284 
Cemeteries close to the epicenter showed less 
consistency in the direction of rotation than 
did those farther away. 
In Figure 12 the prevalent directions of 
rotation of monuments in four cemeteries 
are shown. Four other cemeteries were omit- 
ted because no monuments were rotated in 
them, or because the number of rotated mo- 
numents was too small to yield a reliable 
statistical result. At each of the four ceme- 
teries plotted, the boundaries of the octants 
containing the direction toward the epicenter 
are prolonged. In an area largely west of the 
shoreline, from 2 miles south to 2 miles north 
of the approximate trace of the Kealakekua 
fault, three or more of the four significant 
octants overlap, and it is within this area of 
overlap that the epicenter should be situated. 
LOCATION OF THE EPICENTER 
Because of the dismantling of all but one 
of the seismographs on the island of Hawaii 
during the preliminary phase of the earth- 
quake, it is not possible to locate the origin 
or epicenter instrumentally. The only instru- 
mental datum available is the S — P interval 
of 9.5 seconds given by the Bosch-Omori 
seismograph at the northeast rim of Kilauea 
Caldera (Fig. 13). Using the travel times 
given by Byerly (1942: 210), this gives a 
distance of origin of the earthquake of ap- 
proximately 47 miles from the Bosch-Omori 
instrument. These curves were derived for 
sedimentary and granitic rocks but, over a 
period of several years of use at the Volcano 
Observatory, have yielded more satisfactory 
and reasonable earthquake locations than any 
others. The use of Jones’s (1935: 50) curve 
for duration of the preliminary waves (T*) 
increases the distance to only 49 miles. Taking 
into consideration the area of greatest inten- 
sity of the earthquake, these distances place 
the origin of the quake 3 to 5 miles west of 
the coastline in the vicinity of Napoopoo. 
The depth of origin appears probably to 
have been between 5 and 10 miles. 
Some information bearing on the location 
PACIFIC SCIENCE, -Vol. VI, October, 1952 
of the epicenter can be derived from the 
study of damage by the earthquake. The 
general distribution of damage to roads, 
stone walls, and road cuts along the main 
highway is shown in Figure 3. This is based 
on a count checked against odometer mile- 
age, assigning one unit of damage for each 
1 to 15 feet of collapsed wall or road cut. 
Despite irregularities, the graph shows a dis- 
tinctly symmetrical, bell-shaped distribution 
curve, with its peak about 2.5 miles by road 
southeast of Captain Cook. An average of 
more than 60 items of damage per mile in 
the central 5 miles decreases to only one or 
two per mile more than 9 miles from the cen- 
ter. This point of maximum damage coincides 
closely with the position of the buried inland 
extension of the Kealakekua fault. Other 
types of damage also were most abundant in 
the same general area. Together with the fact 
that most of the aftershocks, located by in- 
strumental means, originated on the Keala- 
kekua fault, it leaves little question that the 
origin of the major earthquake lay on or 
close to this fault, and that the earthquake 
almost certainly resulted from movement on 
it. 
The greatest structural damage was farther 
south, at Hookena, where the destruction of 
the east and west walls of the two stone 
churches suggests an epicenter somewhat far- 
ther south. The possibility of a twin earth- 
quake with one epicenter lying offshore 
nearly west of Hookena has been considered, 
but no other evidence suggests it, and no 
signs of a second earthquake could be de- 
tected from the seismograms either from the 
island of Hawaii stations or from that of the 
Coast and Geodetic Survey at Barbers Point 
on Oahu. 
Throughout the Kona area, the prevalent 
direction of fall of rock slides, stone walls, 
and tombstones was westward, and the next 
commonest direction was eastward. The seis- 
mograms indicate that the first movement of 
the ground was eastward, and it is probable 
that some of the westward fall of objects was 
