Sorghum agronomy in West Kenya : investigations from a farming systems perspective

An adaptive sorghum research and extension programme (1979-82) in West Kenya is reviewed. The focus is on factors operating at farm level. Research results are based on 70 on-farm trials and on some long-term experiments on NPK fertilizer and pest incidence. Extension results are based on two programmes using different approaches. 200 farmers participated in the demonstration approach; in the village approach 100 farmers participated the first year, and 400 the second year.

The impact of the rainfall pattern on local cropping systems is highlighted with respect to 'first rains' (March-June) and 'second rains' (August-November). Until recently, the short unreliable season of second rains played a minor role. However, current constraints on ox-ploughing result in delays in land preparation. This reduces the growing period for cereal crops during the first rains and increases the risk of failure for local late-maturing varieties. Cropping areas are increasingly restricted to what can be cultivated manually; more farmers are tempted to grow cereal crops during the second rains. As an assured food supply dominates farmer decisions, it was hypothesized that introduction of early- maturing sorghum varieties with a potential for ratooning would increase farmers management flexibility.

Initial knowledge of effects of sorghum pests, leaf blight and Striga hermonthica proved insufficient. Crop moisture availability analyses commonly used in Kenya do not permit sufficient quantification of variability to understand farmer response to uncertainty. Three types of sorghum cropping seasons must be distinguished based on the variable start of the first rains. These types differ not only in length of growing period but also in shoot fly and midge incidence. Success of late planted crops varies with type of season. Farmers must play each season by ear; conditional recommendations on variety use and time of planting were developed accordingly. Some early-maturing cultivars were selected. However, their ratooning potential proved limited. Hence, attention shifted to planting these cultivars at the start of the second rains. As such plantings are seriously affected by shoot fly, more research on control by seed treatments and cultural methods is required.

Crop yields and weed composition in farmer fields are related to land use pattern and soil fertility depletion. P is the most limiting nutrient. As N applications often results in negative effects, P/N ratios may play a role. Current socio-economic conditions restrict P applications to first-rain crops to 20 kg/ha P ₂ O ₅ . Higher levels may only be recommended if residual effects of P on second-rain crops are optimized through land preparation. Row planting with 3 plants/hill is advised as it facilitates placement of P and rapid handhoe weedings.

The on-farm demonstration approach proved ineffective in stimulating farmer participation in the sorghum technology development process; it puts too much emphasis on selling preconceived messages. Also, the small isolated plots are threatened by birds. In contrast, the village approach proved effective. Its large, clustered plots generated much discussions among farmers and minimized the bird damage threat. Farmer opinions and crop-cut sampling results were used to fine-tune the recommendations for husbandry practices.

The former compartmentalized research structure in Kenya was not conducive to adaptive research. Insufficient critical mass was available within the small commodity sections at regional level. However, recent interdisciplinary adaptive Regional Research Programmes allow better priority setting. The Farming Systems Approach to Research, Extension and Training is used, which could lead to more demand-driven research.