Using a farming systems approach, a multidisciplinary team in southern Senegal examined three strategies for weed control in rainfed lowland rice: use of herbicide; replacement of broadcast seeding by row seeding with mechanical seeders; and transplanting compared with direct seeding of rice. Thorough evaluation of the agronomic and economic benefits of these strategies for farmers was facilitated by supplementing the standard analysis of on-farm trials with multivariate analysis incorporating agronomic survey data to explain treatment effects better, and by considering technical and economic effects throughout the farming system. The agro-economic analysis led the team to conclude that: where soils are productive, herbicide use is a profitable investment; when converting to direct seeding, farmers should facilitate weeding by seeding in rows with a mechanical seeder; and transplanting should not be discouraged in areas where it is still feasible, since it is less time-consuming, fits in better with other farm activities and requires less fertilizer.

A method is discussed for classifying small scale rice farmers from the Dominican Republic who have similar production systems and access to land, but differ widely in the yields they obtain and in the adoption of new technology. The results are used to define two recommendation domains, for farmers with ‘good’ and ‘poor’ production conditions, and to suggest appropriate technology for each. For farmers working in good production conditions the development of a technological package based on the double cropping of semi-dwarf varieties with high yield potential and production efficiency is recommended; for farmers with poor production conditions, a technological package is suggested based on varieties with a high tolerance of drought, flooding and weed development, which yield adequately at low input levels.

Cropping systems research to improve the productivity of an acid upland rice and maize-based farming system is described and evaluations by co-operating farmers are presented. Farmers rejected introduced patterns. The cropping choices of non-cooperating farmers are described in terms of cost–benefits and the reasons underlying their decisions. An approach that combines an understanding of farmers' systems in order to conduct focused research on the weak points of such systems, together with an evaluation of farmer technology and adaptation, may be more effective than researcher-designed, farmer-implemented cropping pattern trials.

On-farm trials were carried out in the Operation Haute Vallée project, Mali, to assess the possibility of replacing, partially or entirely, imported cotton fertilizer with natural rock phosphate from a local source, and to assess the residual effect on maize and sorghum grown after cotton. The maize/cotton sequence, although not normally used by Malian farmers, was more profitable than either pure cotton grown alone or a cotton/maize sequence when local rock phosphate was used instead of imported fertilizer.

The growing season in the semi-arid region of West Africa is drought prone and of irregular duration. Two experiments were conducted in 1986 and 1987 to test the effects of flat cultivation and tied ridge cultivation (TRC) on the yields of maize and the component crops of a maize and cowpea relay cropping system. The two research sites, with slopes of 0.05% and 3%, were near Sapu, The Gambia, on an Aridic Kandiustalf in the 700 mm rainfall zone.

Both growing seasons had above average rainfall. In both years, maize on level sites showed no response to tillage practices. On the sloped site in 1987, soil moisture 10 and 15 days after the last rain was greater with TRC than with flat cultivation and yields of sole cowpea and maize were 25% and 18% greater, respectively. On the level site, TRC had no effect on residual soil moisture or grain yield. When rainfall was well distributed, tied ridging did not appear to be a viable tillage alternative for maize-based systems on flat land in central Gambia but with modest slopes, tied ridges markedly increased soil water reserves in the 0.15 to 0.60 m depth after maize harvest.

In the wheat–sugarcane rotation, which covers more than half of the growing area in subtropical India, planting of sugarcane is delayed until after the wheat harvest in April–May. This allows little time for the sugarcane to produce tillers and so results in smaller millable cane populations and yields than those from more timely (February–March) plantings. This study considers techniques such as ring and trench planting which might allow a high seed rate to compensate for reduced tillering. Yields obtained using these systems were about 20 t ha−1 greater than from the conventional flat planting method, because of a better distribution of primary shoots from the parent sett and reduced plant competition during the early stage of growth. The cost of cane production was greatest with the ring system and least with the conventional planting method but deep trenches gave the best net return.

Three effects of herbaceous weeds on the productivity of a bush fallow system cropped with maize, cassava and plantain were identified in four experiments carried out in Central Amazonian Peru. First, early weed control is needed to minimize crop yield reductions. Weeds reduce plantain raceme weight and days to flowering, but not cassava yield, if weeding is done early. Second, fallows must be long enough to reduce weed seed banks to levels tolerable to farmers. While fallows of between two and five years eliminate herbaceous weed biomass present at the beginning of the fallow, weed seed banks decline linearly through ten years of tree cover. Third, with successive clean-weedings to reduce weed competition, tree and shrub numbers and their proportion of total non-crop biomass decline, while the amount and proportion of herbaceous weeds increase. Tree and shrub regeneration is reduced by longer cropping periods, more frequent weedings, greater crop cover, and shorter fallow periods.

The influence of planting mixtures of two potato cultivars on the incidence of leafminer fly (Liriomyza huidobrensis (Blanchard)) and on damage to potato foliage, and the effect on growth and yield, was studied in two field experiments in coastal Peru. Climatic differences between the two years of the experiments (1983 and 1984) had a greater effect than did cultivar mixtures on adult fly populations and on the numbers of feeding punctures and pupae per leaflet. Mixtures reduced the number of pupae in the tolerant cultivar, but increased it in the susceptible cultivar; in contrast, mixed planting tended to decrease the adult populations and feeding punctures in the susceptible cultivar of each mixture. In neither year did mixture yields or intercepted radiation exceed those of the highest-yielding sole crop, although land equivalent ratios (LER) reached 1.19. Considering the cultivars' different maturity rates and their tuber growth curves, it is probable that the later maturing cultivar of each mixture was able to exploit available light after, and perhaps even before, senescence of the earlier cultivar. Mixed planting of potato could be beneficial but only to farmers who plant both early and late cultivars, under conditions where the early harvest commands a premium price.

The suitability of charcoal and filter mud as carriers for the survival of Rhizobium isolates at low and high temperatures was investigated. Both carriers were found to be suitable and maintained high rhizobia populations (1.0 × 108 rhizobia g−1). In Sudan, inoculants can safely be stored for 6–10 weeks in the ambient temperature (up to 37°C).

Growth and the partitioning of dry matter and nitrogen were studied in winged bean ‘SLS-40’ grown with single stake supports at two sites in Sri Lanka. Plant parts (leaves, petioles, stems, roots, nodules, flowers, green pods, mature seeds, and mature pod-walls) were collected periodically, if present, for up to 32 weeks after emergence. Slow growth up to the seventh week was followed by rapid growth (2.8 g plant−1 day−1) for the next 15–17 weeks. Little leaf senescence was observed and the nitrogen content of the leaf tissue remained close to 45 mg N g−1 throughout growth. The rate of nitrogen accumulation was 2.3–2.7 kg N ha−1 day−1 between weeks 13 and 32. The final accumulation of nitrogen in the above ground material was 280 to 312 kg N ha−1, only 34% of that being in the mature seed. The harvest index for dry matter was 17.5–20% and the dry seed yield almost 2 t ha−1.

The effects of winged bean (Psophocarpus tetragonolobus) on a following crop of maize were assessed in two agroecological zones of Sri Lanka. Maize was grown after maize with variable rates of nitrogen fertilizer (0–150 kg N ha−1) or after winged bean or soyabean (Glytine max). Where it could be extrapolated from the nitrogen response curve, the apparent nitrogen contribution by winged bean was 70 kg N ha−1. However, in some cases maize after winged bean yielded more than the maximum yield predicted by the Mitscherlich N-response curve, suggesting the effect of some yield-enhancing mechanism in addition to nitrogen contribution. Where no fertilizer nitrogen was used, maize yields on land which had previously grown winged bean were 52–91% higher than those where maize was the previous crop. Winged bean proved to be an effective green manure crop even when all above-ground plant parts were harvested for food and fodder.