and Recording Earthquake-Motions. 203 



tical pendulum is the most convenient, and probably the most 

 accurate, instrument for static records of earthquakes ; and a 

 machine, based on the pendulum principle, was described by 

 me before the Seismological Society of Japan in March 1880; 

 and a description with drawing is given in the ( Proceedings.' 

 A description of that machine, and of some improvements I 

 have since made in it, forms the subject of this part of my 

 paper. 



The machine consisted of an ordinary simple pendulum 

 about 3 feet in length, the bob of which was of considerable 

 mass. From the centre of inertia of the bob three threads 

 radiated, and were attached to three light pulleys arranged at 

 equal distances apart on the circumference of a horizontal 

 circle having its centre at the centre of the bob. To these 

 pulleys very light indices were attached, the points of which 

 turned above graduated arcs, and showed the motion of the 

 earth magnified twenty-five times. The three components 

 were taken for the purpose of showing without ambiguity the 

 direction of the motion, if it had a definite direction, and of 

 giving information as to whether there was a multiplicity of 

 directions of motion. This machine was of course only capable, 

 under the most favourable circumstances, of giving the ampli- 

 tude of the greatest motion, this being indicated by the per- 

 manent displacement of the pulleys and indices. 



The improvements on this machine, to which I now call 

 attention, are, first, a method of rendering the pendulum dead- 

 beat, and, second, a modification of the mode of fixing the 

 pointers to the pulleys. 



A well-known objection to pendulum machines is their ten- 

 dency to acquire a swinging motion during the earthquake, 

 this of course causing their indications of extent of motion to 

 be untrustworthy*. I have found, after a considerable number 

 of trials, that this objection can be almost wholly overcome by 

 a very simple process. Let a rod be adjusted in such a way 

 that it can slide freely in a vertical direction through holes in 

 two plates, one above and the other below the centre of the 

 bob, but as near to its centre as possible. Load this rod until, 

 with its sharp point touching a glass plate, it offers sufficient 

 friction to bring the pendulum to rest after one half-swing, 

 when the initial displacement is about that of the largest 

 earthquake likely to occur. A load of 30 grammes will be 

 found sufficient in Japan for a 3-foot pendulum; the bob of 

 which weighs 10 kilogrammes. With this simple addition to 

 the pendulum its swinging will be almost wholly avoided, and 



* See Professors Ayrton and Perry's paper, u On a Neglected Principle 

 in Earthquake Measurements," Phil. Mag. Aug. 1879. 



