AbstractIn field experiments conducted at the Nyankpala Agricultural Station, Ghana, during 1976–77, soil moisture conditions appeared to be the deciding factor in determining the frequency of handweeding needed to obtain maximum seed cotton yields. In 1976, when planting was early and the late, heavy rains in October prolonged weed growth, two handweedings at 4 and 8 weeks after seeding (w.a.s.) gave comparable yields to continuous weeding and significantly better yields than an alachlor + fluometuron herbicide mixture. Supplementary weeding did not enhance the mixture's performance.In 1977, planting was relatively late and in the peak of the rainy season and so late weed growth was reduced. A single handweeding at 4 w.a.s. gave as good a yield as repeated handweeding treatments under these conditions. The alachlor + fluometuron mixture also gave adequate weed control and again supplementary weeding did not improve it significantly.In herbicide evaluation experiments, the safety and performance of pendimethalin and dinitramine herbicides were much influenced by the rainfall pattern immediately following application. With a relatively drier period following application in 1976, the higher rate of these herbicides depressed yields. This rate produced optimum yields in 1977, however, when a very wet period followed application.Prodiamine and the alachlor + fluometuron mixture were the herbicides least affected by soil moisture and they consistently produced good yields.