Barriers Against Lava — Macdonald 
271 
portant in the operation of lava barriers in 
greatly reducing the hydrostatic pressure ex- 
erted against a barrier. Actually, the thrust 
against a barrier as a result of hydrostatic pres- 
sure is only a small fraction of what it would 
be if the lava were a completely liquid pool 
with the fluidity of water. 
The ability of even loose stone walls to 
withstand the pressure of flows indicates that 
the full theoretical amount of hydrostatic 
pressure is not exerted laterally by the flow. 
Calculations indicate that with fully liquid 
lava resting against a wall of loose rock, slid- 
ing of the wall would result when the depth 
of the liquid against the wall slightly exceeded 
the thickness of the wall. Commonly, how- 
ever, a lava flow piles up behind a wall to a 
depth several times as great as the thickness of 
the wall without displacing the wall. Appar- 
ently the departure of the fluid lava from 
complete liquidity is sufficiently great to pre- 
vent the full theoretical hydrostatic pressure 
within the flow from being transmitted to the 
forward edge. This is further confirmed by the 
frequently observed tendency for a flow to 
stop with only its lowermost edge in contact 
with some natural obstacle, such as a crater 
wall, leaving a moat a few feet wide between 
the obstacle and the higher part of the flow 
margin. 
FACTORS INFLUENCING EFFECTIVENESS 
OF BARRIERS 
The tendency of lava to build up its channel 
to a high level is important to the operation 
of lava barriers in two respects. One is the pos- 
sibility that the flow may build up so high as 
to spill over the barrier. There is little danger 
of this if the angle of the barrier to the flow 
course is not too great — that is, if the barrier 
does not force the flow to turn too sharply. 
A little spill-over may be expected in any case, 
but is unimportant if most of the flow turns 
and follows the barrier. Experience at the old 
railroad embankment near Kapoho and at the 
Yamada barriers, in 1955, clearly indicates that 
the lower part of the flow largely controls the 
direction of movement of the whole flow. A 
well-placed barrier can be confidently expected 
to turn the initial flow of a group, even though 
it is considerably thicker than the barrier is 
high. Once the flow is turned, the main chan- 
nel will develop parallel to the barrier, but 
probably several tens of feet distant from it 
because of the cooling effect of the barrier and 
frictional retardation of the edge of the flow 
against the barrier. 
If the flow continues for a long period, the 
walls confining the main channel may build 
up to form natural levees rising to a level 
higher than the barrier. A breakdown of the 
levee could then release a flood of lava over 
the barrier, possibly establishing a new flow 
course over the barrier in addition to, or even 
instead of, that parallel to the barrier. Such 
breakdowns and lateral floodings are common 
near the vents, especially on steep slopes and 
where the channel makes an abrupt bend, but 
they are very rare on well-established flows at 
a distance from the vents. Provided the angle 
of the barrier to the natural flow course is kept 
small, the danger of such a breakdown of the 
channel levee at a barrier distant from the 
vents is very small. 
More probable is a breakdown of the levee 
near the vents, far up slope from the barrier, 
producing a new major tongue of the flow. 
In early stages of eruptions this is a common 
event, and it sometimes occurs even in late 
stages. It may pose by far the greatest threat to 
the success of a lava barrier. If the new flow 
tongue encounters the barrier on the upslope 
side of the older tongue, which is already 
against the barrier, it may be impounded be- 
tween the barrier and the older tongue, accu- 
mulate until it overtops the barrier, and flow on 
down the mountainside. The effectiveness of 
the barrier is then partly or wholly lost (al- 
though it may continue to divert the first 
tongue and thus reduce the amount of lava 
advancing toward the area under protection). 
Fortunately, it is rare that more than one flow 
tongue reaches a distance from the vents as 
great (12 or more miles) as that of the pro- 
