intervals of a furlong or so a veritable giant towered up. Only
one of these monstrous trees can grow within a given space; it
monopolises the domain, and none but individuals of much inferior
size can find a footing near it. The cylindrical trunks of these
larger trees were generally about twenty to twenty-five feet in
circumference. Von Martius mentions having measured trees in the
Para district belonging to various species (Symphonia coccinea,
Lecythis sp. and Crataeva Tapia), which were fifty to sixty feet
in girth at the point where they become cylindrical. The height
of the vast column-like stems could not be less than 100 feet
from the ground to their lowest branch. Mr. Leavens, at the
sawmills, told me they frequently squared logs for sawing a
hundred feet long, of the Pao d'Arco and the Massaranduba. The
total height of these trees, stem and crown together, may be
estimated at from 180 to 200 feet; where one of them stands, the
vast dome of foliage rises above the other forest trees as a
domed cathedral does above the other buildings in a city.

A very remarkable feature in these trees is the growth of
buttress-shaped projections around the lower part of their stems.
The spaces between these buttresses, which are generally thin
walls of wood, form spacious chambers, and may be compared to
stalls in a stable; some of them are large enough to hold a half-
dozen persons. The purpose of these structures is as obvious, at
the first glance, as that of the similar props of brickwork which
support a high wall. They are not peculiar to one species, but
are common to most of the larger forest trees. Their nature and
manner of growth are explained when a series of young trees of
different ages is examined. It is then seen that they are the
roots which have raised themselves ridge-like out of the earth;