The Early History of Seismometry (to 1900)

James Dewey and Perry Byerly

The Göttingen Seismographs

We return to the history of the seismographs themselves. In 1898, E. Wiechert of Göttingen introduced a seismograph with a viscously-damped pendulum as a sensor (Wiechert, 1899). The damping was added to lessen the effects of the pendulum eigen-oscillations. Wiechert's first seismograph was a horizontal-pendulum instrument, which recorded photographically. After a trip to Italy to study seismometers used in that country, he decided to build a mechanically-recording seismograph. For a sensor, he used an inverted pendulum stabilized by springs and free to oscillate in any direction horizontally (Wiechert, 1904). The seismograph was completed in 1900.

Figure 26. The 1000 kg Wiechert inverted pendulum seismograph (after Wiechert, 1904). The plate P is attached to the frame of the instrument. N is attached to the pendulum mass. The motion of the mass relative to the frame is resolved at A into perpendicular components. Restoring force is applied to the mass M from springs at C, C', by means of the rods B, B'. H, H' are the damping cylinders. The whole inverted pendulum is pivoted at K. In the actual seismometer, the rotation of the pendulum about K takes place in flat springs, which are arranged in a Cardan hinge to permit the pendulum to move in any horizontal direction.

In principle, Wiechert's inverted-pendulum seismometer resembled the inverted-pendulum seismometer of James Forbes, built sixty years earlier. The restoring springs, which kept the pendulum in stable equilibrium for small oscillations, were applied to the top of the inertial mass. At the base of the pendulum was a joint which permitted motion in any horizontal direction. The first description of the seismograph which we have found was published in 1904, after the instrument had undergone alterations in details of its construction (Wiechert, 1904). The 1904 pendulum (Figure 26) carried a mass of 1000 kilograms. [In later years, Wiechert inverted-pendulum seismometers were built with masses as light as 80 kilograms (Berlage, 1930, p. 434-435).] The relative motion of the mass with respect to the ground was resolved into perpendicular components, magnified 200 times by a mechanical lever system, and written on smoked paper. Time marks were put on the paper every minute by lifting the writing index off of the paper. For damping, Wiechert applied the motion of the pendulum mass to pistons which fit closely inside cylinders attached to the seismometer stand. Resistance of the air to the piston motion provided damping for the pendulum; this resistance was controlled by a valve which had the effect of regulating the amount of air space between the piston and the cylinder.

We have mentioned the work of W. Schlüter, a colleague of Wiechert's, who attempted to measure tilt with his klinograph. After he failed to record tilt, Schlüter converted his klinograph into a standard seismograph. By adding mass to one side of the klinograph and thus moving the instrument's center of mass away from the axis of rotation toward that side, he could make the klinograph sensitive to linear motion perpendicular to the line containing the center of mass and the axis of rotation. In particular, by moving the center of mass horizontally away from the axis of rotation, and maintaining it in that position with a vertical spiral spring, he could render the klinograph sensitive to vertical displacement. Schlüter obtained in this way the first long-period vertical-motion seismometer. As he finally developed the instrument, it was used with a period of sixteen seconds and a static magnification of 160. The klinograph was damped; it recorded photographically. Schlüter used compensation devices to remove drift due to temperature changes and changes in the stiffness of the spiral spring: most of the records reproduced by him show remarkably little drift (Figure 27).

Schlüter's klinograph obtained the first satisfactory records of longer-period vertical motion. Schlüter noted that, in general appearance, vertical-motion seismograms resembled horizontal-motion seismograms. There were differences, however, particularly in the onsets of earthquakes, which were much sharper on the vertical-motion seismograms than on horizontal-motion seismograms of the same earthquakes. For very distant earthquakes, further than 90 degrees away, the onset on the vertical-motion records occurred several minutes earlier than the onset on the horizontal-motion seismogram. Schlüter accordingly suggested that travel times might better be obtained from vertical-motion records than from horizontal-motion records.

Schlüter's klinograph seems to have been used only in Göttingen and only for a relatively short time. The high cost of photographic registration may have limited the use of the klinograph (Wiechert, 1907, p. 629). Because of the difficulties inherent in the design of a mechanically-recording vertical seismometer of medium period, Wiechert postponed the construction of such an instrument until after he had perfected his horizontal seismograph (Wiechert, 1904, p. 436). A mechanically-recording vertical seismograph with a period of five seconds was installed in Göttingen in 1905 (Wiechert, 1907, p. 629).

Figure 27. Vertical-motion record obtained with Schlüter's klinograph, on September 30, 1901 (reproduced from Beitrage zur Geophysik, 5. 1903, plate 2). To obtain this record, the klinograph was modified from the form shown in Figure 25 to make it sensitive to vertical displacements (see text).

From the Bulletin of the Seismological Society of America. Vol. 59, No. 1, pp. 183-227. February, 1969.