IMPROVING MAIZE (ZEA-MAYS) YIELDS IN SEMIARID HIGHLANDS OF KENYA - AGROFORESTRY OR INORGANIC FERTILIZERS

Farmers' maize (Zea mays) yields in semiarid Kenya are low because of low and erratic rainfall and soil nutrient depletion. An experiment wa s conducted in Machakos, Kenya, during November 1989-July 1995 to expl ore the potential of hedgerow intercropping (HI) with Leucaena leucoce phala as an alternative strategy to the use of inorganic fertilizers f or improving maize yields in semiarid bimodal highlands. The study, co nducted in a randomized block design with three replications, evaluate d on an operational scale with four treatments: sole maize with or wit hout fertilizer; and HI with prunings of L. leucocephala hedges used a s green manure or with prunings and maize stover fed to oxen and manur e returned to soil. About 1.4 t ha(-1) of L. leucocephala foliar bioma ss and 1.2 t ha(-1) manure were incorporated each season in green-manu re and ox-manure systems, respectively. Although sole maize responded to inorganic fertilizer in six of 11 seasons, maize yield was improved by HI, with prunings used as green manure, only once, 5 yrs after the commencement of the study, when rainfall exceeded 500 mm. Over the ex perimental period, maize yields were 15% and 30% less in green-manure and ox-manure systems, respectively, than in unfertilized maize, prima rily because of the competition of hedgerows with crops for water. The fodder system of HI recycled only 27% of N and 50% of P exported from the field in fodder, grain and stover, and depleted nutrients more th an the other systems. The green-manure system made a positive N balanc e, with similar P depletion as in the fodder system. No changes were d etected in soil organic carbon or cations in any of the systems during the experimental period while extractable P was significantly improve d by fertilizing maize. Net present value of returns from fertilizing maize at 40 kg N and 18 kg P ha(-1) was 10% higher than from the unfer tilized control. Fertilizer use would be 25% more profitable, however, if P were applied only to match the amount of P removed by crop (9 kg P ha(-1) per season) and if maize were not topdressed (amounting to 5 0% of the recommended N) in poor rainfall seasons. Of the two HI syste ms, only the fodder system was more profitable than either system of s ole-cropped maize because the high value of fodder more than compensat ed the smaller maize yield obtained in HI. However, the fodder system may not give higher economic returns than a sole L. leucocephala fodde r bank. Hence, HI using competitive species such as L. leucocephala an d with the level of productivity observed in this study, is not approp riate for semiarid climates to increase cereal yields, either directly through green manuring or indirectly by fodder production and nutrien t cycling through livestock. (C) 1998 Elsevier Science B.V.