ving shot out
from its perpendicular. As these roots increased, they sent out side
rootlets, which, still running on the face of the wall, by and by came
in contact with another of the primary roots. Then, instead of creeping
_over_ it, as the roots of other trees would have done, the soft tip of
the rootlet actually united with the substance of the root at the point
of contact, the fibres of the two becoming interlaced, and their united
surfaces gradually becoming covered with a common bark. The repetition
of this process had produced the very curious wooden net which I have
attempted to describe.

A still more remarkable example of this parasitic mode of growth I have
seen in the same island. By the side of a mountain road was a large
fig-tree, the base of whose trunk was about thirty feet from the
ground. Thence it reared itself up pillar-like towards the heavens, and
spread abroad its vast horizontal array of branches across the road.
From the same point there descended to the earth a hollow cone of roots,
interwoven and anastomosed, especially at the upper parts, in the same
manner as those of the boiling-house wall, but forming towards the
bottom only three or four flattened irregular columns. Into the area
inclosed by this network of roots a person might enter, for it was about
six feet wide, and, looking up, behold the base of the trunk eight or
ten yards above his head.

The explanation of this curious phenomenon depends upon the tendency
just mentioned. On this site once stood a large tree of some other
species, probably a cotton-tree (_Eriodendron_), or some other
soft-timbered kind. The little scarlet berry of a Fig-tree was carried
by some vagrant Banana-bird or Pigeon to its boughs, and there devoured.
After the little truant had finished his morsel, he perhaps wiped his
beak against the rough bark of the trunk, beside the branch on which he
was seated. Some of the minute seeds, enveloped in mucilage, were thus
left on the tree, which the rain presently washed down into the broad
concavity of the forks, where, among moss and rotten leaves, it soon
germinated and grew. The roots gradually crept down the trunk of the
supporting tree, closely clinging to its bark, and by their
interlacement at length formed a living case, enveloping it on every
side, and penetrating the earth around its base. The growth of these,
and also of the inclosed tree, daily induced a tighter and tighter
pressure on the latter, whic