3 oo 



NATURE 



[July 26, 1888 



would overthrow the narrower columns, the broadest that 

 fell serving to measure its severity, and that the columns 

 would fall in a direction which would point to the place 

 of origin of the disturbance. In fact, however, such 

 columns fell most capriciously when they fell at all, and 

 it was impossible to learn anything positive from their 

 behaviour in an earthquake. The reason was that there 

 was no single outstanding impulse : an earthquake con- 

 sisted of a confused multitudinous jumble of irregular 

 oscillations, which shifted their direction with such 

 rapidity that a point on the earth's surface wriggled 

 through a path like the form a loose coil of string might 

 take if it were ravelled into a state of the utmost 

 confusion. The mechanical problem in seismometry was 

 to find a steady-point — to suspend a body so that some 

 point in it, at least, should not move while this compli- 

 cated wriggling was going on. The steady-point would 

 then serve as a datum with respect to which the movement 

 of the ground might be recorded and measured. The 

 simple pendulum had often been suggested as a steady- 

 point seismometer, but in the protracted series of oscilla- 

 tions which made up an earthquake the bob of a pendulum 

 might, and often did, acquire so much oscillation that, far 

 from remaining at rest, it moved much more than the 

 ground itself. The lecturer illustrated this by showing 

 the cumulative effect of a succession of small impulses 

 on a pendulum when these happened to agree in period 

 with the pendulum's swing. The fault of the pendulum, 

 from the seismometric point of view, was its too great 

 stability, and its consequently short period of free oscilla- 

 tion. To prevent the body whose inertia was to furnish a 

 steady-point from acquiring independent oscillation, the 

 body must be suspended or supported astatically ; in 

 other words, its equilibrium must be very nearly neutral. 

 Methods of astatic suspension which had been used in 

 seismometry were described and illustrated by diagrams 

 and models, in particular the ball and block seismometer 

 of Dr. Verbeck, the horizontal pendulum, and a method 

 of suspension by crossed cords based on the Tchebicheff 

 straight-line link-work. 



The complete analysis of the ground's motion was 

 effected by a seismograph which resolved it into three com- 

 ponents, two horizontal and one vertical, and recorded each 

 of these separately, with respect to an appropriate steady- 

 point, by means of a multiplying lever, on a sheet of 

 smoked glass which was caused to revolve at a uniform 

 rate by clock-work. The clock was started into motion 

 by the action of the earliest tremors of the earthquake on 

 a very delicate electric seismoscope, the construction of 

 which was shown by a diagram. In this way a record 

 was deposited upon the revolving plate which gave every 

 possible particular regarding the character of the earth's 

 motion at the observing-station. A complete set of the 

 instruments as now manufactured by the Cambridge 

 Scientific Instrument Company was shown in action. 

 Prof. Ewing also described his duplex pendulum seismo- 

 graph, which draws on a fixed plate of smoked glass 

 a magnified picture of the horizontal motion of the ground 

 during an earthquake. Apparatus was shown for testing 

 the accuracy of the seismographs by means of imitation 

 earthquakes, which shook the stand of the instrument, and 

 drew two diagrams side by side upon the glass plate — one 

 the record given by the seismograph itself, and the other 

 the record derived from a fixed piece which was held fast 

 in an independent support. The agreement of the two 

 recordswith one another proved how very nearly motionless 

 the " steady-point " of the seismograph remained during 

 even a prolonged shaking resembling an earthquake. This 

 test was applied to the instruments on the table, and the 

 close agreement of the two diagrams was exhibited by pro- 

 jecting them on the lantern-screen. A large number of 

 autographic records of Japanese earthquakes were thrown 

 on the screen, including several which have been already 

 reproduced in this journal (Nature, vol. xxx. p. 174, vol. 



xxxi. p. 581, vol. xxxvi. p. 107) ; and particulars were given 

 of the extent of the motion, and the velocity and rate of 

 acceleration, in some representative examples. To deter- 

 mine the rate of acceleration was of special interest, 

 because it measured the destructive tendency of the 

 shock. The lecturer explained that some of the seismo- 

 grams exhibited on the screen had been obtained since he 

 had left Japan by his former assistant, Mr. Sekiya, who 

 now held the unique position of Professor of Seismology 

 in the Imperial Japanese University. Prof. Sekiya had 

 recently taken the pains to construct a model representing, 

 by means of a long coil of copper wire carefully bent into 

 the proper form, the actual path pursued by a point on 

 the earth's surface during a prolonged and rather severe 

 shaking. This model of an earthquake had been made 

 by combining the three components of each successive dis- 

 placement as these were recorded by a set of seismographs 

 like those upon the lecture-table. The appearance of 

 Prof. Sekiya's model (a description of which will be found 

 in Nature, vol. xxxvii. p. 297) was shown to the audience 

 by means of the lantern. 



Prof. Ewing drew attention to the small tremors of high 

 frequency which characterized the beginnings of earth- 

 quake motion, and which were apparent in a number of 

 the diagrams he exhibited. These generally disappeared 

 at a comparatively early stage in the disturbance. In the 

 early portion they were generally found at first alone, 

 preceding the larger and and slower principal motions ; 

 and then when the principal motions began, small tremors 

 might still be seen for some time, superposed upon 

 them. In all probability these quick-period tremors 

 were normal vibrations, while the larger motions were 

 transverse vibrations ; and a reference to the theory of 

 the transmission of vibrations in elastic solids served to 

 explain why the quick-period tremors were the first to be 

 felt. The whole disturbance went on for several minutes, 

 with irregular fluctuations in the amplitude of the motion, 

 and with a protracted dying out of the oscillations, the 

 period of which usually lengthened towards the close. 

 The record of a single earthquake comprised some 

 hundreds of successive movements, to and fro, round 

 fantastic loops. Each single movement usually occupied 

 from half a second to two seconds. Earthquakes were 

 quite perceptible in which the greatest extent of motion was 

 no more than 1/100 of an inch. In one case, on the other 

 hand, Prof. Sekiya had obtained a record in which the 

 motion was as much as an inch and three-quarters. Even 

 that was in an earthquake which did comparatively little 

 damage, and there was therefore reason to expect that in 

 a severely destructive shock (such as had not occurred 

 since the present system of seismometry was developed) 

 the motion might be considerably greater. 



Prof. Ewing concluded his lecture by pointing out that 

 seismographs might find practical application in measuring 

 the stiffness of engineering structures. He exhibited, by 

 the lantern, seismographic records he had recently taken 

 on the new Tay Bridge, to examine the shaking of the 

 bridge during the passage of trains. The instrument had 

 been placed on one of the great girders, two-thirds of a 

 mile from the Fife end, at a place where there was reason 

 to expect the vibration would be a maximum. The extent 

 of motion was remarkably small. It was less than an 

 eighth of an inch, even while the train was passing the 

 seismograph — a fact which spoke well for the stiffness 

 of the structure. Nevertheless, by watching the index 

 of the seismograph he had been able to tell whenever 

 a train came on at the Dundee end of the bridge, a 

 distance of 1^ mile from the place where the instrument 

 was standing. One could then detect a vibratory motion, 

 the extent of which was probably not more than 1/500 of 

 an inch. This began in the longitudinal direction, and 

 for some time longitudinal vibration only could be seen. 

 As the train came nearer, lateral vibration also began, and 

 the amplitude of course increased. It reached a maximum 



