A fallow management trial was initiated in March 1990 in Yaounde, Cameroon, with the objective of identifying an efficient method of managing residue derived from a planted fallow of Cajanus cajan. Ten months after establishment, the shrubs were slashed and residues were treated in one of the following ways: burnt, incorporated in the soil, mulched on soil surface or removed. Following that, maize and groundnut were planted. Soil organic matter was fractionated after the residue treatments, and elementary partial budgeting was conducted. After three cycles of fallow and cropping seasons, maize yield was similar, about 3 t ha(-1) in all plots except in the residue-removed plots, where it was 1.4 t ha(-1). Yield trend of,groundnut was also similar. The residue management method did not affect either the nature of fractions or the total content of soil organic matter. The highest net return, US$5945 ha(-1) year(-1), was obtained from the residue incorporated treatment, whereas the highest return to labour, US$11 per manday(-1) was associated with burning of residues. In areas such as the forest zone of Cameroon where labour is a major constraint and climatic conditions allow vigorous vegetation growth, burning appears to be the best method of residues management, at least in the short run.