The Early History of Seismometry (to 1900)
James Dewey and Perry Byerly
We end our history of early seismometers at the beginning of the twentieth century. The Wiechert inverted-pendulum seismometer is probably the earliest seismograph which is still used, in essentially its original form, in some modern seismological observatories. With the introduction of damped seismographs, seismologists had instruments with which it would be possible to calculate fairly accurate ground displacements for all recorded waves, not just short-period waves. Mathematical theory for calculating displacements from recorded waveforms would be forthcoming in the next decade (for example, Wiechert, 1903; Galitzin, 1904; Reid, 1910).
We have seen that a large number of instrumental needs were recognized by the turn of the century, including the need for a strain seismometer. The type of seismometer most obviously missing from our history of early seismometers is the electromagnetic seismograph, introduced by Galitzin (1903). The galvanometer was widely used in the nineteenth century; important improvements were made by some of the very men we have encountered in seismology, including Thomas Gray, W. E. Ayrton, and J. Perry (see references given by A. Gray, 1921). We find Thomas Gray simultaneously publishing papers on galvanometers and seismometers (Gray, 1887a, 1887b). Milne (1882a) even used a galvanometer in a seismoscope in 1879; the galvanometer indicated the output of a microphone with which Milne hoped to detect possible "creaking" of a fault preceding an earthquake. But no one in the nineteenth century seems to have tried to obtain an electrical analogue of earthquake motion and record it with a galvanometer.
We have described the early development of seismological instruments, from seismoscopes capable of giving only a few data to seismographs suitable for detailed studies of earth motion. We have seen that the first important step in seismometry was the use of an inertial system so arranged that it would be caused by an earthquake to move relative to the ground. The value of various types of pendulums for earthquake sensing instruments was realized early, probably as early as 132 A. D. Andrea Bina, in 1751, was the first to build an instrument for recording the relative displacement of the ground and a pendulum bob in an earthquake.
In these early years of seismology, emphasis was placed on measuring a few specific parameters of earthquake motion. Besides detecting the occurrence of an earthquake, almost all the instruments we have discussed gave an indication of the "direction" of the shock. Many were designed to record the relative "size" of an earthquake. Devices to record the time of an earthquake were proposed in the eighteenth century. Palmieri's apparatus, installed in 1856, seems to have been the first to give the "times" of shocks.
Cecchi may be regarded as the inventor, in 1875, of the earliest seismograph. Nevertheless, British scientists working in Japan in the 1880's must be given credit for developing the seismograph as a practical research instrument. At this time, most seismologists tried to obtain the longest possible periods for their instruments. Long common pendulums were used. The horizontal pendulum was introduced into seismology in Japan. The inverted pendulum, combined with a restoring force so that it would be in stable equilibrium, had been used for a seismometer by Forbes in 1840, and was tried again in Japan in the 1880's. A short-period instrument, to have a period less than the periods occurring in earthquake motion, was suggested by Perry and Ayrton in 1877.
The last decade of the nineteenth century saw seismographs designed to record teleseisms as well as local earthquakes. The first record of a distant earthquake had been obtained by von Rebeur-Paschwitz in 1889. Von Rebeur's seismograph recorded optically on a continuously moving photographic surface. Photographic recording made possible small seismographs of high sensitivity. The records given by mechanically-recording seismographs, however, were generally sharper and easier to read than those from photographically-recording instruments. Artificial viscous damping was first used in seismographs by Wiechert in 1898. In 1900, the first Wiechert inverted-pendulum seismograph was built.
The earliest mathematical theory of a seismograph was given by Forbes in 184l. Forbes considered only simple non-oscillatory ground displacements. A theory of the response of a seismograph to arbitrary, periodic, ground motion was written in Japan, by Perry and Ayrton in 1877, but it appears to have had little influence on the development of seismographs in that country. Later theoretical notes by Poincaré and Lippmann on integrating seismograms to obtain displacements seem also to have passed unnoticed by most seismologists in the nineteenth century. The importance of tilting in seismograms were debated; by the end of the century, the once widely-accepted view that teleseismic records represented tilting of the ground was being abandoned by many seismologists.
From the Bulletin of the Seismological Society of America. Vol. 59, No. 1, pp. 183-227. February, 1969.