Crop rotation is known to enhance crop yields. It is therefore recommended, regardless of rainfall and soil type, as a counter measure for the risks associated with monoculture maize ( Zea mays). Experience in the western Highveld where rainfall is low and erratic, has shown that the yield of maize does not necessarily improve as expected when preceded by alternative crops, but in fact, is often reduced. The present study was initiated to determine the effect of crop rotation with cowpea, groundnut, soyabean, sunflower or fallow on the yield and rainfall use efficiency of maize under marginal conditions on the western Highveld. Dryland maize was grown in five crop rotation systems on Mutton type soils at the farms Holfontein (four years) and Noodshulp (five years), both situated close to Ottosdal (2649?S; 2600?E). The soil profiles had an effective depth of >1.5 m at Holfontein and 1.25 m at Noodshulp. Crop rotation systems consisted of two-year rotations of cowpea-, groundnut, soybean-, sunflower-, and fallow-maize; as well as groundnut-, soybean-, and sunflower-fallow. A continuous monoculture maize treatment was included to serve as control. At Noodshulp where the rainfall was more variable, crop rotation induced maize yield deviations from the monoculture control occurred more often than at Holfontein. Apart from yield neutral and positive effects, instances of a decline in maize yield in some years due to crop rotation with cowpea, groundnut and sunflower also occurred. Taking the long-term rotational effect and the possibility of a yield decline into account, fallowing and the rotational crops ranked from best to worse were groundnut, soyabean, fallowing, cowpea and sunflower. The long-term effect of cowpea on the yield of maize was neutral and that of sunflower negative. The mean rainfall use efficiency of monoculture maize was, with the exception of maize preceded by groundnut, similar to that of maize grown in rotation.
