172 Jag gar — Yolcanologic Investigaiioiis at Kilauea. 



mechanism is indeed isostatic, for the uptiltings and upheavals 

 of bench blocks and lake bottom in Halemaumau are nothing 

 more than adjustments of the crust of the semi-solid lava 

 column, in which the lake is a saucer, to the mobile matter 

 beneath which tends to flow and rearrange the surface features 

 whenever overweighting takes place locally. 



Thus heavy flows repeated on the bench northwest for 

 weeks, August to October, 1916, heaved up out of the shallow 

 bottom of the lake a large crag mass, which eventually stood 

 as a hill 70 feet (21 meters) above the lake, but began as a 

 small fl.at islet (PI. \b, figs. 3, 4, 5). Later a similar islet suddenly 

 appeared in the southeast corner of the lake immediately 

 opposite a shore of continuous overflow, and this rapidly rose 

 until it was a pulpit rock 40 feet (12 metei's) high (PL lb 

 and fig. 7). The border bench itself on the east, fissured off from 

 its southeast extension, rose gradually for five months after 

 August, 1916, its surface tilting to the northeast where there 

 was ceaseless overflow, and finally lifted a pinnacle corner clear 

 above the edge of the pit while its back slope was inclined at 

 forty-five degrees (figs, ^a and 10). Most remarkable of all was 

 the adjustment of February 18, 1917, when after a fortnight 

 of subsidence of other features, this crag suddenly parted from 

 its supports against the old wall and in the course of twelve 

 hours or less subsided 30 feet (9 meters) upon its viscous 

 foundation, and at the same time opposite to it in the lake a 

 low flat island of the previous day rose 40 feet (12 meters) 

 above the lake to become a towering, flat-topped, steep-sided 

 mass (fig. 8Z>, see also S.W. islet fig. 5). In appearance this 

 mass was like the lava dome of Tarumai in Japan, which rose 

 in 1909.* The conclusion was inevitable that deep flow from 

 beneath the crag became inflow beneath the island, the distor- 

 tion affecting, not the lake, but the lake bottom. Other rising 

 features corroborated this by survey. The cause of this reverse 

 movement was the unloading of the lake bottom by the faster 

 sinking away of the liquid. The equilibrium, which had 

 balanced lake bottom versus border bench while the basin was 

 full, was disturbed when the lake sank differentially to its 

 saucer. 



Shoals, Sinkholes and Conduits. 



Other evidences that the lake of liquid lava is at all times 

 shallow, whatever its depression within the pit, were furnished 

 by the afore-described subsidence of June 5, 1916, and by rela- 

 tively sudden subsidences at other times. On June 5, 1916, a 

 shoal appeared of glistening black lava flats, not infallen debris, 

 after sudden subsidence of 60 feet (18 meters). In early Feb- 

 ruary of 1917, rapid subsidence from depression 45 feet (14 



* See article bv Simotomai in August number. 



