Procedures for better decision making in agricultural research are considered. Fanning Systems Adaptive Research (FSAR) has evolved to complement the more traditional research approach, allowing better decisions to be made on recommendations to fanners and on priorities for technical research. This is done through the identification of research results most relevant to the development of specific farming systems; the testing of these results by experiments in farmers' fields; and the identification of factors limiting productivity in local farming systems.

FSAR involves interdisciplinary cooperation between production agronomists, farm economists and, where appropriate, animal production scientists, in close cooperation with other production specialists and with farmers.

Five diverse sorghum genotypes were sown at sole crop density and at intercrop density without cowpea and with two contrasting cowpea genotypes, in the post-rainy season at Hyderabad, India. The interaction of sorghum genotype with sowing density was significant for sorghum dry fodder and grain yield, but the interaction of sorghum genotype with cowpea was not, because of compensation between yield components. The likely response of sorghum genotypes to intercropping can therefore be assessed initially from the performance of low density sole crops, followed by assessment in the presence of a single standard cowpea variety.

The cowpea genotypes were affected by the presence of sorghum but not by the sorghum genotype. This suggests that the effect on the cowpea can be ignored when selecting a sorghum genotype for intercropping, and that a cowpea genotype for intercropping can be selected in the presence of a single sorghum genotype. However, these conclusions arc unlikely to apply to rainy season sowings, when sorghum dominates the intercrop more completely.

Branches are an important yield attribute of the mustard (Brassica junceo)crop. However, not all the branches are equally productive and some of those arising from the basal portion of the plant may form parasitic sinks for assimilates. The effect of de-branching at the 50% flowering stage in the apical (AHD) and the basal halves (BHD) of two mustard genotypes, Pusa Bold and Varuna, was examined. The BHD treatment increased the seed yield per plant of Pusa Bold by 36% and of Varuna by 75% over the control. The increase in yield due to the BHD treatment was mainly due to an improvement in the total number of pods per plant in both genotypes. In Pusa Bold, an improvement in both 1000-seed weight and number of seeds per pod, mainly on late formed branches, was also observed. The AHD treatment reduced seed yield but not significantly so. The results suggest that improvement in the yield of mustard cultivars by suppressing branching on the basal portion of plants should be possible.

A range of planting times, with sequential harvests to age 24 months, was used to determine crop growth and yield of cassava at two locations, one having a humid tropical climate and the other humid sub-tropical. Yield was doubled during the second year at the sub-tropical site. At the tropical site yield was very high after one year, but root rots occurred in the second year. The dry matter percentage of storage roots fluctuated seasonally and was highest during the cooler months when canopy vigour was lowest. Known or inferred environmental controls of the yield-determining factors of growth rate, growth duration, and partitioning to harvestable organs were used to interpret the results and thus derive guidelines for timing of planting and harvesting in moist environments.

Two experiments, one irrigated and one dryland, were conducted at a site having a seasonally-dry, tropical climate. A range of planting times, with sequential harvests to age 24 months, was used to study cassava growth and yield. Very high yields of 35 to 45 tonnes dry matter per hectare were recorded for the irrigated experiment. Water limitation restricted yields to 15 to 26 tonnes in the dryland experiment. Known or inferred environmental controls of the yield-determining factors of growth rate, growth duration, and partitioning of assimilates to harvest-able organs were used to interpret the results and thus derive general guidelines for the timing of planting and harvesting in seasonally-dry environments.

The intercropping objective of the subsistence fanner in Botswana was identified as a need to secure some cowpea production without reducing the yield of the staple crop, sorghum. Field experiments in which sorghum and cowpea were intercropped in several different environments over three years met this objective on most occasions, but cowpea yields were low compared with their sole crop performance. Although intercropping out-performed monocropping by producing some cowpea yield, land equivalent ratios greater than one and higher gross income and cash returns to draught-power (but not returns to labour) the differences were never large and seldom statistically significant.

The comparative stability of sorghum–cowpea intercropping and monocropping was assessed by determination of variances, regression against an environmental index, and probability of failure to achieve minimum levels of gross income. Significant differences between the cropping systems were not detected in any of the assessments though intercropping did have a lower probability of failure at the income levels for which risk was more than minimal. The failure of intercropping to show clearly greater stability in this study in which the environmental conditions were harsh provides a conflict with the conventional view.

Studies were carried out at Kalyani during 1982–83 and 1983–84 to evolve a multiple cropping system with cotton as the main crop. Cotton was intercropped with paddy, greengram, blackgram, peanut and soyabean in the rainy season, and these were also grown as solccrops followed by crops of peanut, lentil, chickpea, wheat and mustard, respectively, in the winter season. All plots were subsequently cropped with sesame in the summer season. There was an increase in total grain yield, land equivalent ratio (LER), area time equivalent ratio (ATER) and relative net return (RNR) values when legumes and cereals were intercropped with cotton. The yield of the winter crops grown with the help of residual soil moisture was in the order: chickpea > mustard > lentil > wheat > peanut. Sesame yielded best after the chickpea and blackgram sequence and yielded almost as well after the lentil and greengram. The greatest net return was obtained from the cotton and paddy (1:1) intercrop followed by sesame.

The introduction of forage or hay crops into fallow land represents a means of increasing feed supplies for small ruminants in north Africa and west Asia. Such fallows lie between zones planted with food crops and traditional grazing areas. Mixtures of vetch with barley, oats or triticale, and peas with barley, were cut at the 10% flowering, 100% flowering (100F) and full pod (FP) stages of the legume component. Effects of crop maturity, weather conditions and cutting height on yield, the chemical composition of the standing crop and re-growth were measured. Barley–vetch mixtures yielded 8147 kg dry matter ha−1 of standing crop at the FP stage in 1983 but only 2283 kg dry matter ha−1 in 1984 under drought conditions. Hay yields were 4377 and 1640 kg dry matter ha−1, respectively. The highest yielding mixture, grown in 1983, was oat–vetch which yielded 8670 kg dry matter ha−1 of standing crop and 4285 kg dry matter ha−1 of hay at the FP stage. Cutting at the 100F or FP stage maximized dry matter yield and minimized the risk of rainfall prolonging hay making. Rain caused slight damage to some hays cut at early stages of maturity. Voluntary intakes and digestibilities of field-cured hays were determined. The voluntary intakes of triticale–vetch hay at the FP stage and all barley–pea hays were low compared with other hays. Hays contained sufficient estimated metabolizable energy (ME) for use in sheep diets during pregnancy and middle or late lactation. ME values averaged 9.2 megajoules per kg dry matter in both 1983 and 1984.

The phenology of mung beans is extremely plastic; responsiveness to both photoperiod and temperature is known to modulate flowering. A previous conclusion, perpetuated uncritically now for almost forty years, has been that both (quantitative) short- and long-day flowering responses exist in the mung bean germplasm. However, our re-analysis of the original data leads us to an alternative conclusion: that genotypes of mung bean are quantitative short-day plants with different optimum mean diurnal temperatures for flowering. This alternative interpretation (which is plausible biologically and in evolutionary terms) is discussed.