THE DEATH-RATE OF EARTHQUAKES 241 



per cent, but in the lower part of Messina itself and Reggio 

 di Calabria the rates may well exceed any of those given above. 



Among the conditions which determine whether the death- 

 rate due to an earthquake shall be high or low may be mentioned 

 the time of occurrence, the suddenness with which the shock 

 begins and the rapid succession of strong after-shocks. These 

 are all properties of the earthquake and beyond our control. 

 There are also others of no less consequence, which are governed 

 more or less by our own actions, such as the proximity of towns 

 to well-known seismic centres, the nature of the site selected — 

 whether on sloping or level ground, on a rocky or loose founda- 

 tion—and the nature of the buildings. I propose to consider 

 these conditions in detail, as it is only from a knowledge of such 

 conditions that we can expect to discover means of mitigating, 

 when we cannot altogether prevent, the disastrous effects of 

 great earthquakes. 



The time of occurrence is one of the most important factors. 

 An earthquake which occurs at night is nearly always more 

 disastrous than one in the daytime. Not only are people 

 gathered indoors but, if asleep, they are unable to take 

 advantage of the brief warning that is sometimes given by 

 the preliminary sound or tremor. Among earthquakes with 

 a high death-rate may be mentioned the Messina earthquake 

 of 1908, which occurred at about 5.20 a.m., the Ischian earth- 

 quake of 1883 at 9.25 p.m., the Neapolitan earthquake of 1857 

 at 10.15 p.m., the Kangra and Dharmsala earthquake of 1905 

 shortly after 6 a.m. and the great Indian earthquake of 1737 

 at night. Among those with a low death-rate are the Assam 

 earthquake of 1897, which occurred at 5.15 p.m., the Ischian 

 earthquake of 1881 -at 1.5 p.m., the Kingston earthquake of 

 1907 at 3.30 p.m., the Port Royal earthquake of 1692 and the 

 first Calabrian earthquake of 1783 which happened shortly 

 before and after noon. But even the daytime loses its advantage 

 when, owing to religious celebrations, many people are con- 

 gregated within doors. The Riviera earthquake of 1887, for 

 instance, took place on an Ash Wednesday morning at twenty 

 minutes past six. After a night spent in amusement, many 

 persons had lain down and were sleeping heavily; others had 

 risen early and were gathered together in churches. The 

 Caraccas earthquake of 1812 occurred at 4.7 p.m. on Ascension 

 Day. " The procession of the day," says Humboldt, " had not 



