'JHK GKEAT ERUPTION OF SAKUKA-JIMA IN 1914. 131 



land depression of 804 mm., wliile outer circles distend considerably 

 westwards. The region embraces the plateau of Yoshino, 500 m. 

 high. (3n the opposite eastern shore of the Bay of Kagoshiraa 

 tlie circles are nearly equidistant, though mostly of schematic 

 nature. , 



The outermost extension of disturbance from an hypothetic 

 centre lies at a distance of 52 km. on the east, where the settling 

 of level sinks to minimum with a negative value of -11 mm. which 

 is within tlie bounds of observational error due to refraction. (Jut- 

 side of the above border, the values are rather high from unknown 

 causes, though they run below -20 mm. The corresponding western 

 limit cannot be fixed owing to the want of data at disposal ; but 

 the line probably passes beyond the land. From the above, we 

 see that the outer limit of disturbance is not far away from the 

 postulated boundary of local gravity anomaly assigned by G. K. 

 Gilbert, that is to say 61 km. 



As already stated elsewhere (pp. 57 and 99), the premonitory 

 indications of recent vulcanism were signalized nine months before 

 in a swarm of earthquakes in the Yoshimatsu region (55 km. distant) 

 and the unusual activity of the neighboring Kirishima volcanoes. 

 The hypocentre was later shifted southwards to Ijii-in, and at 

 last the subterranean commotion was satisfied in finding vents in 

 Sakura-jima. All the recent local subterranean agitations confine 

 themselves within a boundary of about the postulated 61 km. from 

 an hypothetic centre of depression in the Bay of Kagbshima. 

 See PI. XIII. 

 Mixo-owAKi jj^ ^jj^^g connection, a few words may be said of 



Eakth- -^ 



'^'i^; "^ the gi'cat earthquake of Central Japan in 1891. The 

 result of renewed levelings by The Headquarter Staff' soon after 

 the event showed the radius of disturbance in the Mino-Owari 



