ioo SCIENCE PROGRESS 



second epoch were marked by ten small eruptions, and were 

 followed by a period of inactivity of 1 1 5 years, during which 

 there was only one small outburst in the year i860. This 

 lasted until the opening in January 19 14 of the third epoch of 

 eruptive violence. 



In each of these great epochs, there occurred great flows 

 of lava from a number of craterlets arranged along nearly 

 opposite flanks of the mountain. The positions of the 19 14 

 •craterlets are shown by round spots in fig. 2. The areas 

 ^covered by the corresponding lava streams are represented by 

 -the shaded portions of the map, the lavas of the first epoch by 

 vertical shading, in the second by horizontal, and in the third 

 by diagonal shading. It will be noticed that, in the first epoch, 

 the outflow took place on the E.N.E. and S.W. flanks of the 

 volcano ; in the second, on the N.N.E. and S. flanks ; and, in 

 the third, on the E.S.E. and W.N.W. sides. In the first epoch, 

 the first outflow occurred in 147 1 on the E.N.E. side, followed 

 by two others, a few years later, in 1475 and 1476, in the south- 

 west. In the second and third epochs, the outflows took place 

 almost simultaneously on opposite sides, the time-intervals 

 between them being only a few seconds and a couple of hours. 

 The total areas covered by the lava-streams above sea-level are 

 respectively 2|, 2J, and 6\ square miles. 



Premonitory Signs. — The earliest heralds of the coming 

 eruption consisted, as usual, of a series of slight earthquake- 

 shocks. On January 11, about 3.41 a.m., people in Kagoshima 

 were roused from sleep by an earthquake evidently proceeding 

 from Sakura-jima. From this time onwards, the Gray-Milne 

 seismograph at the Kagoshima observatory recorded many 

 shocks, which rapidly increased in frequency, the average hourly 

 number being 4*1 from 3 a.m. to 11 a.m. on January 11 ; 12*4 

 from 11 a.m. to 8 p.m. ; and 19*5 from 8 p.m. on the nth to 

 10 a.m. on the 12th. The total number of earthquakes regis- 

 tered during these 32 hours was 418. 



In Sakura-jima itself, the earthquakes must have been 

 many times more numerous. At the village of Saido, on the 

 north-west coast of the island, 66 earthquakes were felt during 

 a single hour, from noon to 1 p.m. on January 11. In the same 

 time, only nine tremors were registered at Kagoshima. 



A day or so later, there were other symptoms of the approach- 

 ing outburst. Early in the morning of January 12, hot-springs 



