NOTES 
231 
of their movement in the eroded topography. 
With the progress of erosion there is simultane- 
ous sedimentation, with the deposition of both 
terrestrial and marine sediments along the shores 
of the island and in valleys submerged by chang- 
ing relative sea levels. 
The precipitation which falls on these vol- 
canic domes is divided into the usual three parts, 
evapo-transpiration, runoff, and infiltration. Be- 
cause of the high average permeability of the 
lava flows making up the bulk of the domes, the 
ratio of infiltration to rainfall is relatively high. 
The water descending through the rocks may 
accumulate in three principal types of ground- 
water bodies: (l) bodies perched on ash beds or 
soils interbedded with the flows, on conglom- 
erates, on unconformities, or on other rocks of 
low dip less permeable than the average lava flows ; 
(2) bodies impounded between the dikes that 
have intruded the lava flows; and (3) bodies 
floating on salt water. The occurrence of these 
bodies is diagrammed in Figure 2. 
The perched and dike springs representing 
the natural discharge from the first two types 
of ground-water bodies are of great importance 
in providing the largest part of the low water 
flow of most Hawaiian streams which, without 
them, would be almost without value as sources 
of water supply. Partly through wildcat experi- 
ment and partly through knowledge of the 
causes of retention at high levels, the natural 
surface discharge from many of these perched 
and dike-retained bodies has been artificially 
increased, steadied, or diverted to economic 
advantage. 
The qualitative relationship between the 
perched ground water and the ash, soil, and 
conglomerate beds upon which it is held was 
first clearly understood in 1920 through the 
work of W. O. Clark (Stearns and Clark, 1930, 
U. S. Geol. Survey, Water-Supply Paper 516) 
who spent many years thereafter planning and 
guiding development of high-level water of this 
sort for the sugar plantations and other agencies 
of the Islands. Further development is still in 
progress and still requires detailed geologic and 
hydrologic advice. An attempt to generalize 
quantitatively on the results expectable from 
this type of development would be useful in 
engineering future developments but would be 
difficult because of the high variability of the 
perching members in their area, thickness, con- 
tinuity, permeability, and slope, and because 
no adequate theory exists covering nonsteady 
flow conditions in this kind of body. The de- 
velopment of such theory would seem to be 
easier in continental areas where experimental 
work could be done in perched bodies with less 
variability. 
The understanding of the relation between 
dikes and the high-level water impounded be- 
tween them seems to have developed gradually 
as a result of the construction of a number of 
tunnels, penetrating saturated compartments be- 
tween dikes, notably the Waiahole tunnel that 
was driven through the Koolau Range of Oahu 
to transport stream water from the windward to 
the leeward side. The gross effects of these tun- 
nels in draining off water stored between the 
dikes were early noted, but the length of time 
necessary to the restoration of equilibrium con- 
ditions was not recognized for many years. With 
it has come a realization that many dike-devel- 
opment tunnels do not have a permanent flow 
appreciably greater than that available from the 
springs representing the original drainage of the 
dike compartments. A considerable volume of 
records of the yield of dike tunnels, some of it 
coupled with the records of nearby spring flow 
and of pressures encountered behind dikes in 
tunneling, awaits analysis and correlation with 
rainfall records and studies of dike distribution. 
The results will be of great local value in the 
study of future tunnel results and also in the 
study of bulkheading to restore the storage 
capacities behind selected dikes, a practice that 
has been carried on to some extent in recent 
years but which has not been thoroughly 
analyzed. 
The largest draft of ground water comes from 
basal bodies, generally floating on salt water, 
fed by direct recharge from the surface or by 
underground leakage from perched-water bodies 
or bodies impounded by dikes. Where the per- 
vious lava flows containing this basal water 
extend to the coast without cover, the head of 
fresh water above sea level is small, increasing 
at the approximate rate of a foot per mile from 
the coast. Where, however, the seepage of fresh 
water to the ocean is restricted by a "cap” of 
less pervious sediments, as on much of Oahu, 
the head under and immediately behind the cap 
may be several tens of feet, higher in fact than 
the surface of the sediments in the belt close to 
the shore, thus creating artesian conditions. No 
lower limits to the zone of permeable lava flows 
are known or suspected within a range of several 
thousand feet below sea level for most of the 
basal aquifers. 
The possibility of contamination of this basal 
fresh water by salt water was early recognized 
through experience and ascribed to the prox- 
