326 
PACIFIC SCIENCE, VoL XIX, July 1965 
magma body that fed the surface eruptions. 
Knowledge of the position and nature of the 
magma chamber is almost wholly implied from 
seismic evidence and the pattern of swelling 
and shrinking of the volcano before and after 
eruptions. Earthquakes originating from a 
depth of about 60 km below sea level, within 
the upper part of the earth’s mantle, are ac- 
companied by "harmonic” tremor that appears 
to be the same as the tremor that is known to 
accompany movement of magma in the volcanic 
conduits close to the surface. From this it is 
implied that the magma that feeds the volcano 
is formed at a depth of about 60 km (Eaton 
and Murata, I960). It has long been known 
that Kilauea volcano gradually tumesces over 
periods ranging from a few weeks to several 
years before eruptions (Jaggar and Finch, 
1929), presumably because of inflation of an 
underlying magma reservoir, and detumesces 
during and after the release of magma by erup- 
tion. Recent analyses of the patterns of ground 
tilting resulting from the tumescence and de- 
tumescence, coupled with the depth distribution 
of the shallow-seated earthquakes that precede 
eruptions, have led to the conclusion that the 
top of the magma body lies at a depth of only 
about 2 km below the summit of Kilauea vol- 
cano (Eaton and Murata, I960). It should be 
emphasized that this is well above the level of 
the surrounding ocean floor and consequently 
well within the mass of the volcanic mountain 
itself. Room for the magma body must have 
been obtained either by lateral displacement of 
its margins or by melting of the rocks that for- 
merly constituted the core of the volcano. 
The lateral extent of the magma body is less 
certain, and indeed it appears doubtful (from 
inductive reasoning) that it has any sharply 
defined margin. More likely there is a gradual 
passage from freely fluid magma to somewhat 
plastic but essentially solid material, and thence 
to the truly solid rocks of the volcano’s exterior. 
There can be no question that the magmatic 
core, or at least material in a very mobile con- 
dition, extends outward from beneath the sum- 
mit of the mountain for many kilometers into 
the rift zones. Thus, the 1955 eruption of Ki- 
lauea from the east rift zone 30 km east of the 
caldera was preceded by a marked swelling of 
the rift- zone, with lifting of the ground surface 
along the crest of the rift-zone arch in the 
eruption area of more than a foot (Macdonald 
and Eaton, 1964:105). 
There is no real evidence of magma cham- 
bers at shallow depths beneath the other Ha- 
waiian volcanoes. Even at Mauna Loa the meas- 
urements of tilting are far from sufficient to 
demonstrate such a chamber. By analogy with 
Kilauea, however, it may be assumed that such 
magma chambers probably exist, or did exist 
when the volcano was active. The collapse of 
the shield to form the caldera is probably de- 
pendent on the existence of this magma body, 
and the fact that caldera formation does not 
occur until near the end of the shield-building 
stage is probably because sufficient remelting 
of the core of the shield to form the chamber 
has not taken place until that stage. 
During recent years... gravity studies in the 
Hawaiian Islands have indicated very high val- 
ues in and near the areas of some of the 
calderas (Woollard, 1951; Woollard et ah, 1964). 
Strange, Woollard, and Rose (p. 381 in this 
issue) reports that whereas the average Bou- 
guer gravity anomaly value along the Ha- 
waiian ridge is about -[-200 mgal, the in- 
tensity in the central parts of the Koolau and 
Walanae calderas, Oahu, is in excess of 310 
mgal, and "the maximum Bouguer anomaly 
values over most of the volcanic centers range 
between +285 and +325 mgal.” He points 
out, as had Woollard (1951), that to explain 
the anomalies it is necessary to assume a very 
high density for the material in the volcanic 
"pipes.” Similarly, Adams and Furumoto (p. 
29 6 in this issue) and Furumoto, Thompson, 
and Woollard (p. 306 in this issue) find 
seismic velocities in the Koolau "plug” greater 
than 7 km/sec, as compared with 4.6 km/sec 
in surrounding material. Adams estimates the 
plug to be about 6 km across, with Its top 
about 1.6 km below sea level. Similar high 
seismic velocities were found by Shor ( I960') 
approaching close to the surface in the north- 
western part of the Hawaiian ridge, near the 
Gardner Pinnacles. To explain the observed 
pattern of high gravity values combined with 
high seismic velocities, one Is forced to assume 
