226 POLYNESIAN RESEARCHES. 
which are concave on the upper side, from the 
top of the tree to the lower branches, and thus 
form a kind of spout, along which the rain that 
falls on the tree descends into their calabashes, or 
other vessels, placed underneath these vegetable 
aqueducts for its reception. By this means, during 
a shower, they often procure a tolerable supply. 
After travelling between three and four miles, 
we reached Keapuana, a large cavern, frequently 
used -as a lodging-place by travellers. The sun 
was nearly down, and the guides proposed to halt 
for the night in the cave, rather than proceed any 
further, and sleep in the open air. The proposal 
was agreed to, and when we had gathered a quan¬ 
tity of fern leaves and grass for our bed, and col¬ 
lected some fuel for the evening fire, we descended 
about fourteen feet to the mouth of the cavern, 
which was probably formed in the same manner as 
those we had explored in the vicinity of Kairua. 
The entrance, which was eight feet wide and five 
high, was formed by an arch of ancient lava, 
several feet in thickness. The interior of the 
cavern was about fifty feet square, and the arch, 
that covered it, ten feet high. There was an 
aperture at the northern end, about three feet in 
diameter, occasioned by the falling in of the lava, 
which admitted a current of keen mountain air 
through the whole of the night. While w r e were 
clearing out the small stones between some of the 
blocks of lava that lay scattered around, a large 
fire was kindled near the entrance, which, throw¬ 
ing its glimmering light on the dark volcanic sides 
of the cavern, and illuminating one side of the 
huge masses of lava, exhibited to our view the 
strange features of our apartment, which resembled, 
in no small degree, scenes described in tales of 
