324 
PACIFIC SCIENCE, Vol XIX, July 1965 
tering at the Kawainui Swamp near Kailua 
(Stearns, 1939: PI. 1). That of east Molokai 
was about 7.2 km long and 5 km wide (Stearns, 
and Macdonald, 1947:19, Ph 1). The larg- 
est of the Hawaiian calderas was that of 
Kauai, some 16 by 20 km across (Stearns, 
1946: Fig. 22; Macdonald, Davis, and Cox, 
I960: Pi. 1). In and close to the east Molokai 
and Koolau calderas, and locally in that of West 
Maui and near the boundary of that of Kauai, 
rising gases brought about chloritization of the 
rocks and deposited secondary minerals, includ- 
ing quartz and chalcedony. The boundary cliffs 
of the calderas, as delineated by mapping, sloped 
inward at angles of 50° to nearly 90°, averaging 
about 75°. The slope of the boundary cliffs is 
not, of course, necessarily the same as that of 
the faults that bound the sunken block at 
depth. Outward-dipping faults would produce 
surficial scarps that are unstable, and slumping 
would quickly form inward-sloping fault-line 
scarps. However, other evidence discussed on a 
later page also suggests that the subsurface 
faults dip inward. 
Nowhere has the bottom of the caldera- 
filling mass been reached by erosion. The sink- 
ing of the original tops of the east Molokai 
and Kauai shields exceeded 1,400 m. 
There is considerable evidence that Mokua- 
weoweo caldera grew in part by the repeated 
inclusion of marginal pit craters (Stearns and 
Clark, 1930:49; Stearns and Macdonald, 1946: 
29). On the floor of Kilauea caldera also, the 
distribution of areas of alteration of the rocks 
by rising gases indicates the presence of a 
series of pits buried by the lavas erupted within 
the last 150 years (Macdonald, 1955*0 . The 
Kilauea Iki and Keanakakoi pit craters nearly 
coalesce with the inner caldera of Kilauea, and 
the outermost caldera faults are beyond them, 
so that they are actually within the major 
sunken area. The pit craters probably have 
formed by sinking of a roof block into an 
underlying magma chamber eaten upward into 
the mass of the volcano by s toping and melting 
of the lavas above the magma body (Macdon- 
ald, 1956:281), but since the sinking is not 
accompanied by a rise of magma along the 
bounding ring fractures into the crater the 
magma must have been withdrawn, into adja- 
cent rift fractures or elsewhere, to make room 
for the sinking block. 
RELATION OF CALDERA FORMATION TO 
VOLCANIC AND MAGMATIC HISTORY 
As a result of his years of study of the areal 
geology of the Hawaiian Islands, H. T. Stearns 
was able to distinguish a series of stages in the 
history of the volcanoes (Stearns, 1940; 1946: 
17-19). These may be summarized briefly as 
follows, with some additions and modifications 
from Stearns’ original statements: 
1. A youthful shield-building stage, during 
which frequent eruptions of very fluid basaltic 
lava build a shield volcano composed almost 
wholly of thin extensive lava flows. Eruptions 
come so frequently that there is not time for 
any appreciable amount of weathering or ero- 
sion between successive flows. Pyroclastic ma- 
terial forms probably less than 1% of the part 
of the shield above sea level. Hydromagmatic 
explosions may have produced a much larger 
proportion of pyroclastics in the zone within 
a few hundred feet below sea level, but at 
greater depths the pressure of overlying water 
probably effectively restrained both explosion 
and vesiculation, and the lava flows were prob- 
ably much denser than those later poured out 
above sea level, as are those formed in deep 
water on the east rift zone of Kilauea (Moore, 
1965). 
2. Late in the period of shield building the 
summit of the shield collapsed to form the 
caldera. There followed the so-called caldera- 
filling period, during which thick massive lava 
flows accumulated within the caldera, gradually 
filling it. It should be noted, however, that 
eruption is not restricted to the caldera, but 
continues to take place on the flanks of the 
volcano also, and the shield continues to build. 
Volcanism continues vigorous and eruptions 
frequent; the activity remains almost purely 
effusive and pyroclastic material still is formed 
in only very small amount. 
3. Eventually the caldera becomes filled and 
the volcano enters on the post-caldera, or old- 
age, stage. Eruptions become less frequent, and 
many of them are more explosive, partly be- 
cause of greater viscosity of the erupting magma 
but partly also because of a greater abundance 
