— 138 — Crossing between unlike individuals has proceeded unchecked and uncontrolled for years in the commercial crop, so that the hopelessness of ever obtaining a uniform seed supply from such a source is apparent.

Methods Employed.—The first step to be taken when attempting the improvement of the crop, is the purification of existing strains until individuals are obtained which breed true in every character. This result is only to be secured by constant self-fertilisation extending over a number of years, and by excluding all possible sources of contamination.

Work was begun on the farm in 1912, when a large number of apparently desirable types were isolated from the field crop, and self-fertilised by the simple expedient of covering the flowers before opening with gauze bags. The seed from these plants was sown in 1913 and the progeny compared with the detailed records which had been made of the parents during the previous year. Purity in all the characters was of course out of the question, so that particular stress was laid on such features as determine the commercial value of the crop.

Plants which approximated most closely to the parent types were selfed, and the resulting seed sown next year.

It is obvious that the number of individuals dealt with each successive year must increase very largely, as does the labour of selfing and comparing characters. In 1914, during which year the operations described above were repeated in detail, over 2,000 bags were in use every morning.

Sufiicient seed was accumulated to sow this year 1 feddan of each of the most desirable and uniform types isolated in 1912.

Further selections were constantly made throughout the three years, and the purification of these will be continued separately. A supply of pure seed for field purposes will be maintained by selfing a sufficient number of plants.

It is hoped that cages will be available in order to render this work more certain and to lessen the ever-increasing labour of hand-bagging.

The combination of desirable characters will be attempted by crossing, but as some years must necessarily elapse before the resulting hybrid is definitely fixed, this work will continue as a branch of the main task of furnishing a supply of seed which may be relied on to produce a uniform crop.

Practical Application of Methods.—To demonstrate the practical effects of the methods described above, it will be well to disregard the question of Inter-Egyptian crosses and consider the case of a hybrid which furnished an interesting example of the laws of heredity.

In 1912 a small plot was sown with Makhsous cotton, this being a strain which had shown promise in the Northern Sudan. One hundred plants were selected for examination, but owing to the irregularity of the lint, only one plant was finally chosen to propagate the strain. Of all the external features recorded