292 MIDDLEMISS : KANGRA EARTHQUAKE. 



maintained from the epicentre, would give a time of commencement of 

 6h. 9m. 59s. and the latter Gh. 8m. 52s. 



The latter very closely agrees with the time provisionally taken 

 as correct at the epicentre, namely, Cm. 9m. being based on the time 

 given as 6h. 9£m. at the Simla Meteorological Office and on the large 

 amount of evidence for 6h. 10m. in the cities lying about isoseismal 

 VII between Lahore and Saharanpur. 



Besides the above time of 6h. 17m. 52s. recorded at Calcutta, we 



have further corroborative times afforded by Mud 



Clocks stopped in Point and Saugor Island, which agree in giving 



6h. 17m. 0s. Other time evidence in Calcutta that 



may be mentioned here has been derived from the stoppage of clocks. 



At St. Xavier's College the electric clock stopped at 6h. 20m. 0s. ; 



at Alipur Observatory the astronomical' clock stopped at 6h. 19m. 0s. ; 



but it is important to remember that so far from the centre as Calcutta, 



where the shock was necessarily very weak, it is probable that these 



clocks did not stop all at once, and therefore that they considerably 



overstate the time 



Evidence derived from the examination of the magnetograms 

 at Dehra Dun, Bombay, Barrackpore and Toungoo 



Magnetograms. ° 



by the Survey of India Officers in collaboration with 

 Prof. Omori show, however, a possible modification of this which would 

 fix the time at the epicentre as 6h. 10m. 43s. 



The following is a report by Captain R. H. R. Thomas, R.E., on the 



subject : — 



On the ttme of the principal Earthquake shock on the 4th April 1905 as 



registered on Magnetograph Curves.— By R. H. R. Thomas, Captain, 



r.e. 



For the purposes of the magnetic records no minute accuracy in time is re- 



The accuracy oi mea- quired. The shortness of the time scales, viz., about 0-6 



surements. inch for one hour precludes accurate measurements to more 



than one minute, -01 inch being equivalent to that amount. 



The process of measuring the exact moment of time of a particular point on 

 the trace is burdened with several sources of error in the measurement of small 

 quantities. These are as follows :— 



(1) Error of zero of measurement — the verticals on the glass scale being 

 lines of a certain thickness, it is difficult to place the vertical line 



