238 
PACIFIC SCIENCE, Vol. 1, October, 1947 
maumau had to exceed the pressure of the 
talus plug above the seat of the explosions. 
If the explosions took place 500 feet above 
sea level, that would mean that they occurred 
about 1,900 feet below the bottom of Hale- 
maumau. The pressure of 1,900 feet of talus 
material with a specific gravity of 2.0 would 
be about 1,600 pounds per square inch. 
To produce the explosions there must have 
been either a gradual build-up of pressure 
or a rapid generation of the same. If the 
pressure build-up were gradual, then an ef¬ 
fective seal was required. A seal made up of 
talus blocks even with an appreciable per¬ 
centage of dust would appear to be inade¬ 
quate. There were no indications of fusing 
or cementing of the talus plug, and because 
the plug developed above the retreating lava 
column, there w r as no molten lava to fill in 
the voids between the blocks and help de¬ 
velop a seal. A slow build-up of pressure 
would necessarily limit the amount of water 
available for any one explosion. The pres¬ 
sure build-up would soon prevent the 
further entry of water. A rapid generation 
of steam pressure, as in a "flash boiler/' 
would appear to be the most plausible ex¬ 
planation of the explosions. Any explana¬ 
tion of the explosions must consider the 
periods of quiet between explosion and 
series of explosive bursts (Jaggar, 1924: 
44), often less than a minute apart. 
The rapid sinking of the lava column 
from a depth below the Halemaumau rim of 
about 500 feet on May 1 to a depth exceed¬ 
ing 1,300 feet in the middle of the explo¬ 
sive period, as well as the increase in diam¬ 
eter of the crater from 1,500 to 3,000 feet 
just prior to and during the explosions, indi¬ 
cates a progressive rupturing of the conduit 
walls. A common interval between the ex¬ 
plosions was 6 to 8 hours. Finch (1943: 
1-3) has suggested that intermittent ruptur¬ 
ing of the conduit walls and the resulting in¬ 
troduction of water may have been influ¬ 
enced by surging in the retreating lava 
column. 
The lack of chlorides in the material ejected 
indicates that the steam came from fresh 
ground water. Stearns (1946: 44) has sug¬ 
gested that the explosions may have been 
caused by ground water trapped between 
dikes at relatively high levels. G. A. Macdon¬ 
ald of the U. S. Geological Survey has sug¬ 
gested (oral statement, 1943) that ground 
water confined between dikes may stand as 
much as 1,500 feet above sea level in the vici¬ 
nity of Kilauea caldera. The lag in time, 
often 2 or 3 minutes, between the premoni¬ 
tory symptoms of an explosion as shown by 
peculiar earthquakes both recorded and felt 
and the appearance of the explosion cloud in 
Halemaumau indicates that the seat of the 
explosions may well have been but compara¬ 
tively little above sea level—say, 500 feet. Es¬ 
timates of the depth of engulfment (Jaggar, 
1924: 118) make 500 feet above sea level 
seem a more probable seat of the explosions 
than 1,500 feet. The relation between the 
amount of material engulfed and the diam¬ 
eter of the conduit indicated a withdrawal 
to at least 300 feet below sea level. With 
the greater depth, more and larger trapped 
pockets of water might be expected. At or 
slightly above sea level the large body of 
basal ground water would be available. 
However, some of the first small explosions 
may have been due to water trapped about 
as high as Macdonald suggests. The retreat 
of the lava column and progressive engulf¬ 
ment indicate the possibility of a consider¬ 
able vertical range in the seat of the explo¬ 
sions. 
Collapsing in Halemaumau and harmonic 
tremor (Finch, 1943: 1-2) indicated that 
there was a retreat of the live lava column 
both before and during the explosions. A 
retreat of the lava column would leave the 
conduit walls and much of the material 
within at a temperature of about 1700° F. 
