ON SEISMOLOGICAL INVESTIGATION. 



145 



destroyed or avoided. Often they nitay be destroyed by giving the room 

 in which they are situated a copious and even draughty ventilation. If 

 this does not succeed, the instrument must have a new installation. They 

 are seldom met with in a badly constructed hut or beneath a tent. 



&c., &c., &c., up to the end of the week. 



From the above records it will be observed that the light has been 

 removed or extinguished twice a day. The times at which this is done is 

 very roughly noted with a pocket-watch. In the morning the lamp is 

 refilled, the eclipse watch wound, and, if necessary, the pendulum, which 

 may have wandered too much on one side, is reset. 



The error of the eclipse watch must, relatively to some standard time, 

 be noted accurately. For meaning of ' period ' and ' sensibility,' which 

 only need be determined once a week, and which can be expressed in 

 seconds of arc, see pp. 139, 140. 



From the mark shown on the developed film when the light is eclipsed 

 the time at which the watch commences to make an eclipse mark can be 

 calculated. These times, as shown on the dial of the eclipse watch, should 

 always be the same, and therefore in order to guard against accident they 

 are only made occasionally. By adding or subtracting the error of the 

 eclipse watch to the time at which an eclipse mark has been made, the 

 exact G.M.T. of this mark is obtained, from which any particular phase of 

 an earth movement may be computed. 



Weekly Bejwrt. 



At the end of the week a report is drawn up of the records, the 

 form of which largely depends upon the movements which have been 

 recorded. 



All times must be expressed in Greenwich mean time (civil), the day 

 commencing after 24 hours or midnight. Thus the ordinary notation of 

 June 16, 1.30 a.m., and June 16, 11.30 p.m., becomes June 16, 1.30, and 

 June 16, 23.30. 



The most important elements to be noted about an earthquake 

 disturbance are : — 



1. The exact time at which preliminary tremors commence. 



2. The duration of those tremors. 

 1897. 



