Dryland agriculture is critically important to food security and rural livelihoods in Tanzania, but crop production is seriously constrained throughout the semi-arid lowlands by the rainfall regime. A major challenge is to develop improved cropping systems to alleviate the moisture constraint. Experimental evidence indicates that adoption of rainwater harvesting systems can bring benefits, but the restricted spatial and temporal extent of the experimental work leads to difficulties in extrapolation. This paper shows how the PARCHED-THIRST model can add value to the experimental results and provide important insights into their transferability. The model is seen as an aid to researchers, planners and extensionists in interpreting experimental results and designing locally appropriate interventions. Simulation based on 30 years of daily meteorological data provides an opportunity for temporal extrapolation. The long-term simulation allows an objective assessment of the risks and benefits associated with alternative rainwater harvesting systems. Simulation for different soils and modified rainfall regimes permits objective analysis of spatial transferability of experimental results to any other site for which rainwater harvesting interventions might be considered. It is shown that macrocatchment rainwater harvesting reduces drought risk within the target area, but may bring a serious risk of erosion due to excessively high flow rates. The overall assessment of the twin-track approach (experimentation+simulation) is that rainwater harvesting has potential for increasing productivity and sustainability of maize cropping systems in semi-arid Tanzania provided that the innovations are properly matched to the site-specific environmental conditions.

Cassava, a crop widely adapted in the tropics, has the important attribute of withstanding adverse environmental conditions better than do many other staple crops. The performance of an individual genotype, however, is influenced by the environment in which it grows. In Malawi, the heterogeneity of agro-ecologies requires the cumbersome and costly assessment of new cassava genotypes at many sites. This study was conducted, therefore, to test the feasibility of selecting only a few locations for cassava evaluation that would be representative of all the agro-ecologies in which cassava is grown in Malawi. Enormous environmental effects, largely contributed by the interaction between season and location, were manifested. Genotype×environment interaction, due largely to a third level interaction (genotype×season×location), was highly significant for all the traits studied. A principal component analysis scatter plot showed no particular grouping of environments, but a pair-wise comparison showed that some of the locations had limited genotype×environment interaction, indicating that it would be sufficient to use one of these sites for evaluating these traits. The value of the residual was often large, probably as an effect of environmental heterogeneity in the test sites. The authors conclude that cassava genetic improvement will continue to be slow if Malawi is used as a single breeding zone. They recommend a much finer grouping of the locations and the use of smaller plot sizes to allow more clones to be tested at more sites for the same cost. Locations may be selected for intensive cassava breeding work from those that give the best discrimination between genotypes while having insignificant genotype×environment interactions in a relatively large number of environments.

In a field experiment initiated in the 1972/73 rabi season (November–March) on an acidic soil in the western Himalayas of India, continuous cropping for 25 years with nitrogen fertilizer alone reduced the grain yields of maize (Zea mays) and wheat (Triticum aestivum) to almost zero. Based on five-year moving average values, the application of 100% NP resulted in a reduction of grain yields of maize and wheat crops by 67 and 19% respectively during 1994 to1998 compared with 1973 to 1977. The crops did not respond to K application during the initial years of cropping, but a significant reduction in crop yields was observed in the absence of K after five years of continuous cropping. With the integrated use of chemical fertilizers (100% NPK) and farmyard manure, the pooled grain yields of 25 years were about 4600 and 3300 kg ha−1 for maize and wheat respectively. Corresponding yields with the use of chemical fertilizers and lime (100% NPK + lime) were 4000 and 3100 kg ha−1 respectively. The use of nitrogenous fertilizer alone aggravated the problem of soil acidity by lowering the pH from 5.8 to 4.7 after 25 years. In almost all the treatments, the organic carbon content increased: a marginal decrease occurred in the 100% N and control plots. Even under optimum application rates, the available N and K status decreased with time. Again with the exception of the control plots, the contents of the Diphenyl Triamine Penta Acetic Acid-extractable micronutrients were sufficiently high for the crops in all the treatments. The available sulphur decreased with the use of di-ammonium phosphate as the P source, which drastically reduced the yields of both crops. In the N-alone treatment the weed flora also changed with the change in soil environment. There is a need to review fertilizer recommendations in this region of India.

In Mozambique the sweet potato (Ipomoea batatas) is grown as both a leafy vegetable, the terminal shoots or vines being progressively harvested during the season, and as a root crop. This paper reports the development of crop yield/water-use production functions using data from two irrigation experiments designed to evaluate the effects of water availability and vine harvesting practices on the productivity and water use (ETc) of sweet potato (cv. TIS 2534) during two contrasting seasons in the south of the country. As the frequency of vine harvesting increased, the water-use efficiency (WUE-ETc) for vine production (dry mass) increased from 1–2 to 4–5 kg ha−1 mm−1 during the rainy season, and from 1 to 9 kg ha−1 mm−1 during the dry season. By contrast, there was a corresponding reduction in the WUE-ETc for storage root production from 14 to 8–9 kg ha−1 mm−1 during the rainy season, and from 23 to 15–17 kg ha−1 mm−1 during the dry season. For the total yields (vines plus storage roots) the WUE-ETc during both seasons were independent of the vine harvesting treatment. Separate lines represented each season, the slopes of which were 13 kg ha−1 mm−1 in the rains, and 24 kg ha−1 mm−1 in the dry season. When, however, crop water-use was normalised using either the seasonal mean daily total of incoming solar radiation, or reference crop evapotranspiration (ETo), but not the saturation deficit of the air, a single common, linear relationship with yield resulted. Possible explanations for this are considered. Plotting relative yields against relative rates of water-use, also gave consistent results that could have general application for predicting the effects of water availability on productivity. Vine yields, when harvested frequently, were less sensitive to drought (yield response factor, Ky=0.7–0.9) than storage roots (Ky=1.2). For total production (vines plus roots), the sensitivity to water stress (Ky) increased, from 0.9–1.0 to 1.2, as the interval between vine harvests increased. Farmers in southern Mozambique trying to maximize total yield during the season, under conditions of water uncertainty, should harvest vines at intervals of not more than 14 days.

The first ever on-farm soil fertility research in the rainfed lowland rice (Oryza sativa) fields of Sukumaland, in northwest Tanzania, was carried out between 1990 and 1996 in response to farmers' complaints about declining rice yields. From diagnosis to extension, the research approach followed that of the International Maize and Wheat Improvement Center (CIMMYT). In 1990/91, rice yields in the Maswa district increased sharply when nitrogen at a rate of 30 kg ha−1 in the form of urea was broadcast in flooded rice fields at tillering. Similar research was subsequently conducted in other parts of Sukumaland to evaluate this type of low-dose nitrogen application under varying circumstances. In 1995/96, higher doses of nitrogen (60 and 120 kg ha−1) and a high dose of phosphorus (17.5 kg ha−1) were applied for comparison in Sengerema district. Between 1990 and 1996, the average increase in rice yield from the application of 30 kg N ha−1 varied between 463 and 986 kg ha−1. In 1995/96, the same application of N was more economical than both 60 and 120 kg N ha−1, and no phosphorus deficiency was found. The deteriorating ratio between the price of rice at the farm gate and that of urea, however, threatens the adoption of this technology by farmers. Adaptability analysis showed that the relatively small differences in response per field (environment) in all years did not justify a need for multiple different extension messages. Until more detailed recommendations can be made, therefore, a single dose of 30 kg N ha−1, in the form of urea, applied to rice at tillering is recommended for the whole of Sukumaland to reverse the decline in yields. Further on-farm research should concentrate on improving the efficiency of nitrogen fertilization and on determining the optimum rates of other major nutrients to refine this initial recommendation.

The small amount of on-farm research that has been carried out with fodder trees in Africa has mainly involved exotic species selected by researchers rather than farmers. In this study, farmers who had participated in a pre-planting survey chose, during feedback meetings, seedlings of indigenous, naturalized and some exotic fodder trees from a nursery. They planted them on-farm, where the planting niches, management, biomass production, and animal response were evaluated over a short period, and compared with the results of the earlier survey. The farmers came from three agro-ecological zones in central Kenya with respective mean annual rainfall totals of 775, 950 and 1300 mm. Data were collected at two months, one year and two years after planting. The feedback meetings proved important to the understanding of discrepancies between the earlier survey results and farmers' practices. Survival and growth of seedlings, manure application by farmers, and percentage of trees pruned for fodder, differed among species. The farmers showed a strong interest in experimenting with and actively cultivating local fodder trees.

This study reports the results of model simulations of the potential yield of tea in north-east India. The CUPPA-Tea model, developed using data from a high-altitude site close to the equator in East Africa, was validated against the yield data from irrigation experiments conducted on contrasting soil types at Siliguri and Tezpur in the tea growing region of north-east India. The close correspondence between observed and predicted yield and yield distribution suggests that the model is applicable in north-east India. The model was used to simulate the yield response of tea to drought and irrigation using daily weather data for seven years (1983–89) at Siliguri and 14 years (1974–85) at Tezpur. On a clay loam soil at Siliguri, with an available water capacity of about 200 mm m−1, the predicted mean reduction in yield was 1.5 kg ha−1 for each 1 mm reduction in evapotranspiration. However, there appeared to be no reduction in evapotranspiration until the soil water deficit reached about 240 mm. By contrast, at Tezpur on loamy sand, with an available water capacity of about 100 mm m−1, evapotranspiration was reduced once soil water deficits exceeded about 85 mm. There was then a mean reduction in yield of 2.2 kg ha−1 for each 1 mm reduction in evapotranspiration. At both sites, even when soil water deficits were not limiting, the predicted year-to-year variation in yields was about 500 kg ha−1. The results highlight the soil-related differences in response to irrigation and the benefits of using process-based simulation models to investigate the potential yields over long periods.

Phosphorus (P) deficiency is a major constraint to pearl millet (Pennisetum glaucum L.) growth on acid sandy soils of the West African Sahel. To develop cost-effective fertilization strategies for cash poor farmers, experiments with pearl millet were conducted in southwestern Niger. Treatments comprised single superphosphate hill-placed at rates of 1, 3, 5 or 7 kg P ha−1 factorially combined with broadcast P at a rate of 13 kg ha−1. Nitrogen was applied as calcium ammonium nitrate at rates of 30 and 45 kg ha−1. At low soil moisture, placement of single superphosphate in immediate proximity to the seed reduced seedling emergence. Despite these negative effects on germination, P placement resulted in much faster growth of millet seedlings than did broadcast P. With P application, potassium nutrition of millet was improved and seedling nitrogen uptake increased two- to three-fold, indicating that nitrogen was not limiting early millet growth. Averaged over the 1995 and 1996 cropping seasons, placed applications of 3, 5 and 7 kg P ha−1 led to 72%, 81% and 88% respectively, of the grain yield produced by broadcasting 13 kg P ha−1. Nitrogen application did not show major effects on grain yield unless P requirements were met. A simple economic analysis revealed that the profitability of P application, defined as additional income per unit of fertilizer, was highest for P placement at 3 and 5 kg ha−1.

This paper presents the results of:

(a) On-farm trials (eight) over a two-year period designed to test the effectiveness of leguminous cover crops in terms of increasing maize yields in Igalaland, Nigeria.

(b) A survey designed to monitor the extent of, and reasons behind, adoption of the leguminous cover crop technology in subsequent years by farmers involved, to varying degrees, in the trial programme.

Particular emphasis was placed on comparing adoption of leguminous cover crops with that of new crop varieties released by a non-governmental organization in the same area since the mid 1980s. While the leguminous cover crop technology boosted maize grain yields by 127 to 136% above an untreated control yield of between 141 and 171 kg ha−1, the adoption rate (number of farmers adopting) was only 18%. By way of contrast, new crop varieties had a highly variable benefit in terms of yield advantage over local varieties, with the best average increase of around 20%. Adoption rates for new crop varieties, assessed as both the number of farmers growing the varieties and the number of plots planted to the varieties, were 40% on average. The paper discusses some key factors influencing adoption of the leguminous cover crop technology, including seed availability. Implications of these results for a local non-governmental organization, the Diocesan Development Services, concerned with promoting the leguminous cover crop technology are also discussed.

Striga gesnerioides is a serious parasite of cowpea (Vigna unguiculata) on the Abomey plateau of southern Bénin. S. gesnerioides in southern Bénin has been characterized as a race that is different from those found in the dry savanna of West Africa. IITA breeding line IT93KZ-4-5-6-1-5 was used in a trial to study fertilizer response of cowpea on two farmers' fields in 1998. No emerged S. gesnerioides plants were observed on this breeding line while heavy parasite loads were observed on local cultivars growing in adjacent fields. In 1999, IT93KZ-4-5-6-1-5 was grown again on seven fields and no emerged S. gesnerioides plants were observed. In 2000 the same line was grown on 18 farmers' fields with local cowpea cultivars in adjacent control plots. The mean number of S. gesnerioides plants on the local cultivars was 0.92 per cowpea plant (s.d. 0.72) while the improved line carried none. The mean yield gain from the improved cowpea line was 156% and varied from 20 to 680 kg pods ha−1. Thus it appears that progress has been made in developing cowpea lines that are resistant to S. gesnerioides and the potential impact of adoption of resistant cowpea is high.

India is the world's dominant producer of crude Mentha arvensis oil, and exporter of its processed derivative, natural menthol. This status has been achieved through superior product price competitiveness resulting from progressive agronomic improvements that include a transition from cultivation of this mint as a semi-perennial to an annual crop. This paper describes and analyses recent developments in Indian M. arvensis cultivation.

Observations were made during 1998/99 and 1999/2000 in the fields of 295 farmers in the central Indo-Gangetic plains area. The criteria recorded included: mint cultivars used; the planting, weeding, irrigation, pesticide application and harvesting schedules; crop health in relation to diseases and pests; yields of the harvested herb and the distilled essential oil; the oil menthol content; and the net income.

Most of the farmers used the cultivar, Kosi. The performance and returns of Kosi transplanted in April had proved superior to Kosi and Shivalik suckers and to transplanted Shivalik. The returns from late-transplanted Kosi were in the range of Rupees (Rs) 25 000 to Rs 40 000 ha−1 (Rs 50≈US$1), similar to those from Kosi suckers and much higher than for both transplanted and suckers of Shivalik. This survey revealed that rotations of rice (Oryza sativa) with chickpea (Cicer arietinum) and mint, and pigeonpea (Cajanus cajan) with wheat (Triticum aestivum) and mint could be profitable in the central Indo-Gangetic plains. They also conserved inputs and were preferable to the conventional rice-wheat rotation. A re-survey in the summer of 2002 confirmed that recommendations made to farmers have led to the adoption of the Kosi and Himalaya cultivars in 80% of the survey region and that 70% of mint cultivation in the area occurs in the summer season.

The effects of different seed treatment methods and durations of exposure on germination of wild okra (Corchorus olitorius), a traditional leafy vegetable consumed in many parts of Africa, were investigated. Seeds were exposed to seven treatments (hot water at 80 °C, hot water at 100 °C, dry heat at 80 °C, dry heat at 100 °C, concentrated sulphuric acid, 10% hydrogen peroxide and water at room temperature) for periods of up to 30 minutes. Exposure for five to 15 minutes in hot water at 80 °C was the most effective treatment for enhancing germination (>90%), followed by five minutes in hot water at 100 °C (80%) and 30 minutes in sulphuric acid (57%). Other treatments were less effective. Soaking seeds in hot water at approximately 80 °C for about 10 minutes can be recommended, therefore, to farmers as a simple, cheap and very effective way of germinating wild okra.

On-farm trials were conducted in the northern Guinea savanna of Nigeria. The objective was to compare integrated Striga hermonthica control measures (soyabean or cowpea trap crop in the first year followed by maize resistant to Striga in the second year) with farmers' traditional practices (cereal-based cropping systems) under farmer-managed conditions.

Integrated control proved to be highly effective in terms of reducing Striga incidence both in terms of reduced seed density in the soil and decreased infection in maize. Resistant maize following the soyabean trap crop yielded 1.58 t ha−1 of grain and out-yielded local maize following traditional practices by more than 80%. Similarly, the overall productivity over the period of the experiment was highest with the integrated control treatment using soyabean. Conversely, resistant maize after the cowpea trap crop yielded only 0.92 t ha−1 (possibly due to the poor performance of the cowpea crop in the first year), and maize yields were similar to those obtained with farmer practices. Initial Striga seed density in the soil was negatively correlated (r = −0.33) with soil nitrogen, but nitrogen-fertilizer application rates did not seem to affect the level of Striga infection in maize.

A reliable means to induce sprouting in dormant seed tubers of yams (Dioscorea species) is required to enhance flexibility in planting date and rate of propagation of the crop. Experiments were conducted to assess the potential of two gibberellin inhibitors, Uniconazole-P and Prohexadione-calcium, to induce sprouting in tubers from three varieties of D. rotundata and four of D. alata. Uniconazole-P and Prohexadione-calcium shortened the period of dormancy in tubers of some varieties. In others, they either had no effect or extended dormancy. The varietal responses were also influenced by whether the tubers were treated at harvest (before shoot senescence) or four weeks after harvest by which time shoots had senesced fully. Tubers stored in darkness at a constant temperature of 30 °C sprouted earlier but lost weight faster than did those stored under natural daylight and ambient temperature. The apparent slow action of Uniconazole-P and Prohexadione-calcium, and their variable effectiveness in relation to variety, would limit their usefulness in seed or breeding programmes, or in determining treatment timing and the best storage environment for the tubers after treatment.

Readers may be interested to know about the above publication received but not reviewed because of its limited relevance to the majority of readers of Experimental Agriculture.