388 
PACIFIC SCIENCE, Vol. XIX, July 1965 
rift zone were studied. Rough calculations 
showed that, by assuming somewhat different 
geometric shapes and depths of burial and mak- 
ing slight changes in densities of less than 0.1 
g/cc, density models which would have the 
same basic structure as those derived seismi- 
cally for the Koolau caldera could be con- 
structed that would satisfy the gravity field of 
the other intrusive complexes. The final model 
chosen to represent the volcanic pipe associated 
with the Koolau caldera is illustrated in Figure 
5. This cross-sectional view implies a symmetri- 
cal model made up of cylinders or vertical 
prisms. Actually, the model used was not quite 
symmetrical. The model consisted of a number 
of vertical prisms 0.5 X 0.5 km in horizontal 
dimensions, with the gravity effect of each 
prism computed assuming that its mass was 
concentrated along a line element at its center. 
The cross-sectional shape of the volcanic pipe 
was chosen to conform in its upper parts to 
geologic knowledge of the dimensions of the 
caldera and to the seismic information. Because 
of the complications due to the gravitational 
effects of the rift zones to the north and east, 
Fig. 5. Density model for the Koolau volcano, 
along line B-B' of Figure 1. 
this model Is chosen essentially to fit gravity 
profiles taken from the center of the caldera 
in a western or southern direction and corre- 
sponds to the line of seismic measurements 
reported by Adams and Furumoto (p. 296 in 
this issue). As indicated, it assumes the same 
densities external to the pipe as those assumed 
for the Ridge as a whole and illustrated in 
Figure 4. 
INTERPRETATION OF RESULTS 
The composite section across the Hawaiian 
Swell based on seismic and gravity evidence is 
shown in Figure 4. This section seems compati- 
ble with most of the available evidence on the 
Hawaiian Ridge as well as on other oceanic 
areas. 
As indicated by the gravity map in Figure 1, 
and as shown even more convincingly by the 
magnetic work of Malahoff and Woollard (in 
a forthcoming issue of Pacific Science ), the lavas 
which built the Hawaiian Islands were extruded 
primarily along faults oriented either east-west 
and associated with the Molokai fracture system 
or northwest-southeast and associated with the 
trend along which He the Koolau dike complex 
and the Musician Seamounts. The volcanic pipes 
may have formed at points of intersection of 
rifts of the two fracture systems. After an appre- 
ciable build-up of extrusive material, the weight 
of the extruded lavas caused the ridge to sink 
in order to re-establish isostatic equilibrium. 
Since the depth to Moho is about 15 km on 
the Ridge and about 11 km in the normal 
ocean basin, a thickening of the crust of some 
4 km is indicated. Whether or not this also 
represents the degree of crustal subsidence is 
not definite as yet. The material with velocities 
in the range 2. 5-4.0 km/sec and assumed den- 
sities of less than 2.6 v g/cc is from 2 to 3 km 
thick and, judging from the work of Moore 
(unpublished), must have been erupted above 
or near sea level. These data, therefore, suggest 
roughly 2-3 km of subsidence. How much sub- 
sidence occurred prior to the build-up of the 
Ridge to sea level can only be surmised. It 
does appear significant, though, that the upper 
face of the basal crustal layer also indicates 
only 2-3 km of subsidence. 
