The International Benchmark Sites Network for Agrotechnology Transfer (IBSNAT) project has established an international network of co-operating farming systems researchers. A decision support system that includes a number of crop models and databases has been developed to assist in the evaluation of new agrotechnology packages in the tropics and sub-tropics. Standards have been established for the recording of ‘minumum data sets’ from field experiments to allow the crop models to be validated for specific sites. The decision support system, still under active development, has the potential to reduce substantially the time and cost of field experimentation necessary for adequate evaluation of new cultivars and new crop management systems.

On-farm research was conducted in Zambia between 1984 and 1987 to examine the performance of an imported bean variety and identify management practices that could result in higher and more stable yields and to examine their economic feasibility for risk averse smallholder farmers. Results showed that the Brazilian bean variety Carioca was superior to the local variety tested in terms of yield, net economic returns and reliability of performance, especially when used in combination with fertilizer and insecticide.

Three pearl millet and three cowpea genotypes were examined in intercropping systems on deep sandy soils in Niger between 1984 and 1987. The intercropped pearl millet genotypes usually produced at least 80% of their sole crop yields when the cowpea was planted two or more weeks after the pearl millet. The early maturing and spreading cowpea genotype contributed most to LER. All cowpea genotypes yielded more when planted in association with dwarf rather than the traditional tall pearl millet genotypes because they received more light.

The physiological basis for responses to daylength of a photoperiod sensitive groundnut genotype (NC Ac 17090) was investigated by comparing its growth and development in natural daylength with that in an artificially manipulated photoperiod in three field experiments. Photoperiod did not influence the thermal time to flowering, or the subsequent appearance of flowers until 900–950 flowers m−2 had appeared. Thereafter flowers continued to appear in short, but not in long, days. In each experiment, long days increased the thermal time between the initiation of each peg and pod, and the thermal time required for each fruit to mature. These fruit initiation and developmental changes were reflected in the partitioning of assimilates to pods, this being substantially less in long days than in short. Changes in pod initiation rate, partitioning co-efficient, and the thermal time taken for a pod to mature were related to mean daylength.

The effects of intercropping roselle with blackgram, cowpea, soyabean, groundnut and sesame under single, double and triple row planting patterns and at three different intercrop sowing dates were studied from 1983 to 1985. The fibre yield of roselle was reduced by intercropping. The reduction was least with groundnut and greatest with cowpea but was compensated for by the yield of the intercrops. Double and triple row planting systems favoured the productivity of both roselle and the intercrops, increasing the Land Equivalent Ratio (LER) of the cropping system. The best yields of roselle were obtained when the intercrops were sown 15 days after the roselle.

The productivity of each component of a sorghum/groundnut intercrop and its constituent sole crops is determined in terms of a ‘Crop Performance Ratio’ (CPR) defined as the productivity of an intercrop per unit area of ground compared with that expected from sole crops sown in the same proportions. The CPR allows productivity, intercepted radiation and seasonal transpiration to be compared so that conversion coefficients for radiation (e; g MJ−1) and dry matter/water ratios (q; g kg−1) can be calculated for each intercrop component and its constituent sole crops. In this experiment, CPR for total dry weight in the intercrop was 1.08 and that for reproductive yield was 1.27. These advantages in overall productivity and yield were typical of those reported elsewhere for sorghum/groundnut intercrops. The proportional increase in total dry matter in the intercrop was largely a result of its greater interception of radiation. The further advantage in reproductive yield was a consequence of an improved harvest index in the sorghum component of the intercrop (0.64) compared with that of its sole crop counterpart (0.55).

Topsoil depth and yields of sorghum, pearl millet and castor bean were found to be positively correlated when monitored over thirteen seasons under widely varying rainfall conditions on an Alfisol in Andhra Pradesh, India. Yields responded up to 2.5 times more to topsoil depth when rainfall in the critical period exceeded evapotranspiration than under drier conditions. Yield variations were explained by topsoil depth in cereals, and by the interaction between rainfall during the critical period and topsoil depth in cereals and castor bean (R2<0.72). The expected yield losses as a result of erosion, based on 56 years of rainfall data, were 138, 84 and 51 kg ha−1 cm−1 for sorghum, pearl millet and castor bean, respectively. Quadratic models explained the yield losses due to cumulative erosion under conditions of both poor and good rainfall during the critical period. Sorghum and pearl millet tended to produce no yield when erosion exceeded 34 cm and castor bean when it exceeded 40 cm. There was less reduction in yield when rainfall was poor, up to about 10 cm of erosion, than when rainfall was good. The results indicate that pearl millet and castor bean are more suitable than sorghum for low rainfall and eroded conditions.

Soyabean (Glycine max) is a relatively new crop for small-scale farmers in Zambia which has been adopted following the introduction of new cultivars, greater opportunity to obtain credit, easier marketing and an attractive guaranteed price. However, low yields limit production partly due to the lack of a planting method that establishes optimal populations. The present method is to plough and plant in the same operation, dribbling the seed behind the ox-plough. This often leads to uneven depth of planting, and hence to poor seedling emergence and erratic stands. Alternative planting techniques evaluated on farmers' fields for three seasons (1985/86–1987/88) suggest that farmers should replace their practice of planting behind the plough with either hand seeding following a plough–harrow operation or the use of a modified ox-drawn planter (Taparia).