154 POLYNESIAN RESEARCHES. 
was collected. The former were small huts, built 
with cocoa-nut leaves, or boughs and green ti 
leaves, which each party or family erected for their 
own accommodation, around that of their chief; 
and thus formed a small encampment by them¬ 
selves. The latter was a large open building, 
constructed with the same materials, in which the 
chief and his warriors all dwelt together. 
Their camp was near an open space, and they 
generally selected the most broken and uneven 
ground, frequently rugged tracts of lava, as their 
fields of battle. Sometimes they encamped on 
the banks of a river, or deep ravine, which, lying 
between them and their enemies, secured them 
from sudden attack. But they do not appear to 
have thrown up lines or other artificial barriers 
around their camp ; they did not, however, neglect 
to station piquets at all the passes by which they 
were likely to be approached. Each party usually 
had a pari or pakaua , natural or artificial fortress, 
where they left their wives and children, and to 
which they fled if vanquished in the field. These 
fortresses were either eminences of difficult ascent, 
and, by walling up the avenues leading to them, 
sometimes rendered inaccessible; or they were 
extensive enclosures, including a cave, or spring, 
or other natural means of sustenance or security. 
The stone walls around the forts were composed 
of large blocks of lava, laid up solid, but without 
cement, sometimes eighteen feet high, and nearly 
twenty feet thick. On the tops of these walls the 
warriors fought with slings and stones, or with 
spears and clubs repelled their assailants. When 
their pari was an eminence, after they had closed 
the avenues, they collected large stones and frag¬ 
ments of rock on the edges of the precipices over- 
