88 ART. 3. — B. KOTÔ : 



The Nabé-yama lava was fast moving under water, but the 

 heads of the lava-front rose up and formed a fuming series of 

 equatorial islets parallel to the stream-line of the 8éto channel. 

 The venthole No. 4 (Text-fig. 19) seemed to have fpvicn in sue, 

 this, the writer thinks, was due probably to the ash ejections of 

 geeatLata fj^f, i^fgi three days. The Java fiou' then received ahund- 



FLOWS IN ^ 



yama'^iTnts ^^^^ supplies from the widened pit. It was on this day 

 that the lava-free, cliffy Akashi Gongen hill (103 7??.) between 

 Waki and 8éto became completely overflooded (PI. IX. Figs. 2-3, 

 PL X. Fig. 1, Geologic Map), and the lava tongue around Nabé- 

 yama pushed considerably forwards toward Kurokami. By morn- 

 ing, the channel, which was formerly 400 m. wide and 8-0 fathoms 

 deep, was open only about 24 m. The lava flow on this side was 

 so rapid that on the next day the channel became almost land-locked. 



A 



Ash fell here and stifling gases swept the opposite coast of Osumi. 

 It was the first time that the writer heard of suffocating (jases in 

 the present eruption, indicative of the declining phase of activity. 



Jan. 24th r^\^Q activity was gi^adually diminishing. During the 

 morning it roared periodically and ash was still precipitating ; only 

 white fumes were seen on the city side ; the lava moved little by 

 little into the sea (29^ C.) with saline fumes. 



The east side was also rather quiet and exhausted, though 

 enveloped in white clouds. The channel of Séto remained in the 

 same condition, though becoming a little narrower. PI. X. Fig. 1. 



Jan. 25th rpj^^^ activity bccame feeble, the Yunohira vent being 

 apparently dead, and the lower ones were only making little 

 secondary eruptions. 



The east side was, ho'v\fever, still roaring and exhaling, and 

 the whole area was entirely enveloped in vapor. Little ash fell, 

 and odious gases swept down. 



