596 



NA TURE 



[April 23, 1908 



of which arc symmetrical hi form and geological material. 

 All idea, but one which is not very probable, which this 

 suggests is that at some very early period in the world's 

 history two Rift Valle_vs, one parallel to the eastern sub- 

 merged backbone of the Atlantic, and the other parallel 

 to its western frontier, were formed. Separation subse- 

 quently took place along these faults, and these, under the 

 inlluence of surface and underground activities, have con- 

 tinually increased. If, then, the Atlantic had an origin 

 due to Rift \"alley formation rather than to folding or 

 contraction, then the greatest earthquake in the history of 

 the world may have taken place when east became east 

 and west became west, and our woi-ld was cracked from 

 pole to pole. 



Just as the frequency of earthquakes has fluctuated 

 during geological time, similar fluctuations have taken place 

 during historical time. In central Japan earthquake fre- 

 quency had a maximum in the ninth century, and since 

 that time, century after century, violent shakings have 

 become less and less. In January, 1844, at Conirie, in 

 Perthshire, twelve earthquakes were recorded. Now there 

 may not be one per annum. At the present time, in con- 

 sequence of the destruction of several large cities, the 

 popular idea is that earthquakes are on the increase. As 

 a matter of fact, the world as an earthquake-producing 

 machine has a steady output. On the average, about sixty 

 very large disturbances are recorded, and the greater 

 number of these, fortunately for humanity, have their 

 origins beneath ocean beds or in sparsely inhabited regions. 

 In addition to these mcgaseismic efforts, it is estimated that 

 about 30,000 small earthquakes take place per year, 

 England's annual contribution to this number being about 

 half a dozen. If we had records like these extending back- 

 vifards through several ages, we might readily estim'ate thi! 

 time when seismic activity would cease. When this ceases, 

 rock folding will also cease, and the degrading processes 

 resulting in surface denudation will be unopposed. Bit by 

 hit land areas will be reduced to sea-level, and the habit- 

 able surfaces, as we now see them, will be no more. 



.\n interesting observation bearing upon megaseismic 

 frequency is found in the analyses of registers relating to 

 the North Pacific. On the west side of that ocean seismic 

 frequency is greatest in the summer, while on the east 

 side it is greatest in the winter. An explanation for this 

 is sought for in the seasonal alteration in the flow of 

 ocean currents, the oscillations of sea-level, and changes 

 in the direction of barometric gradients, which phenomen.1 

 are interrelated. In summer, off the coast of Japan, tha 

 Rlack Stream runs pcrhaps_ 500 miles farther north thai 

 it does in winter, while Dr. Omori points out that, although 

 barometric pressure may on the Japan side of the Pacific 

 be low in summer, this decrease in load is more than com- 

 pensated for by the increased height of ocean-level ; the 

 inference is that the pressure on the ocean bed is greater 

 in summer than in winter, and this is the time of the 

 greatest seismic frequency. 



.■\nothcr factor bearing upon earthquake frequency mav 

 perhaps be found in the change in position of the earth's 

 pole. A chart showing the path of the earth's north pole 

 indicates that its movements are by no means alwavs 

 uniform. Although at times these may be nearly circular, 

 it also shows sharp changes in the direction of its motion. 

 It has even been retrograde. If on a chart showing these 

 pole displacements we mark the time positions of world- 

 shaking earthquakes, it is seen that these are grouped 

 round the sharper bends of the pole-path. World-shaking 

 earthquakes have, in fact, been most numerous when the 

 Dole-pnth has deviated farthest from its mean position. 

 The observations embrace a period of thirteen years, during 

 which 750 large earthquakes were recorded. Although 

 these earthquakes represent large mass displacements, it is 

 not supposed that they would be sufficient to produce the 

 observed pole movement. The pole movement, however, 

 may liave given relief to seismic strain, or both effects mav 

 arise from some common cause. 



Mass displacements accompanying a megaseismic effort 

 must, however, tend to produce, some pole displacement, 

 and thus set up strains. From time to time these should 

 hnd relief in the weaker portions of the earth's crust, 

 l-arge earthquakes should therefore occur in pairs, triplets. 

 ^■^ 111 groups, after which we should expect a period of 



NO. 2nnS. vni,. 7-/^ 



quiescence. This idea is due to the Rev. H. V. Gill, S.J. 

 I find that the British Association registers lend consider- 

 able support to the hypothesis. The author of the idea, 

 however, goes a step farther, and points out that if all 

 matter within our globe or that which constitutes its crust 

 was equally free to move, the secondary displacement 

 should, with regard to the earth's axis of rotation, be 

 symmetrically located in regard to the position of the 

 primary disturbance. Out of 126 large earthquakes re- 

 corded between 1899 and 1905, I find that twenty of these 

 appear as ten pairs, the members of each pair being in 

 symmetrically located districts. This may or may not 

 have been a matter of chance. The observation that a 

 marked relief of seismic strain in one part of the world 

 has frequently been followed by a smaller relief in some 

 distant region also suggests the idea that earthquake begets 

 earthquake. In my own mind the relationship of earth- 

 quake to earthquake has been fairly well demonstrated, but 

 to place the matter beyond the borderland of doubt large 

 earthquakes must be compared in regard to space and time 

 with their kind, with small earthquakes, and with volcanic 

 eruptions. All the volcanic eruptions of the West Indies 

 have closely followed on the heels of great earthquakes 

 which have originated, not in the West Indies, but on the 

 neighbouring coasts of Central and South America. One 

 general inference is that the faultings and freckles on the 

 face of our world should have a distribution as symmetri- 

 cally disposed as wrinkles are on the face of an elderly 

 person. 



.Mready when speaking about the length of faults which 

 have been created at the time of large earthquakes, w'e 

 have indicated at least one dimension of the earth block 

 which has been disturbed. For instance, the earth block 

 which was disturbed at the time of the San Francisco 

 earthquake may have had a length of 400 miles ; its 

 breadth might be determined by the width of the country 

 which had been broken up by branching and parallel faults. 

 Harboe suggests that in a meizoseismic area hidden faults 

 may be assumed to exist along lines drawn half-way 

 between pairs of groups of places which have been struck 

 at about the same time. R. D. Oldham attributes 

 the Assam earthquake of 1897 'o the sudden shifting of 

 10,000 square miles of territory over a thrust plain. The 

 molar displacement determined by the method suggested 

 bv Harboe would be that 50,000 square miles had been 

 disturbed. The fact that so many earthquakes shake the 

 whole world, or will agitate an ocean like the Pacific for 

 many hours, indicates that the initial impulse must have 

 been delivered over a large area, or that sudden altera- 

 tions have taken place in the contour of ocean beds. 

 With regard to the magnitude of the latter changes, we 

 have learnt much from cable engineers, who have given 

 us many instances where cables lying in parallel lines, ten 

 or fifteen miles apart, have been simultaneously inter- 

 rupted, and ocean depths over considerable areas have 

 been increased. The depth to which these large faults 

 extend is .a m.atter of inference. We may well imagine 

 them as passing through the whole thickness of the earth's 

 crust, and the displaced block falling to give up its energy 

 to a nucleus which we know transmits undulatory move- 

 ments all over our globe with uniform velocity. If we 

 take this crust to be thirty miles in thickness, then with 

 Harboe's area for the superficial disturbance, the block 

 which was disturbed at the time of the Assam earthquake 

 would be represented by \\ million cubic miles. 



Following the initial impulse of a large earthquake, it 

 frequently happens a few minutes later that a second 

 severe movement is felt. In Japan this is popularlv spoken 

 of as the Yuri Kaishi, or the return shaking. This may 

 be a second yielding within the disturbed district, but from 

 its resemblance to the main shock it suggests an echo- 

 like reflection. If we drop a bullet into a large tub of 

 water, waves travel outwards to the sides of the tub, where 

 thev are reflected, and converge at the centre from which 

 they set out. With the earthquake waves, the reflecting 

 surface may be represented by the roots of mountain 

 ranges. If these are at varying distances from the origin, 

 the reflected waves would give rise to complications at the 

 focus. The transmitting medium for these waves I t.-ike 

 to be the more or less homogeneous material which lies 

 beneath the heterogeneous crust of our world. This 



