to give a rough sketch of the present state of knowledge in this complicated subject.
Although history abounds with more or less complete accounts of earthquakes, it is remarkable that hardly ten years have yet elapsed since an accurate record was first obtained of what actually occurs during an earthquake. The combination of circumstances is curious, by which a knot of Scotch students, working in Japan, has secured so considerable an advance in seismology. The incessant but usually non-destructive earthquakes by which Japan is visited, the strange Japanese renaissance, and the importation of foreign professors, thoroughly trained at the Scotch universities in an accurate perception of mechanical principles, are the three factors which have co-operated to bring about this result.
The Scoto-Japanese professors, of whom the most eminent are Ewing, Gray, and Milne, have studied their subject with admirable persistence, and have by their ingenuity placed seismologists in possession of instruments by which the motion of the ground during an earthquake is recorded on an accurate scale of time. Such instruments are called seismographs, or recording seismometers. During an earthquake the ground and all that is fixed to it move together, and at first sight it seems impossible to get anything to stay still during the vibration. An exact description of a scientific instrument would be out of place here, but a general notion of these seismographs may be easily grasped.
The horizontal pendulum of Zöllner, and a suggestion of Chaplin (also a professor in Japan), are the sources from which "the horizontal pendulum seismograph" of Ewing originated. The principles according to which it is constructed may be explained as follows: If we consider an open door which can swing on its hinges, and imagine that a sudden horizontal movement is given to the door-post, at right angles to the position in which the door is hanging, then it is clear that the outer edge of the door will begin to move with a sort of recoil in the direction opposite to that of the movement imparted to the door-post. Since the door-post moves in one direction, while the edge of the door recoils, somewhere in the door there is a vertical line which remains still. The exact position of this line depends on the proportion which the amount of the recoil of the outer edge bears to the direct motion of the door-post. Now, if the sudden movement is imparted to the door-post by means of the floor to which it is attached, it is clear that a pencil attached to the door at this vertical line will write on the floor the displacement of the door-post, notwithstanding that the floor has moved. If next we suppose that there are two such doors hanging at right angles to one another from the same door-post, and that a sudden horizontal movement in any direction is given to the floor, each pencil will write on the floor that part of the movement which was at right angles to its door. Lastly, if the floor or surface on which the record
