Seismological Investigations. 107 



or three instruments of this class upon which I have experi- 

 mented, this theoretical advantage is not realized in practice 

 The distance between the two platinum surfaces is so ex- 

 tremely small, when the instrument is sensitively adjusted, that 

 very trifling and generally unknown causes will considerably 

 modify the position of the divided head when contact occurs. 

 The setting of the instrument is therefore a matter of great 

 uncertainty. 



This instrument possesses what at first appears to be an ad- 

 vantage, in that it is u self-setting ;" but a reference to what 

 has gone before will show that this is not considered a desirable 

 feature of a seismoscope. In fact, it is very undesirable in my 

 judgment, except where devices of a special character are made 

 use of for securing time records. If the- epoch is established 

 by stopping or starting a clock or by Milne's printing device, a 

 self- setting instrument would be objectionable. If the record 

 is made upon the revolving drum of a chronograph, or on a 

 continuously moving strip of paper, this form of seismoscope 

 might fix the time of all sensible movements during the distur- 

 bance, and might, indeed, play the part of a seismometer in some 

 degree by distinguishing the more violent motions. But it 

 would be likely to fail in this respect, as well as in its general 

 performance as a seismoscope, owing to the peculiarity already 

 referred to. To vibrations of moderately long period it does 

 not promptly respond. Earthquakes seldom, if ever, begin 

 with a sharp and sudden movement. The maximum vibration 

 is nearly always preceded by several of less amplitude and 

 nearly the same period, which are themselves preceded by os- 

 cillations of extremely small amplitude but of great frequency. 

 The intensity of these preliminary tremors is not, in general, 

 sufficient to cause a mercury seismoscope to act, and to the 

 succeeding movements of greater amplitude but longer period, 

 (often as long as one second) it is not sensitive. Further ex- 

 periment and investigation is needed, however, to determine 

 the relative merits of these and other instruments. 



Time Apparatus. 



Various methods of time-registration have been made use of, 

 and most of them are generally well known. A choice among 

 them must depend largely on the probable frequency of earth- 

 quake phenomena. In Japan and some other parts of the 

 world rarely more than a few days pass without a sensible dis- 

 turbance. 



A continuously operating chronograph would be desirable 

 and profitable under such circumstances and various modifica- 

 tions of the ordinary astronomical chronograph will at once 

 suggest themselves as suitable for this work. When an earth- 



