34 REPORTS ON THE STATE OP SCIENCE. 



Bidston, San Fernando, Shide, and Kew, the distances of which from 

 the epicentre lie between 27° and 79°, the average duration of move- 

 ment was 2h. 32m. ; whilst at Christchurch, Tokio, Irkutsk, Capetown, 

 Calcutta, Bombay, Perth, Kodaikanal, and Batavia the epicentral 

 distances of which lie between 104° and 160°, the average dura- 

 tion was 2h. 51m. ; the durations for Capetown and Christchurch were 

 respectively 4h. 52m. and 3h. 19m. — the longest recorded in the world. 

 One explanation for these observations is the assumption that the flood 

 of motion set free in Guatemala had been augmented as it travelled. 1 



Amplitudes point in the same direction. At Wellington, New 

 Zealand, the pointers were driven off the recording surface, and the 

 seismogram suggested that the earthquake had originated near to that 

 place rather than in Guatemala, which is 103° distant. Although 

 this particular record is exceptional, the seismograms from quadrantal 

 regions do not exhibit the great falling off in amplitude which we should 

 anticipate. 



Another indication of reinforcement is in the repeated maxima seen 

 in seismograms obtained at different stations. 



The records from Manila 2.46, Tokio 2.51, Irkutsk 2.51, and Wel- 

 lington 3.39 may be explained by the assumption that at about 2.36 an 

 earthquake originated at 113° E. and 27° N. 



The Wellington maximum which exceeded 15 min., it will be 

 observed, roughly agrees with each of the three epicentres which have 

 been considered — that is to say, maxima movements from three centres 

 reached Wellington at nearly the same time. 



All the observations to which I have referred were made with similar 

 installed instruments. 



Example 2. — The Arica Earthquake, December 26, 1906. 



On December 26, 1906, at 5.56 a.m. G.M.T., an earthquake wrecked 

 several houses in Arica and was felt severely in Iquique and Pisagua. 

 Prom local observations and observations made in Cordova and Trinidad 

 I should place the origin about 3° or 4°' from Arica, and conclude that 

 the disturbance originated a few minutes before it was recorded in that 

 city. A certain number of minutes after the Arica earthquake seis- 

 mograms were obtained at observatories in the West Indies, North 

 and South America, Great Britain, Europe, Siberia, India, Australia, 

 Batavia, China, Manila, and Capetown. These are more than fifty in 

 number, and because their commencements are approximately at times 

 we should expect P 1 to reach these various stations from an origin off 

 the North coast of Chile, seismologists have attributed these various 

 records to the Arica earthquake. They however indicate that the time 

 at which it originated was at 5.51 or 5.52, which is about 2 min. 

 earlier than the time I have just suggested. 3 



A close examination of these numerous records, however, shows that 

 those relating to P 2 , P 3 , and, I may add, amplitudes, do not lead to the 

 same conclusions as those derived from a discussion of those referring 



1 See Brit. Assoc. Reports, 1908, p. 72. 



2 Ibid., 1907, p. 86, fig. 1323 on the map (Milne) ; Bollettino delta Socield Sis- 

 mologica Italiana, Appendice, vol. 13, p. 511 (Martinelli) ; and the Publications du 

 Bureau Central de 1' Association Internationale de Sismologie, Catalogues 1906, 

 p. 84 (Szirtes). 



