258 HAWAIIAN ISLANDS. 
twin of volcanic summits like the peninsulas of Maui, or like Loa and 
Kea on the island of Hawaii. 
The question next arises, where were the centres of these moun- 
tains situated ’—what remains can be found of the great craters ?— 
How was the volcanic dome changed into a narrow mountain ridge? 
Whether these points admit of explanation or not, the fact that such 
summits existed, cannot be doubted. Examples have so frequently 
been observed of volcanic summits without distinct craters, owing to 
the peculiarity of the closing eruption, or its final phase, that they are 
no longer considered a necessary feature. Moreover, the degradation 
which has made the many valleys, each vastly more extensive than 
the great pit of Mount Loa, was sufficient to obliterate any trace of a 
mountain crater. But we are led by the facts to another important 
consideration bearing upon this point. 
The sloping layers of basaltic lava, whose accumulation raised the 
eastern mountains, rise gently from the south and west, and termi- 
nate abruptly in the face of the great Pali or precipice. Instead of a 
mountain, with opposite declivities on opposite sides, as in Kea, Loa, 
and the sammits of Maui, there are only the declivities of one side; 
these are cut short by a section upwards of twenty miles long, which 
is at the present time, notwithstanding all the extensive degradation 
that the valleys indicate, from two to four thousand feet in height, 
—a mountain wall of an extent seldom witnessed, even among the 
more famous Alps and Andes. Can this wall have been a part of the 
outline of the great crater? Its twenty miles of length, with so small a 
curve, would indicate that the crater had been more than sixty miles 
in circuit. We have no warrant for such an extravagant conjecture 
in any observed facts. 
The only plausible hypothesis, the reader has probably anticipated : 
that this wall was the course of a rupture in the former volcanic cone, 
—a fissure along which the mountain was rent asunder. A section 
through Mount Loa, across its southern slopes, in the same direction, 
would give us almost a fac simile of the effect here supposed to have 
taken place. The plains at the foot of the precipice have been de- 
scribed as very narrow; and Kaneohe Bay is a deep indentation in 
the coast, situated not far from the axis of the original vent. 
With regard to the extent of the whole dome before the rupture took 
place, we may obtain some data from the size of the sezment remain- 
ing, and by comparison with Loa, Kea, and Hale-a-kala. It might be 
thence inferred that the fissure passed through the cone two-thirds of 
