II 
abundance. This is no theory but has been practised for several 
years with success. There are no complete failures in any hill, 
except where planting has been done in ground which is too sour. 
In such places we may have too replant in small patches after 
proper draining has been prepared for, and in such cases we use 
excess plants growing in the neighbourhood. During the first few 
months seedlings backward, misshaped, or where crowding is 
observed, are cut out at the time of each weeding. Generally we cut 
at least 50 per cent, of all the seedlings during the first six months. 
Detailed rules have been worked out for this first thinning and they 
will be applied in the field by the assistants in charge. Great care 
is exercised in the selection of the plants allowed to remain, and 
many points have to be attended to. Thus if a plant branches in a 
way that does not promise regularity, or if it has grown twisted or 
bent it is cut out. In selection those plants that show a tapering 
shape with thick stem are preferred to plants growing mainly in 
height, and plants with large dark green leaves are preferred to 
those with small leaves of a lighter or yellowish colour. 
Advantage of Close Planting. 
One of the great advantages of this system of close planting 
and successive thinning is the opportunity for selection, a matter 
which is generally overlooked, but which, no doubt, will be found to* 
be a most profitable policy. In Castilla cultivation we often get 
trees which produce very little or sometimes no rubber. It is 
vm^ ent SU °k trees are an unnecessary expense on the plantation.- 
Why should we waste time on valueless trees ? Can this in any 
way be avoided ? With our present limited.knowledge of the nature 
of the tree we cannot say with absolute certainty whether a seedling 
will become a good “ milker ,l * or not*. But in the course of my 
physiological investigations of Castilla , I have already been able to 
draw some conclusions, which give us certain indications on which 
we can judge in this matter. Thar is to;say, we can in some 
instances say definitely that a seedling, Which shows certain 
characters, will never produce a large amoiint of latex. On the 
other hand, we cannot guarantee that a number of non-producing 
trees would not be passed during the selection time. But we have 
found a way to reduce their number, and I hope that further 
investigation will considerably improve the method. There is, 
however, another important point which can be applied in this 
process of selection. It is the choice of rapid growing, healthy 
plants of a certain desirable type. It is of the greatest value to 
the planter to have trees which are as rapid growing as possible. 
iNow with the transplanting system a certain amdunt of selection 
can be done, but this is before the transplanting process, during 
which the plant is always more or less injured. There is jio selection 
possible after the transplanting. With the “at stake” planting 
system mentioned we do our most important selection after sowing 
the plantin ^ W1 * P resen ^7 re fer to the selection which precedes 
With the growth of the young rubber trees, successive thinnings 
become necessary to prevent crowding. It is here the men in 
c arge ot the plantation have to exercise their best judgment. 
