r bare or occupied by low and very prickly bushes. The
ground is parched, and one can get no shade, except by standing close
under a trunk somewhat thicker than its neighbours. Still farther north
the timber is hardly larger, though the general aspect of the woods is
improved by the more frequent occurrence of flowering trees, some
sweet-scented, with glossy leaves and small white flowers, some with
gorgeous clusters of blossoms. Three are particularly handsome. One,
usually called the Kafir-boom, has large flowers of a brilliant crimson.
Another (_Lonchocarpus speciosus_[4]), for which no English name seems
to exist, shows lovely pendulous flowers of a bluish lilac, resembling
in colour those of the _wistaria_. The third is an arboraceous St.
John's wort (_Hypericum Schimperi_[4]), which I found growing in a
valley of Manicaland, at a height of nearly 4000 feet above the sea. All
three would be great ornaments to a south European shrubbery could they
be induced to bear the climate, which, in the case of the two latter
(for I hardly think the Kafir-boom would suit a colder air), seems not
impossible. In Manicaland, among the mountains which form the eastern
edge of the plateau, the trees are taller, handsomer, and more tropical
in their character, and palms, though of no great height, are sometimes
seen. But not even in the most humid of the valleys and on the lower
spurs of the range, where it sinks into the coast plain, nor along the
swampy banks of the Pungwe River, did I see any tree more than sixty
feet high, and few more than thirty. Neither was there any of that
luxuriant undergrowth which makes some tropical forests, like those at
the foot of the Nilghiri Hills in India or in some of the isles of the
Pacific, so impressive as evidences of the power and ceaseless activity
of nature.

The poverty of the woods in Bechuanaland and Matabililand seems to be
due not merely to the dryness of the soil and to the thin and sandy
character which so often marks it, but also to the constant grass-fires.
The grass is generally short, so that these fires do not kill the trees;
nor does one hear of such great forest conflagrations as are frequent
and ruinous in Western America and by no means unknown in the south of
Cape Colony. But these fires doubtless injure the younger trees
sufficiently to stunt their growth, and this mischief is, of course, all
the greater when an exceptionally dry year occurs. In such years the
grass-fire