The agronomic and economic performance of banana, bean and tree intercropping in the highlands of Burundi: an interim assessment.
To meet their wood, fodder and fruit needs, resource-poor farmers with only small land holdings are forced to mix trees in their food crop plots. An experiment was conducted to study the effect of 9 tree species planted as 5-month-old seedlings in Dec. 1989 at a density of 312.5 trees/ha (spacing 4×8 m) on the yield of bananas (a local cultivar of Musa, planted in January 1990 at 625 stools/ha, spacing 4×4 m) and beans (Phaseolus vulgaris, undersown each season from Oct. 1990 to Feb. 1992 at 80 000 plants/ha), as well as the wood production of the intercropped trees. In addition, an economic analysis was done to compare the different tree/banana/bean associations. After 3.5 yr, wood volume (in m3/ha) of Grevillea robusta (18.1) was highest, and that of Erythrina poeppigiana (2.7), Cedrela odorata (2.4) and Markhamia lutea (0.8) lowest. Volume of Cedrela serrata (13.7) was not significantly different from that of Albizia chinensis (12.8) but was significantly higher than that of Leucaena diversifolia (6.8), Acrocarpus fraxinifolius (6.7) and Calliandra calothyrsus (6.0). None of the tree species had a significant influence on the yields of the bananas and none affected the yield of the bean crops until the 7th cropping season, 3 yr after the trees were planted. In that year, Grevillea reduced bean yield by 29%, Albizia by 34% and Leucaena by 36%. From the economic analyses, all the treatments except Leucaena and Markhamia had positive net benefits relative to the control (banana/bean) but the results were highly variable. C. serrata was the best tree to be intercropped in a banana/bean system.