Low levels of phosphorus limit the yield of pearl millet (Pennisetum glaucum) grown on acid sandy soils of the West African Sahel. The application of phosphorus and the recycling of millet residues produces large increases in the total dry matter and grain yields of millet but little is known about the effects of these amendments on the by-products of millet fields which include weeds and millet leaves at harvest. Phosphorus applied at 13 kg ha−1 as single superphosphate and millet residues applied at 500 or 2000 kg ha−1 as a mulch or as the ash from 2000 kg millet residues ha−1 that had been burned, increased dry matter yields of millet leaves and weeds leading to large increments in the total in vitro metabolizable energy harvested per unit area. These increases in dry matter production were only accompanied by minor decreases in the concentration of crude protein. The increased availability of millet by-products should be taken into account when evaluating the effects of phosphorus and crop residue application on the economy and sustainability of agro-pastoral farming systems in the West African Sahel.

The use of tropical grass–legume mixtures containing the grasses Brachiaria decumbens or Andropogon gayanus associated with a mixture of several herbaceous legumes was tested on farms in the rainforest area of Pucallpa, Peru. Pastures were established by farmers using manual labour and without fertilizer in previously deforested areas covered with secondary growth. The pastures, together with their grass-alone control, were incorporated by farmers into the normal paddock rotation, and were grazed by dual-purpose cattle of mixed breeding. Cows were milked once daily by hand and with their calves at foot. Following a training period, farmers kept records of milk yields. Frequent and unannounced visits at milking time by the research team were used to verify milk yields and to annotate events such as calving, drying out, sales and others. Across farms and years, grass–legume mixtures significantly outyielded grass-alone pastures by 9%. Milk yields on the mixed pastures were examined further by stepwise regression. A positive and significant effect of the proportion of crossbred cows in the herd was established, whereas soil quality (evaluated by the percentage of aluminium saturation) had a negative impact on milk yields. Two of the farms, which for different circumstances had poorer than average animal and pasture management, also had a significantly negative effect on milk yield. The interactions between animal and pasture management, and regional infrastructure are discussed. It is concluded that grass–legume mixtures increased milk yields on farms. Nevertheless, it is also hypothesized that they may have a more restricted niche than anticipated and that their adoption may be highly sensitive to the overall economic context of the region.

The strengths and weaknesses of the evolving techniques used by researchers in Bolivia to evaluate potato germplasm with farmers are assessed. When detailed knowledge of farmers' criteria is required, questionnaires or forms are most appropriate; otherwise techniques from participatory rural appraisal could be used. Initially, when evaluating large numbers of clones, scientists' and farmers' choices tended to coincide, although subsequently they diverged. This suggested that it may be more efficient to involve farmers later in the selection process, ensuring that an interesting amount of diversity still exists, and work to improve breeders' knowledge of farmers' criteria. Integration between social scientists and breeders was less than ideal at the start of the study, reducing the impact of farmer evaluations on the breeding programme.

Employing data from an intercropping field experiment which measured the yields achieved by a surviving crop when the other crop failed at 0, 6 or 10 weeks after sowing, the use of a simple computer model to study the long-term yield effects of compensation in intercropping is described. Assuming different probabilities of crop failure for one or both of the crops, the model provided a comparison of long-term intercrop and sole crop yields from intercropping systems of oat–mustard, mustard–bean and oat–bean. It showed that if failure was restricted to only one of the crop species, intercropping suffered less yield decline with increasing probability of failure than did sole cropping only if there was a sufficiently high yield of the surviving crop. However, in the more realistic situation where each crop was subjected to the possibility of failure, this decline in yield was less in intercropping than sole cropping for all three combinations because poor yields when one species was the surviving crop were offset by good yields when the other species was the survivor. Mustard–oat and mustard–bean gave no yield advantage in the absence of failure (Land Equivalent Ratios less than 1) but they outyielded sole cropping for failure probabilities above about 5% and 15%, respectively. All combinations gave large intercropping advantages at high probabilities of failure. Oat–bean always had a strong yield advantage, even with no failure.

The model is very simplistic but it illustrates the potential for modelling long-term effects from field data. Limitations of the model are discussed and it is emphasized that these can be overcome if information to make the underlying assumptions more realistic is available.

The potential of pigeonpea (Cajanus cajan (L.) Millsp.) as a simultaneous fallow component in cropping systems is unique in that, being a shrubby grain legume, it combines food production with ease of establishment, fast growth and high biomass productivity. A study was carried out under on-farm conditions at three different sites in southern Bénin, West Africa over two years to evaluate the biomass productivity and recycled nutrients of a local pigeonpea cultivar, managed as annual hedgerows. Pigeonpea was sown between a standard cassava–maize intercrop and compared with two other agroforestry systems and annual intercrops with and without mineral fertilizer. The number of cuts taken at a height of 1 m was doubled from two in 1991–92 to four in the 1992–93 season, leading to an increase in total cut dry matter by a factor of eight (1908 g m−2) and cut leaf dry matter by a factor of fourteen (1317 g m−2) in 1992–93. There were no trade-offs in subsequent dry grain (9.5 g m−2) and firewood yields (96.2 g m−2) for cutting hedges earlier and more often during the second year, despite much lower precipitation. In pigeonpea recycled nutrients, N, P, Ca, Mg and K, increased proportionally to cut dry matter yields. Yield increases over two years with pigeonpea were highest among all evaluated cropping systems for maize (+150%) and significant for cassava (+66%). Pigeonpea as a simultaneous fallow component in cassava–maize intercropping, can help to sustain moderate yields of maize and cassava, provided that insects, nematodes and diseases do not lower its high biomass productivity with continued cropping after two years.

The results of a survey showed that one of the most critical constraints in sesame production in southern Tanzania is sesame flea beetle (Alocypha bimaculata, Jacoby). Local methods used by farmers to control, avoid or compensate for the damage caused by flea beetles were identified. Farmers regarded resistance against the flea beetle as a very desirable objective in a sesame breeding programme, but a number of other important characteristics, such as disease and drought resistance, was also mentioned. At three locations in southern Tanzania twenty sesame varieties were screened for resistance to flea beetles by counting the number of surviving plants at the 1–2 true pairs of leaves growth stage. The results suggest that there were significant differences between varieties in terms of their susceptibility to this pest, which suggests that a resistance breeding programme may be successful.

Topsoil nitrate was measured every few days under grass, sorghum and bare fallow at two sites in lower Embu District, Kenya. Concentrations were highest soon after the start of rain and within a week most nitrate had been lost. Under bare fallow and sorghum, nitrate concentration increased during dry weather and losses occurred when soil water content exceeded 10%. Modelling of soil water movement indicated that there was no leaching below the rooting zone, so denitrification may be the main route for loss of soil nitrogen. Application of nitrogen fertilizer early in the season is not recommended as it would be lost with the native soil nitrate. The use of phosphorus fertilizer at the phosphorus-deficient site reduced soil nitrate concentration throughout the season and increased nitrogen uptake by sorghum. At the phosphorus-rich site, vigorous sorghum growth kept the soil nitrate concentration low, even though much nitrate was produced and lost in bare fallow. Adequate soil phosphorus is considered essential if the natural nitrogen resources of the soil are to be managed efficiently.

To enhance the diversification of rice varieties for the specific needs of farmers through a participatory varietal selection (PVS) approach, four pre-release and one released chaite rice varieties were distributed in 20 villages of the Western Hills of Nepal in 1991. A survey conducted during June 1993 found that varietal diversity increased in all locations and in 80% of the study area at least two new rice varieties were reported where only CH 45 was grown before the distribution. On-farm varietal diversity was further enhanced by farmer-to-farmer dissemination of new rice varieties. All the rice varieties tested were adopted, but the adoption level varied between locations. Of the households surveyed 37% were growing the new rice varieties and a further 57% were aware of those varieties within two years of introduction. The PVS approach provided farmers with the benefits of new genetic materials five to six years in advance of the formal system and with minimum effort. Institutionalization of the PVS approach and the use of the farmers' network of information and seed exchange, involving relevant grassroots level institutions, can improve the effectiveness of the variety evaluation system.

Two glasshouse experiments were conducted at the Department of Plant Production, University of Helsinki, Finland, to study the effect of foliar application of aqueous 0.1 and 0.3M glycinebetaine on leaf growth of drought-stressed tobacco (Nicotiana tabacum L.) cv. Samsun. Leaf nitrogen and glycinebetaine concentrations were estimated periodically. After 16 days of treatment, leaves were harvested, and fresh and dry weights, and area determined. The number of total and green leaves was also recorded. Significant increases in leaf fresh and dry weights, and area were associated with glycinebetaine application, especially when drought was induced early. The application of 0.3M glycinebetaine reduced leaf area duration. Residual leaf glycinebetaine content remained high 16 days after application. The results indicate that exogenous glycinebetaine has the potential to improve the drought tolerance of leaf tobacco.

Four contrasting varieties of cassava (Manihot esculenta Crantz) were grown for ten months over two consecutive seasons in hillside Colombia with or without nitrogen, phosphorus and potassium fertilization. Sequential harvests were every two months. The two varieties well adapted to acid soils (CM 489-1 and CM 523-7) had higher yields and exhibited the strongest response to fertilization, highest solar radiation use efficiency, earlier canopy ground cover and higher leaf litter accumulation. The export of potassium, however, was no higher than in the other genotypes, mainly as a consequence of higher potassium use efficiency in the adapted varieties. Significant amounts of total calcium, magnesium and nitrogen accumulated in the whole plant were recycled in the leaf litter during growth. High-yielding varieties adapted to acid soils did not exacerbate soil fertility problems of cassava crops; rather, they were beneficial. Certain growth parameters could be used to select for varieties with low impact on soil fertility.

The effects of temperature (14–19°C) and mean soil water deficit (up to 90 mm) on the shoot extension and development rates of six contrasting tea clones in southern Tanzania were studied 10–28 months after field-planting. An exponential function with two constants, an initial shoot length and a relative extension rate (r), provided a reliable description of the length of axillary shoots for the period from the release of apical dominance to the unfurling of three true leaves. Apparent base temperatures (Tbe) and thermal extension rates (ρe), derived from linear relations between r and the mean air temperature, were able to explain 85–92% of the seasonal variation in r for each fully irrigated clone. The mean soil water deficit was used to explain 80% of the remaining variation in r across three drought treatments.

Clonal differences in the values of Tbe (8.9–11.3°C) and ρe (7.8–11.8 × 10−3 mm mm−1 (°Cd)−1) can be used to explain differences in the seasonal distributions of yield. The apparent base temperatures for development (Tbd) were consistently 1.7–3.4°C below those for extension; an observation supported by the effects of temperature on the length of shoot internodes. The results are compared with those obtained in other studies and are discussed in terms of clonal selection, the choice of harvest intervals and the prediction of green-leaf quality.