An alley cropping experiment was conducted from 1986 to 1992 near Cotonou, Benin Republic on a field previously used as four-year bush-fallow. A no-tree control was compared with Leucaena leucocephala (Lam.) de Wit and Cajanus cajan (L.) Millsp. grown in 4 m distant hedgerows. In the same experiment, application of 90-39-75 kg ha(-1) NPK fertiliser vs. no fertiliser and maize-cassava intercropping vs. rotation of sole crops was also tested. Topsoil samples (0-30 cm) were taken initially and again after four and six years. Alley cropping enhanced nitrogen recycling relative to the no-alley control by 253 kg ha(-1) yr(-1) for Leucaena and 131 kg ha(-1) yr(-1) for Cajanus in the overall average but a maximum of 22 kg N ha(-1) yr(-1) was recovered in total crop exports. Apparent fertiliser-N recovery increased from a low average of 14% in the first year to 66% in the second year up to a maximum of 109% in the no-tree control in the last year. Soil nutrient depletion during the six years was significant for all nutrients except P, K falling from 0.5 to 0.1 cmol(c) kg(-1) soil and N from 0.08 to 0.05%. Hedgerow accumulation of N in the topsoil-plant system calculated as difference to a no-tree control was only 18 to 37 kg ha(-1) yr(-1) during the first four years after clearing, increasing to 208 kg in Leucaena and 111 kg ha(-1) yr(-1) in Cajanus without NPK during the last two years. During the first four years, average N, K, Ca and Mg losses from the topsoil-plant system of 223 kg, 160 kg, 176 kg and 80 kg ha(-1) yr(-1) were high but probably stored in the subsoil because corresponding leaching losses were only 10, 1, 15 and 6 kg ha(-1) yr(-1) at 150 cm depth during the second to fourth cropping year. Beyond the fourth year, markedly reduced nutrient losses from the topsoil and stable food crop yields in no-tree plots suggest that continuous maize-cassava cropping using balanced rates of mineral fertiliser is producing sustainable yields with almost closed nutrient cycles of the soil-plant system on our site.
