In 92 field experiments in Tanzania's major cotton-growing area south and east of Lake Victoria mean yield without fertilizer was nearly 900 kg/ha of seed cotton. The mean direct effect of 224 kg/ha of single superphosphate with 202 kg/ha of ammonium sulphate was 380 kg/ha of seed cotton; with twice as much fertilizer the increment was 525 kg. First and second residual responses made important contributions to the total effect of superphosphate. In most zones the response to each fertilizer was large but in Nassa response to superphosphate was small; in Nzega yields were small and there was no response to ammonium sulphate, perhaps because Calidea dregei caused abscission of bolls; on the acid soils of Ukerewe Island yields and responses to superphosphate were small.

Data are given to show the average amounts of plant nutrients in the soils most used for cotton. Correlations of yields with soil chemical data suggest that some soils lack enough calcium for large yields. A long-term experiment at the Ukiriguru research farm (Le Mare, 1972) showed that productivity of impoverished soil growing cotton every year was recovered and improved by lime with phosphorus and nitrogen fertilizers: the experiments throughout the cotton area indicated that those results were widely relevant.