116 
a good-yielder is to be produced there must be a favourable combi- 
nation of the necessary hereditary characters. Therelative scarcity 
of high-yielders (1-2 per cent.), and possibly the very gradual 
range from these to non-yielders, lends support to this view. 
hen once the stock has become more homogeneous, favour- 
able combinations will be more easily obtained, the isolation of 
any particular character could then be brought about by self- 
fertilisation. 
There is yet a third difficulty, since Hevea is fertilised by 
insects and appears to be practically self-sterile. Petch records a 
tree which stood entirely isolated and had never been known to 
bear fruit. A similar case, though not absolutely reliable, is 
known from the east coast of Sumatra. Cramer reported in 1914 
that artificial self-fertilisation, as described above, applied to over: 
300 flowers of one tree failed completely. The fruit-setting was 
likewise poor (7 per cent.), after artificial fertilisation with pollen 
from other trees. This may be comparable to the facts well-known 
to fruit growers, that some trees, though they flower profusely set 
no fruit, while others are highly fertile. 
e case of one of these fertile trees, four out of 95 self- 
fertilised flowers set their fruit, whilst of 78 cross-fertilised 
flowers 19 set fruit. Self-fertilisation, therefore, is possible in 
fertile trees as Mr. J. G@. J. A. Maas demonstrated in 1917. 
Further trials will, however, be needed to prove that enough seed 
can be obtained by this method for selection on a large scale, and 
results one of the two individual trees can be removed. 
ropagation by seed is thus a lengthy process, but it must not 
be neglected, and should, if possible, be undertaken in experi- 
ment stations. 
1. Nurseries should be established with trees grafted from two 
female parent trees, and natural cross-fertilisation should be relied 
n 
upon. 
2. Artificial cross-fertilisation (eventually self-fertilisation) of 
good-yielders should be carried out where it is not possible to form 
isolated nurseries. 
The same rules which apply to the selection of the character of 
high latex production may be followed for any other advantageous 
qualities, such as resistance to disease, thickness of bark, ete., in 
so far as these may be hereditary. 
Vegetative selection.—A rapid improvement in the strain of 
Hevea as regards latex yield may be obtained by vegetative 
methods of propagation. 
The stock of a high-yielder may be multiplied by grafting or 
* Van Helten: The Grafting of H Adehined tials ee 
No. 1, 2nd Vol. 1918, p. 637. aren, ives for Rubber culture 
