78 ART. 3. — B. KOTO : 



Jan. 16th Early on the morning the writer was at the lookout 



of the Customhouse and saw the lava-front overriding the entire 

 ash-covered village of Yokoyama, and a half of Akamizu. A photo 

 taken (PL VI. Fig. 2, also Text-fig. 19) at 10 a.m. shows the ver}^ 

 Lava momeut of a lava edge reaching the sea. From this 



REACHED 



Sea moment the activity on the west side of Mt. Sakura- 



jima entered into the Saumiian stage in the eruption of Matavu of 

 1905. The scene was then beginning to change, white curdy 

 clouds of salt fumes enveloped the entire lava-front (PI. VI. Figs. 

 2, 3 ; PI. VII. Fig. 1). For convenience sake, the writer hereafter 

 employs the term ' Sawaiian type ' for all the phenomena involved 

 in the interaction of hot lava with sea water. 



The lava of tlie Séto branch from the Nabé-yama 



Eastebn '' 



Vents vcuts also scenied to have entered the sea, as plenty 



of half -boiled fishes shoaled about the coast of Futagawa in Osumi 

 on this day. The vents in the rear were very active in contrast 

 to those of the western, front sides, throwing out black ash clouds 

 to a height of 2,000 m. 



The writer was on board the ' Nishiki Maru,' which sailed 

 along the northwest coast of Sakura-jima rescuing cattle still 

 left on the island, and on this occasion he was enabled to approach 

 the ' live ' lava, which was still in motion, and to have a nearer 

 view of volcanic display periodically detonating and emitting black 

 Spatter ^^^ wliito clouds from Spatter cones in the still uncon- 

 ^^"""^ solidated lava tunnel (Text-fig. 20, s). Cursory ob- 



servers took these minor outbursts for the genuine eruptions from 

 the vents newly opened in the body of the mountain. 



The writer saw from on board threads of white fumes rising 

 from fissures above the vent No. 1. (Geologic Map). Later on com- 



