The Chinese philosopher Zhang Heng invented the earliest known seismoscope in 132 A.D. The instrument was said to resemble a wine jar of six-foot diameter. On the outside of the vessel there were eight dragon heads, facing the eight principal directions of the compass. Below each of the dragon heads was a toad, with its mouth open toward the dragon. The mouth of each dragon held a ball. At the occurrence of an earthquake, one of the eight dragon-mouths would release a ball into the open mouth of the toad situated below. The direction of the shaking determined which of the dragons released its ball. The instrument is reported to have detected a four-hundred-mile distant earthquake which was not felt at the location of the seismoscope.

The inside of the Chinese seismoscope is unknown. Seismologists of the nineteenth and twentieth centuries had speculated on mechanisms which would duplicate the behavior of Zhang Heng's seismoscope, but would not be beyond the Chinese technology of Zhang Heng's time. All assume the use of some kind of pendulum as the primary sensing element, the motion of which would activate one of the dragons. In his translation of the original Chinese description of Zhang Heng's seismoscope, English seismologist, Milne implied that the pendulum was a suspended mass, a common pendulum. Seismologist, Imamura thought an inverted pendulum was more probable. Hagiwara constructed an inverted-pendulum seismoscope which behaved nearly as Zhang Heng's was reported to have behaved. The model designed by Hagiwara, however, responded most frequently to transverse motion, and indicated a direction normal to the azimuth between observer and epicenter, whereas the Chinese seismoscope was reported to have indicated the azimuth of the earthquake. It has been suggested that Zhang Heng's "earthquake weathercock" was calibrated empirically for its direction-determining properties.