ON SEISMOLOGICAL INVESTIGATION. 219 



give r-ise to a sensation not unlike that felt upon a floating stage rising 

 and falling upon a swell. The movement of hanging pictures and that of 

 seismographs indicate that an intermittent tilting is taking place. The 

 heavy masses of metal in bracket seismographs, conical pendulums, and 

 other instruments, no longer act as steady points, but swing fitfully with 

 varying amplitudes from side to side, and, rather than giving records of 

 horizontal displacement, they are roughly recording the maximum slopes 

 of the earth waves which tilt the supporting piers. 



Beyond the 300-mile limit nothing is felt, and it is seldom that an 

 ordinary seismograph, writing with frictional indices, gives a record. 

 Now and then, where the friction of writing pointers has been exceedingly 

 low, records of unfelt earthquakes have been obtained from ordinary 

 seismographs. It was the magnitude of these diagrams obtained by the 

 writer, coupled with numerous observations made by astronomers on 

 the movement of the bubbles in levels, the tilting of water in ponds, 

 and kindred observations, which enabled him, in 1883, to venture the 

 opinion that with suitable instruments the movement of all large earth- 

 quakes might be recorded in any portion of the world (see ' Earthquakes 

 and other Earth Movements,' Int. Sci. Series, pp. 226 and 342). The 

 ample manner in which this has been confirmed is known to all seismo- 

 logists. 



Preliminary Tremors. — The period of these, as recorded on seismo- 

 graphs with frictional indices, has varied between \ and 4^ of a second. 

 Along paths of from 1 to 4 geographical degrees (111 to 444 kms.) the 

 velocity is apparently about 2 kms. per second. This, however, is the 

 velocity of the larger waves, which the preliminary tremors most certainly 

 outrace. Strange to say, we know less about the difierence in rate of 

 propagation of these small movements and their larger followers over 

 short ranges than we do over long ranges. As a working hypothesis, 

 founded on the interval of time that elapses between the screaming of 

 pheasants and the arrival of sensible motion and the records of seismo- 

 grams, I anticipate that this interval will be found to be about 10 seconds 

 for about every 100 kms. of travel ; that is, if a shock originates at a dis- 

 tance of, say, 200 kms., these preliminary tremors may be noticed 20 seconds 

 before the arrival of pronounced motion. If this is so, then the velocity 

 of propagation for preliminary tremors over short ranges will be about 

 2-5 kms. per second. 



If, for the time being, we accept this factor, then if I is the length of 

 a wave, t its period, and v its velocity, because 



l=.vt 



with a period of -'y second, the length of a wave is about -125 km. 

 (410 feet). 



Their amplitudes, as shown on seismograms, are exceedingly small, 

 say -j'^ mm. 



Large Waves. — The large waves have periods of from 1 to 2-5 seconds, 

 which, with velocities of 2 kms. per second, would indicate lengths of 

 2 to 5 kms. (6,560 to 16,400 feet). The maximum amplitudes of these, 

 which represent shocks which will shatter ill-constructed chimneys, lie 

 between 20 and 70 mm. 



Concluding Vibrations. — Seismogi-ams clearly show waves having 

 periods of from 3 to 5 seconds, the lengths of which may therefore reach 

 as much as 10 kms. (32,800 feet). 



