FORESTS OF THE LORD AUCKLAND GROUP. 131 
(Coprosma foetidissima), the inaka (Dracophyllum longifolium), the 
divaricating shrubs Coprosma ciliata, C. parviflora, Suttonia divari- 
cata, and, where the forest changes into scrub, the yellow-leaved 
mountain-cottonwood (Cassinia Vawvilliersii). The extraordinary 
manner of growth of this association; the close, even foliage of its 
roof; the twisted, far-reaching branches, semi-prostrate and arching 
trunks, and consequent lowness of the trees, are plainly the expres- 
sions of the tempestuous climate—rigorous enough in many ways, 
but never really cold. Within the shelter of the forest quite other 
conditions exist, so here flourish those plants that can exist only in 
an atmosphere saturated with moisture. As for the affinities of the 
forest, they are subtropical and not subantarctic. Here, of all places, 
where a southern-beech (Nothofagus) forest might be expected, it is 
absent. Before leaving the forest it must be pointed out that at 
one spot it contains a certain amount of the tree-fern Hemitelia 
Smithii (fig. 86), and that this locality is the most southerly point 
in the world reached by tree-ferns. 
Above the forest-line comes a scrub-association of the most 
extreme density, varying in height from 8 to 3 feet. The dominant 
shrub is Suttonia divaricata. With this is stunted southern rata 
(Metrosideros lucida), the coprosmas already mentioned, together with 
Coprosma cuneata, and stunted inaka (Dracophyllum longifolium). 
It is thus largely forest stunted to scrub. 
In some few parts of these subantarctic islands—namely, on Ewing 
Island of the Lord Auckland Group, to an extremely limited extent 
on the north of Lord Auckland Island itself, and especially on the 
Snares—are small woods of another character. These are composed 
of the truly magnificent tree-daisy (Olearia Lyallii), found only in 
these islands, but closely related to O. Colensoi of Stewart Island, 
the Southern Alps, and the North Island mountains. O. Lyall has 
great leathery leaves, up to 8 inches in length, which are green on 
the upper surface but pure white beneath, thus affording a delightful 
contrast when they are stirred by the wind. Perhaps for this tree- 
community the term “scrub-forest’’ would be more fitting than 
“forest.” Probably this association is a modified remnant of the 
ancient forest of that latitude and farther to the south, which during 
the problematical great expansion southwards of New Zealand in 
later Tertiary times was driven into its present narrow limits by 
the invading and more vigorous rata forest of the north. The 
community of herbaceous plants, too, is possibly to be similarly 
