs made by Professor A. Heilprin, who gives an excellent ac- 
 count of the history of the obelisk and its gradual disintegration. 
 If these forces are suddenly arrested, as they are in all earthquakes 
 which do not produce immediate eruptions, the shock is taken up by 
 the surrounding earth and we have a violent earthquake which may 
 be felt all over the world. 
 
 If the explosions are no larger than those involved in the erup- 
 tions of geysers the results are mere microseisms or earth tremors, 
 interesting enough to be sure, but of an unimportant character. A 
 really serious earthquake, to be felt all over the world, implies the 
 exertion of the most tremendous forces, and the way in which these 
 forces set the earth particles vibrating shows that they must depend 
 primarily upon the explosive power of steam-saturated lava at great 
 temperature. The result is a violent shaking of the whole overlying 
 layers of rock and the occasional upheaval of volcanoes where the 
 strata are fractured and weak. The very way in which the earth 
 twists, heaves, labors and vibrates shows the awfulness of the pent-up 
 
igo6.] 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 FIG. 15. 
 
 381 
 
 The shattered obelisk of Mt. Pelee, photograph taken by Professor 
 Angelo Heilprin. 
 
 forces; and it is no wonder that such shaking should throw down 
 buildings, settle soft " made " ground, leaving fissures where it is 
 shaken down, and even cause faults in the overlying mountains to 
 slip or open. 
 
 50. But the movements of faults are the effects of the earth- 
 quake, not the cause. 
 
 Such small slips as are usually observed would not account for 
 the enormous forces shown in the accompanying earthquakes. Be- 
 sides the horizontal form of the vibrations, so frequently shown, as 
 more fully pointed out in 51, could not be explained by simple 
 
382 
 
 SEE THE CAUSE OF EARTHQUAKES. 
 
 [October 19, 
 
 subsidence. Yet it is evident that these faults may easily result from 
 the mighty steam pressure which has shaken the earth sufficiently to 
 break the layers of rock, and give relief to the pent-up forces. In 
 so-called tectonic earthquakes the forces usually are greater than 
 those involved in volcanic action, because the whole body of the 
 overlying strata must be shaken to afford the smallest relief, but are 
 seldom so concentrated as would be required for the upheaval of a new 
 volcano through the entire depth of the unbroken crust. If the crust 
 is already badly fractured or breaks more easily than it gives relief 
 
 FIG. 1 6. 
 
 when shaken, sometimes new volcanoes actually break forth, especi- 
 ally under the sea, where the explosive forces are greatest. 
 
 When we consider the terrific shock required to break all the 
 horizontal strata in a situation where the imprisoned forces are 
 deep-seated we can readily imagine that yielding will often occur 
 through condensation of soft rocks, through tightening up of crev- 
 ices, joints and faults, and sometimes by uplifts of all the strata, 
 affording room for the injection of a layer of lava of large extent, 
 or its diffusion by spreading into surrounding areas. The move- 
 ment of the fluid is shown by the arrows in the figure. In that case 
 
i 9 o6 1 SEE THE CAUSE OF EARTHQUAKES. 383 
 
 the eruptions may be either subterranean or submarine, as in the 
 elevation of the Chilean coast noticed by Darwin and Fitzroy in 1835. 
 
 On the theory of faults, as now held, it is difficult if not im- 
 possible to account for the observed elevations. It is perhaps true 
 that elevations apparently are rarer than subsidences, yet they 
 are much more significant and furnish a better criterion of the forces 
 at work, since many subsidences on land are due to settling of soft 
 ground which has never been consolidated under pressure. But 
 we must remember that only very few absolute levels are accurately 
 known and still fewer remeasured after an earthquake ; and therefore 
 while subsidences appear to be the more general phenomena, espe- 
 cially in regions of soft earth, one may well be very doubtful whether, 
 in the case of hard ground, elevations, though mostly unnoticed 
 because there is no easy means of measurement, do not really pre- 
 dominate. Reasons connected with the mode of formation of moun- 
 tains and the elevation of coast lines, given heretofore, point strongly 
 to elevation as the more general movement in nature. For it is this 
 movement which has uplifted both mountains and continents, and we 
 cannot suppose that it has ceased to be the dominant influence, 
 though it generally escapes notice, because we have no means of 
 detecting it, while local subsidences frequently are easily recognized, 
 and we naturally look for it because of the frequent shaking down of 
 " made " ground. 
 
 In most earthquakes the heaving force is distributed over a 
 considerable area, and when the stress becomes great enough a 
 movement takes place along the nearest fault line the path of least 
 resistance and the observer who sees the slip says the movement 
 of the fault caused the earthquake. 
 
 The fact that most earthquakes are found to originate at a depth 
 of from ten to twenty miles shows that the epicentrum is below the 
 depth at which the strata have any opportunity of moving; and the 
 proof that the shock usually comes from an area and not from a 
 point or from a line, shows that the shock depends on an explosive 
 stress spread over a considerable region, and in no way depends on 
 dislocational or fault movements, which are always quite superficial. 
 
 51. Explanation of the rotatory motion observed during an 
 earthquake. 
 
384 SEE THE CAUSE OF EARTHQUAKES. [October i 9f 
 
 If we consider with attention how simple the motion of sub- 
 sidence or elevation, in a supposed fault, really is, we shall per- 
 ceive that such displacement cannot give the earth particle an ap- 
 preciable rotatory motion. For the two sides of a fault are con- 
 ceived as continuous and unbroken parts of the earth's crust, and 
 thus securely fixed at their backs, and moving only at their faces, 
 where the fracture exists. The motion is essentially like that of 
 a double cellar door opening or closing very slightly in the middle, 
 with rigid hinges at their backs. They can only go up or down 
 together, or one up and one down, with little or no horizontal 
 motion; and thus cannot produce a revolving tremor when opened 
 or shut. 
 
 It is exactly so with the slipping of a supposed fault in the 
 earth's crust. It is mechanically inconceivable how the vertical sub- 
 sidence of a fault could give the earth particle a revolving motion in 
 the plane of the horizon. At most, such a slip could produce an 
 up-and-down oscillation, with the path of the ellipse described by 
 the particle very nearly in the vertical plane. 
 
 Now the earthquake which destroyed San Francisco gave the 
 earth particle a large and conspicuous rotation almost parallel to 
 the horizon ; it was " a twister," and this rotatory character was 
 so marked as to attract instant attention. The conspicuous rota- 
 tion and the difficulty of explaining such a motion by the theory of 
 faults, as ordinarily stated, led the writer to question the validity 
 of that theory. As most of the violent earthquakes have both rota- 
 tion and vertical movement, it is evident that the difficulty felt in 
 explaining the San Francisco earthquake is very generally encoun- 
 tered in earthquake phenomena. 
 
 It seems to have escaped the attention of seismologists that 
 rotatory earthquakes require explanation, and the theory of subsi- 
 dences of rocks and faults is incapable of furnishing it. How then 
 can motion of rotation be accounted for? 
 
 To answer this question in the simplest way we may recall that 
 there are many impulses which can give a motion of rotation, either 
 by direct impact or by the " kick back," or recoil of reaction. If, 
 for example, we suppose an orifice to be forced through an under- 
 lying or overlying layer of rock so that lava escapes under great 
 
I 9 o6.] 
 
 SEE-THE CAUSE OF EARTHQUAKES. 
 
 385 
 
 pressure, but the resistance is unequal in the different directions, 
 it is evident that as the imprisoned matter expands or explodes, the 
 reactions in the different directions will be inversely as the corre- 
 sponding resistances to the escaping fluid. In general the explosion 
 will give a rotation to the surrounding particles, since the moment 
 
 FIG. 17. 
 
 Seismographic record of the San Francisco Earthquake, April 18, 1906, 
 taken at the Chabot Observatory, Oakland, California, Professor Charles 
 Burkhalter, Director. 
 
 of the resistances will not pass exactly through the center of the 
 exploding mass, and the rotation is as likely or more likely to be 
 in the horizontal plane than in any other. If the explosion or rapid 
 diffusion of steam pressure takes place in stratified rock, the strata 
 are fairly sure to be approximately horizontal ; but no doubt most all 
 
386 SEE THE CAUSE OF EARTHQUAKES. [October 19 
 
 violent earthquakes occur at greater depths, and no one can predict 
 the inequalities of the encountered resistance. It would all depend 
 on the shape and relative situation of the expanding matter in rela- 
 tion to the surrounding rock, which would also be at high tem- 
 perature. 
 
 As the reservoir of steam-saturated lava would rarely be spherical 
 and would usually be a layer, as we have seen in part IV, and several 
 separate and distinct reservoirs might develop near one another, 
 a streaming of the released matter, when the walls yielded and the 
 strain is released and the fluid quickly adjusted itself to the new 
 surroundings, would almost always produce some kind of rotatory 
 motion, and it would always lie in a plane between the horizon and 
 the zenith. If the layers were complex or of irregular figure this 
 movement might be made up of several parts and the adjustment 
 occupy several seconds, and possibly minutes, of time, even when 
 the forces are enormous and the motion correspondingly rapid. 
 
 The successive powerful impulses or blows imparted to the sur- 
 rounding earth might be of unequal intensity and not all in the same 
 plane, and moreover the vibration would continue for a short time 
 after the internal movement had ceased, on account of the elasticity 
 of the rocks of the earth's crust. Lava saturated with superheated 
 steam would behave essentially like steam in an exploding boiler, 
 because it would give body and momentum to the spreading steam 
 and be capable of transmitting shocks of appalling violence. 
 
 This gives us a conception not only of the process involved in an 
 earthquake, but also of how the irregularities noted by seismographs 
 might be accounted for; and when we recall that the subterranean 
 boiler might surpass the largest mountain in size, or be flattened 
 into an immense disc 1 of slight thickness, with vent chiefly or 
 wholly at the sides, we can easily understand the terrific forces 
 which shake the whole earth when once the surrounding walls give 
 way or a fault moves, so that an explosion and diffusion of the lava 
 is effected. 
 
 It must be assumed, for reasons already fully developed, that 
 sensible readjustment and motion of large masses of steam-saturated 
 
 1 It is found by investigation that many of the tectonic earthquakes origi- 
 nate in an area of considerable extent. 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 387 
 
 lava or molten rock take place in every important earthquake ; 
 confining walls and caverns of unequal pressure are reduced to one 
 common strain, and the resulting motion involves rapidly acting 
 and enormously powerful forces, which may shake the whole earth 
 and sometimes crack or derange the overlying strata of rocks many 
 kilometers deep. 
 
 52. Explanation of the New Madrid earthquake and some 
 other earthquakes often classed as tectonic. 
 
 In his excellent work on " Earthquakes " Major Button remarks 
 that the New Madrid earthquake could not be said to have any real 
 depth, although he says it was strongly felt as far east as Boston, 
 and in fact nearly all over the United States. His reasoning is 
 based upon the supposition that it was due to the sinking of some 
 bottom land, which seems to be quite unjustifiable. The subsidences 
 happened, indeed, but we shall find the best reasons for holding 
 that they were produced by explosions within the earth, which 
 must therefore have been at least ten miles deep. Unless it had 
 been of about this depth this earthquake could not have been so 
 strongly felt throughout so wide an area of country. 
 
 It seems probable that the water seeped down from the Mis- 
 sissippi River, which always overflowed badly here, and worse in 
 prehistoric times than now, because the country was then much 
 more heavily timbered; or the seepage was a survival effect of the 
 fractures of the Ozark Mountains, from the time when the Gulf of 
 Mexico extended far up the river, and thus was beginning to form 
 a sea valley about parallel to the general trend of the Ozark 
 Mountains which were left unfinished. 
 
 Thus the cause producing this earthquake is apparently not dif- 
 ferent from those at Charleston, San Francisco and many other 
 places which have experienced so-called tectonic earthquakes. 
 
 The most trustworthy accounts of the New Madrid earthquake 
 show that it was the most violent and destructive earthquake ever 
 felt in the United States. Lyell has given an excellent account of 
 this earthquake, based upon observations gathered during a per- 
 sonal visit in 1846, but the records here followed are of still earlier 
 date. 
 
 Over a region of 300 miles in length, but especially from the 
 
388 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 mouth of the Ohio to that of the St. Francis River, the ground rose 
 and sank in great undulations ; trees had to be felled across the 
 fissures for the preservation of the inhabitants, and lakes were 
 formed and drained again, and many large streams changed their 
 channels and even their courses. Physical evidences of this terrible 
 convulsion are still seen near New Madrid, where the loose land 
 had settled so much as to form swamps, leaving great cypress and 
 other trees so deep in the water that they died (some were entirely 
 submerged), and in many places still remain to show where the 
 water rose in 1811. Little Prairie, now called Caruthersville, 20 
 miles below New Madrid, was considered the center of greatest 
 violence. The first severe earthquake occurred on the night of De- 
 cember 1 6, 2:15 a. m., -and shook severely the whole region of the 
 Ohio, Mississippi and Missouri rivers, as far north as the Lakes, as 
 far east as the Alleghanies, south to the gulf, and west to the 
 Rocky Mountains. 
 
 One of the best accounts of this great earthquake, drawn from 
 contemporary sources of information, is that given by Professor G. 
 C. Broadhead, of the University of Missouri, in the American Geolo- 
 gist for August, 1902. The leading points may be summarized thus : 
 
 1. After the first severe shock at 2:15 in the night, smaller 
 shocks followed, and at 7 a. m., December 16, 1811, came a much 
 more severe shock ; then came lighter shocks daily or oftener, until 
 January 23, 1812, when an extremely severe shock was felt; con- 
 tinual agitations were felt till February 4, which brought another 
 severe shock, and four more followed next day ; and on February 
 7, at 4 p. m., came one so much more violent than the preceding that 
 it was called " the hard shock." Hundreds and even thousands of 
 smaller after-shocks have continued at irregular intervals to the 
 present time. " Two series of the (original) concussions were par- 
 ticularly terrible." 
 
 2. Thus there were at least eight very severe shocks, each of 
 which did great damage and spread devastation far and wide. 
 
 3. The eye-witnesses generally agree that these severe shocks 
 were preceded by heavy subterranean thunder ; " a loud roaring 
 sound like steam escaping from a boiler " ; " distant rumbling sounds 
 succeeded by discharges as if a thousand pieces of artillery were 
 
i 9 o6.] SEE THE CAUSE OF EARTHQUAKES. 389 
 
 suddenly exploded " ; "a rumbling like distant thunder " ; " an aw- 
 ful noise resembling loud and distant thunder, but more hoarse 
 and vibrating," are some of the descriptions. 
 
 4. All the severe shocks were accompanied by the escape of 
 sulphurous vapors from the great fissures which opened in the 
 earth; and on December 16, after 7 a. m., the severe shock was 
 accompanied by total darkness till sunrise ; in one of the shocks 
 witnessed by J. J. Audubon near Henderson, Kentucky, he took 
 the cloud of vapor on the western horizon for a rising storm; and 
 in general the eye-witnesses in Missouri agree that flashes were 
 frequently observed as if these vapors were generating electric 
 discharges in the air. 
 
 5. The severe earthquake shock of February 7, 1812, lasted four 
 minutes, according to Jarred Brooks of Louisville, Kentucky, who 
 seems to have given much attention to these phenomena, and 
 classified them carefully. He said : " It seemed as if the surface of 
 the earth was afloat and set in motion by a slight application of im- 
 mense power, then a boiling action succeeded, houses oscillate, 
 gables and chimneys of many houses are thrown down." 
 
 6. Great fissures running nearly north and south (Lyell's meas- 
 urement in 1846 made the direction from 10 to 45 W. of N.) were 
 formed five miles long, ten feet wide, and four feet deep, by the 
 great undulations which came from the west ; hills were sunk, forests 
 inundated, lakes drained, swamps formed, the bed of the Mississippi 
 River upheaved so that its waters backed up for a time, then came 
 booming on, broke over and swept everything before them, and 
 nearly all the shipping was destroyed; hills rose where lakes and 
 swamps formerly stood; the waters of the Mississippi receded from 
 its banks and then returned as a wall fifteen or twenty feet high, 
 tearing the boats from their moorings and carrying them up a 
 creek a quarter of a mile. 
 
 7. Water and sand, and some coal or lignite, and sulphur was 
 ejected from the fissures, the materials being thrown forty feet 
 high, which aided in filling the air with noxious vapors. 
 
 The ejection of this material from the ground would explain 
 some of the noises which accompanied these earthquakes, but not 
 all. The deep subterranean thunder preceding the shocks cannot 
 
390 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 be explained except by explosions in the earth. Some of the land 
 was 'no doubt simply settled by shaking, but this will not account for 
 the upheaval of the bed of the Mississippi River so as to make its 
 waters flow upstream ; and we seem obliged to admit that in addition 
 to subsidence there was also elevation of ground. 
 
 The New Madrid earthquake is extremely remarkable for its nu- 
 merous severe shocks and the long intervals at which they occurred, 
 the last great shock of February 7, 1812, being the worst of all. The 
 order of events does not harmonize so well in the tectonic theory of 
 rock slipping as it does with Strabo's account of a volcanic out- 
 break near Chalcis, in Euboea, in which the shocks ceased only 
 after a fissure opened on the Lelantine Plain and ejected a fiery 
 river of mud. To be sure, no lava was ejected in the New Madrid 
 earthquake, but the phenomena resemble those described by Strabo, 
 in which elastic vapors within the earth were seeking release. If 
 no single outburst of lava and steam could break forth above ground 
 the agitation might on that account continue all the longer, and an 
 impartial study of this remarkable earthquake strongly suggests 
 this explanation. The whole course of events is singularly incon- 
 sistent with any suggestion that mere slipping of rock or subsidence 
 was involved. 
 
 We have examined this earthquake in detail, because it is little 
 understood and so complex and extraordinary in respect to location 
 and duration that it constitutes a severe test of any theory ; and yet 
 a comprehensive theory must be able to account for such phenomena 
 without prejudice to the historical facts. 
 
 53. Other important earthquakes. 
 
 If we examine attentively the available details of various other 
 earthquakes which are classed as tectonic we shall find that most 
 if not all of them give evidence of high explosive power within 
 the earth. Under ordinary conditions it may be supposed that these 
 forces are exerted long enough to afford partial or complete release 
 from the strain, and then the agitation ceases or continues in the 
 form of slight after-shocks. It is only when unlifts occur that 
 we have direct proof of an uplifting force, the visible effect of 
 subterranean explosive power; but it would seem that the heaving 
 of the earth accompanying the rending of strata and breaking of 
 
X9 o6.] SEE THE CAUSE OF EARTHQUAKES. 391 
 
 rocks is a fairly obvious indication of imprisoned vapors seeking 
 release from contact with molten rock. 
 
 In fact nearly all the - recent important earthquakes of great 
 violence, and inappropriately classed as tectonic, present strong 
 indications of being due to steam power. It probably is not too 
 much to say that it is doubtful if one of them could be fairly ex- 
 plained by mere subsidence due to the slipping of rock faults. 
 
 In almost every case it will be found that the rock slipping 
 noted at the surface is small, while it is conceded that the chief 
 effect must depend upon great forces exerted from deep down. This 
 seems to be substantially admitted by most writers, who have a sus- 
 picion that the forces are singularly deep-seated and otherwise act 
 in a strange way. The Charleston earthquake originated twelve 
 miles below the surface, while the San Francisco earthquake will 
 not prove to be of less depth. The Bengal-Assam earthquake of 
 June 12, 1897, investigated by Mr. R. D. Oldham, must have had 
 great depth, owing to the wide extent of country over which is was 
 felt and the great intensity of the shocks. The same is true of the 
 Mino-Owari earthquake of October 18, 1891, and all other earth- 
 quakes of high intensity which were widely felt. The Valparaiso 
 earthquake of August 16, 1906, was among the most terrible of 
 modern times, and it was so widely felt that the depth could hardly 
 be less than fifteen or twenty miles. 
 
 It seems desirable to direct attention to the fact that probably 
 not one of these great earthquakes has occurred in a region where 
 seeping of water and formation of steam may not have been the 
 dynamic cause. The center of the violent Bengal-Assam earthquake 
 was under the great Bramaputra River, where it spreads to great 
 width and drains an immense volume of water from the Himalayas. 
 This region was originally a deep trough in the sea and has since 
 been filled in ; and owing to the great surface drainage is still 
 essentially an inland sea not far from the Bay of Bengal, which 
 receives also the Ganges as well as the Bramaputra. The situation 
 is very similar to that of New Madrid, but is nearer the sea, and 
 has the Himalayas on the north, from which the drainage is enor- 
 mous. Besides there is in that region a natural survival of the forces 
 which uplifted the Himalayas, and this ancient sea valley did most 
 of the work involved in the raising of these mighty mountains. 
 
 PROC. AMER. PHIL. SOC., XLV. 184 Y, PRINTED FEBRUARY 25, 1907. 
 
392 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 If we examine the violent Japanese earthquakes we shall find 
 that most of them, as Professor Milne says, arise under the sea near 
 the seashore or at the heads of bays, where the influence of the sea 
 would predominate. .It seems impossible to doubt that they are due 
 to the influence of steam formed in the earth beneath ; for the island 
 of Nipon is essentially narrow, mountainous and broken by great 
 irregularity of topography, so that the seepage of water is easily 
 accounted for, and more especially since the worst earthquakes are 
 on the east coast and partly under the sea on the edge of the Tus- 
 carora Deep, where the shore is steepest and the sea pressure 
 greatest. Moreover, it-4s"Shown in-~ -^^-that the whole island of 
 Nipon has been uplifted by injections of matter expelled largely 
 from under the bed of the Tuscarora Deep ; and what is more natT 
 ural than to suppose that this process is still going on ? This infer- 
 ence in fact is confirmed by the secular elevation of the east coast 
 observed within the historical period. 
 
 The regions of Iberian Peninsula visited by violent earthquakes 
 are those similarly exposed to the sea Lisbon and the southern prov- 
 inces, such as Andalusia, which are also broken and mountainous, and 
 have been recently rising from the sea. In Italy the region of great- 
 est and most violent disturbance is Calabria, the " toe of the boot," 
 which is a long peninsula nearly surrounded by the sea and of re- 
 markably fractured and broken topography. The islands of Sicily, 
 Ischia, Lapari, Stromboli and the coast near Vesuvius is similar, 
 and all these regions are still rising from the sea. Nearly all Greece 
 is very broken and mountainous and it has always suffered severely 
 from earthquakes. In Hungary, where the severe earthquake of 
 Agram, Croatia, occurred November 9, 1880, the country is full of 
 hot springs, indicating an abundance of underground water, and 
 while the volcanoes have died out the earthquakes still survive. The 
 earthquake in Silesia, June n, 1895, occurred in a region of the 
 same kind. 
 
 If now we turn to Central America we find it a narrow broken 
 country with sea on both sides ; and on the other side of the Pacific, 
 New Zealand is a narrow island like Japan, and the presence of 
 violent earthquakes there is not strange. The same may be said 
 of the Aleutian Islands and the whole East Indian Archipelago, 
 including the Philippines. 
 
I90 .6] SEE THE CAUSE OF EARTHQUAKES. 393 
 
 The same reasoning applies to all the South American earth- 
 quakes, including those which have destroyed Caracas, Cumana, and 
 other places in Venezuela. This coast lies between the eastern spur 
 of the Andes and the waters of the Caribbean Sea, which is one 
 of the deepest parts of the Atlantic, and the intensity of the sub- 
 terranean forces is shown by the violence of the volcanic outbreaks 
 in the Lesser Antilles. 
 
 _J 54. Are any earthquakes really tectonic? 
 
 \It is held that many earthquakes are tectonic because there is 
 externally only a movement of the strata, as if they were seeking 
 release from strain, and obvious volcanic forces do not appear. 
 Boussingault long ago concluded that many of the earthquakes in 
 the Andes depended on the settling of the strata in these mountains, 
 and this was the beginning of the tectonic theory that most earth- 
 quakes are due to collapse or movement for release of strain. To 
 test the validity of this theory, it is advisable to apply similar reason- 
 ing to the mountains of a country remote from the seacoast. We 
 choose for this purpose the Rocky Mountains in Colorado, which 
 were formerly near but are now remote from the sea. If what 
 Boussingault witnessed in the Andes was really the settling of the 
 mountains, the same effects ought to be going on in the Rocky 
 Mountains of Colorado. So far as the records of history go, it may 
 be safely said that not a single serious earthquake has ever visited 
 Colorado, and yet many should have been felt if Boussingault's 
 theory of the settling of the mountains is correct. As change is 
 usually ascribed to secular cooling, why should the Andes settle 
 and not the Rocky Mountains? From the absence of earthquakes 
 in Colorado, it is evident that tectonic movements have ceased in 
 that region, though secular cooling has not; and thus we see that 
 these movements after all do not depend on settlement of the moun- 
 tains due to the shrinkage of the earth. For, if so, it is incredible 
 that the Rocky Mountains can have already attained a perfectly 
 stable position. They ought to be still collapsing like the Andes, 
 since secular cooling is always going on. Thus we see that all 
 earthquakes must depend on underlying explosive forces, and not 
 on mere adjustments of strata to secure release of strain or stability 
 of position, required by the progress of secular cooling^j 
 
394 SEE-THE CAUSE OF EARTHQUAKES. [October 19, 
 
 Another way to reach this result is to recall that the investiga- 
 tion of earthquakes in Hungary, Croatia, Bohemia and other coun- 
 tries shows that most frequently the impulse proceeds from an 
 underlying area of considerable extent, and not from a point nor 
 line, though the surface movement may be chiefly along the nearest 
 fault. If the underlying area from which the shock proceeds is 
 elliptical, as usually happens, we may be quite sure that this diffusion 
 of the impulse over an area indicates an underlying outspread sheet 
 of lava saturated with steam, which finally acquires such tension that 
 it is enabled to shake the overlying crust and cause a movement 
 along the fault lines where the crust is broken and movement is 
 easiest. The lava sheet seeks readjustment, and in the process of 
 equalizing the strain, movement of the molten rock takes place, and 
 the faults not only move vertically, but often also horizontally. 
 This is a natural and simple explanation of fault movement, and it 
 accounts for the rotatory motion so frequently noticed in earth- 
 quakes. The great earthquakes of Lisbon, Arica and Iquique have 
 usually been classed as tectonic, but in view of the sinking of the 
 bed of the sea shown by the accompanying seismic sea waves, it is 
 clear that all these terrible disturbances were due to the expulsion 
 of lava from under the ocean. 
 
 55. The geological significance of earthquakes. 
 
 In their new work on " Geology " (Vol. I, p. 534) Chamberlin 
 and Salisbury remark that " Earthquakes are of much less im- 
 portance, geologically, than many gentler movements and activities. 
 Disastrous as they sometimes are to human affairs, they leave few 
 distinct and readily identifiable marks which are more than tem- 
 porary." Mention is made of the effects of earthquakes in fractur- 
 ing the rocks of the earth's crust, but the fractures, it is pointed out, 
 do not show at the surface when the soil is deep. These authors 
 also remark (p. 527) that " the most prevalent (source of earth- 
 quakes) is probably the fracture of rocks and the slipping of strata 
 on each other in the process of faulting." 
 
 Let us now examine these remarks a little more carefully. If 
 earthquakes are due to fracture of rocks and the slipping of strata, 
 it follows that the forces involved here have played no part in the 
 original formation of the globe, but are the effects of collapse after 
 
19o6 .) SEE THE CAUSE OF EARTHQUAKES. 395 
 
 the mountains were formed. On this view, earthquakes would be 
 of very small consequence geologically, because they are associated 
 only with destructive and not with constructive processes. But is 
 such a view tenable ? If earthquakes are not due to the same forces 
 which upheaved the mountains, what other forces besides these have 
 been active in the development of the globe? The only forces of 
 construction now felt upon the earth are those exerted in earth- 
 quakes. So far as we can see, no other constructive forces are at 
 work. Therefore, the forces felt in earthquakes are identical with 
 those which formed the mountains, and this is sometimes admitted, 
 though mountain formation itself is assigned to the wrong cause. 
 
 Destructive forces such as erosion are wearing down the struc- 
 ture of the globe, while earthquakes are the only known forces which 
 are building it up. We take it, therefore, that so far from being of 
 little importance geologically the forces felt in earthquakes are of 
 the greatest importance, and most of the constructive forces 'in the 
 development of the earth are due to this cause. The destructive 
 effects of earthquakes are only incidental to the more fundamental 
 constructive purpose which underlies the operation of these forces. 
 When an earthquake occurs rocks in unstable positions fall, loose 
 sediment is shaken down, and other settlements occur, but the real 
 constructive work consists in upheavals, little by little it may be, 
 of mountains, islands, coasts, plateaus and larger areas. These 
 elevations are actually witnessed in certain earthquakes, and could 
 not possibly arise from any processes of collapse. Sometimes these 
 constructive forces work slowly and quietly, but usually with more 
 or less violence; and the usual method of elevation is by the injec- 
 tion of lava saturated with steam. 
 
 What has been taken to be the cause of earthquakes, namely, 
 the slipping of rocks, is really the effects of more deep-seated ex- 
 plosive forces. Earthquakes, therefore, are not due to the effects 
 of secular cooling, but to the vapor of steam arising from the pene- 
 tration of water into the heated layers just beneath the crust. If 
 earthquakes were due to cooling they ought to be as frequent in 
 desert regions as in deep seas along the shores of continents, where 
 they really are abundant. 
 
396 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 XL EARTH TREMORS AND TELESEISMIC DISTURBANCES. 
 
 56. Slight movements of the earth. 
 
 A great many questions recently have been discussed relating to 
 slight slow oscillations shown by horizontal pendulums and other 
 instruments designed to detect delicate changes of level. No doubt 
 very many of these changes are correctly ascribed to the yielding 
 of the solid earth, due to variable loading of the soil by tidal, sea- 
 sonal and meteorological influences. But may not others be due 
 to the slow but steadily varying influence of subterranean forces 
 as discussed in this paper? Some of the strains thus constantly 
 arising would be released by microseismic disturbances which show 
 no periodicity' or regularity ; while others would be cumulative and 
 have at length a small secular effect. The fact that the ground in 
 most places is comparatively so stable under the test of astronomical 
 observations seems to show that these effects usually are slight, 
 -except in the neighborhood of the sea ; but many small irregular 
 disturbances occur, and it is not improbable that a considerable 
 number of them may have their origin hidden deeper in the earth 
 than heretofore has been suspected. If great earthquakes originate 
 .at depths of from eight to twenty miles, we may be sure that the 
 forces there at work produce some surface changes of level even if 
 mo violent outbreaks occur. 
 
 57. Humboldt' s views on earthquakes. 
 
 Jn the fifth volume of the Cosmos (p. 288) Humboldt justly 
 remarks how Charles Darwin, " with his peculiar generalizing view, 
 has grasped the connection of the phenomena of earthquakes and 
 eruptions of volcanoes under one point of view." 
 
 In his "Views of Nature," Vol. I, p. 361, Humboldt alludes to 
 the nearly simultaneous occurrence of volcanic and seismic phe- 
 nomena in places widely separated, and says : 
 
 " All these phenomena prove that subterranean forces are manifested 
 either dynamically, explosively, and attended by commotion, in earthquakes; 
 or possess the property of producing, or of chemically modifying a substance 
 in volcanoes; and they further show, that these forces are not seated near 
 . the surface in the thin crust of the earth, but deep in the interior of our 
 planet, whence through fissures and unfilled veins they act simultaneously at 
 widely distant points of the earth's surface." 
 
 And in his " Travels," Vol. I, p. 172-3, he adds: 
 
I9 o6.] SEE THE CAUSE OF EARTHQUAKES. 397 
 
 " Everything in earthquakes seems to indicate the action of elastic fluids 
 seeking an outlet to diffuse themselves in the atmosphere. Often on the 
 coasts of the Pacific, the action is almost simultaneously communicated from 
 Chili to the Gulf of Guayaquil, a distance of six hundred leagues; and, what 
 is very remarkable, the shocks appear to be the stronger in proportion as 
 the country is more distant from burning volcanoes. The granitic mountains 
 of Calabria, covered with very recent breccias, the calcareous chain of the 
 Apennines, the country of Pignerol, the coasts of Portugal and Greece, those 
 of Peru and Terra Firma, afford striking proofs of this fact. The globe, it 
 may be said, is agitated with the greater force, in proportion as the surface 
 has a smaller number of channels communicating with the caverns of the 
 interior. At Naples and at Messina, at the foot of Cotopaxi and of Tun- 
 guragua, earthquakes are dreaded only when vapor and flames do not issue 
 from the craters. In the kingdom of Quito, the great catastrophe of Riobamba 
 led several well-informed persons to think that that country would be less 
 frequently disturbed, if the subterranean fire should break the porphyritic 
 dome of Chimborazo; and if that colossal mountain should become a burning 
 volcano. At all times analogous facts have led to the same hypotheses. 
 The Greeks, who like ourselves, attributed the oscillations of the ground to 
 the tension of elastic fluids, cited in favour of their opinion, the total cessa- 
 tion of the shocks at the island of Eubcea, by the opening of a crevice in the 
 Lelantine plain." 1 
 
 58. Views of Charles Darwin. 
 
 In the " Voyage of a Naturalist," Chapter xiv, Darwin says : 
 
 " The forces which slowly and by little starts uplift continents, and 
 those which at successive periods pour forth volcanic matter from open 
 orifices, are identical. For many reasons, I believe that the frequent quak- 
 ings of the earth on this line of coast (Chili) are caused by the rending of 
 the strata, necessarily consequent on the tension of the land when upraised, 
 and their injection by fluidified rock. I believe that the solid axis of a moun- 
 tain differs in its manner of formation from a volcanic hill, only in the 
 molten stone having been repeatedly injected, instead of having been re- 
 peatedly ejected. 
 
 "Moreover, I believe that it is impossible to explain the structure of 
 great mountain-chains, such as that of the Cordillera, where the strata, 
 capping the injected axis of plutonic rock, have been thrown on their edges 
 along several parallel and neighboring lines of elevation, except on this 
 view of the rock of the axis having been repeatedly injected, after intervals 
 sufficiently long to allow the upper parts or wedges to cool and become solid ; 
 for if the strata had been thrown into their present highly inclined, vertical, 
 and even inverted positions, by a single blow, the very bowels of the earth 
 would have gushed out; and instead of beholding abrupt mountain axes of 
 rock solidified under great pressure, deluges of lava would have flowed out 
 at innumerable points of every line." 
 
 "The shocks ceased only when a crevice, which ejected a river of fiery 
 mud, opened in the plain of Lelantum, near Chalcis." Strabo. 
 
398 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 This view of mountain formation is essentially identical with 
 that here adopted, except that I conceive the interior part of a 
 mountain, whether a peak or a chain, to be filled underneath with 
 porous lava, which explains the feebleness of the attraction of 
 mountains and the readiness with which they are converted into 
 volcanoes when once their tops are burst open during the paroxysms 
 of an earthquake. The present views therefore confirm and some- 
 what extend those Held by the elder Darwin seventy years ago. In 
 Chapter VI of his " Geological Observations on Volcanic Islands " 
 the great naturalist says : 
 
 " Some authors have remarked that volcanic islands occur scattered, 
 though at very unequal distances, along the shores of the great continents, 
 as if in some measure connected with them. In the case of Juan Fernandez, 
 situated 330 miles from the coast of Chile, there was undoubtedly a con- 
 nection between the volcanic forces acting under this island and under the 
 continent, as was shown during the earthquake of 1835. The islands, more- 
 over, of some of the small volcanic groups which thus border continents, 
 are placed in lines, related to those along which the adjoining shores of the 
 continents trend; I may instance the lines of intersection at the Galapagos, 
 and at the Cape de Verde Archipelagoes, and the best marked line of the 
 Canary Islands. If these facts are not merely accidental, we see that many 
 scattered volcanic islands and small groups are related not only by proximity, 
 but in the direction of the fissures of eruption to the neighboring continents 
 a relation which Von Buch considers characteristic of his great volcanic 
 chains. 
 
 " We ought not, however, to suppose, in hardly any instance, that the 
 whole body of matter, forming a volcanic island, has been erupted at the 
 level on which it now stands ; the number of dikes, which seem invariably to 
 intersect the interior parts of every volcano, show, on the principles of M. 
 Elie de Beaumont, that the whole mass has been uplifted and fissured. A 
 connection, moreover, between volcanic eruptions and contemporaneous ele- 
 vations in mass 1 has, I think, been shown to exist in my work on Coral 
 Reefs, both from the frequent presence of upraised oreanic remains and 
 from the structure of the accompanying coral reefs. Finally, I may remark, 
 that in the same Archipelago (Galagapos), eruptions have taken place within 
 the historical period on more than one of the parallel lines of fissure: thus, 
 at the Galapagos Archipelago eruptions have taken place from a vent on 
 Narborough Island, and from one on Albemarle Island, which vents do not 
 fall on the same line; at the Canary Islands, eruptions have taken place in 
 Teneriffe and Lanzarote; and at the Azores, on the three parallel lines of 
 Pico, St. Jorge, and Terciera. Believing that a mountain axis differs essen- 
 
 1 A similar conclusion is forced on us by the phenomena which accom- 
 panied the earthquake of 1835, at Conception, and which are detailed in my 
 paper (Vol. V, p. 601) in the Geological Transactions. 
 
I906 .] SEE THE CAUSE OF EARTHQUAKES. 399 
 
 tially from a volcano, only in pultonic rocks having been injected, instead of 
 volcanic matter having been ejected, this appears to me an interesting circum- 
 stance; for we may infer from it as provable, that in the elevation of a 
 mountain chain, two or more of the parallel lines forming it may be upraised 
 and injected within the same geological period." 
 
 59- Views of Professor Milne. 
 
 Professor Milne has recently expressed views of somewhat 
 similar character, many of which agree closely with those reached in 
 this paper. In the recent Bakerian Lecture before the Royal Society, 
 he says: 
 
 " But if, instead of confining our attention to a relationship between 
 earthquakes, we consider the question of the relief of volcanic strain, many 
 illustrations may be adduced which indicate a close connection between such 
 activities. For example, all the known volcanic eruptions which have oc- 
 curred in the Antilles, from the first which took place in 1692, have been 
 heralded or closely accompanied by large earthquakes in that region, but 
 more frequently by like disturbances in neighboring rock-folds, particularly 
 that of the Cordilleras. This was notably the case in 1902. On April 19 
 of that year an unusually large earthquake devastated cities in Guatemala. 
 Small local shocks were felt in the West Indies, and on April 25 it was no- 
 ticed that steam was escaping from the crater on Mont Pelee, in Martinique. 
 These activities continued to increase until May 8, when they terminated 
 with terrific explosions, submarine disturbances, and the devastation of great 
 portions of the islands of Martinique and 'St. Vincent. 
 
 " The last illustration of hypogene relationship between these regions 
 occurred on Januarv 31 of the present year. On that date a heavy earth- 
 quake originated off the mouth of the Esmeralda River, in Columbia. Sea- 
 waves inundated the coast, islands sank, and a volcano erupted. The news- 
 papers of February 2 announced that cables between Jamaica and Puerto 
 Rico had been interrupted, and on later dates it was reported that severe 
 shocks had been felt among the West Indian islands, that six or seven sub- 
 marine cables had been broken, and that Mont Pelee and La Soufriere, in 
 St. Vincent, were again active." * 
 
 We have quoted these views of Humboldt, Darwin and Milne 
 in order to exhibit fairly the beliefs of all these great investigators 
 in the development or the possibility of the development of seismic 
 action at a distance when a disturbance is once started. 
 60. Tele seismic disturbances. 
 
 From the theory developed in this paper we take is that when a 
 
 severe earthquake is started at one place the tremors may cause 
 
 disturbances to spread into neighboring regions or to break out at 
 
 a distant point if the conditions of the steam pressure underlying 
 
 1 Proc. Roy. Soc., Vol. 77, 1906, p. 374. 
 
400 SEE THE CAUSE OF EARTHQUAKES. [October i 9( 
 
 the crust are already highly unstable ; but there is probably no com- 
 munication through the depths of the planet, except a wave motion 
 which spreads in every direction, the earth acting as an aeolotropic 
 elastic solid, because of the great pressure to which the matter is 
 subjected. In adopting the view that, because sympathetic effects 
 may be aroused at great distances, the disturbances are therefore 
 very deep-seated, Humboldt appears to have been somewhat misled 
 by the ordinary effects of tremors on unstable conditions even at 
 great distances. 
 
 Yet there can scarcely be any doubt that the connection of the 
 Andean trough from Valdivia to Guayaquil is so intimate that a dis- 
 turbance once started under parts of it may easily be propagated 
 over the whole length of this great trough, or even to another part 
 of the globe; and thus we conclude that seismic activity easily 
 extends, and has a widespread effect which was formerly supposed 
 to be transmitted through the deep interior of the earth. It is 
 impossible to doubt that Charles Darwin was entirely correct in 
 concluding that a subterranean connection generally exists between 
 a continent and its outlying islands, for both are often on the 
 borders of the same continental trough. Occasionally the extent 
 of this connection may be even wider, and sometimes cover a whole 
 region or run from one region into another, as in the events men- 
 tioned above by Professor Milne ; but the disturbances are trans- 
 mitted principally by waves and by strains through the crust, and 
 not by means of any currents through the deep interior of the earth. 
 As regards the general question of slight disturbances bringing on 
 greater catastrophes when the conditions of subterranean steam 
 pressure are already highly unstable we may go even further. We 
 occasionally read in History of the West Indies, especially the 
 group including St. Vincent, Martinique, St. Lucia, Guadaloupe, 
 Barbadoes and Trinidad, being visited by a terrible hurricane, fol- 
 lowed by an earthquake and a " tidal wave " ; so that it seems as 
 if all the worst elements in nature were suddenly let loose to devas- 
 tate these islands. In view of the cause assigned in this paper, it 
 is evident that unstable conditions of subterranean steam pressure 
 may not require anything more violent than the raging of a hurri- 
 cane to bring on an earthquake, which in turn may be followed 
 
Jgo6 j SEE THE CAUSE OF EARTHQUAKES. 401 
 
 by a " tidal wave," for the reasons and in the manner above ex- 
 plained; and thus to the afflicted inhabitants all nature seems to be 
 convulsed at once. From the observed course of events, we cannot 
 doubt that this connection actually exists ; and unfortunately it 
 seems to be abundantly illustrated in the annals of the East as well 
 as of the West Indies. 
 
 In the same way we explain Alexis Perrey's well-known laws 
 of earthquakes. The forces assigned to account for these disturb- 
 ances, however, are not the cause of the convulsions of nature, but 
 only the occasions for outbreaks of highly unstable conditions de- 
 pending on subterranean forces easily set off. 
 
 In like manner the violent outbreak of a volcano or the occur- 
 rence of a great earthquake in one part of the world may tend to 
 bring on similar phenomena in another remote region. 
 
 The order of events often observed in the development of the 
 volcanic and seismic phenomena following great outbreaks seems 
 to support this view. While we do not regard it as proved that 
 an eruption like the recent great outbreak of Vesuvius, for example, 
 could indirectly bring on the earthquake in California and other 
 similar disturbances, yet we do not regard such an influence as 
 at all impossible. Conditions of instability once existing are con- 
 tagious and tend to spread like a conflagration. As there are on 
 the average over 60 world-shaking earthquakes every year, or 
 more than one a week, it is evident that if one should break out in 
 a region where it might accelerate the outbreak of others, several 
 might be grouped into a small space of time, and these in turn might 
 exercise wide influence on unstable conditions throughout the world, 
 as often seems to be the case. 1 Our knowledge of these teleseismic 
 effects, however, is still far from complete, and the settlement of the 
 question must be left to the future. 
 
 61. Internal state of the earth. 
 
 The investigations of Lord Kelvin and Professor Sir G. H. 
 
 1 In his " Journal of Researches during the Voyage of the Beagle," 1845, 
 p. 291, Charles Darwin mentions this remarkable coincidence of phenomena: 
 After a long slumber, Conseguina in Central America, and Aconcagua and 
 Corcovado (S. lat. 32^ and 43^2) in Chile, broke out the same day! His 
 suspicion that this coincidence was not accidental would be confirmed by 
 our knowledge of the Andean trough, in the line of which all these volcanoes 
 are situated. 
 
402 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 Darwin have shown that the earth is highly rigid, and in a recent 
 paper on the rigidity of the heavenly bodies (cf. Astron. Nachr., 
 No. 4104) the writer has shown that no motion of currents deep 
 down in our planet is really possible, because of the enormous 
 friction due to the pressure at great depths. Thus no currents of 
 fluids or gases exist in the earth, except just under the crust where 
 the explosive strain is terrific. And even near the surface, where 
 the lava is forced to move under the thin crust, in the building of 
 mountain chains like the Andes, the motion is usually accomplished 
 only by the dreadful paroxysm of an earthquake, which expels the 
 molten rock from under the bed of the sea. The suspicion of Capt. 
 Fitzroy and Charles Darwin that in the three months following the 
 earthquake of 1835 the Chilean coast partially subsided to its former 
 level seems not only possible, but extremely probable. Under great 
 strain the viscous mass may have yielded somewhat, and thus there 
 may have been a slow creeping tendency towards the former level. 
 
 Judging by the thickness of the sides of Aconcagua, Cotopaxi 
 and other typical volcanoes of the Andes, one would probably be 
 justified in concluding that the thickness of the crust under which 
 the lava moves when expelled is not less than five miles, and it may 
 be as much as fifteen or twenty, but ordinarily it could not well lie 
 outside of these limits. For if. we suppose it to be thicker, the leak- 
 age of the water would present greater difficulty if the temperature 
 is low ; and such thickness would not be required if adequate steam 
 developed nearer the surface. On the other hand, the thickness 
 could hardly be much less than ten miles without enfeebling the 
 layers which must support great strain in the expulsion of matter 
 from the broad trough of the ocean bed. The most probable thick- 
 ness is from ten to twenty, and for most purposes we shall be safe 
 in adopting the simple mean of fifteen miles. 
 
 There are other considerations which lead to substantially the 
 same conclusion. It is found for example by the critical investiga- 
 tion of great earthquakes that most of these disturbances proceed 
 from an average depth of something like ten to fifteen miles. Now, 
 if the theory here developed be admissible, it will follow that these 
 disturbances usually are in or near the lower stratum of the earth's 
 crust, which thus fixes the thickness of the layer at about the same 
 
19 o6.] SEE THE CAUSE OF EARTHQUAKES. 403 
 
 figure. It will be evident from this consideration and others that 
 the thickness of the crust is by no means uniform throughout the 
 globe. Determinations of the depths of large earthquake disturbances 
 are probably the best means of approximating the thickness of the 
 earth's crust, since data of this kind depend wholly upon observation 
 and are indepenedent of any hypothesis. 
 
 XII. CONCLUSIONS. 
 62. Summary of results. 
 
 , i . We have seen that deposits of sediment on the continental 
 shelves could not possibly produce anything but the most gradual 
 increase of weight on these portions of the earth's crust; and since 
 such rocks as marble are proved to be fluids of great viscosity and 
 therefore capable of slow secular bending without rupture, we may 
 feel sure that any stresses thus arising in the earth's crust would 
 be relieved by gradual yielding, and that no violent earthquake shock 
 could ever arise from such a cause. 
 
 2. The theory that earthquakes are due to fracture and slipping 
 of rocks is disproved by the great depth (ten to twenty miles) 
 at which world-shaking earthquakes are found to originate, and by 
 virtue of the fact that they come not from a point nor from a line, 
 but from an area; and moreover earthquakes follow the seashore, 
 seldom occurring far inland, and never in desert countries, though 
 abundant in the bed of the ocean. 
 
 3. It therefore follows that earthquakes must depend upon ex- 
 plosive forces within or just under the earth's crust, and frequently 
 spread over a considerable area, and the preponderance of disturb- 
 ances in the sea and along the shores of continents shows that the 
 forces depend in some way upon the sea water. These explosive 
 forces can be best studied in connection with the eruption of vol- 
 canoes, since volcanic outbreaks are also accompanied by earth- 
 quakes often felt over large areas. 
 
 4. Not all earthquakes lead to eruptions, but if the shocks in a 
 given region cease on the eruption of a neighboring volcano, we may 
 feel sure that the forces producing the eruption also produced the 
 antecedent earthquake shocks felt by the surrounding country. 
 
 5. That steam is the cause of volcanic eruptions is proved by the 
 
404- SEE THE CAUSE OF EARTHQUAKES. [October i 9r 
 
 distribution of active volcanoes about the seashores and by the 
 innumerable eruptions which occur in the depths of the ocean, 
 whereas such vents always die out inland ; and moreover by the fact 
 that of the vapors emitted from volcanoes 999 parts in 1,000 is 
 estimated to be steam, the remaining one-thousandth part being by- 
 products incidental to the moisture and high temperature. 
 
 6. The vera causa of volcanic action and of certain earthquakes 
 thus established for some particular cases must be held to be the 
 universal cause in all cases whatsoever, according to Newton's rule 
 of philosophy. 
 
 7. The heaving of steam accumulating within or just beneath 
 the earth's crust is therefore the true cause of all world-shaking 
 earthquakes, and volcanic outbreaks occur only when an outlet is 
 forced through to the surface, which usually happens in mountains, 
 where the earth's crust is already badly fractured and upheaved. 
 
 8. When the subterranean steam pressure becomes great enough 
 to shake the earth's crust, it naturally moves at the nearest fault line, 
 where the rocks are broken, but the movement observed is the result, 
 not the cause of the earthquake. 
 
 9. Volcanoes are particular mountains blown open by steam 
 pressure under the throes of earthquakes (cracks in the rocks appear 
 to be the beginning of some few volcanoes), and since all volcanoes 
 blow out pumice and ashes, these materials must be held to exist 
 in all mountains, and are made by the inflation of molten rock with 
 steam and other vapors. 
 
 10. Any mountain peak, therefore, is capable of becoming a 
 volcano if the subterranean steam pressure be sufficiently powerful 
 to break open an orifice. But orifices close up and volcanoes die 
 out inland and elsewhere if the supply of steam is inadequate to 
 keep open the vents upon which the activity depends. Even if 
 stopped up for a time later heaving of the earth may give the vol- 
 cano renewed activity, and when the mountain has been dormant 
 for a long time it is found that the violence of the eruption is greatly 
 increased. The violence of the subterranean pressure in such a case 
 approaches that of a region which has no vent at all, and hence we 
 see why earthquakes in non-volcanic regions frequently become so- 
 terrible, because the forces accumulate to frightful fury before any 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 405 
 
 relief whatever is afforded, and the result is a most terrible earth- 
 quake. 
 
 11. The mountains are formed by the injection of steam- 
 saturated lava under the coast, which breaks the overlying surface 
 rocks and gives rise to a ridge parallel to the sea. This is why all 
 mountains are formed parallel to the seashores. 
 
 12. By continually injecting the land with lava from under the 
 bed of the sea the coast is raised and the mountains upheaved, and 
 some of them usually break out into volcanoes; while at the same 
 time the support of the sea bottom is undermined by the thinning 
 out of the fluid substratum, and at intervals the bottom sinks down 
 to restore stability. 
 
 13. The sinking of the sea bottom in this natural process of 
 earthquake injection of the land is the cause of that class of sea 
 waves found to follow violent earthquakes, in which the water first 
 withdraws from the shore and then returns as a huge wave. Those 
 waves noticed to rise suddenly without previous recession of the 
 water usually are due to submarine upheavals and eruptions in the 
 bed of the sea. 
 
 14. Islands are built up by injection from the sea, and hence 
 have their mountains as veritable backbones, because the injection 
 is symmetrical from both sides. In many cases the sea bottom is 
 thus undermined and finally sinks down, making a hole beside the 
 island, or a trench. The fact that all islands are not accompanied 
 by such sinks is no argument against the theory, because the subsi- 
 dence has not always taken place ; it is the occurrence of even a con- 
 siderable number of such sinks beside islands which proves the 
 validity of the theory. Such intimate associations between elevation 
 and depression could not be the result of chance. 
 
 15. In the repair of ocean cables broken by earthquakes, Subsi- 
 dence of the sea bottom is frequently found to follow these dis- 
 turbances. This is a direct observation of the above effects in certain 
 cases which are established by actual measurement, the subsidences 
 frequently amounting to hundreds of fathoms. 
 
 1 6. The sea bottom does not subside without the lava under the 
 crust being forced out into some other place, as into islands, sub- 
 marine ridges, or shores; none of this movement is due to the 
 
406 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 secular cooling of the earth, but is all to be explained by the under- 
 mining effect of steam accumulating under the earth's crust. 
 
 17. Mountains in the interior of a dry country, as the Rocky 
 Mountains in Colorado, exhibit no important movements, while 
 those on the coast, like the Andes, are always heaving. This shows 
 that the sea is the cause, and not the secular cooling of the globe, 
 which is wholly insensible. 
 
 1 8. The only, countries which are free from earthquakes are the 
 deserts, and therefore practically uninhabitable ; there is accordingly 
 no escape from earthquakes, and buildings designed for permanency 
 should be framed to withstand them without material injury. 
 
 19. While in the long run the elevation of the land predominates, 
 there is also subsidence, due to the non-concurrence of the forces 
 in certain regions beneath the crust. It is idle to deny these oscilla- 
 tory movements of the crust, and many good illustrations of both 
 are clearly established. Every island which is thrown up in the 
 sea is a witness to one of the most general laws of nature. 
 
 20. As water is taken up in the crust both in the crystallization 
 of rocks and in the processes of earthquake movements, and only a 
 part of this vapor is restored to the surface through volcanic action, 
 there is a secular desiccation of the oceans, but the process is exces- 
 sively slow and not certainly recognizable during the historical 
 period, though a part of the lowering of the strand line in later 
 geological ages is no doubt traceable to this cause. 
 
 21. The elevation of the plateaus depends on the same cause 
 which upheaved the mountains ; and all plateaus, like the mountains, 
 are underlaid with various forms of pumice, which accounts for 
 their feeble attraction as shown by geodetic observations. 
 
 22. No doubt various chemical changes go on under the earth's 
 crust where the water has penetrated the lava and the steam becomes 
 superheated, but the predominance of water vapor in volcanoes 
 shows that the other gases are only by-products, incidental to the 
 moisture and great heat. Dissociation of water vapor is one of these 
 effects. 
 
 23. The details of mountain structure admit of explanation on 
 the present hypothesis, while heretofore no such explanation was 
 forthcoming. A theorv which accounts for the position of the 
 
19 o6.] SEE-THE CAUSE OF EARTHQUAKES. 407 
 
 ranges relatively to the sea, the slopes of the ranges, and the side 
 spurs, and the relation of mountains to earthquake and volcanic 
 phenomena, should have a strong claim to acceptance. This theory 
 was partially foreshadowed by the Arabian astronomer Avicenna, in 
 the tenth century of our era. 
 
 24. The theory of the penetration of sea water into the crust 
 of the earth and its connection with volcanoes and earthquakes 
 dates back to Lucretius and Aristotle, while the upheaval of the land 
 is distinctly announced by Strabo. We have, therefore, been simply 
 verifying and extending the impressions of the ancients formed 
 from the general aspects of nature long before the sciences had 
 become exact. 1 
 
 63. General considerations. 
 
 1 Since finishing this paper the writer has been much impressed with the 
 following passage in the article Poseidon, Encyclopedia Britannica, ninth 
 edition : 
 
 " POSEIDON, the ancient Greek god of the sea and of water generally. . . . 
 He was the god of navigation, adored by all who sailed the sea. His temples 
 stood especially on headlands and isthmuses. As god of the sea he disputed 
 with other deities for the possession of the land with Athene for Athens 
 and Troezen, with Helios for Corinth, with Hera for Argos, with Zeus for 
 Aegina, etc. Earthquakes were thought to be produced by Poseidon shaking 
 the earth, hence his epithet of ' Earth-shaker,' and hence he was worshipped 
 even in inland places, like Apamea in Phrygia, which had suffered from 
 earthquakes, flence also may have arisen the custom in some places of 
 sacrificing moles to him. The great sea-wave which often accompanies an 
 earthquake was also his work; the destruction of Helice in Achaia by such 
 a wave (373 B. C.) was attributed to his wrath. Once when an earthquake 
 shook the ground where a Spartan army was encamped, the whole army 
 sung a hymn to Poseidon. The island of Delos was thought to have been 
 raised by him from the bottom of the sea, and in 237 B. C, when a new 
 island appeared between Thera and Therasia, the Rhodians founded a temple 
 of Poseidon on it. Thessaly was said to have been a lake until this god 
 opened a way for the waters through the Vale of Tempe. Poseidon was also 
 the god of springs, which he produced by striking the rock with his trident, 
 as he did on the acropolis of Athens when he was disputing with Athene for 
 the sovereignty of Athens. This dispute was represented on the western pedi- 
 ment of the Parthenon. . . . There were collossal statues of him at Helice 
 in Achaia, on the Isthmus of Corinth (set up by the Greeks after the Persian 
 wars) and at Tenos." 
 
 It is very remarkable to find that at an early age the Greeks had so 
 directly connected earthquakes with the sea, probably through the seismic 
 sea waves, which they often observed in this part of the Mediterranean. 
 (Note added December 17, 1906.) 
 
408 SEE- THE CAUSE OF EARTHQUAKES. [October 19, 
 
 The way in which several different classes of phenomena find 
 explanation by the most simple of causes may be considered not the 
 least remarkable result of the present investigation. In the minds 
 of those who follow Newton's first rule of philosophy, " to admit 
 no more causes of natural things than are both true and sufficient 
 to explain their appearances," this will probably tell strongly in 
 favor of the truth of an hypothesis which explains easily and natur- 
 ally such diversified phenomena as earthquakes and volcanoes, the 
 formation of mountains and the deficiency in their attractions, the 
 origin of cordilleras from the ocean trenches near continents, and of 
 the great sea waves which frequently accompany violent earthquakes. 
 But there will doubtless be others who will prefer a variety of 
 causes, and -will be slow to believe that the laws and order of nature 
 ire so simple. 
 
 In an investigation of such considerable extent it would not 
 be surprising if many difficulties should require further elucidation 
 than they have received in this imperfect outline ; for the writer has 
 not the geological learning required for the full treatment of many 
 of the great problems of the earth's crust. But if, on the one hand, 
 some defects or omissions should be found, and no doubt many of 
 them will appear in special branches of the extensive lines of 
 thought here traversed, may it not be thought, on the other, that the 
 harmony established with geodetical measurements on the attrac- 
 tions of mountains and the deviations of the plumb line, combined 
 with the explanation of the equilibrium of the terrestrial spheroid 
 between the land and water hemispheres, is not wholly without a 
 certain degree of compensation to those interested in the numerous 
 and related problems of the physics of the earth? 
 
 That the existing theory of earthquakes and volcanoes and 
 mountain formation is embarrassed by many difficulties has often 
 been frankly admitted. While the theory here suggested may re- 
 quire extension or modification, one may confidently submit the 
 question whether it is really possible to doubt its essential truth. 1 
 
 1 The sinking of the sea bottom after great earthquakes have heaved up 
 the coast, as along the west shore of South America, furnishes a truly 
 remarkable criterion for unraveling the unseen processes of nature, hidden 
 beneath the earth's crust The sea bottom could not sink unless it was in 
 some way undermined, and the adjacent coast could not be upraised unless 
 
X9 o6.] SEE THE CAUSE OF EARTHQUAKES, 409 
 
 If this inference should prove to be well founded, may it not be 
 possible to conclude the contest long waged between astronomy and 
 geology with credit and advantage to both of these great sciences? 
 For it seems to be proved that owing to the great pressure acting 
 throughout the earth's interior, and the solidity of the rocks which 
 cover its surface, the terrestrial spheroid, when subjected to great 
 strain, behaves and vibrates as an seolotropic elastic solid; yet the 
 fluid medium long contended for by geologists to explain crump- 
 lings and foldings of the crust is really found to exist or develop in 
 a thin layer just beneath the outer crust, and we have explained 
 its simple and somewhat automatic mode of operation. On several 
 grounds it appears that this layer is extremely thin, certainly not 
 much thicker than the crust itself; and, moreover, it too remains 
 essentially quiescent and rigid, except when set in motion by the 
 recurring paroxysms of steam pressure seeking release, which gives 
 rise to an earthquake and the movements of molten matter essential 
 to geological processes. The original conceptions of astronomy are 
 thus apparently verified, while at the same time the long-standing 
 needs of geology are amply met. And although no effort has been 
 made in this paper to harmonize the conflicting views long prevailing 
 in the two sciences, it would seem that this unexpected result may 
 prove to be not the least interesting among several conclusions 
 respecting the constitution of the globe. 
 
 something was pushed under it: both these movements are observed to 
 occur, and always in the same order. Along the shore, just parallel to the 
 sea coast, the crust is pushed upward to form mountains, while the adjacent 
 sea bottom is correspondingly sunk down. Can anyone doubt that we have 
 here a true cycle of events? One of these events necessitates the other, and 
 all are so connected that one can begin at either end of the chain of phe- 
 nomena and reach the other. Thus, if we begin with volcanoes we are led 
 downward through these vents to the earthquake and mountain forming 
 forces at work under the surrounding land; and finally we come out beneath 
 the ocean where world-shaking earthquakes and seismic sea waves originate. 
 On the other hand, if we begin with seismic sea waves, we are led downward 
 through the observed subsidence of the sea bottom and pass under the crust 
 to the mountains raised along the coast, by earthquakes occurring mainly 
 under the sea, and thus finally reach the volcanoes at the top of the range, 
 erupted by the pressure of subterranean steam, and blowing out lava, pumice, 
 ashes, and vapors. All the grandest phenomena relating to the earth's sur- 
 face are thus connected in an endless chain or circuit. Could such perfect 
 order and harmony be established among all these varied phenomena without 
 the discovery of the true physical cause? (Note added December 3, 1906.) 
 
410 SEE THE CAUSE OF EARTHQUAKES. [October 19 
 
 The details of the processes involved in the elevation of conti- 
 nents, and just how these great land areas originated, are yet to be 
 worked out, and appear to be involved in considerable obscurity, 
 most probably connected with the terrestrial origin of the moon. 
 Perhaps our knowledge is still too incomplete for a satisfactory 
 attempt at this inquiry; but it seems not improbable that the lines 
 of thought here struck out may. eventually prove fruitful of further 
 discovery. 
 
 64. Origin of the present investigation. 
 
 It will be seen from the foregoing account that the present in- 
 vestigation grew out of the difficulty of explaining satisfactorily 
 the rotatory motion of the San Francisco earthquake. When the 
 writer very soon became convinced that the accepted theory, which 
 explained these phenomena by dislocations, subsidences and faults, 
 was not well founded, the investigation was extended little by little 
 till it covered a number of important natural phenomena not usually 
 correlated. The order in which this extension took place is perhaps 
 of no importance, but to enable anyone who might be curious to 
 understand the original line of thought, it may not be wholly out of 
 place to mention that the next step was to assign the explosive 
 activity of volcanoes to steam; and then the traditional reasoning 
 about these vents being safety valves was intelligible, and it became 
 clear why violent earthquakes do not predominate in volcanic dis- 
 tricts and why large tectonic earthquakes are always so violent. 
 
 While considering the mode of release of steam pressure in 
 tectonic earthquakes, where the strata are unbroken, I was led to 
 inquire into the origin of mountains. The only explanation I could 
 find was by the theory of contraction producing folding along lines 
 of weakness in the earth's crust. It seemed to me remarkable that 
 the lines of weakness should nearly always follow the coast lines 
 of continents. Such an arrangement should not occur by chance. 
 It was easy to grasp the connection of earthquakes with volcanic 
 activity when the volcano is once formed ; and the question naturally 
 arose, how did the volcanoes and mountain chains originate? I 
 recalled that volcanic islands are always being thrown up, and 
 numerous islands had been formed in this way on the bed of the 
 sea, which are mountains under water. On the other hand most 
 
I9o6 .] SEE THE CAUSE OF EARTHQUAKES. 411 
 
 geological theories regarded the mountains as formed in the past, 
 while it seemed to me that some of the mountains are near enough 
 to the seashore to be still in the process of upheaval. In reflecting 
 over these questions, on May 2Oth, I recalled, from impressions ob- 
 tained at the Coast Survey in 1899, that some of the islands in the 
 sea had depressions on the side of them as if the material had been 
 thrown up from the adjacent bed, and that deep trenches in the 
 oceans also ran parallel to the great ranges such as the Andes. I 
 noticed also that the volume of the Andes and of the islands were 
 severally not very far from equal to those of the adjacent trough 
 and sinks. Could it be possible that the Andes also had been raised 
 by matter pushed up under the crust from the bed of the adjacent 
 oceanic trough? On recalling that the sea bottom near Pelee has 
 been found by observations to have actually sunk after the great 
 ejection of ashes and vapors in 1902, and that the hollow form of 
 the Andean trough would prevent vertical eruptions from beneath, 
 and cause any explosions which might originate under it to seek 
 release at the sides, thus pushing up the mountains through col- 
 umns or rather layers of lava advanced little by little, by "lateral 
 thrusts," as the geologists say, from under the sea, I did not hesitate 
 to conclude that the Andes had originated in that way. The ex- 
 istence of the long trough, so deep and so exactly parallel to the 
 mountains for so great a distance, proved it; and it was easy to see 
 also that later the resistance toward the land side would become so 
 great, owing to the great distance and height of the mountains, that 
 release would eventually come more easily by the other side of the 
 trough folding up into a parallel range, which would thus emerge 
 from the sea as the New Andes, and eventually cause a further re- 
 cession of the west coast line of South America. 
 
 The provisional entertainment and private acceptance of these 
 views seemed daring enough, but it appeared to me there was no 
 escape from them. On the following day, May 21, I recalled that 
 geodetic observations indicated a feebler attraction for mountains 
 than if they were solid and of the same mean density as the crust ; 
 and it was evident that my earlier conception of pumice-filled cones 
 and chains had not been altogether too rash. The rest of the work 
 consisted in nothing but verification and elaboration of these 
 intuitions. 
 
412 SEE-THE CAUSE OF EARTHQUAKES. [October ,, 
 
 Moreover, the cause of the great sea waves was now perfectly 
 clear. It was impossible to doubt that these phenomena, of the 
 class in which the water recedes from the shore in the first few 
 minutes following an earthquake, generally arise from a subsidence 
 of the sea bottom after the expulsion of porous lava in the natural 
 process of elevating the mountains and coasts. , 
 
 After these ideas had been worked out in such a way as to es- 
 tablish harmony among the widely diversified phenomena it was 
 gratifying to find that such views were not inconsistent with the 
 thoughtful remarks of Humboldt, and that Charles Darwin seventy 
 years before had reached conclusions in most respects remarkably 
 similar. It seemed, therefore, justifiable to suppose that the process 
 of mountain formation thus outlined and verified was in all proba- 
 bility the correct one. Such was the order of the writer's thought, 
 which he records with reluctance and hesitation, but chiefly in the 
 hope that it may not be wholly without value to others. 
 
 In the final arrangement of the paper it was deemed best to intro- 
 duce the citations from Humboldt and Darwin under the discussion 
 of the several topics rather than as extremely long foot-notes. And 
 this arrangement was the more agreeable on account of the great 
 esteem in which the writer has always held these illustrious investi- 
 gators. As Professor Milne is justly recognized to be the most 
 eminent living authority on earthquakes, to whom we owe extensive 
 international cooperation and the present widespread interest in the 
 subject, which has contributed so greatly to the advancement of our 
 knowledge, the same principle has naturally been employed in the 
 exposition of the results of his researches. 
 
 While it is impossible to estimate too highly the great strides 
 recently made in seismology under the leadership of Professor 
 Milne, Dr. Davidson, Montessus de Ballore, Major Dutton, Von 
 Rebeur-Paschwitz, Dr. Agamennone, Dr. Rudolph, Professor 
 Omori, and others, who have reduced seismology to a science of 
 precise observation and measurement, it is perhaps unfortunate that 
 the views of Humboldt and Charles Darwin as to the causes involved 
 in earthquakes were ever abandoned by modern investigators, though 
 this probably has not retarded the progress of observation and 
 experimentation on the constitution of the globe. 
 
1906.] SEE THE CAUSE OF EARTHQUAKES. 413 
 
 In certain respects the theory here outlined may be capable of 
 observational tests. The importance of the subject would appear 
 to be such that it may be worthy of consideration whether geodetic 
 investigators of the different nations might not advantageously 
 arrange to establish earthquake and tidal observatories and more 
 precise levels on coasts such as those of Chili, Peru, California, Japan, 
 Italy, Greece and other countries, for the more exact study of pro- 
 gressive secular movements. Has not the time come to test geo- 
 logical and seismological theories by the accumulation of exact 
 empirical data, and is not this a debt which we of this generation 
 owe to the future ? 
 
 65. Seismic activity a maximum along the coasts of deep seas 
 and a minimum in the great inland deserts. 
 
 If the cause assigned for the explanation of earthquakes be 
 confirmed by time and experience, it is evident that no place on 
 the earth can be said to be wholly removed from the danger of these 
 disturbances; yet the dangers will be a maximum on the coasts of 
 deep seas where the shores are of leaky character and the troughs 
 are at work, 1 and grow less and less along the coasts of the shallower 
 waters where the troughs are absent and the stratification of the 
 rocks is more secure. Thus northern and central Europe and the 
 eastern coast of the United States are comparatively safe. Yet 
 sooner or later, but fortunately to be reckoned in many cases by 
 intervals of thousands of years, every place (except the great in- 
 land deserts which are nearly uninhabitable) is likely to be shaken 
 by an earthquake of considerable severity; and those works of man 
 which are to be preserved and to stand through the centuries should 
 be built accordingly. 2 The great layer of water covering the earth, 
 
 1 In his thoughtful article on " Geology," Encyc. Brit., p. 255, Sir Archibald 
 Geikie justly remarks: "Some of the most terrible earthquakes within human 
 experience have been those which have affected the western seaboard of South 
 America." The cause of this is now plain, viz., the Andean Trough is 
 probably the largest and most powerful in the world. And in general the 
 seimicity of a region in the production of world-shaking earthquakes depends 
 on the extent and power of its ocean troughs. 
 
 2 Humboldt laments the destruction wrought by earthquakes upon works 
 of art, architecture, monuments and inscriptions of the classic period, which 
 were developed in a region of such high seismic activity as to render their 
 preservation difficult. 
 
414 SEE THE CAUSE OF EARTHQUAKES. [October 19, 
 
 which gives life to animals and plants, and in the form of steam is 
 the greatest mechanical agent of man, when sunk into the crust be- 
 comes also one of his worst destroyers, on account of the explosive 
 vapor generated beneath by the internal heat of the globe. 
 
 Finally, it may not be without interest to note that, so far as our 
 knowledge extends, the earth alone among the encrusted planets 
 of the solar system has an abundance of water and mountains, with 
 volcanoes and earthquakes. Mars seems to have only a trace of 
 water and no sensible mountains or volcanoes. And while Venus 
 is largely obscured from our view, the chances are that its veil is 
 due to clouds, indicating an abundance of water, and hence that its 
 conditions of evolution may approach those of the earth. 
 
 BLUE RIDGE ON LOUTRE, MONTGOMERY CITY, MISSOURI, 
 September 23, 1906. 
 
 NOTE. For kind permission to use Figs. 6, 7, 8, and 9 we are 
 indebted to the courtesy of Rand, McNally and Co., of Chicago. 
 
LD 21-40m-10,'65 
 (F7763slO)476 
 
 General Library 
 
 University of California 
 
 Berkeley 
 

 'tftw