56 , REPORTS ON THE STATE OF SCIENCE. 



correspond to those of records obtained at Shide and published in 

 British Association circulars. The small numbers which appear on the 

 map (Plate II.) by their positions approximately indicate those of origins. 

 The greater number of these, it will be observed, have been submarine. 

 To make these determinations for each earthquake a list of the prin- 

 cipal stations at which it had been recorded had to be drawn up. This 

 showed the times at which P,, P 2 , P 3 , and the maximum had been 

 noted, together with the amplitudes recorded at each station. In the 

 publications of the International Seismological Association, the Earth- 

 quake Investigation Committee of Japan, and in papers on particular 

 earthquakes this fundamental data has for similar work always been 

 published, but here it has been omitted. The reason for this is that 

 it would occupy some three hundred pages, together with the fact 

 that it can at any time be reproduced by consulting British Association 

 and other Eegisters. 



The names of places where an earthquake has been felt are followed 

 by the letter ' F, ' whilst those at which destruction has taken place 

 are indicated by the letter ' D. ' In these instances local observations 

 in or near to an epifocal region have been used to determine the 

 approximate position of a district from which an earthquake radiated 

 and the time of its origin. 



In cases where we have been without this local information, origins 

 have been determined by selecting five or six stations at which the earth- 

 quake arrived first and where their amplitudes were large, and with this 

 data we have computed the position sought, by methods well-known 

 to seismologists. The method I find most satisfactory is that of circles 

 (see British Association Eeport, 1900, p. 79). This has been applied to 

 the differences in time at which P,, P 2 , P 3 or the maximum were recorded 

 at selected stations. Another method which gives the distance of an 

 origin from a station is the difference of time between the arrival of any 

 two of these phases of motion. The results as to the times of occur- 

 rence at and position of an origin have been checked by comparing the 

 computed times at which the earthquake should be noted at stations 

 not included in the group of selected stations with the times actually 

 recorded at the non-selected stations. In consequence of this method of 

 working I have been led to the idea that certain seismograms which have 

 hitherto been referred to as a single disturbance may refer to two or more 

 disturbances (see p. 32). When an indicated time is followed by plus 

 or minus two or more minutes, this also means that there is a corre- 

 sponding uncertainty as to the position of an origin. All other times 

 given may be read to within plus or minus one minute, unless they refer 

 to records made in inhabited districts. These latter are probably correct 

 to within 30 seconds. 



The dotted lines on the map which are parallel to mountain ranges 

 or oceanic ridges and troughs are the axes of districts from which many 

 large earthquakes have originated. They are indicated by the letters 

 A, B, 0, &c. In the list two of these capital letters indicate that the 

 earthquake originated near to the junction of two ridges. 



The materials chiefly used for these investigations have been those 

 obtained from stations co-operating with the British Association using 



