481 
VII. — Native Trees Attacked. 
Among native trees T . gestroi appears mainly to attack 
species of the genus Macaranga, which, it is significant to note, 
belongs to the same natural order as Hevea brasiliensis. Jelutong 
(Dyera costulata ) is also affected, but not to any serious extent, 
and so are various species of Mango, both wild and cultivated. 
Grown under artificial conditions, tembusu (Fagraea ft agrans) 
has been damaged, especially in the neighbourhood of Seremban 
and near Tampin. On one estate, where Grcvillea robusta has 
been planted as a shade tree, the ants have practically extermi- 
nated it. Coconuts are also affected: but not to anything like the 
same extent as Para rubber. Rambong is but slightly attacked, 
and when it is, generally survives, a result probably due in part 
to its constitutional powers of resistance and in part to its habit 
of growth. 
VIII. — Age of Para when Attacked. 
It is very rare for any material damage to be done before the 
tree is 3^ years old, but I can find no satisfactory explanation of 
this undoubted fact. It may possibly be that Termes gestroi is 
naturally so scarce and scattered that it takes an appreciable 
time for a sufficient stock to accumulate, but more probably the 
species cannot work without a minimum amount of shade, which 
is not afforded until the age of 3 to 4 years is attained. A third 
theory, to which a certain amount of plausibility attaches, is that 
the nests of the ants, as with certain South African species, are 
normally situated at a considerable depth beneath the surface, to 
which the roots of the rubber do not extend until the age of about 
four years, though this again is discounted by the situation of the 
nests as actually found. 
IX. — Influence of Son. upon T. Gestroi. 
It has been demonstrated from returns sent in and from various 
typical estates visited, that, other things being equal, the damage 
caused by this pest varies more or less directly with the heaviness 
of the soil, light sandy loams being almost unaffected, while heavy 
clay soils have sometimes as many as 25 per cent, of the trees 
attacked. As regards drainage it is impossible to lay down any 
general rule. Swamp land is almost free from any species of 
termite, but, on the other hand, where the natural water level lies 
far beneath the surface termites can penetrate to a correspondingly 
great depth, and speaking broadly it may be said that the hill 
country estates in Negri Sembilan are more badly affected than 
those of the flats in Selangor, though possibly other causes may 
also have contributed to this. 
Taking all considerations into account, it has, I think been 
fairly demonstrated that the land best suited to Para rubber is a 
fairly stiff soil, capable of retaining a certain amount of moisture 
during drought, which is provided with a thoroughly efficient 
system of drainage. The slightly increased percentage of trees 
