Long-term agroforestry demonstrations/trials using Acacia albida and other nitrogen fixing multipurpose trees/shrubs were initiated in mid-1982 to assess soil and crop productivity at a coastal lowland site characterized by low soil fertility, weed problems and consequent poor crop yields. Growth performance (height and diameter at breast height, dbh) of Acacia albida under eight densities rotationally intercropped with maize (Zea mays) and green gram (Phaseolus aureus), crop grain yields, soil fertility changes and weed control were assessed for a 5-year period (May 1982 to March 1987). A parallel-row systematic spacing field layout was used. Intercropped Acacia albida mean hight and dbh were 140 and 24% respectively higher than tree-only controls by the fifth year. Growth rate was low during the first year but increased in subsequent years to mean height and dbh of 9 m and 10 cm respectively by March 1987. While differences in dbh were significant, those between stand heights were not. Crop yields, especially under higher tree densities, declined considerably due to unexpected shade which also caused significant reductions in weed biomass. Soil fertility levels remained unchanged during the experimental period relative to the initial status, and differences between the intercropped Acacia albida plots and the tree — or crop — only control appeared not to be significant. We conclude that an understanding of the mechanism regulating leaf fall/retention phenomena of Acacia albida is crucial towards determining the intercropping potentials of the species.