Picking also is considerably easier with a wide ridge ; fewer branctes and bolls are broken ofE during the operation, and it is easier to keep the picked cotton clean.

It naay, therefore, be provisionally stated that both from the point of yield and facility of cultivation a spacing of one metre between the rows, with 45 cm.

between the hills of two plants, is the most suitable spacing and arrangement for cotton in this district.

Manuring.—Desert soils are remarkable for the high percentage of phosphates and potash which they contain and for the virtual absence of humus and nitrates in any form. The problem which confronts every agriculturist in the Sudan is consequently that of increasing the nitrogen content of the soil. Areas of 10 feddans are cultivated annually as a manure experiment. The feddans are bisected, half of each constituting a check plot while the other half is manured.

The following combinations of manures are applied : Plot 1a receives 10 tons farmyard manure, 300 lbs. superphosphate and 200 nitrate of soda.

Plot 1b check.

Plot 2a receives 10 tons farmyard manure.

Plot 2b check.

Plot 3a receives 100 lbs. nitrate of soda.

Plot 3b check.

Plot 4:A receives 300 lbs. superphosphate and 200 lbs. nitrate of soda.

Plot 4b check.

Plot 5a receives 300 lbs. superphosphate.

Plot 5b check.

Plot 6a receives 200 lbs. nitrate of soda.

Plot 6b check.

Plot 7a receives 150 lbs. superphosphate and 100 lbs. nitrate of soda.

Plot 7b check.

Plot 8a receives 150 lbs. superphosphate.

Plot 8b check.

Plot 9a receives 100 lbs. nitrate of soda.

Plot 9b check.

Plot 10a receives 50 lbs. nitrate of soda.

Plot 10b check.

The results were remarka|)ly similar in each year {vide Chart No. 7). They clearly indicate that the addition of superphosphate to the soil is not only unnecessary but has a limiting efEect on the yield. In each case where superphosphate has been applied alone, the yield has fallen below that of the check plot.