218 Jacjgar — Yolcanologic Livestigations at Kilauea. 



drawal the pipe cannot be pulled up hand over hand, because 

 it emerges from the lava red-hot. The only way to pull it out 

 is by walking away from the rampart and leaving the hot por- 

 tion to trail over the bank. The end of the steel pi}ie emerged 

 clean and showed no fusion or especial oxidation, nothing to 

 indicate that the depths of tlie lake were comparable in tem- 

 perature with the flaming cones. This second sounding appara- 

 tus was not equipped with Seger cones. 



Confirmation of Soundings hy subsequent Subsidence. 



The second sounding test, like the first one, but at a different 

 place, indicated a depth of liquid lava, in the lake opposite the 

 northeast rampart, of 50 feet (15 meters), with the lower few feet 

 more viscous than the fluid above, and consequently of presum- 

 ably lower temperature. This place, like the east point which 

 was the scene of the first test, became during the following 

 month, February 1917, a vertical cliff over the subsided lake. 

 The sounding was made January 23, and on January 19 the lake 

 had been 68 feet (21 meters) below the rim and was rising 

 approximately two feet (61"^"') per day, so that on January 23, 

 the depression was 60 feet (18 meters). On February 22 the 

 depression of the lake was 106 feet (32 meters) or ^Q feet (11: 

 meters) lower, and the day before, February 21, a cascade devel- 

 oped from the lake over the sibbmerged ledge of the lake bottom 

 into a cavernous recess northeast, nearly under the locality 

 where the second sounding test was made. 



In other words, the revelation of a month of subsidence was 

 the actual outcropping through the lake at its margin of a por- 

 tion of its bottom 16 feet below the lake level of the previous 

 month. And the soundings in two places at this margin in 

 the previous month revealed about 50 feet (15 meters) of depth. 

 Here we have close accordance between the results of experi- 

 ment and the evidence afforded by changes in time. 



Incidentally this development of a cascade, at a marginal sink- 

 hole after about 50 feet of subsidence, is nothing new, but has 

 happened repeatedly before as one of the characteristics of 

 subsidence. The evidence of the existence of a ledge over 

 which such a cataract could fall (flg. 2, see section of sinkhole) 

 has in the past been puzzling, but this is fully explained when 

 we realize that the bottom builds up with the rising pool, that 

 the latter is habitually shallow and that when subsidence sets 

 in, the lake magma becomes even shallower, by sinking more 

 rapidly than the semi-solid bench magma. The lake at such a 

 time assumes the appearance of a rapidly streaming river. 



These preliminary experiments provide a method and open 

 the way to interesting further work relative to gas composition, 

 viscosity and temperature in depth, direction of the subsurface 



