246 SCIENCE PROGRESS 



earthquake of 1887, it is estimated that more than 90 per cent, 

 of the dead bodies were found crushed beneath fallen arches. 



Of the six conditions which govern the high death-rate of 

 earthquakes, we are chiefly concerned only with the last three. 

 We cannot in any way limit the time of occurrence of a great 

 earthquake nor can we prevent the rapid succession of strong 

 after-shocks. Fore-shocks when they occur and the preliminary 

 sound may provide early notice of the coming shock but, unfortun- 

 ately, they are characteristic of slight as well as of disastrous 

 earthquakes. Weak shocks may come alone and we cannot dis- 

 tinguish between such isolated tremors and the forerunners of a 

 catastrophe. Moreover, when they assume the latter aspect, the 

 interval that may elapse before the great shock comes is of 

 uncertain duration. It may be a few minutes or hours, it may 

 amount to days or weeks. The preliminary sound and tremor 

 differ in this respect. Both precede the shock by a few seconds ; 

 and except in large and lofty buildings a warning of even five 

 seconds may be sufficient. Pheasants and other birds are often 

 terrified by the early tremors of an earthquake ; but when kept 

 by the late Prof. Sekiya for the purpose they failed to serve as 

 satisfactory heralds. The deep earthquake-sound, again, is not 

 equally audible to all persons. It is so low that to some, who 

 are not in the least deaf to ordinary sounds, it is quite inaudible. 

 There is also reason to believe that races differ in their capacity 

 for. hearing the earthquake-sound ; and it is possible that a 

 general deafness towards the earthquake-sound may result in 

 raising the death-rate. When this defect exists, it might perhaps 

 be remedied by the use of sensitive flames adjusted so as to 

 respond to the deepest sounds alone. 



All attempts to issue earthquake-warnings have failed and 

 have deserved to fail, for the supposed forecasts have been 

 based on insufficient data. Without some knowledge of the 

 origin of earthquakes and of the movements which precede the 

 final catastrophe, such attempts were of necessity futile. But, 

 with the recent growth of our knowledge, it seems by no means 

 impossible that we may in time be able to provide rough fore- 

 casts of a coming shock. To be of service, such forecasts should 

 give the approximate time at which an earthquake may be ex- 

 pected and the region in which its severity will be chiefly 

 concentrated. To furnish both elements is at present beyond 

 our powers. But to give one only may be useful and of the 



