Seismological Investigations. 113 



The records of a seismic station, even when equipped in the 

 best possible manner, will be unquestionably affected by what 

 may be called " accidental errors." That is to say, it is be- , 

 lieved that the best of seismoscopes will fail to satisfy entirely 

 the conditions given above and will occasionally "go off" whe,n 

 there has been no real seismic disturbance. This may be due 

 to a variety of causes, among which may be mentioned : changes 

 in the position, level, etc., of the support of the instrument, 

 through a slow movement of the wall or pier to which it may 

 be attached ; occasional movement of heavy bodies in close 

 proximity to the instrument ; very violent winds which may 

 disturb the foundation of the building in which it is placed, 

 etc. It is often difficult to distinguish a false alarm thus given 

 from a true seismic disturbance and in the records of the seis- 

 moscopes at Washington and Terre Haute already given, two 

 or three doubtful instances occur. 



When the earthquake is sufficiently violent to be detected 

 without instrumental aid, confirmatory evidence is furnished ; 

 but one of the principal objects in making use of instruments 

 is to detect movements which would otherwise pass unnoticed 

 and to extend the range of observation of the more violent far 

 beyond what it can be at present. As an illustration of the 

 great importance of being able to know a genuine record I may 

 invite attention to one of the records made at Washington a 

 few hours later than the Italian earthquake of last year. The 

 epoch of the disturbance is such that there is nothing unreason- 

 able in supposing that it was really a wave which had been 

 transmitted across the sea or through the crust of the earth. 

 Interesting as this fact would be, it would be rash to make such 

 an assumption in view of the possibility of a purely accidental 

 disturbance of the instrument. 



A very similar state of affairs exists in connection with one 

 of the records of the Terre Haute seismoscope. Had there 

 been several instruments along the Atlantic coast they might 

 have confirmed or confuted the record made at Washington. 

 It is highly probable, however, that disturbances frequently oc- 

 cur which do not affect large areas. In arranging for the 

 systematic observation of phenomena so rare and so uncertain 

 as to time and place as earthquakes, it will hardly be possible 

 to place stations sufficiently near each other to insure the de- 

 tection of these minor movements, although they are for many 

 reasons of prime importance. With the possibility of accidental 

 disturbance and with seismoscopes somewhat widely separated, 

 it would seem impossible ever to certainly detect such minor 

 movements in the surface of the earth. There is a remedy, 

 however, and it lies in the precaution, the truth of which ex- 



Am. Jour. Scr.— Third Series, Vol. XXXV, No. 206.— Feb., 1888. 



7 



