Sequential sowings were carried out at Dodoma, Tanzania, to examine the effect of changing climatic parameters on the growth and yield of bambara groundnut (Vigna subterranea). Sowings took place on 4 January, 4 February and 4 March 1994; 4 and 24 January, and 13 February 1995; 4 and 21 January, and 7 February 1996. Rainfall during the crop life cycle varied from 163 to 611 mm, mean photoperiod from 11.82 to 12.09 h d−1 and mean temperature from 22.6 to 24.4 °C. In 1994, the highest pod yields were achieved at the earliest sowing date, with a maximum of 2.87 and 1.42 t ha−1 for the red- and cream-seeded landraces, representing pod harvest indices of 0.56 and 0.34 respectively. A 30-d delay in sowing caused >60% reduction in pod yield, and a further 30-d delay resulted in no pods at all. Similarly, in 1995 successive delays in sowing caused dramatic yield declines, and the maximum yield was much lower, at 0.44 t ha−1. In 1996 there was no significant difference in pod yields between the two early sowing dates for the red-seeded landrace and yields were again lower than in 1994 with a maximum of 1.02 t ha−1. Differences in dry matter production between sowings and years were attributed mainly to differences in the amount and distribution of rainfall and to declining temperatures towards the end of the season; however, partitioning to pods was remarkably consistent across sowings.

A study was conducted to determine the population sizes of indigenous pigeonpea (Cajanus cajan)-nodulating rhizobia and responses of the crop to rhizobial inoculation in soils under smallholder management. Rhizobia populations were determined in 21 soils from three different agro-ecological regions of Zimbabwe using the plant infection most-probable-number technique. Pigeonpea response to rhizobial inoculation was tested in five soils representative of low, medium and high rhizobia populations. Pigeonpea rhizobia ranged from undetectable to 121 cells per g soil compared with 16 to 159 cells per g soil for cowpea (Vigna unguiculata) which was used for reference. Soils with high cowpea rhizobia counts had relatively low counts of pigeonpea rhizobia and vice versa, showing that the two legumes associate with different subgroups of rhizobia. Poor soil organic matter, low soil moisture at sampling, low pH and low clay content of the soils had a significant negative effect on rhizobial counts. Organic matter appeared critical for maintenance of high populations of indigenous rhizobia in the mostly sandy soils sampled. Lack of pigeonpea response to inoculation in all the soils tested despite the low initial rhizobial populations could be the result of within-season proliferation of indigenous populations which are competitive and effective. There was evidence of rapid build-up of pigeonpea-compatible rhizobia within one growing season when the crop was first introduced. It was concluded that effective pigeonpea rhizobia occur in many arable soils of Zimbabwe. However, to fully exploit biological nitrogen fixation and maximize yields of pigeonpea, highly efficient, adapted and competitive indigenous rhizobial isolates must be identified and evaluated.

Two pigeonpea (Cajanus cajan) cultivars GT-100 (determinate type) and BDN-2 (indeterminate type) were planted on three sowing dates (30 June, 20 July and 9 August) in 1993 and 1994. Pigeonpea sown on the earliest date attained the highest leaf area index (LAI), absorbed the largest amount of photosynthetically active radiation (PAR) and produced the highest total dry matter (DM). The differences in biomass and seed yield among sowing dates were largely ascribed to totals of PAR absorbed and dry matter produced, especially in the reproductive phase. The high LAI persistence and PAR interception, coinciding with the podding phase, appeared to be mainly responsible for the increased yield in early sowings. Radiation use efficiency decreased as sowing was delayed, but did not have much effect on DM accumulation in various phases nor on final yields. Although the extinction coefficient was not influenced by sowing dates, it was inversely related to leaf area index in both cultivars. Between the cultivars, the differences in biomass reflected the differences in PAR absorbed and DM accumulation, depending upon leaf area development and growth duration. The cultivar GT-100 had a higher seed yield and harvest index than BDN-2 due to more of the DM produced being partitioned into pods during the reproductive phase on account of its determinate growth habit. Early sowings of determinate cultivars could maximize both vegetative and reproductive growth, capture more light and produce more seed yields under rainfed conditions.

Drawing on ideas advanced during the course of an international workshop, this paper reviews agricultural research priorities in the dry areas of the developing world, where economic and demographic pressures are driving a rapid intensification of production, even on very fragile lands. Current priorities for research and research funding reflect a strongly felt need, among governments and donor agencies, for rapid problem-solving to relieve poverty. In this context, traditional long-term agronomic research trials, with essentially strategic and basic objectives, are often seen as irrelevant. Yet the need to anticipate problems impending in environments where intensification is consuming the natural resource base and degrading the production potential is an urgent one. To accommodate this need to the perceptions of the different stakeholders, a new research paradigm, ‘Anticipatory Long-Term Research for Sustainable Productivity’, is proposed. It invokes a holistic approach, building a longer view and greater dynamism into traditional agronomy, enhancing ties to socio-economic and resource management disciplines, and reconciling the immediate demand for high-yielding technology with the strategic issues of sustainable production.

The aims and design features of historic long-term experiments are discussed and modern perceptions are presented of long-term studies to address sustainability and research impact in farming communities. A research strategy is proposed combining on-station and on-farm research whose results influence dynamically the directions of the component studies. Guidelines are given for choosing well-controlled designs, appropriate data and sampling procedures and for maintaining quality control in data collection and indicator formation over anticipated lengthy periods of time.

High temperature is a major problem in groundnut (Arachis hypogaea) producing areas of arid and semi-arid tropics whether rainfed or irrigated. Knowledge of the maximum air temperature that plants can tolerate will help in selecting genotypes adapted to high temperature. The objectives of this study were to evaluate the growth, development, yield and yield components of different groundnut genotypes grown under elevated temperatures. Five groundnut cultivars and lines were grown in the field from 1984 to 1986 in Vertisols. Measurements on flowering, growth and development were carried out frequently. Thermal units required for groundnut cultivars to reach maturity were 2247 °C days. Larger kernel size, higher pod:total growth ratio and longer pod-filling period were the most important parameters for yield increase in heat-stressed environments. Early Bunch gave a stable yield across environments and was found to be more tolerant of high temperature. Cultivars differed in their pod:total growth ratio and Early Bunch had the highest ratio.

Yields of rainfed barley (Hordeum vulgare) for feed and forage in small-ruminant production systems in low-rainfall areas of North Africa and West Asia are limited by low and erratic water availability. Therefore, the testing of conservation farming techniques, effective in dryland systems elsewhere, is often suggested. Seven-year results from a typical site in northern Syria showed that zero-till (direct-drill) systems with cereal residue retention may marginally enhance soil moisture status, but the yield effect on barley, either monocropped or rotated with vetch (Vicia sativa), was small and non-significant. In the vetch–barley rotation, a small and fairly consistent benefit to vetch was observed, amounting to a 20% yield increase in vetch hay. Given smallholders' strong preference for barley and reluctance to grow vetch as an alternate crop, there is little in this result to encourage the promotion of zero-till conservation techniques in these farming systems.

This study, designed to investigate trade-offs between forage yields, feed quality and labour use in fitting a newly introduced crop, Lablab purpureus, into an existing farming system, adopted distinct applied and adaptive research approaches. The applied research component compared forage yields, nutrient contents and feeding qualities of L. purpureus harvested at 100, 114, 128 or 142 d after planting in an on-farm trial. The adaptive research investigated how farmers incorporated L. purpureus cultivaton into their farming calendars given the constraints of labour for planting and harvesting.

Harvested dry matter yields from the on-farm experiment were 1.28, 1.98, 1.74 and 1.44 t ha−1. The resultant hays were fed to four groups each of five Bunaji lactating cows in an on-farm feeding trial over a 10-week period. Ranking the various harvests for yield, the hays for feed value, animal responses to the feed consumed and the quality of animal products on a linear scale (1 to 4) to estimate a composite utility index, showed that harvesting at 114 d after planting gave the most desirable overall benefits.

The survey for adaptive research showed bimodal rainfed and irrigated cultivation patterns for L. purpureus. Peaks were observed in September (31% of farmers) and November (24%) for planting, and in December (39.4%) and March (15%) for harvesting. Forage yields were, however, highest for the 3% of farmers who afforded labour to plant in July and harvest in October. The lack of congruence between highest forage yield periods and the peaks of cultivation of L. purpureus by farmers in the study area was mostly explained by very significant differences in the cost of hired and household labour in different months. Thus, socio-economic and cultural factors, including stakeholders' participation, need to be considered along with applied research results when recommending forage species for farmers' adoption.

Isolation of Colletotrichum gloeosporioides (causal agent of yam anthracnose) from soil is difficult because of the presence of fast growing micro-organisms like Rhizoctonia spp., Rhizopus spp., Penicillium spp. and bacteria. Therefore, a semi-selective medium consisting of antimicrobial agents such as pencycuron 50 mg L−1, tolclofos-methyl 10 mg L−1, streptomycin sulphate 100 mg L−1, chloramphenicol 100 mg L−1 and chlortetracycline 100 mg L−1 was developed using potato dextrose agar as the basal medium. The procedure involved screening of candidate fungicides, developing a suitable combination of antimicrobial agents, assessing the proposed semi-selective medium and adjustment of pH of the medium to 5. The effectiveness of the medium is due to selective inhibition of micro-organisms by the antimicrobial agents and the development of distinct salmon-pink colonies of C. gloeosporioides.

Seeds of watermelon (Citrullus lanatus) with a moisture content of 7.2% were heat treated at 60°C for 15 or 20 h, and their performances compared with that of untreated seeds. Radicle emergence of treated seeds was only slightly lower than that of the control when germinated at 25°C. To ascertain the effect of the treatments on hypocotyl growth (emergence) per se, radicle emergence was synchronized in all treatments and emergence from deep plantings in sand determined. Small differences in emergence were observed when seeds were germinated at 25°C, but seed treatment resulted in large reductions in emergence at 15°C. This demonstrates that the hypocotyls developing from heat-treated, germinated seeds are less able to withstand low germination temperatures, which could be an important factor in emergence failure. Levels of free abscisic acid (ABA) were higher at 15°C than at 25°C, with hypocotyls of treated seeds containing more free ABA and its breakdown products than the control at the lower temperature.

Photosynthetic photon flux density (PPFD) has variable effects on cotton (Gossypium hirsutum) plant growth, development and lint yield, but little is known about the responses of yield, yield components and fibre quality to low PPFD at different developmental stages. Field studies were conducted in 1993–95 to determine the effects of an 8-d period of shade (63% reduction in PPFD) imposed at different developmental stages (pinhead square (PHS), first flower (FF), peak flower (PF) and boll development (BD)) on cotton growth, lint yield, yield components and fibre quality. Shade for 8 d at the early square stage did not affect cotton growth and yield. Shade during FF, PF and BD increased fruit abscission, and decreased specific leaf weight, dry matter accumulation, lint yield, fibre micronaire, and fibre strength compared with the unshaded control. The detrimental effect of shade on yield increased with later growth stages. Shade at the FF, PF and BD stages decreased lint yield by 18, 34 and 52% in 1993, and by 18, 21 and 29% in 1994 respectively compared with the unshaded control. Shade caused the greatest decrease in the dry weight (DW) of fruits (squares + flowers + bolls) and the smallest decrease in leaf DW among leaves, stems and fruits. Of the three yield components (boll number, boll weight and lint percentage), boll number was the most sensitive to shade. Shade during squaring and early flowering had little effect on average boll weight, and did not affect lint percentage. Decreased yield of cotton shaded during FF and PF appeared to be related to decreased boll retention. Shade during BD decreased both boll retention and average boll weight. Under shade conditions, fibre micronaire and fibre strength were decreased, while other fibre properties were unaffected. These results help explain the yield response of field-grown cotton to low PPFD stress.

Trends over time in annual crop yields potentially provide measures of the likely long-term sustainability of cropping systems. However, where large annual variability in the growth environment is responsible for most of the large year-to-year yield differences, appropriate analytical techniques must be developed to distinguish real long-term trends from the ‘background noise’. This paper presents models for the estimation of time trends in the yield data from crop rotation systems and discusses the results of applying these models to yield values from two types of long-term trial involving barley, each conducted at two sites in northern Syria.

The models used were linear with respect to time (years) and allowed for seasonal effects by means of a quadratic relationship on total rainfall and a linear relationship on planting date. A more complex model might account for more of the variance, but restrictions were imposed by the limited number of degrees of freedom (number of years of data less one) and the choice of meaningful single-valued parameters of growth-season conditions. For many experimental treatments the model accounted for less of the total variance at the wetter site. This may be due to seasonal bufferring by soil moisture stored at depth from one year to the next, and future iterations of the analysis will try to allow for this. The appropriateness of the linear time function is also questioned, and alternative functions will be tested along with alternative structures for plot errors over time.

A participatory on-farm research project was started at Tah village, 100 km south of Aleppo, Syria in 1984. The objective was to introduce annual medics (Medicago spp.) as a replacement for fallow in dryland wheat (Triticum durum)/fallow farming systems. About 50 farmers were directly involved in the project while an equal number were passive participants, receiving seed but no technical advice. The project did not attain its original intention of introducing a ley-farming system into Syria. The Syrian Government decided to eliminate fallow – which was to receive medics – in the agricultural planning for 1989–90. An array of profitable crops such as lentil (Lens culinaris), cumin (Cuminum cyminum), sesame (Sesamum indicum) and watermelon (Citrullus lanatus), together with price supports on wheat presented formidable economic competition to the use of pasture on cultivated land. Technical constraints include insufficient farm size, lack of a medic phase in every year, deep ploughing, overgrazing and uncontrolled grazing. The Tah project experience also suggests that feed production projects must include an assessment of how the feed benefits animals and be designed to demonstrate the economic advantages of integrating livestock and crops in dryland farming systems.

A study was conducted to examine the effects of the time and severity of cutting lablab (Lablab purpureus), both in sole crops and intercropped with maize (Zea mays). The effects of cutting management on the dry matter (DM) yield, leaf area index (LAI), nodulation, the nitrogen (N) yield, and N fixation by lablab were measured, as were the total DM yield, grain yield, and a yield component of maize.

Two dates of lablab cutting (40 and 60 d after sowing (DAS)), and two heights of cutting (30 and 20 cm above the soil surface) were applied in factorial combinations as well as a no-cutting treatment. Maize was not cut either in monoculture or intercrops.

Intercropping lablab with maize increased the grain yields of maize from 3.2 t ha−1 to 3.8 t ha−1 if lablab was cut at 40 DAS, but delaying cutting until 60 DAS had no effect, while uncut lablab reduced the yield of maize by 10%. Intercropping lablab with maize reduced the total N yield of lablab by over 50%, from 302 to 132 kg N ha−1. Regardless of cropping system, the total N yield of lablab at 130 DAS was 24% less when lablab was cut at 60 DAS compared with 40 DAS. Cutting height did not significantly affect the total N yield of lablab. Although both cropping system and cutting time significantly affected the total N fixation by lablab at 130 DAS, there were no significant effects of cropping system, cutting time or severity of cutting on the proportion of N in lablab derived from N fixation.

Values of Land Equivalent Ratio (LER) which ranged between 1.2 and 1.6, and of Area Time Equivalent Ratio (ATER) which ranged between 1.0 and 1.4 for intercrops of maize with lablab indicated a definite advantage of intercropping. LER and ATER values were greatest when lablab was uncut, and least when lablab was cut at 60 DAS.

A field experiment was conducted to study the influence of 8- to 9-year-old nitrogen-fixing tree species (Faidherbia albida, Acacia ferruginea, Albizia lebbeck) and fertilizer nitrogen levels on yield and nitrogen uptake of intercropped cowpea (Vigna unguiculata). Grain and dry fodder yields of cowpea were reduced by 9% to 27% and by 11% to 35% respectively compared with the sole crop. Application of fertilizer nitrogen produced significantly higher yields. Nitrogen uptakes by leaves, stems and grain of cowpea were significantly reduced under A. ferruginea and A. lebbeck, but cowpea as sole crop and under F. albida showed no differences. Nitrogen uptake in different plant parts increased with increase in fertilizer nitrogen level.

As cassava is grown mostly by small resource-limited farmers throughout the tropics on low-fertility soils with little fertilization and, due to the large potassium (K) export in harvested roots, genotypes that tolerate low-K soils and respond to K fertilization are warranted. The objective of this study was to evaluate cassava germplasm and identify such genotypes.

Fourteen cultivars of cassava (Manihot esculenta) selected from the core germplasm at CIAT were grown under rainfed conditions for ten months over five consecutive seasons in Inceptisols either with no K application or with 50, 100 or 200 kg K ha−1 applied annually together with adequate nitrogen and phosphorus. All cultivars responded to K application both in terms of root and shoot biomass with the highest yields obtained by CM 507-37 and M Ven 25 in the absence of K application and at high K levels. These cultivars had the highest adaptation indices to low K and the highest K use efficiency for total biomass production. CM 507-37 had the highest K use efficiency for root production at all K levels. Thus, it is desirable to use this material for breeding cassava which is adapted to low-K soils and is able to respond to fertilizer application. However, because of the high hydrocyanic acid (HCN) content in their roots, these cultivars should be crossed with genotypes that are low in HCN, such as HMC-1 and M Cub 74, in order to select lines with both high K use efficiency and low HCN for fresh cassava consumption. Root HCN contents were significantly reduced by application of K across all cultivars. Application of K fertilizer to low-K soils is warranted to minimize health hazards when fresh cassava is used for human consumption.

Field plots of sesame (Sesamum indicum) with six different levels of seed infection with Alternaria sesami were monitored for Alternaria leaf spot severity at Kibwezi, eastern Kenya. The aim of the study was to determine the effect of seed transmission of the pathogen on yield and tolerance level of the fungus in sesame seed. Increase in percentage leaf area diseased and percentage defoliation fitted the Gompertz model more closely than the logistic model. Areas under disease progress curves (AUDPC), infection and defoliation rates varied among the six infection levels. Disease severity increased with increase in seed infection and was least and most severe in plots established with seeds with 0 and 8% infection levels respectively. Yields ranged from 234.9 to 300.1 kg ha−1 compared with 312.5 kg ha−1 for the control, and losses due to seed infection ranged from 4% to 25%. Disease severity was negatively correlated with seed yield, 1000-seed weight and seeds per capsule. Alternaria leaf spot severity had a major effect on the seed weight component of yield. Tolerance level of A. sesami in sesame seed was determined to be less than 2%.

Over a period of seven years farmers participated in trials in which common vetch (Vicia sativa (V)) or chickling (Lathyrus sativus (C)) replaced the fallow in a barley (Hordeum vulgare)–fallow rotation (F) or were introduced into continuous barley cropping (B) – giving a total of four rotations, B–F, B–B, B–V and B–C. Trials on 4 ha, 2 ha per phase of each rotation, were replicated on 6–8 farms. Some vetch and chickling crops were grazed in spring. Mean seven-year dry matter yields were 2.91 t ha−1 for B–F, 4.82 t for B–B, 5.02 t for B–C and 5.32 t for B–V; total crude protein outputs were twice as high from rotations including legumes; and the B–V rotation yielded most metabolizable energy. Realizing the benefit, farmers started to adopt vetch. In 1991 three farmers were growing vetch on 7 ha but by 1997 174 farmers in 15 villages were growing vetch on 420 ha. Forage legumes will not, however, become more widely grown until inexpensive and efficient mechanized methods of harvesting the mature crop are available in order to avoid the high cost of hand labour. Drought and cold tolerance, early maturation and high harvest index may also enhance farmers' interest in forage legumes.

Infection of groundnut by pathogens causing early and late leaf spot diseases is strongly affected by accumulated daily leaf wetness periods and, in the rainy season, temperature is unlikely to severely limit infection. Earlier work relating patterns of leaf wetness to infection, was used to define a daily Wetness Index (WI) which was compared with infection on inoculated plants exposed in the crop for periods of 7 d. Infection was only severe when the 7-d WI total exceeded a value of 2.3. The proportion of leaves with one or more lesions on the main stem was used to assess the amount of inoculum in the crop. When the proportion of diseased leaves exceeded 10% and the WI total exceeded the threshold, application of a fungicide was advised. Successive sprays were separated by at least 14 d and a maximum of three sprays were applied in the growing season. Field trials showed that three sprays gave limited benefit where the disease pressure was severe, but substantial increases in pod and haulm yield were possible with only one or two fungicide applications in locations with less disease pressure.

Increased use of prime agricultural land for cropping and non-agricultural uses in many tropical countries implies that crop residues become more important as a source of feed for livestock. Traditionally, much research on crop residue feeding was done by focusing on laboratory measurements of feed quality but neglect of farmers' perceptions led to disappointing results in the transfer of straw feeding technologies based upon laboratory results. Farming Systems Research (FSR) provides methodologies and concepts that bridge the gap between formal commodity research (including crop residues and by-products) and field application. This paper reports the experiences of a project in India that changed the emphasis from a commodity research approach on improved crop residue feeding to a system approach by using three types of FSR. The paper first reviews the achievements of on-station research on feeding systems for crop residues in terms of treatments (biological, chemical, physical) and in terms of breeding and managing cereal crops for more and better straw. Next, it discusses definitions and problems of FSR as encountered in the project's reorientation of livestock research and development programmes. Finally it summarizes the overall results of the three FSR approaches used in the project.