220 HAWATIAN ISLANDS. 
to me at the time that this rounded, well-shaped cavity might be the 
site of a former boiling pool, like the small one then near me. It 
extended from below the upper layers of the black ledge to a depth of 
about one hundred and fifty or two hundred feet, and if once actually 
a boiling cavity, it was probably so just previous to the eruption 
of 1840, when the black ledge was overflowed by the lavas. This 
is mentioned as a probable case favouring the view that small pools 
may contract below. This fact, however, hardly authorizes us to 
infer that the same basin character may be the condition of the main 
pools of the craters. 
If we admit a union below, (which may be considered somewhat 
doubtful,) the waters that gain access to the fires, and promote the 
action of the volcano, must operate on the conduits above the point of 
bifurcation, or certainly there would be a sympathy in their move-. 
ments. ‘The same is true of the conduits to separate pools in Kilauea, 
which also act in some degree independently. ‘The water, in sucha 
case, would necessarily tend to escape by ascending, and so exert its 
action upwards, inflating the lavas more and more as it approached 
the upper surface and found diminished pressure. Its influence, i 
giving activity to the fires, would thus gradually increase, and would 
be greatest above, where the pressure counteracts, in a less and less 
degree, its mechanical as well as any chemical effects. ‘The lavas 
above, thus inflated by vapours, will therefore be lhghter than those 
below. In order that the two legs of the syphon should not vibrate 
together, it is necessary that the point of junction of the legs should 
be at so great a depth that the difference of length between the legs 
would be but afraction of their whole length. In this case, the difference 
of height might be as great even as between Kilauea and Mokua-weo- 
weo, and still the two act without any mutual influence. From the 
sketch given on the preceding page, it is evident that this union must 
take place, if at all, at a great distance below the surface; and at any 
probable point of bifurcation, the actual difference of length 
will in fact be relatively small. If, in the annexed figure, 
K and S be the two craters (Kilauea, and that at the 
summit) and the lines Kc, Sc, represent the convergence of 
the conduits, the point c will be forty-four miles below the 
surface, and Sf the difference in the height of the two 
craters above the sea, will be about one twenty-third Sc; 
orif c’ be the point of convergence, the distance Sc is thirty- 
three miles, and one leg of the syphon is one-eighteenth 
