264 METROSIDEROS ROBUSTA. 
with plant-life which is more directly suggestive of intelligence. When two or 
more root-stems are developed, they are frequently connected by laterals, and 
in many cases they give off oblique laterals, which usually become united by a 
process of natural grafting, and the supporting tree is sooner or later killed by 
their iron embrace. | w 
When several roots are given off they ultimately become inosculated, and 
constitute an uneven gigantic cylinder, but never form a solid trunk as is com- 
monly supposed. On felling a tree of this kind the form of the original aérial 
roots can be made out; the enormous pressure has forced the central column of 
pith much closer to the side in contact with the supporting tree, and distorted 
the stem itself. It is often found that a second series of root-stems is given off 
from giant specimens, and overlies the original series: this is only seen in very 
old specimens, and may arise from the original root-stems being unable to obtain 
sufficient nourishment for the increased bulk. This phenomenon is chiefly 
exhibited by gigantic specimens with large heads: a grand specimen on Mount 
Egmont has given off a secondary root-stem above 3ft. in diameter, which is 
nowhere in contact with the old trunk, although parallel with it for its entire 
length. 
Large stems are only produced by single roots, and in some places they are 
so abundant that it is evident a remarkable form of replacement has been 
effected » in these localities the forest consists almost exclusively of ratas which 
commenced life high up on the stems of the kahikatea, tawa, pukatea, or other 
trees which formerly occupied the site, but have passed away, their places being 
taken by the slow-growing but more gigantic ratas, which germinated not on the 
ground, but on the trees which they have replaced. In localities where the rata 
is abundant it occasionally forms a natural arch, raft. to 2oft. high, and from 
Toit. to 18ft. wide at the base, crowned with the erect stem and spreading head, 
and affording ample space for a buggy to be driven underneath. This peculiar 
growth has evidently been caused by the bifurcation of the descending root, 
and the subsequent divergence of the limbs. 
There is one tree, and only one, on which the rata can make no impression 
—-the puriri, Vitex littoralis (P|. CV.): its iron grain sets the epiphyte at defiance; 
slowly but surely it forces the epiphytic root-stems apart, and sometimes dis- 
lodges the trespasser entirely. Instances may occasionally be seen in which a 
purin is growing between three or four large root-stems of rata which it has 
gradually forced apart by its tremendous energy. 
It has been asserted that the northern rata must invariably originate as an 
epiphyte; but there is no real foundation for the statement, as terrestrial speci- 
mens are often met with, but they never attain such large dimensions as those 
which begin life as epiphytes. Terrestrial specimens give off several ascending 
stems, and many years elapse before one of them takes a decided lead: 
numerous cultivated specimens from twelve to twenty years old may be seen, 
which are little more than bushy shrubs, having made less than rft. of growth 
for every year of their existence. A plant with this peculiar habit, and with such 
a slow rate of growth, could never attain large dimensions in an ordinary forest 
unless able to obtain some advantage over other trees in early life: this is gained 
by its epiphytic habit, which affords ample compensation. 
It stands alone amongst Australasian forest-trees in developing stems of 
large bulk, affording durable timber, from aérial roots. 
In very rare cases the same remarkable habit is exhibited by the pohutu- 
kawa (Metrosideros tomentosa) and by the southern rata (M. lucida) when growing 
on the margin of cliffs, but specimens of this kind are extremely rare. | 
