— 129 — III.—100 lb. nitrate of soda.

IV.—200 lb. nitrate of soda and 300 lb. superphosphate.

V. 300 lb. superphosphate.

VI. 200 lb. nitrate of soda.

VII. 150 lb. superphosphate and 100 lb. nitrate of soda.

VIII. 150 lb. superphosphate.

IX. 100 lb. nitrate of soda.

X. 50 lb. nitrate of soda.

The superphosphate was incorporated before ridging, the stable manure before ploughing, while the nitrate was added mixed with sand after the crop was well established, and worked in with the first inter-row cultivation.

The curves for 1915 are given alone on account of lack of space ; but it may be said that greater agreement was obtained with the records of previous years than in the case of any other experiment. The results set out above may therefore be taken as fairly representative.

The flowering curve is not so useful as the boUing curve when comparison of the yields is desired, but two points are brought out strongly. The first is the increase in the rate of flowering caused by manuring with nitrate of soda, and the second the depressing effect of superphosphate even when applied with nitrate. The antagonism between superphosphate and nitrate is more apparent than real, for only when superphosphate is applied alone is any reduction in yield obtained.

The boiling curve is more illustrative of final results than the flowering curve. With the exception of the superphosphate plots, the addition of manure has always resulted in a heavier yield. The contrast between the manured plots and the controls was most marked in the fields. The more luxuriant growth and the deeper green of the manured portion spoke eloquently to the need for nitrogen.

Turning to effects produced on the cotton itself, the records obtained in the laboratory during the past three years will be found on the accompanying table.

The uniformity of the cotton derived from the check plots permitted a pooling of the figures, which thus furnished a standard for the effect of the manures.