Important Details Of Construction

Exact dimensions are unimportant. The drum A is one foot in diameter. The following points, however, are of vital importance:

1. The wall from which the pendulum is hung must be exceedingly solid. If possible it should be below ground, and not subject to great and changing strains. A lengthwise displacement of the millionth part of an inch in the upper part of the wall makes a perceptible jog in the record. Short-period tremors, however, such as machinery or cars near by, do no very serious harm.

2. The steel bearing plates, D and D', after being shaped with a smooth, slightly concave surface, should be tempered file-hard, and then the bearing face should be highly polished with leather and fine emery.

3. The bearings of the lever I, the link L, and the stylus R S must be very perfect. The points of the needles, K and V' and of the link L must be perfectly sharp and smooth. The conical depressions in which they rest may he made by pressing into the metal a sharp-pointed awl with a whirling motion. In regions where sharp earthquakes are sometimes felt these depressions should be rather deep, to prevent the points flying out. The needle K must be exactly vertical.

4. As most of the friction of the seismograph is at the point of the stylus S, it is of the utmost importance that that stylus should rest very lightly on the paper, only heavily enough to scratch through a moderately thin soot layer. The broad part of the crosspiece Q should be bent upward, so as to prevent the stylus dropping too far when the pen swings off the paper in a great earthquake.

5. If there are drafts in the room where the seismograph is installed, the instrument must be well protected from them.

Time Marking

To be of scientific value, the records should have exact time marked on them at frequent intervals. This can easily be clone if a reliable clock is available. A bit of platinum wire soldered to the second-hand wheel makes a short contact once each minute with a fixed platinum wire. These contact points are connected, through two dry cells, to the magnet of an electric bell. (Directly, not through the vibrator.) The bell, with gong removed, is rigidly attached to the post P, so that the strike of the armature is at right angles to the lever /. Thus at the end of each minute there is a sharp click against the post, which causes, as it were, a miniature earthquake, which is plainly visible in the record. The effect is improved if the clapper of the bell be replaced with a lead weight of two or three ounces.

Records

Once in twenty-four hours, after marking on the smoked paper the exact time at the last minute mark, the paper is carefully removed, a fresh sheet put in place and smoked, and the clock wound. First the beginning of each hour is marked, and on the top line a mark is made at every tenth minute. The date, ratio of magnification, and clock error are also noted. All these are scratched in the soot on the sheet. The record is then fixed by brushing rather thin varnish over the back of the sheet.

If a register is kept, at least the following data should be entered in it: I. Time of the beginning of first preliminary tremors. P'. 2. Beginning of second preliminary tremors. P".

3. Beginning of the first group of large or principal waves, P'".

4. Time of maximum motion. 5. Amplitude of maximum motion. ( Measured from position of rest of pen to extreme of motion to either side. This should he divided by the ratio of magnification of the lever I. ) 6. Period at time of maximum. (I. e., time from one crest to the next of the largest waves.) 7. Time of end of principal portion. 8. End of succeeding tremor-.

Locating A Distant Earthquake

The writer has been able, in the case of large, distant earthquakes, to announce the general location of the shock at once, from the records of the seismograph. Two elements are needed for this - the distance and the direction. As the first preliminary tremors travel much faster than the main, large waves, the difference in time of their arrival gives a measure of the distance of the origin. Various formulae have been computed for this, some of them very complicated. The writer has found, however, that a uniform rate of three degrees per minute is not far from the truth, for all distances; that is, for every minute that elapses between the beginning of the first tremor P' and the beginning of the first group of large waves P'", measure three degrees of distance on a great circle of the globe. That will generally give within ten per cent of the correct distance.

To determine the direction of an origin, a single horizontal pendulum is inadequate. There must he two, set at right angles to one another, so that by compounding the two co-ordinates thus given, the actual direction of the earth's movements may be seen The small diagram shows one method of bringing the records of two pendulums on one recording drum. The short arm of each recording lever is set at an angle of 135 deg. to the long arm. thus bringing the long arms parallel, as shown. One pendulum hangs north and south, and records motions of the earth east and west, while the other records motions north and south. To determine the direction of an earthquake origin, attention need be given only to the very first one or two waves of the preliminary tremors. It is known that the first preliminary tremors are waves which, like sound waves, move in a direction parallel to the line of propagation, while the main waves have a motion at right angles to this, like light. The latter, however, are exceedingly complicated waves, while, so far as the writer has observed, the first preliminary tremors always begin with a very -light motion away from the point of origin, followed by a considerably larger swing toward the origin. So that whenever the beginning of these tremors is strongly recorded it is possible, by comparing the north-south and east-west components of these first motions, to ascertain the direction from which the waves have come. This, with the distance, marked out (on the great circle) on a globe, gives the approximate location of the earthquake.