The Early History of Seismometry (to 1900)
James Dewey and Perry Byerly
The "Seismometer" of James Forbes
The interest in seismic instruments generated by the Calabrian earthquakes was not sustained in succeeding years. In 1839, however, a series of small earthquakes began near Comrie, in Perthshire, Scotland. Scores of shocks were felt over a period of several years. A direct result of the Comrie earthquakes was the establishment of a Special Committee of the British Association for the Advancement of Science, the purpose of the committee being to obtain "instruments and registers to record shocks in Great Britain" (Milne, 1842).
The most significant instrument resulting from the committee's work was an inverted-pendulum "seismometer", designed by James Forbes (Forbes, 1844). The seismometer is shown in Figure 2. It consisted of a vertical metal rod having a mass C moveable upon it. The rod was supported on a vertical cylindrical steel wire. The wire could be made more or less stiff by pinching it at a greater or lesser height by means of a screw S. By adjusting the stiffness of the wire, or the height of the ball, the free period of the pendulum might be altered. A pencil L placed on the prolongation of the metal rod wrote a record on a stationary, paper-lined, spherical dome I. By placing the pencil sufficiently far above the mass, a magnification of the motion of the mass by a factor of two or three could be obtained.
Figure 2. Forbes' seismometer (after Forbes, 1844). The screws E, acting on the support D, are used to help set the pendulum in an upright position.
Forbes used the inverted pendulum, mounted on a stiff wire, to provide a sensing element in weakly-stable equilibrium, without having to use a very long common pendulum. An upright inverted pendulum, by itself, is in unstable equilibrium. A slight motion topples it over. When the inverted pendulum is mounted on a suitably stiff wire, the apparatus may be rendered stable, so that the inverted pendulum returns to an upright position after having been disturbed. Such a combination had been used already by a Mr. Hardy for the purpose of detecting vibrations set up in a clock frame by the beating of the clock (Kater, 1818).
Forbes was probably the first to attempt explicitly to give a seismological instrument a "long" period. [Forbes speaks of requiring, for earthquake recording, a common pendulum, ten or twenty feet long, which would have a period of four or five seconds (Forbes, 1844, p. 219). Later, in a discussion of the size of his seismometer, he states that the sensibility of seismometers of different size should be the same - "say one second" (Forbes, 1844, p. 221). We suspect, however, that he hoped for a period nearer five seconds than one second, since his seismometer offered no advantage in compactness over a common pendulum with a one-second period. We have found no indication of the period actually obtained with the instrument.] He was also the first to try to avoid the clumsiness of a long common pendulum in obtaining long periods. As we shall see, his method of approaching neutral equilibrium was similar to that used by Wiechert in 1900. Forbes desired a long period in order that the pendulum remain stationary as the Earth moved beneath it. He clearly wanted to measure ground displacement in an earthquake. However, he considered only the effect on his instrument of an "earthquake" consisting of a single, horizontal displacement of uniform velocity, beginning and ending suddenly. For this reason, we can't be sure if Forbes knew that a long-period pendulum would function as a displacement meter for very-short-period oscillations of the ground. For the type of motion he considered, Forbes expected his seismometer to show a straight line, corresponding to the sudden displacement of the Earth, which would be easily distinguished from the ellipsoidal traces caused by the pendulum oscillating about its new equilibrium position.
A mathematical theory of the instrument accompanies its description (Forbes, 1844). Forbes was the first to describe mathematically the behavior of a seismic instrument in an "earthquake". The assumed earthquake, again, is a single motion of uniform velocity, starting and stopping abruptly.
Six of the inverted-pendulum seismometers were set up at Comrie (Milne, 1842, 1843). One was thirty-nine inches long and another was ten feet long. The lengths of the remaining four were not noted. Neither were the periods of the instruments noted. They gave disappointing performances. In one year, for example, the seismometers recorded only three of sixty earthquakes felt at Comrie (Milne, 1843). The nature of the records is difficult to decipher from the descriptions given by the investigators, who mention the indicating pencil being displaced by the quake, but don't mention any records being written by it. They conclude that the motion in the recorded earthquakes consisted of a sudden movement of the ground, with a maximum displacement of one half of an inch horizontally (Milne, 1842, 1843).
Friction between the writing pencil and the recording surface must take some of the blame for the disappointing results given by the seismometer. Forbes was aware of the problem of friction. He attempted to overcome it by using a heavy mass, but he does not seem to have been successful. We will see that friction would severely limit the sensitivity of many later seismographs which had mechanical registration.
Common pendulums and an instrument for measuring vertical motion were also used at Comrie. The vertical motion seismoscope consisted of a horizontal metal bar, loaded with a weight at one end, and fastened to the wall with a flat spring. In an earthquake, the weight was expected to lag behind the ground motion and move a light straw to indicate the extent of vertical motion. The vertical-motion instrument functioned on several occasions, indicating vertical "displacements" as great as half an inch.
The earthquakes at Comrie diminished in number, and after several years the committee on earthquake instruments was allowed to dissolve.
From the Bulletin of the Seismological Society of America. Vol. 59, No. 1, pp. 183-227. February, 1969.