THE GEEAT ERUPTION OF SAKURA-JIMA IN 1914. 89 



■Tax. 20th rjij^^ activity was fast declining. It roared at 7 p.m., 



11.42 r.ii., 11.51 P.M., and at 0.15 a.m., 0.42 a.m., 1.23 a.m., 2.15 a.m., 

 and 5.50 a.m. of the next day. The lava was moving a little 

 under the sea. The east was active, throwing up smoke to 

 a height of 2,000 m. The entire region was wrapped in white 

 fumes. The Séto channel was still open (18 m..) though becoming 

 narrower by 6 m. One could not row through the boiling strait 

 (60^' C). 



.Tax. 27th ijij^q activity of the Yunoliira ventholes passed aicay 

 with this day, though that of the Nahè-yama craterlets still lingered 

 on fo>' 6 days. Feebly active in the Yunohira area, with occasional 

 minor secondary eruptions and ' trombe.' The entire field had 

 a melancholy aspect (PI. IX. Fig. 1) like that of the hot rubbish of 



Apparent ^ \yjcr towu after a great fire. The lava-flow in the sea 



Movement ^ ^ 



laTa^IIo^v ^^'^^ ^^^'^ at a standstill before the city of Kagoshima, 

 siAXDSTM. the new lava -field, 1.4 sq. km., extended into the sea 

 950 ?7i. from shore with a breadth of 1^ hw. (Text-fig. 19). 



The Séto channel on the east was agitated. It was now re- 

 duced to 9 m. in breadth, and 3.6 m. deep with a brown pumiceous 

 bottom, observed at 3 p.m. at low tide. The sea w^as hot, but only 

 for the upper 90 cm., while at lower depth the water was cold. 



Jan. 28th rj:^^ FouRTH Phase. 27ie third p/irtst' of eruption ex- 

 tending from the 21st to the 27th icas now closed, and the activity 

 entered into the fourth phase. Tremors occurred only at 5.49 p.m. ; 

 roarings w^ero heard IG times in the city from 4 p.m. tiU 10 a.m. 

 of the 29th ; glares of lava at night became weak and were seen 

 in 5 spots. The lava of the Yunohira vents apparently moved no 

 more. On the east side, the channel was still not yet closed by 

 lava wdiich was moving fast forwards, being 9 m. in height on the 



