Between October 1988 and August 1992, field experiments were carried out in West Kenya to evaluate the suitability of Leucaena leucocephala, L. collinsii, Gliricidia sepium, Calliandra calothyrsus, Sesbania sesban, S. grandiflora, Senna siamea and S. spectabilis to provide a range of agroforestry products and services. The initial objective was to establish the growth rates and wood and leaf yields of these tree species, when planted in single rows. After the initial evaluation, it was evident that valuable additional information could be collected if the trees were converted to hedges and their effect on intercropped maize and soils was studied. At 21 months after planting, different species and provenances ranged in height between 3.5 and 6 m and varied considerably in phenotypic appearance. Wood production (1988–1990) varied from 3 to 33.8 t ha −1 and leaf production varied from 0.62 to 10.1 t ha −1. During intercropping (1990–1992), leaf production varied from 0 to 10.9 t ha −1. Maize yields were higher in association with Leucaena and Gliricidia than with Calliandra, Sesbania and Senna. Cumulative maize grain and stover yields over four seasons were positively correlated with the total amount of tree leaves applied ( r 2 range, 0.70–0.95). The effect of tree leaf mulch on crop yields decreased over time for all species. Leaves with high nutrient contents, which decompose fast ( Leucaena, Gliricidia, Sesbania) are likely to have been more effective in sustaining crop yields than leaves with lower nutrient contents ( Senna) or more complex decomposition patterns ( Calliandra). Simple “leaf input-crop output” budgets to calculate the reserves for N, P and K in different systems explained crop yield differences in some cases. Compared to the fertility status of “zero-mulch” control plots, the status of soil C, N, P, K, Ca, Mg and S was to varying degrees improved under Leucaena, Gliricidia and Sesbania, much less under Calliandra but not under Senna. First season grain yields were related to the soil fertility status at the end of the tree fallow. The results of these experiments suggest that under subhumid tropical conditions with soils of relatively poor nutrient status, where light and water are not likely to be the major limiting factors to crop production, the application of sufficient quantities of high quality tree mulch may positively influence maize yields. When agroforestry tree species with contrasting decomposition and nutrient release patterns are evaluated jointly, it is more difficult to demonstrate a general relationship between quantities of mulch applied and improvements in crop yields and soil fertility levels. Therefore, further chemical, physiological and phenotypic characterization of free species with potential for fallow and intercropping systems is required.
Read full abstract