SEISMOMETER. This name was originally given to instruments designed to measure the movement of the ground during earthquakes. Recent observations have shown that, in addition to the comparatively great and sudden displacements wliich occur in earthquakes, the ground is subject to other movements. Some of these, which may be called " earth-tremors," resemble earthquakes in the rapidity with which they occur, but differ from earthquakes in being imperceptible (owing to the smallness of the motion) until instrumental means are used to detect them. Others, which may be called " earth-tiltings," show themselves by a slow bending and unbending of the surface, so that a post stuck in the ground, vertical to begin with, does not remain vertical, but inclines now to one side and now to another, the plane of the ground in which it stands shifting relatively to the horizon. No sharp distinction can be drawn between these classes of movements. Earthquakes and earth-tremors grade into one another, and in almost every earthquake there is some tilting of the surface. The term " seismometer " may con-veniently be extended (and will here be understood) to cover all instruments which are designed to measure move-ments of the ground.
Measurements of earth-movements are of two distinct types. In one type, which is applicable to ordinary earthquakes and earth-tremors, the thing measured is the displacement of a point in the earth's crust. In the second type, which is applicable to slow tiltings, the thing measured, is any change in the plane of the earth's surface relatively to the vertical. Under EARTHQUAKE mention is made of instruments designed by Palmieri and others to register the occurrence of earthquakes, and in some cases to give a general idea of their severity. While some of those instruments act well as seismoscopes, none of them serve to determine with precision the character or the magnitude of the motion. In this article notice will be taken only of instruments intended for exact measurement.
Earthquake displacements are in general vertical as well as horizontal. For the purpose of measurement it is con-venient to treat the vertical component separately, and in some cases to resolve the horizontal motion into two com-ponents at right angles to each other.
Inertia Method. In the first type of measurements what may be called the " inertia " method is followed. A mass is suspended with freedom to move in the direction of that component of the earth's motion which is to be measured. When an impulse occurs the supports move, but the mass is prevented by its inertia from accompanying them. It supplies a steady point, to be used as a standard of reference in determining the extent through which the ground has moved in the direction in question. But, in order that the suspended mass shall not acquire motion when its supports move, one essential condition must be satisfied. Its equilibrium must be neutral, or J S.
nearly so, in order that, when the supports are displaced, little or no force may be brought into operation tending to bring the mass into the same position relative to the supports as it occupied before dis-turbance. This can be made plain by considering the case of a common pendulum hung from a support which is rigidly fixed to the ground. When the ground moves in any horizontal direction the pendulum's inertia causes a certain point in it (the centre of percussion) to remain for the instant at rest. But this contrivance does not yield a steady point, because the stability of the pendulum makes the bob swing down to recover its place directly under the support ; and in fact, if a succession of oscillations of the ground occur, the bob acquires a motion often much greater than the motion of the support itself. This tendency may be corrected, and the pendulum made fit to act as a seismo-meter, by any contri-vance which (without , introducing friction)
FlG- 1-Duplex pendulum seismograph.
will reduce its stability so much as to make the equilibrium of the bob very nearly neutral. In all instruments designed to furnish a steady point the suspended mass must have some small stability, else it would be unmanageable; but its period of free oscillation must be much greater than that of the earthquake-motions which it is employed to measure. Even a simple pendulum can have its stability reduced sufficiently to fit it for seismometric work by making it very long. The same result is, however, much more conveniently achieved by combining a common pendulum with an inverted pendulum placed just beneath it. The common pendulum being stable and the inverted pendulum unstable, if the bobs are jointed so that they must move together, the combination can be made as nearly astatic as may be desired. Figs. 1 and 2 illustrate how this combination is applied in seismometry. The stable bob re, hung from a fixed support above by three parallel wires, is connected with the inverted pendulum i by a ball-and-tube joint. A lever c, carried by a gimbal joint in the fixed bracket d, is geared also by a ball-and-tube joint to the upper bob. Its long arm carries a jointed index e, which projects out and touches a smoked-glass plate /, held on a fixed shelf. Any horizontal motion of the ground acts on the lever by the bracket d, and causes the index to trace a magnified record on the smoked-glass plate. Fig. 1 is taken from a photograph of an instrument of this kind, constructed to give a much magnified record of small movements. When large earthquakes are to be recorded the mul-tiplying lever is dispensed with, and the index is attached directly to one of the bobs. Observations with instruments of this class exhibit well the very complicated motion which the earth's surface undergoes during an earthquake. In small earthquakes (such as are only slightly or not at all destructive) the greatest amplitude of motion is often less than a millimetre, and rarely more than a centimetre ; the disturbance nevertheless consists of a multitude of successive movements, quite irregular in amplitude, period, and direction. Fig. 3 is a facsimile of the record given by a duplex pendulum seismograph during one of the earthquakes which occur frequently in the plain of Yedo, Japan. The record, as engraved, is three and a half times the earth's actual motion. Instead of two pendulums, a single inverted pen-dulum has been used, with a spring stretched between it and a fixed support above. By adjusting the spring so that a proper proportion of the weight is borne by it and the remainder by the rigid stem of the pendulum, an approach to neutral equilibrium can be made. In Forbes's inverted pendulum seismometer a somewhat similar plan was adopted : the foot of the pendulum was attached to an elastic wire which tended to restore it to its normal vertical position when displaced.
Another group of instruments designed to furnish two degrees of freedom for the purpose of recording all motions in a horizontal plane, but much less satisfactory on account of their friction, is that in which a rolling sphere either itself supplies inertia or forms a support for a second inertia-giving mass. Probably the earliest was one used in Japan by Dr G. F. Verbeck in 1876 (see fig. 4). On a marble table, ground plane and carefully levelled, four balls of rock-crystal were placed, carrying a massive block of hard wood. A pencil, sliding in a hole in the block, re-gistered the relative motion of the table and the block on a sheet of paper fixed below. The motion registered is (or would be, if there were no friction) somewhat larger than the true motion of the table, for the system is kinetically equivalent to four upright pieces whose centres of percussion lie in a plane nearly, but not quite, as high as the tops of the balls. This forms what may be called the steady plane ; its position depends on the relative masses of block and balls, and is easily calculated. When the ground moves in any direction the block moves through a short distance in the opposite direction, and the record is magnified in a fixed ratio. Various forms of rolling-sphere seismometers have been proposed by Mr. T. Gray, Mr. C. A. Stevenson, and others. Probably the best form would be that of a light spherical segment rolling on a level plane base and carrying a heavy bob fixed to it. To give some stability the bob should be placed so as to bring the centre of gravity a little under the centre of curvature. The centre of percussion, somewhat higher than this, would of course be the steady point, and a multiplying pointer might take the motion either from it or from any other convenient part of the rolling piece. All rolling seismometersincluding rolling cylinders, which have been proposed by Mr Gray as single-freedom instruments, to register one component of horizontal motionfail to act well, partly because of the comparatively great frictional or quasi-fric-tional resistance which is presented to the motion of the free mass, and partly because, owing to imperfections in the construc-tion and want of perfect rigidity in the materials, the ball or cylinder takes up a position in which there is an objectionably great stability as regards very small displacements. These objections make the use of rolling seismometers unadvisable, except perhaps for the rough measurement of violent earthquakes.
The seismographs which have been described draw a horizontal plan of the path pursued during an earthquake by a point on the earth's surface. They take no note of the relation of the displacement to time,an element which is required if we are to form any estimate of the violence of an earthquake from the record. With this view a different method of registration is also followed. The whole movement is resolved into rectilinear components, and these are separately recorded (by single-freedom seismometers) on a plate or drum which is kept in continuous movement, so that the record of each component takes the form of an undulating line, from which the number, succession, amplitude, velocity, and acceleration of the component movements can be deduced and the resultant motion determined. A single steady mass with two degrees of freedom may still be employed to record, separately, two components of horizontal motion ; but it is generally preferable to provide two distinct masses, each with one degree of freedom. The principal instrument of this class is the horizontal pendulum seismograph, which has been used to record Japanese earthquakes since 1880. It consists of two horizontal pendulums, set at right angles to each other, each supplying a steady point with respect to horizontal motions transverse to its own length. Each pendulum is pivoted about two points, on an axis which is nearly vertical, but inclined slightly forwards to give a suitable degree of stability. In some forms of the instrument the pivoted frame of the pendulum is light, and the inertia is practically all furnished by a second piece or bob pivoted on the frame about a vertical axis through the centre of percussion of the frame. This construction has the advantage of compactness and of making the position of the steady point at once determinate. But a simpler construction is to at-tach the bob rigidly to the frame. This shifts the steady point a little way outwards from the position it would have if the bob were pivoted. In either construction a prolongation of the pendu-lum beyond the bob forms a convenient multiplying index. Fig. 5 shows a complete horizontal pendulum seismograph (with pivoted bobs). Two rectangular components of earthquake motion are r-corded radially on a revolving plate of smoked glass, which receives its motion through a friction roller from a clock furnished with a fluid-friction centrifugal governor. The clock may either be kept going continuously, in expectation of an earthquake at any moment, or be started into motion by an electric seismoscope when the earliest indications of an earthquake are felt. The former plan is practicable only when the instrument can receive careful attendance and where earthquakes occur often. It has the drawback that the circle which is drawn by each pointer as the plate revolves below it gradually broadens, partly because of warping and temperature changes in the supports and partly because of actual tilting of the ground. As an earthquake generally begins with comparatively insignificant movements, there is not much to object to in having the plate at rest to begin with, provided a sufficiently sensitive starting seismoscope be used. A suitable arrangement for this purpose is one due to Palmieri: a short pendulum hangs over a cup of mercury, in the centre of which a depression is formed by an iron pin, whose top is a little lower than the surface of the mercury. The pendulum ends in a platinum point, which stands clear in the centre of this depression, but touches the edge whenever a horizontal movement of the ground takes place, thereby closing the circuit of an electro-magnet, which starts the clock. In the most recent form of the horizontal pendulum seismograph the bobs are fixed to the pivoted frames, and the pointers are arranged to trace their records side by side. Records with instruments of this class, besides giving much additional information, agree with those of the duplex pendulum in showing that earthquake motion is a tangle of waves in all azimuths. This will be seen by reference to fig. 6, which shows a small portion of
an earthquake registered by a pair of horizontal pendulums. Contemporary parts of the two records are shown together, the straight radial lines marking seconds of time. The phases of the two components are continually changing, and when the two are compounded the result is a path having the same characteristics as those of the diagram in fig. 3. Fig. 7 gives the result of compounding " the records of fig. 6 during three seconds, while the range of move-ment was a maximum.
-Result of compounding the record of fig. 6.
To register the vertical component of earthquake motions we require to suspend a mass
with vertical freedom. Most ways of doing this give too much stability, as, for instance, when a weight is hung from a spiral spring or carried by a horizontal bar that is fixed to a wall or table by a flexible spring joint. This last is the vertical motion seismometer which was used by the British Association Committee at Comrie in 1842. Another form, mechanically equivalent to this, is a weighted horizontal bar, pivot-ed on a fixed horizontal fulcrum, and held up by a spiral spring, stretched from a point near the fulcrum to a fixed support above. This mode of suspension is still too stable, though less so than if the spring were
FlG- 8.Principle of directly loaded. To vertlcal motlon make it nearly a- seismograph, static
Mr T. Gray proposed the use of a tube containing mercury, connected with the bar in such a manner that when the bar goes down the mercury, running to-wards one end of the tube, has the effect of increasing the weight, and when the bar goes up an opposite effect occurs. This plan is open to the objection that the mercury is disturbed by horizontal movements of the ground. A simpler plan is shown in fig. 8.8 There the pull of the spring is applied at a short distance v below the plane of the bar. Hence when the weight goes down the spring, which then pulls with more force, pulls with a smaller leverage, and it is easy to adjust the distance v so that the moment of the pull of the spring remains sensibly equal to the moment of the weight,the condition necessary to make the bar astatic. This is _ V
crum to the point at which the spring acts, and I the length by which the spring is stretched when the bar is undeflected. Stability is given by making v somewhat less than this. A vertical-motion seismograph, constructed on the principle which fig. 8 illustrates diagrammatically, is arranged to trace its record on a revolving glass plate. This, along with a pair of horizontal pendulums recording on the same plate, completes a three-component seismograph.
An interesting mode of suspension, by which a mass is hung in neutral or nearly neutral equilibrium, with one degree of horizontal freedom, is shown in fig. 9. It is based on the approximate straight line linkwork of Tchebicheff. When a bar is hung from fixed supports by crossed ties, at a distance below the supports equal to the distance between the supports, the length of the bar being equal to half that distance, its middle point moves in very nearly a straight line. By fixing a weight at the centre of the bar and adding a suitable recording apparatus, we have a very friction-less form of one-component horizontal seismometer. When a displacement of the ground occurs in the line of the bar, the bar is tilted through an angle which is proportional to the linear displacement, and the centre of the bar consequently shares, in a small and definite proportion, the motion of the ground,a fact which is to be borne in mind in estimating the degree of multiplication given by the recording apparatus.
FlG 9--Astatic suspension,
The instruments which have been described afford complete and satisfactory means of determining the motion which a point of the ground undergoes during any disturbance which would be recog-nized as an earthquake. For minute earth-tremors, however, a larger multiplication is necessary, and the absence of friction is of even more importance than in the measurement of earthquakes proper. Optical methods of magnifying the motion are accordingly resorted to. In the " normal tromometer " of Bertelli, used in Italy to detect earth-tremors, the bob of a pendulum, suspended by a fine wire from a fixed support, is viewed through a reflecting prism and its motion in any azimuth measured by a micrometer microscope. The great stability of the pendulum, which is only 1J metres long, prevents it from behaving as a steady-point seismometer; and, if successive earth-movements were by chance to occur with a period equal or nearly equal to its own free period, its acquired swing would altogether mask the legitimate indications. This kind of action has, in fact, been turned to account as a means of detecting very minute earth-tremors by Rossi, who has devised a micro-seismoscope, consist-ing of a number of pendu-lums of various lengths, one or other of which is likely to be set swinging when the ground shakes to and fro re-peatedly, through even the minutest range. To measure tremors, however, the instru-ments of Bertelli and Rossi are inappropriate ; for that purpose, just as for the pur-pose of measuring larger motions, the suspended mass must be in nearly neutral equilibrium. To find a mode of suspension which is at once astatic and extremely frictionless is a matter of some difficulty ; the crossed - link suspension, which has been already described, is probably the most satisfactory means hitherto suggested. It has been adopted in the micro-seismometer sketched in sec-tion in fig. 10. Two bobs are separately suspended, in the manner shown by fig. 9, at right angles to each other, one above the other, in a cast-iron case. A microscope, fixed to the top of the case and furnished with a micrometer eye-piece, is focused on a hair, which is stretched transversely across a vertical tube in the upper bob a. This serves to measure horizontal motion in the plane of the drawing. Motion at right angles to this is shown by the lower bob c (drawn in section), which carries a similar transverse hair. A fixed lens b between the bobs gives an image of the lower hair in the plane of the upper hair, so that both appear crossed in the field of the microscope, thereby allowing both components of horizontal motion to be observed together.
Equilibrium Method.In observing slow earth-til tings an entirely different process is followed. The problem then is, not to measure displacements by aid of the inertia of a body which tends to preserve its original position, but to compare the direction of a line or plane fixed to the earth with the direction of the vertical. The earliest observations of earth-tiltings were made by the aid of spirit-levels. If a level be set on a table fixed to the rock, its bubble, watched through a microscope, will be seen to move slowly now to one side and now to another. The movements are so slow that the inertia of the fluid is unimportant. Observations with pairs of levels, set at right angles to each other, have been carried on systematically for some years by M. P. Plantamour.' This is the simplest method of measuring earth-tiltings, but it is liable to errors which are not easily excluded. Another method of investigating changes in the direction of the vertical was initiated in 1868 by M. A. dAbbadie, who had before that observed the movements of level-bubbles. Light from a fixed source is made to fall on a reflecting basin of mercury about 10 metres below it. Above the basin is a large lens of long focus, which brings the rays into parallelism dur-ing their passage to the mercury, and causes them to converge after reflexion, so that an image of the source is formed at a convenient distance from it, and in the same horizontal plane. The interval between the source and the image is measured (in amount and azimuth) at least twice a day by a micrometer microscope. The accuracy of the method depends on the fixity of the source of light relatively to the lens and to the surface of the ground, and to secure this M. d'Abbadie built a massive hollow cone of concrete for the support of his apparatus. His observations have shown that the earth's surface undergoes almost incessant slow tilting through angles which, in the course of a year, have been found to range over four seconds. He has also noticed the occurrence of earth-tremors by the occasional blurring of the image through agitation of the mercury. An improvement on his apparatus sug-gested by M. Wolf is shown in fig. 11. The light, instead of being all reflected from the free surface of mercury (a), is partly reflected from that and partly from a plane mirror (b) fixed to the rock. Two images are therefore formed, whose rela-tive position measures the tilting of the ' surface. The advantage of this is that the position of the source of light need no longer be fixed, and the accuracy of the method depends only on the fixity of the mirror b with respect to the rock. Further, to avoid having the source and image at a great height above the surface, M. Wolf allows the light to reach and leave the apparatus horizontally, in the manner indicated in the sketch, by using a plane mirror inclined at 45° to the horizon. Still another mode of investigating slow changes of the vertical was followed (at the suggestion of Sir William Thomson) by Messrs G. H. and H. Darwin, in observations made by them with the view of measuring the lunar disturbance of gravity. The Reports of the British Association for 1881 and 1882 contain a full account of their apparatus, as well as notices of the work of other observers and a discussion of the cause of earth-tiltings. Their instrument was a short pendulum hung in a viscous fluid, from a fixed support, by two wires arranged V-wise to leave the pendulum only one degree of freedom. Below the bob was a small mirror hung by two threads, one of which was attached to the pendulum bob and the other to a fixed support. The pendulum was free to swing at right angles to the plane of the threads, and any movement of this kind caused the mirror to rotate through an angle which was measured in the usual way by a telescope and scale. The method is susceptible of very great delicacy, but Messrs Darwin found that when the instrument was adjusted to be specially sensitive its manipulation became extremely difficult. Wolfs modification of D'Abbadie's method appears to furnish, on the whole, the most promising apparatus for measurements of this type. The apparatus represented in fig. 10 is also applicable. The method of measurement employed in the case of slow tiltings may be called the equilibrium method in contradistinction to the inertia method, which is used to measure comparatively sudden displacements. The two methods are applicable to two widely different classes of move-ments. It is at least possible that between these classes there may he other modes of motion,displacements which are too slow for the inertia method, and which give rise to too little change of slope for the equilibrium method. How to measure them is, and must apparently remain, an unsolved problem in seismometry.
References.The Report of the British Association for 1858 contains an account by Mallet of some of the older and now obsolete forms of seismometers (see also EARTHQUAKE). For accounts of modern instruments of the inertia class, see the Transactions of the Seismological Society of Japan from 1880, also Prof. Ewing's Memoir on Earthquake Measurement, published by the university of Tokio (1883). References to papers on the equilibrium method of measurement have been made in the text. (J. A. E.)
4 Gray, Trans. Seis. Soc. Jap., vol. iii. p. 137.
5 Ewing, Trans. Seis. Soc. Jap., vol. iii. p. 140.
1 Ewing, " On certain Methods of Astatic Suspension," in Trans. Seis. Soc. Jap., vol. vi. p. 25.
The above article was written by: J. A. Dundee, B.Sc., F.R.S., Professor of Engineering, University College, Dundee.