In the early days of rubber planting, seed for nurseries had to be taken where it could be got and selection was out of the question; but in the future it ought to be possible to lay considerable stress on the selection of seed. Among planters a marked predilection exists for the selection of seed from old trees. Probably this preference rests upon some foundation of fact, hence, other things being equal, seed should be taken from old trees so long as these have not been tapped too heavily. Much more important, however, is the selection of seed from trees of known yield. We have already had occasion to point out the marked differences which exist between individual trees in the matter of yield. It is of the utmost importance that seed for future planting should be taken from the best yielding trees. This is not such a simple matter as might appear at first sight, 6n account of the scattering of the ripe seeds which takes place when the fruits burst. This phenomenon places a serious difficulty in the way of the collection of seed from particular trees, except on a very small scale. On estates therefore which are going in for wide extensions and are in possession of old trees already in bearing, the following procedure may be recommended. A few acres should be set apart definitely for seed bearing, as is done for example in the case of tea. All the trees on this area should be tapped in the same way for a definite period say for fifty tappings and a record should be kept of the yield of each individual tree. All except the best yielding trees should then be ruthlessly cut down and the stumps extracted in order to avoid the danger of root disease. Tapping should then be stopped for the whole period during which seed is required. In this way not more than 20 per cent, of the original trees should be left standing. If the distance of planting was originally 15 x 20 feet this would leave about thirty trees to the acre. With closer planting the selection might be still more stringent. In felling, regard should be had only to yield and not to size or position. Not only would such a group of seed-bearers yield a return from the sale of seed to neighbouring estates, but when the demand for seed is over, tapping may be resumed, and it is not impossible that the yield per acre from the seed-bearers will be found to be as large as or larger than that from any part of the estate not planted from selected seed.

To those in charge of government plantations and experiment stations a further course of selection may be recommended which is scarcely practicable on individual estates. This is the method of selection by progeny which has been practised with great success in the case of many annual crops. The method consists in planting a number of definite areas, each of an acre or less, with seed taken from the best yielding individual trees selected over as wide an area as possible. If practicable the seed of at least fifty trees should be separately planted in this way. When the different plots come into bearing their yields should be compared, and a few of the best plots retained for seed bearing. On the selected plots a comparison of individuals should again be made by the only reliable test, namely individual tapping, and the best yielding trees only should be allowed to stand. The analogy of cinchona would lead us to believe that if this method is carried far enough a very marked increase in the average yield of latex may be obtained. In Java the proportion of alkaloid in the bark of the introduced cinchona plants has been very nearly doubled by careful selection. There is good evidence to show that the variation in the yielding capacity of Hevea trees is considerably greater than the normal variation in richness of cinchona bark.