108 T. C. Mendenhall — Seismoscopes and 



quake is a rare phenomenon, years instead of days elapsing be- 

 tween two successive disturbances, all appliances for their 

 observation must be of the utmost simplicity of construction 

 consistent with certainty and accuracy of performance. A 

 fair]y good clock is a necessity and that being provided it re- 

 mains to determine how it may be used in recording the instant 

 of disturbance. One of three different schemes may be adopted : 

 a running clock may be stopped ; a clock at rest with hands set 

 in a known position may be started, or the position of the hands 

 may be registered without interfering with the going of the 

 clock. Either of these can be easily accomplished through the 

 instrumentality of the electric circuit closed by the eifect of the 

 disturbance on the seismoscope. The simplest method is that 

 of stopping or starting the clock, and considerable difference of 

 opinion has existed as to the relative merits of these two pro- 

 cesses. It will be seen at once that, the plan of starting a clock 

 from rest in a known position would have many advantages, 

 and one of them is certainly of great importance. It is that 

 after the happening of the earthquake the started clock can be 

 allowed to run until -a comparison with some standard time is 

 possible and if necessary its rate may be determined, so that the 

 exact epoch of the disturbance can be ascertained with consid- 

 erable accuracy. On the other hand, if a clock be stopped our 

 knowledge of the exact time of the occurrence will depend on 

 our knowledge of its error at the time of stopping, which can 

 only be known through previous observations of error and clock 

 rate. Thus it will appear that if the method of clock stopping 

 is to be resorted to, the clock must be under constant surveil- 

 lance and frequent comparisons must be made with some stand- 

 ard time. The clock-starting method is also open to some rather 

 serious objections. A clock which has been at rest for months 

 or even years is hardly likely to be constant in its rate during 

 the first few hours after starting. This objection can be in a 

 great degree removed, however, by carrying out suggestions to 

 be given later. Another, and more important, is that the re- 

 cord is liable to be lost entirely through the subsequent stopping 

 of the clock by the violence of the earthquake. The clock-stop- 

 ping method is not open to this objection as it is difficult to 

 imagine the starting of a clock by an earthquake, particularly if 

 the pendulum is held somewhat firmly by the stopping appa- 

 ratus. The third plan, that of registering the position of the 

 hands of the clock at the moment of closing the circuit without 

 interfering in any way with the movement of the clock com 

 bines many of the advantages of both the others, but has the 

 disadvantage of being more complicated and difficult to accom- 

 plish. It is easy to expose for an instant a quick photographic 

 plate, upon which an image of the clock face is projected, but 



