[Illustration: FIELD OF OLD RUBBER TREES IN WHICH THINNING HAD BEEN DELAYED TOO LONG. Note height and comparative lack of girth.] FUTURE YIELDS PER TREE.--It has been shown that by selective methods based on yields, poor trees can be eliminated. Whether by a process of seed-selection or by means of propagation based on bud-grafting and marcotting, it needs no great stretch of imagination to forecast future conditions under which trees may be bred which will be capable eventually of giving an average yield of 25 lbs. per annum over any given area. Yields of 1,000 lbs. per acre per annum should be obtained easily. TREES PER ACRE.--This brings us to the question as to how many trees one should leave to the acre after thinning operations. Figures have been given by various authorities, but it appears to the writer at the present time to be impossible to lay down a general rule. So much depends upon conditions. In certain cases where the soil is admittedly poor, the average growth below normal, and thinning has been postponed too long, the writer has been forced to the conclusion that it would be most inadvisable, and commercially unsound, to reduce the stand of trees below 120 per acre. In such instances the average yield per tree equalled only 3 lbs. per annum, and although the trees were upwards of nine or ten years old the crowns were small and sparse. It is doubtful whether such trees will ever exhibit any further development, and to thin them further would probably lead only to a diminution in the crop per acre. Under normal conditions of growth an arbitrary figure of eighty trees per acre has been selected as a standard by many estates. In these cases it would probably be correct to state that thinning was undertaken on almost purely empirical lines--_i.e._, that trees were not selected by tests of individual yields. As far as such a method retained the apparently most vigorous trees it was successful; but in view of what has been written it might explain some of the disappointing results which have followed upon such a system of thinning. It will be clear that any decision regarding the number of trees to be retained must be derived from a study of the detailed results of individual tests. If the large majority of the trees appear to be fairly uniform in yields the first thinning must be confined to comparatively few trees. Where there is, on the other hand, a good percentage of high-yielding trees the final stand per acre may be appreciably less. Unless and until such information is available, one cannot give any definite opinion as to the requisite number of trees to be retained per acre.