THE GKEAT ERUPTION OF SAKÜRÄ.-JIMA IN 1914. 135 



From the new values of longitude and latitude, X and Y, the 

 resultants are gi'aphically estimated which express the directions 

 and relative measure of horizontal shifting of ground on the posi- 

 tion of bench marks, as shown with arrowheads on Map, PI. XIII. 

 An inspection of arrowheads shows that they also point at the 

 hypothetic centre of subsidence within the Bay of Kagosliima, only 

 Nos. 17, 18 and 10 on the southwest slope^^ of Sakura-jima, 

 however, point at the opposite direction, and those on the high flat 

 of Sakkabira on the southeast are directed toward a secondary 

 centre near the bench mark No. 12. The writer cannot give any 

 adequate explanation on these apparent anomalies. 



In summaritinçi what is stated above, our Headquarter Staft' 

 did good service in the cause of science in laboriously 

 working out the actual occurrence of both the vertical negative 

 movement of the , crust and the horizontal shifting of geographic 

 elements around and in Sakura-jima after its paroxysmal eruption. 



A novel feature on vulcanology was brought to hght in the 

 concentric circle of equal subsidence and the horizontal shifting of 

 ground toward a common centre in the bottom of the Bay of 

 Kagoshima, where the maximum subsidence of the crust is to be 

 expected. These movements may well be explained as a local 

 disturbance of the lithosphère on the principle of isostasy. Gilbert 

 (p. 12!)) postulated the radius of local disturbance of gravity 

 anomalies to be 01 /.???., and these anomalies arise from irregular 

 vertical distribution of density with a defect above and excess 

 below, all above the zone of compensation at 122 hn. The in- 

 equality is adjusted by the undertow of a shallow layer made mobile 



1) Ihe abnormal shifting of bench marks seems to tho writer to be due to the tilting and 

 slipping of a portion of mountain mass through the shocks given by volcanic spasms. The 

 same explanation can be applied to the bench mark No. 21. See p. 132. 



