ON SEISMOLOGICAL INVESTIGATIONS. 



36 



than ±50' were excluded. This is far from a satisfactory procedure, 

 but it has been applied uniformly to all stations in Table II. All 

 observations for A> 120° have been omitted. 



Table II. 



VII. The Stereographic Method of Finding an Epicentre. 



If a large and accurate globe is available, distances between epicentre 

 (E) and observing station (S) can be read from it with considerable 

 accuracy ; and the quickest way of finding an epicentre (approximately) 

 is to describe arcs with centres at two or three stations for which A is 

 known (the radii being the known values of A ), and to note the common 

 point, or small area, of intersection. It may be worth remarking that 

 ibefore attempting to draw such arcs it is well to examine which stations 

 give consistent records, as shown by the time at origin. 



Thus for the quake on 1914 May 28'^ ll'i'S we have: — 



From the observed differences S — P the distances A from the epi- 

 centre can be inferred, and hence the whole time of transmission of P. 

 Applying this to the observed P we get the time at epicentre O. From 

 these figures for 0, which can thus be written down from the tables 

 alone, it is clear that the Barcelona and Padova results will not in this 



D 2 



