136 ART. 3. — B. KOTÔ: 



by reduction of density as a consequence of crustal dislocation. 

 The undercurrent so started finally causes vulcanism. 



The limit of the subsiding area in the present case is at a radial 

 distance of 52 km., while in the Mino-Owari plain at the time of 

 the earthquake of 1891 it was 70 km., approaching in both cases 

 the postulated limit of local disturbance of gravity anomalies. The 

 premonitory wandering earthquakes of Yoshimatsu and Ijû-in, and 

 the unwonted activity of the Kirishima volcanoes for nine months 

 prior to the Sakura-jima eruption all happened in the locally dis- 

 turbed area within the prescribed radius, {sse Premonifory Symp- 

 toms, p. 99). 



A few points escaped the notice of the writers on the recent 

 eruption, which ought not to be passed over without comment. 



First of all, it is to be remarked that during the destructive 

 earthquake on January 12th, 6,39 p.m., the first day of eruption, a 



A 



solid basalt cliif, 400 ???. high, collapsed near Cape Osaki (sse Map, 

 PI. XIII.), where the railway track and telegraph posts were 

 heavily damaged. All communication with the outer world was 

 then cut off and caused a rumor to spread all round the globe 

 that the city of Kagoshima was buried under lava (p. 70). It seems 

 evident to tlie writer that the accident was not merely due to 

 gravity, but the actual tossing of the subsiding ground induced the 

 heavy slip of the chif. As it will be remembered, Osaki is tlie 

 spot of maximal subsidence on the land side, amounting to 

 894 mm. 



Simultaneously with the landslump, the village Game, 12 J>m. 

 northwest from Cape Osaki, was strongly shaken, causing collapse 

 of bridges and cracking of house walls, besides gushing of ground- 

 water from fissures many cho (1 clio=:109 7îî) in length. Cape 

 Osaki is an edge of the Yoshino plateau, and Camô lies on the 



