The Early History of Seismometry (to 1900)
James Dewey and Perry Byerly
Further Studies with Seismoscopes
The years between 1850 and 1870 saw several significant contributions to seismological instrumentation. These included Palmieri's seismoscope for recording the time of an earthquake, and a suggestion by Zöllner that the horizontal pendulum might be used in a seismometer. Mallet studied earthquake motion by observing the effects of earthquakes and by measuring the velocity of elastic waves generated by explosions. A true seismograph still eluded seismologists.
Figure 3. Mallet's seismoscope (after Mallet, 1852). The image of a cross-hairs in C is reflected from the surface of mercury in the basin B and viewed through a magnifier, D.
Explosion seismology was born in 1851, when Robert Mallet used dynamite explosions to measure the speed of elastic waves in surface rocks (Mallet, 1852, 1862a). He wished to obtain approximate values for the velocities with which earthquake waves were likely to travel. To detect the waves from the explosions, Mallet looked through an eleven-power magnifier at the image of a cross-hairs reflected in the surface of mercury in a container (see Figure 3). A slight shaking caused the image to blur or disappear. Transit velocities were measured over distances of the order of a thousand feet. For granite, Mallet obtained velocities of about 1600 feet per second. He had expected to find velocities of 8000 feet per second. The unexpectedly low elastic-wave velocity was attributed to the heterogeneity of the rock through which the wave traveled. Later investigators (Abbot, 1878; Fouqué and Levy, 1888), using instruments similar to Mallet's, have found higher velocities and have suggested that Mallet may not have detected the earliest arrivals in his experiments.
Mallet advocated the use of fallen objects and cracks in buildings as aids in the study of earthquakes. He made a detailed investigation of the Neapolitan earthquake of 1857, in which he paid particular attention to the way buildings were cracked, walls overthrown, and soft ground fissured (Mallet, 1862b). Mallet believed that an earthquake consisted primarily of a compression followed by a dilatation. For such a shaking, he suggested, the resulting cracks in structures would be transverse to the direction of wave propagation. Overturned objects would fall along the horizontal projection of the direction of wave propagation. By observing the directions of arrival from a number of different points, he plotted an origin from which the wave seemed to spread. Mallet also published a set of formulae for calculating the velocities necessary to overturn structures of various simple shapes. From these, and observations of overturned objects, he estimated the velocity of particle motion at different sites.
Mallet's assumption that earthquakes consisted mainly of longitudinal motion was proven invalid as soon as seismometers were built which recorded the large transverse component of ground motion. Interest remained in the possibility of using overturned columns and other ruins as seismoscopes for measuring certain parameters of earthquakes, such as particle velocity and acceleration. In the 1880's, investigators in Japan checked Mallet's method of calculating ground velocity from observations of bodies overturned or displaced by earthquake motion. Velocities calculated from Mallet's formulae did not in general agree with those calculated from seismograms (Milne, 1885b). J. Milne and F. Omori introduced a formula for the acceleration necessary to overturn columns, which they checked by putting columns on a cart and shaking them with a sinusoidal motion (Milne and Omori, 1893). Omori applied the formula to gravestones overturned in the epicentral region of the Mino-Owari earthquake of October 28, 1891 (Milne, 1893c). He found accelerations greater than 0.4 g.
A design of a pendulum seismoscope was reported by Kreil (1855). In this instrument, the pendulum mass was to be a cylinder, on which recording paper was to be wrapped. The recording stylus, fixed to the ground, would write on the pendulum mass as it moved in an earthquake. The mass was to be rotated by a clock at a rate of once every twenty-four hours. In this way, the time of an earthquake could be noted. There is no evidence that Kreil's machine was built, although its design seems to have been considered significant by seismologists of the day (Mallet, 1859, p. 76).
In 1858, P. G. M. Cavalleri (1858, 1860) reported the construction of a common-pendulum seismometer similar to that described by Bina, more than one hundred years earlier. As in Bina's instrument, a pointer on the pendulum bob traced a record of the motion of the pendulum in fine powder. Observations of felt earthquakes suggest to Calvalleri that the frequency of earthquake waves would be three cycles per second. For this frequency, the 1.25 meter-long common pendulum would function approximately as a displacement meter, as its inventor intended.
Cavalleri described other instruments. A mass on a spiral spring was intended to detect vertical motion. It had a period of one second - long enough, Cavalleri felt, to record the rapid pulse-like vertical displacement of the ground believed to occur in an earthquake. The mass of this instrument was connected to the short arm of an indicating lever. The lever was constructed so that it would remain at the position of its maximum excursion.
Finally, in order to study the frequency content of earthquake waves, Cavalleri constructed six short pendulums of different periods, each of which traced the record of its motion in fine powder, as the larger pendulum did. Assuming that a range of frequencies from two to four cycles per second would be "sufficient to embrace every undulation occasioned by any earthquake" (Cavalleri, 1860, p. 113), Cavalleri expected that the pendulum whose period was closed to the predominant period of the earthquake would resonate and show a larger amplitude than the other pendulums. This apparatus was new to Europeans. Jared Brooks of Louisville, Kentucky, had constructed pendulums of different lengths to observe the New Madrid earthquakes of 1811 and 1812 (Fuller, 1912, p. 32).
In 1856, Luigi Palmieri installed his "sismografo elettro-magnetico" in the volcanic observatory on Mount Vesuvius (Palmieri, 1871, 1874). This instrument was intended to give the direction, intensity, and duration of an earthquake, and was capable of responding to both horizontal and vertical motions. It was not a "seismograph" in the sense in which we are using the word, but rather a collection of seismoscopes, each intended to record particular parameters of an earthquake (Figure 4).
Figure 4. Palmieri's "sismografo elettro-magnetico" (reproduced from The Engineer, 33, 1877, p. 407). Vertical motion is detected by a mass on a spiral spring E. The U-tubes n detect horizontal motion. Paper is unrolled from the drum i and a pencil mark put on the paper at m. The speed of the paper is regulated by the clock B. The clock A is stopped by the earthquake to give the time of the shock.
The seismoscope for detecting vertical motion consisted of a conical mass on a spiral spring. The mass was suspended just over a basin of mercury. When a slight motion caused the tip of the cone to touch the mercury, an electric circuit was completed, which caused a clock to stop, indicating the time of the shock. The spiral spring was constructed so that thermal changes in the length of the spring were balanced by thermal changes in the length of the frame to which the spring was attached.
Horizontal motion was detected with common pendulums, whose swinging completed the same circuit as that completed by the mass-spring seismoscope. In addition, U-tubes filled with mercury were used to detect horizontal motion.
The closing of the above-mentioned electric circuit, besides stopping the clock, started a paper recording surface and caused a pencil to be pressed against the surface. The recorder, once started, continued running until the paper was used up. Every time the circuit was completed, a pencil dash would be left on the moving paper. The duration of the quake was thereby recorded. The size of the earthquake was indicated by the amplitude of oscillations suffered by a mass on a spring and by the amplitudes of the oscillations of the mercury in the U-tubes. The size of the earthquake was measured in "degrees".
Palmieri's "sismografo" seems to have been an effective earthquake detector for its time. Palmieri used it for many years on Mount Vesuvius and detected numerous shocks with the instrument (Palmieri, 1862a, 1862b, 1864, 1866, 1867, 1869, 1870, 1876). Disturbingly, however, the instrument was unable to detect many shocks which were felt in the nearby city of Naples. Palmieri, in fact, believed that the apparatus functioned better as a predictor of earthquakes and volcanic eruptions. He observed that many instances before the eruption of Vesuvius or other Mediterranean volcanoes, or before large earthquakes in the Mediterranean area, his seismoscope would detect "shocks". But when the earthquakes occurred, even if they were felt in Naples, the instrument would not detect them. (Fortunately, an identical instrument located in Naples did detect the earthquakes felt there.) Palmieri observed that before Vesuvius was going to erupt, the "shocks are more frequent; or to express it better, the ground trembles in a continuous manner with diverse phases" (Palmieri, 1867).
Palmieri's "sismografo" was later used by seismologists in Japan. For ten years it was used to detect earthquakes in Tokyo; 565 earthquakes were detected from October, 1875 to March, 1885 (Milne, 1880c, 1883b, 1885c). For most of these earthquakes, "force" (the size of the earthquake in "degrees") and direction of motion, as well as time, are catalogued, indicating that the whole "sismografo" was functioning. It should be noted that some workers in Japan did not believe the "force" to be necessarily even an approximate indication of the relative "intensity" of different earthquakes (see, for example, Ewing, 1883a, p. 72).
After 1885, routine earthquake recording in Tokyo was taken over by the new seismographs just developed in Japan, but Palmieri's circuit-closing seismoscopes were used past the turn of the century as triggering devices to start recording systems in other seismographs (Holden, 1898). One such seismoscope was among the seismographic equipment at Mount Hamilton, California, at the time of the California earthquake of April 18, 1906. The seismoscope did not trigger the other seismographs until thirty-three seconds after the first tremors were felt at Mount Hamilton (Read, 1910, p. 64).
The horizontal pendulum appears to have been independently invented several times in the nineteenth century (Darwin, 1882; Davison, 1896). In 1869, Zöllner described a horizontal pendulum with the suspension which has since been associated with his name (Zöllner, 1869, 1872). Zöllner's suspension is shown in Figure 5. The rod R, with a mass on one end, rotates about the axis AC, which is inclined at an angle i to the vertical, V. The rod is supported by wires AB and CD, which are attached to the rod at points B and D some distance from each other. A mirror on the pendulum was used to reflect a light beam from a lamp to a scale, where the motion of the pendulum, as magnified by the optical lever, was directly observed. The instrument was installed in the cellar of the university in Leipzig; Zöller could detect significant movement of the pendulum due to the filling up of the auditorium on the second floor of the building. The pendulum was built in order to observe changes in the direction of gravity due to tidal forces, but Zöllner suggested that it might also be valuable as a seismometer.
Figure 5. Zöllner's horizontal-pendulum suspension.
Horizontal pendulums were to be widely used in seismographs after 1880, because they could be given long periods and could still be compact. There were several different suspensions used in these later horizontal-pendulum instruments. The Zöllner suspension was used in the Galitzin horizontal seismograph. The suspension of the Wood-Anderson torsion seismometer may be considered a limiting case of the Zöllner suspension.
From the Bulletin of the Seismological Society of America. Vol. 59, No. 1, pp. 183-227. February, 1969.