Hawaiian Calderas — MACDONALD 
329 
Reynolds’ (1956) suggestion that the Scot- 
tish cauldron subsidences and the calderas that 
presumably lay above them were formed by 
the gas-coring mechanism suggested by Escher 
(1929), with "fluidization,” intrusion, and ejec- 
tion of an ignimbritic solid-gas emulsion, can 
have no bearing on Hawaiian calderas because 
of the complete absence in Hawaii of ignim- 
britic material and the extreme paucity of frag- 
mental explosive material of any sort. 
Before considering further these older hy- 
potheses, let us look briefly at a new one. The 
very high density of the material in pipe-like 
masses beneath the summit areas of Hawaiian 
volcanoes has led to the suggestion that the 
formation of a caldera might be the result of 
the isostatic sinking of the heavy column, car- 
rying the overlying mountain top down with it. 
In this connection it is necessary to consider 
the gravity field found by Kinoshita and others 
(1963) on the island of Hawaii. Their Figure 
1 shows that the gravity high for Mauna Loa 
(a Bouguer anomaly reaching between 330 
and 340 mgal) is offset several kilometers to 
the southeast of Mokuaweoweo caldera, farther 
to one side than is the presumed top of the 
dense material below the surface. There is no 
sign whatever of sinking of the mountain sur- 
face above the center of the gravity high. The 
high for Kilauea (reaching about 315 mgal) 
also is excentric to the caldera, the center of 
the high lying some 2 km or more to the south- 
west of the center of the caldera. Thus, it ap- 
pears unlikely that the caldera formation can 
have resulted simply from isostatic sinking, 
unless the improbable mechanism of a highly 
oblique subsidence is invoked. Furthermore, no 
discernible sinking has disrupted the post- 
caldera cap on volcanoes such as Mauna Kea, 
beneath which a markedly high gravity anom- 
aly still exists (Kinoshita et al., 1963). Iso- 
static sinking in the ordinary sense, therefore, 
appears improbable as an explanation for the 
Hawaiian calderas. 
The high-density column beneath the vol- 
canoes may have another effect, however. Ris- 
ing, as it apparently does, some 5 km or more 
above the base of the volcanic mountain, the 
base of the mass must have a considerable 
tendency to spread, and must exert a consid- 
erable lateral thrust on the lighter material 
adjacent to it. This must be particularly true 
when the mass is still somewhat mushy. Does 
the tendency for the heavy mass to spread result 
in spasmodic lateral movements of its lower 
part into the proximal ends of the rift zones, 
causing a wedging open of the rift zone and a 
distension of the volcanic edifice? 
The principal reason given by Stearns and 
Macdonald (1946:29-33) for the rejection of 
Williams’ proposed mechanism for formation 
of calderas of Kilauean type was the fact that 
flank eruptions with voluminous drainage of 
magma are frequent throughout the period of 
building of the shield, whereas the formation 
of the caldera takes place only near the end of 
it. Actually, however, if the formation of the 
caldera depends on the existence of a magma 
chamber in the core of the volcano, the absence 
of such a chamber in the earlier stages would 
account for the absence of a caldera. Consider- 
able time must be necessary for enlargement of 
the chamber to the point where its roof is too 
broad and thin to support itself. 
A more conclusive argument can be made 
against the application of the Glen Coe mech- 
anism to Hawaiian calderas: namely, the fact 
that the sinking caldera block must displace an 
equal volume of magma. Where does this 
magma go? Does it rise into the ring fractures 
around the sinking block, as in the classical 
interpretation of the cauldron subsidence of 
Glen Coe (Clough, Maufe, and Bailey, 1909: 
Fig. 14)? There is no evidence to suggest it. 
Dikes are nearly absent along the caldera 
boundaries at levels exposed by erosion. Erup- 
tive vents on the caldera-boundary fractures are 
very rare, and the few that are found appear to 
have no fundamental relationship to the frac- 
ture. Thus, although the main cone of the 1949 
eruption of Mauna Loa lies on the caldera 
boundary, the eruptive fissure was not the 
caldera-boundary fissure, but one that extended 
across the center of the sunken caldera block, 
up over the caldera wall, and several kilometers 
down the flank of the mountain. But if the dis- 
placed magma does not rise around the sinking 
block, where does it go? Out into the rift zones? 
This brings us right back to Williams’ Kilau- 
ean mechanism! 
