The Early History of Seismometry (to 1900)
James Dewey and Perry Byerly
The Seismograph Becomes an International Instrument
News of the successes of the British in Japan began to affect European seismology in the middle 1880's. It did not take long for most European seismologists to appreciate the advantages of the British instruments and begin to improve on them. In 1886, E. Brassart of Rome constructed common-pendulum seismographs after studying the seismographs made in Japan (Brassart, 1886; Agamennone, 1906, p. 94). The motion of a meter-long pendulum was resolved into perpendicular components and traced on smoked paper which was mounted on a cylinder. This instrument and modifications introduced later by Brassart and Agamennone (Ehlert, 1897a, p. 434; Agamennone, 1906) were used rather widely in Italy.
A most dramatic increase of seismological activity in Europe followed the confirmation in 1889 that waves from large earthquakes could be detected by sensitive instruments located halfway around the world from the earthquakes' epicenters. We have studied, so far, only seismographs intended to record local earthquakes. For many years, however, there had been indications that the vibrations from large earthquakes traveled far from the regions where the earthquakes were felt, causing otherwise inexplicable disturbances on delicate instruments, such as astronomical levels and magnetometers (Baratta, 1895, 1897; Darwin, 1882; Fouqué, 1888b). Some observers had attempted to calculate the amplitude and/or the propagation velocity of the disturbances affecting their instruments, under the assumption that these disturbances originated from certain earthquakes (Oriani, 1783; Nyren, 1878; Fouqué, 1888b).
Figure 18. One of the first known recordings of a distant earthquake, obtained with von Rebeur's horizontal pendulum (reproduced from Nature, 40, 1889, p. 295).
The first known recordings of a distant earthquake, which were identified as such, were made in 1889, also with astronomical instruments (Figure 18) (von Rebeur-Paschwitz, 1889). The instruments were horizontal pendulums, designed by Ernst von Rebeur-Paschwitz to measure slight changes in the direction of the vertical. Two of these pendulums, located in Potsdam and Wilhelmshaven, recorded a large earthquake on April 17, 1889. The earthquake had been felt in Japan about an hour before it was recorded in Germany.
Von Rebeur's pendulum was of the form used earlier by Ewing in his horizontal-pendulum seismometer (von Rebeur-Paschwitz, 1894). It is shown in Figure 19. The instrument consisted of a rigid frame, rotating about two bearings A and B, each consisting of a point pressing into a socket. To the frame was attached a mirror M, which reflected light from a lamp, through a cylindrical lens, to a rotating drum which was covered with photographic paper. The drum turned 11 millimeters in an hour. Von Rebeur's pendulum was only 10 centimeters long, and carried a mass of only 42 grams. It was usually used with a period of from 12 to 17 seconds and a static magnification of 100 (Ehlert, 1897a, p. 404-407). Time was obtained with a second fixed light trace which wrote on the same photographic paper. Every hour, this second trace was eclipsed for five minutes.
Von Rebeur was the first to use a photographically-recording instrument for continuous seismological observations. [Fouqué and Levy (1888) measured propagation velocities of explosion-generated waves by using a modification of Mallet's seismoscope in which a ray of light was reflected off of a surface of mercury onto a moving photographic plate. The plate was started into motion shortly before the explosive was detonated. Earlier, earthquakes had sometimes been registered on photographically-recording magnetographs (Fouqué, 1888b).] The advantage of photographic recording was the complete absence of friction in magnifying and recording the relative motion of the pendulum and the Earth. The only sources of friction in von Rebeur's apparatus were the points where the pendulum arm was pivoted. The effect of this friction on the dynamic behavior of the pendulum was independent of the magnification of the instrument. In contrast, with mechanical registration, the friction between the indicator and the recording surface exerts a greater force on the pendulum according as the mechanical magnification increases. With mechanical registration, there exists a limit of magnification, above which the inertia of the pendulum cannot overcome the frictional forces between the indicator and the recording surface.
There were, however, disadvantages to photographic registration as compared with mechanical registration. Photographic records weren't as sharp as smoked paper records. Rapid, high amplitude oscillations did not record on photographic paper. The photographic paper was expensive. And, because of expensive paper, many investigators recorded at such slow speeds that accurate timing and detailed studies of the recorded waveforms were impossible. Mechanical registration thus continued to be widely used in seismographs. We shall find, in fact, that by using heavy masses and reducing friction to a minimum, mechanically recording seismographs were built which rivalled the photographically recording instruments in sensitivity.
Figure 19. The horizontal pendulum of von Rebeur Paschwitz (after Davison, 1896).
Von Rebeur kept careful records of "earth-tremblings" recorded by his pendulums for several years, through a period of an illness of which he died in 1895, at the age of thirty-four (von Rebeur-Paschwitz, 1895a; Davison, 1927). In 1892 and 1893, he had instruments set up in Strassburg and Nicolajew, 1800 kilometers apart. About half of the disturbances on his pendulums were recorded at both sites. At each location, only one pendulum was used. After von Rebeur's death, his instrument was modified by Ehlert (1897a, p. 406, 1897b), who increased the recording speed of the instrument and increased its weight, to make the pendulum less susceptible to movement by air currents.
Contemporaneously with the development of the horizontal pendulum seismometers in Germany, some Italian seismologists extensively developed the long common-pendulum seismometer. G. Agamennone and A. Cancani made particularly important improvements in this type of seismometer. They were aware of the work being done in Germany, but they considered the common pendulum superior to the light horizontal pendulum for purely seismological research. At this time, Agamennone believed the long common pendulum to remain nearly stationary for short-period oscillations, in contrast to the German light horizontal pendulums which he did not think would remain stationary under rapid vibrations of the ground (Agamennone, 1894). Most of the Italian seismographs used mechanical registration.
The first of the long common-pendulum seismographs was constructed by Agamennone (1893). It was designed as an improvement on a 1.5 meter long, 10 kilogram, Brassart-type seismograph (Agamennone, 1894). The smaller instrument had too slow a recording speed, so that individual oscillations could not be resolved in the seismograms. In addition, the mass was so light it couldn't overcome the friction of the writing stylus. It would remain displaced from the zero line after being disturbed These problems were both corrected in the new seismograph. It had a length of 6 meters and carried a mass of 75 kilograms. The success of this instrument prompted Agamennone to build a still larger common pendulum. In 1894, a 16 meter long pendulum, with a mass of 200 kilograms, was constructed (Agamennone, 1894). A similar instrument built at Catania had a length of over 25 meters (Milne, 1899, p. 259). Almost all of the common-pendulum seismometers employed a device to resolve the ground motion into mutually-perpendicular components. The records were written with pen and ink on relatively rapidly moving surfaces (Milne, 1899).
Figure 20. A Japanese earthquake recorded in Italy on a Cancani common-pendulum seismograph (reproduced from Atti. Accad. naz. Lincei Rc. vol. 3, ser. 5, sem. 1, 1894, p. 554).
The early Italian pendulums gave important records of large teleseisms. A glance at a teleseism recorded by a 7-meter long, common-pendulum seismometer of Cancani (Figure 20) shows that the two body-wave groups and the surface-wave train were fairly well recorded. These instruments, in fact, were the earliest to make such a separation of phases for teleseisms. Cancani proposed that the first two wave groups, corresponding to our P and S waves, both represented compressional waves, and the third, our surface waves, represented distortional waves (Cancani, 1893). In 1899, Oldham (1900) presented a thorough study of teleseismic waves, which concluded that the first wave group represented compressional waves, the second represented distortional waves, and the third represented surface waves. He relied heavily on data from the low-magnification Italian common-pendulum seismometers, believing the German horizontal-pendulum seismometers to be less trustworthy for his purposes.
Disappointingly, however, the sensitivity of these early, long, common-pendulum seismometers did not approach that of the German light horizontal pendulums. Friction limited the Italian instruments to static magnifications of ten or so.
In 1895, Vicentini and Pacher constructed the Vicentini "microsismografo", a mechanically-recording seismograph with a magnification nearly equal to that of the German machines (Pacher, 1897). The seismograph is shown in Figure 21. A 100 kilogram mass M was suspended in a 1.5 meter-long pendulum. The relative motion of the bob and the ground was first magnified by a mechanical lever L. The motion of this lever was resolved into perpendicular components at V and, in the process, the pendulum motion was magnified again. The total magnification was 80. The traces of the two horizontal components were written side by side on smoked paper, along with a time trace. In 1896, a large-scale version of the "microsismografo" was constructed, with a pendulum length of 10.68 meters and a bob of weight 400 kilograms (Pacher, 1897).
A vertical-component seismometer was later introduced by Vicentini and Pacher (1898). Rather than employing suspension with a spiral spring for restoring force, Vicentini and Pacher used a flat spring, clamped to the wall at one end and loaded with a weight at the other end. The instrument was clamped so that the loaded end of the flat spring, bent under the weight of the mass, was horizontal. As finally developed, the mass oscillated vertically with a fundamental period of 1.2 seconds. The vertical seismometer wrote with a static magnification of 130 on a smoked-paper record which was constantly in motion.
Figure 21. Vicentini's "microsismografo" (modified from Galitzin, 1914). The recording part of the apparatus is shown enlarged three times relative to the rest of the instrument.
In 1895, John Milne left Japan and returned to England, where he established a seismological observatory on the Isle of Wight (Davison, 1927). He concentrated now on the study of unfelt earth movements, both microseisms and teleseisms. Milne made extensive use of a horizontal-pendulum seismograph which he designed in 1894, while he was still in Tokyo (Figure 22) (Milne, 1894b). The instrument recorded photographically. Instead of having light reflected onto photographic paper with a mirror fastened to the frame of the pendulum, Milne had light shine onto the paper through the intersection of two mutually-perpendicular slits. One of the slits was fastened to the pier. The other slit was fastened to the pendulum, and moved with the pendulum, thus causing the spot of light to move on the paper. Leveling screws in the base of the apparatus made possible a determination of the static tilt sensitivity of the pendulum by giving the base of the instrument a small known tilt and observing the resulting displacement of the trace. Static magnification may be calculated from an instrument with known period and known static tilt sensitivity. Milne's instruments usually had a period of about fifteen seconds and a static magnification of six.
Figure 22. The Milne horizontal seismograph (modified from Milne, 1898a). Light from L is reflected by M through the intersection of two crossed slits onto photographic paper. The lower illustration is a top view of the instrument with its outer case removed. T is a flexible wire holding up the boom. The weight W is pivoted on the boom.
Milne pressed for the establishment of a world-wide network of seismographic stations, with standard instruments (Milne, 1897). Milne's photographically-recording horizontal-pendulum seismograph was selected by a committee of the British Association for the Advancement of Science to be the standard instrument for such an undertaking. By 1900, similar Milne seismographs were established on all of the inhabited continents (Milne, 1901). [The desirability of having a station in the Antarctic region was also apparent at this time. In 1902, a Milne instrument was operated near the shore of the Ross Sea, at 77 degrees 50 minutes south latitude as part of the British national antarctic expedition of 1901-1904. The seismograph recorded over one hundred teleseisms in the period of months in which it was operated (Milne, 1905).] Sixteen stations were regularly sending records to Milne (Figure 23). Using data from the Milne seismographs, and published data from German and Italian observatories, Milne plotted travel-time curves for teleseisms with known epicenters. The first curve (Milne, 1898a) gave only the transit time of the phase of maximum amplitude, which, on seismograms made with the Milne seismograph, usually corresponded to our surface waves. A year later, the transit times of both the "preliminary tremors" and the phase of maximum amplitude were plotted (Milne, 1899). The transit times of the "second preliminary tremors" were plotted by Oldham (1900), who correctly inferred that the wave group was composed of transverse waves. Milne began systematic location of large teleseisms, using arrival times of the maximum phases, felt reports, and the time intervals between the arrival of preliminary tremors and the maximum phase (Milne, 1900b).
Figure 23. A record obtained with a Milne horizontal seismograph on April 5 1901 (reproduced from Rep. Bril. Ass. Advmt. Sci. 1901, p. 50). As may be seen, the usefulness of Milne's instrument was diminished by its lack of damping.
Italian seismologists began using horizontal-pendulum seismometers toward the end of the nineteenth century. In 1895, G. Grablovitz built a seismometer of the type earlier devised by Gray (above) - a horizontal pendulum pivoting on a single point and held up by a flexible wire (Grablovitz, 1896a). At first, Grablovitz used three horizontal pendulums, no two of them colinear, recording simultaneously at a single location. This was so that the direction of propagation could be determined from the amplitude of the recorded waves alone without knowledge of the phase of the wave for each component. This method of determining the direction of propagation assumes that the ground particles are known to vibrate either longitudinally or transversely: it still results in an ambiguity of 180 degrees for the direction of propagation. The method was required by the slow recording speed used by Grablovitz, which made it impossible to match up an individual oscillation from one component with an individual oscillation on another component. The simultaneous use of three horizontal components had been suggested by other authors (Ehlert, 1897a, p. 358, 403, 406). In a later version of his instrument, Grablovitz (1896b) increased the recording speed and used only two pendulums at right angles to each other. The improved version of Grablovitz's horizontal pendulum carried a mass of twelve kilograms. It had a static magnification of eight, and was used with a period of around seventeen seconds. A. Cancani built a larger horizontal-pendulum apparatus (Cancani, 1897). Cancani's seismometer used a suspension similar to that used by Ewing, with two bearing points. This seismometer wrote with pen and ink on a surface moving at the relatively rapid rate of sixty centimeters per hour. The period of the instrument was usually about twenty-four seconds. The Stiattesi horizontal pendulum, introduced in 1900, was a larger seismometer built on the same principle as the Cancani horizontal-pendulum seismometer (Agamennone, 1906, p. 115).
F. Omori, a pupil and colleague of Milne's, constructed horizontal-pendulum seismographs. As with the instrument of Grablovitz, each pendulum consisted of a mass on a rod, pivoting about a socket, with the mass held up by a flexible wire (Omori, 1899). The static magnifications of Omori's instruments were about ten, and they were given periods of about twenty seconds. These instruments were prototypes of the Bosch-Omori seismograph which was widely used throughout the world in the early twentieth century. The Bosch-Omori seismograph was built by the firm of J.A. Bosch, of Strassburg. In 1907, Bosch added damping to Omori's originally undamped seismometer (Sieberg, 1923, p. 442).
From the Bulletin of the Seismological Society of America. Vol. 59, No. 1, pp. 183-227. February, 1969.