Interactions between upland rice and three phenologically distinct pigeonpea cultivars were examined on a medium deep Vertisol. The productivity of each intercrop component and its respective sole crop was determined in terms of a crop performance ratio (CPR). The extra-early pigeonpea cultivar recorded the largest partial CPR of grain followed by early and medium genotypes. Spreading genotypes had a larger partial CPR than semi-compact genotypes. However, the CPR of intercropped rice was less (0.65–0.69) with spreading pigeonpeas but exceeded unity with compact types. The canopy structure of pigeonpea appeared to be more important than differences in phenology. A large range of light transmission coefficients (K) existed in pigeonpea (from 0.45 to 0.78) but it is argued that a further reduction in K may not be necessary since intercropped rice yield was unaffected even with a K of 0.64. The relative height of intercropped pigeonpea and upland rice may also determine competitive ability since rice is very sensitive to low light and shading, particularly during the reproductive phase.

The production of lentil (Lens culinaris) was examined in 24 on-farm trials conducted over three seasons (1985/86 to 1987/88) in northwest Syria to examine the main effects and interaction of sowing date, weed control, Sitona weevil control, and fertilizer application (phosphorus and nitrogen). Advancing the date of sowing from late winter (mid January) to early winter (mid-November) resulted in an increased yield of grain and straw in all years across locations. Responses to phosphorus application were also significant in all years. Controlling Sitona weevil increased grain yield by 20% in early sown crops and by 14% in those that were sown late. Weed control by the pre-emergence application of cyanazine and pronamide increased grain yield by 59% in early sowings and by 20% in late sowings. Economic analysis of the improved production practices showed that early sowing combined with the control of Sitona and weeds, and phosphorus application, gave the highest probability of an increase in net revenue with virtually no risk of economic loss to the farmer.

In two field experiments conducted over five successive sugarcane crops (including four ratoon crops), yield decline was associated with a decrease in the organic carbon content of the soil. Reduction in cane yield was minimized by the combined application of farmyard manure and urea, which resulted in an increase in soil organic carbon of 0.023% over the initial status within five years. Incorporation of farm waste materials such as cane trash, rice straw and legume husk into the soil before cane planting also increased the soil organic carbon content. Maximum benefits, however, were produced by the incorporation of legume husk with 150 kg N ha−1. A judicious mixture of chemical fertilizer and organic manures is therefore suggested as a means of sustaining the productivity of sugarcane.

A shortage of available labour early in the growing season, when effort is directed to the weeding ofcereal crops, appeared to be a primary constraint to increased forage cropping by smallholder farmers in the semi-arid zone of central Mali. In experiments with millet, the labour requirement for weeding was reduced by an average of 45% when the crop was weeded at two rather than four weeks after emergence. A combination of weeding by hand and by ox-drawn plough or harrow only slightly increased labour productivity compared with hand weeding alone. Both lablab and cowpea produced good yields of forage when sown in mid July, but average yield declined by 71 kg ha−l for each day's delay in sowing between mid July and mid-August. Green gram was also productive over a two month growing period and appeared to have potential as a late-sown forage crop.

The leaf mulch of the tree legumes Acacia cunninghamii, A. fimbriata, Calliandra calothyrsus, Gliricidia sepium, Leucaena leucocephala and Sesbania sesban was evaluated as a source of nitrogen for maize growth in glasshouse pot experiments. The mulch of Sesbania sesban resulted in the largest yields of maize stover but yields were less than those that resulted from application of equivalent rates of fertilizer nitrogen. The mulches of Gliricidia and Leucaena also produced a yield response in maize but those of Calliandra and the two Acacia species were ineffective as sources of nitrogen in the short term. Sesbania mulch at the equivalent of 5 t dry matter ha-1 produced a maize yield similar to that obtained with 75 kg N ha-1. The poor response to Calliandra and Acacia mulch may be associated with the high polyphenol and/or lignin content of the leaf of these species.

Inoculation with vesicular-arbuscular mycorrhizal fungi (VAMF) improved growth of chick-pea (Cicer arielinum L.) and doubled phosphorus (P) uptake at low and intermediate levels of P fertilization in a pot experiment on sterilized low-P calcareous soil. In field experiments at Tel Hadya, northern Syria, growth, shoot P concentration and seed yield of spring-sown chickpea remained unaffected by inoculation with VAMF or by P fertilization. The mycorrhizal infection of chickpea was high (approximately 75% of root length mycorrhizal at the flowering stage) irrespective of inoculation with VAMF or P fertilization and may ensure efficient P uptake under field conditions.

The traditional crops of Oceania are asexually propagated and are losing their positions in cropping systems because of their genetic vulnerability. This paper reviews genetic reasons explaining the deterioration of their agronomic performance, using information mostly derived from surveys of genetic resources conducted in more than 50 Pacific islands, coupled with genetic investigations. More than 2000 cultivars have been analysed for isozyme variation. Cultivars and wild forms showed greater isozymic variation in Melanesia than in Polynesia and Micronesia but the genetic base for many of these crops was much narrower than morphological and chemical variation had suggested. No correlations were found between zymotypes, chemotypes, and morphotypes. Cultivars exhibited variation in qualitative traits as a result of human selection of somatic mutations. Accumulation of viral particles in propagules and build-up of pest and pathogen populations are often responsible for poor agronomic performance. If breeding of these traditional crops is to have any future, it is important to exchange genotypes to broaden the base of existing programmes.

The effects of site and rate of application of fertilizer on the grain and straw production of barley were investigated in 75 trials on representative farmers' fields. Grain and straw production were strongly but curvilinearly correlated with growth-period rainfall over the range 136–568 mm, almost irrespective of soil type, previous crop or fertilizer rate, and responded positively to applied nitrogen and/or phosphorus in 74 of the trials. Responses to nitrogen increased and those to phosphorus decreased with increasing rainfall. Yields tended to be lower but responses to nitrogen were higher where barley followed barley and were influenced by the availability of phosphate and mineral nitrogen in the soil at planting time.

These results are summarized in regression equations, which express yield quadratically in terms of fertilizer rates, seasonal rainfall and their interactions over the full 75-trial data set and within representative sub-sets. The wide differences in crop response to fertilizer highlight the problem of recommending fertilizer rates for a low and variable rainfall environment. Simple second-order multiple regressions for different geographical areas, rainfall zones, crop rotations etc., based on mean rainfall values, account for around 40% of the variance in yield response to nitrogen and phosphorus fertilizer; only a small proportion of the remaining variance appears attributable to rainfall variability. Alternatively, yield may be expressed as a function of fertilizer rate and rainfall using rainfall probability values derived from long-term records. This can be used to estimate the probabilities of predetermined yield responses for specified sites and fertilizer rates and to produce maps of those probabilities.

Because of shortages of oxen for mouldboard ploughing, delayed planting of maize is common in Mangwende, Zimbabwe and reduces grain yield by 32%. On-farm experiments over four years tested the possibility of using a ripper tine, with atrazine herbicide, to allow smallholders to plant maize earlier. Compared to mouldboard ploughing, tine cultivation increased grain yield at 13 out of 18 sites. All these sites had less than 240 mm of rainfall in January (which coincided with crop anthesis) and long term rainfall records suggest tine cultivation should raise yields in two out of three years. Handweeding and atrazine treatments gave similar grain yields when used with tine cultivation, provided weeds were controlled within 14 days of crop emergence. Tine cultivation was economic at sites where the yield was greater than that with the mouldboard plough. The farmers could manage the combination of tine and atrazine use on a field scale.

This paper addresses the current and future contributions of plant genetic resources and plant improvement to sustainable agriculture with reference to the activities of the International Center for Agricultural Research in the Dry Areas (ICARDA) in association with national programmes in West Asia and North Africa. These regions constitute the primary centres of diversity of crops such as wheat, barley, chickpea and lentil. Genetic erosion is being curtailed by germplasm collection and preservation. Selection for low-input cultivars of barley is conducted under low input conditions, and new cultivars of lentil and barley are often intentionally heterogeneous to stabilize their performance in dry rainfed areas. The importance of genetic differences in the cultivars on subsequent crops in the rotation and on straw quality for livestock is under study. Insect pests and diseases contribute to yield instability. Because of the potential adverse impact of pesticides on the fragile ecosystems of the region, integrated control strategies based on agronomic management, host plant resistance, biological control agents and strategic use of selective insecticides are being developed.

On-farm research in remote areas demands methods that keep down costs and reduce the need for subsequent research and extension inputs and yet that still produce useful and sustainable recommendations. An important part of the process is to identify existing resource user groups and use them, rather than individual farmers, as a starting point for on-farm research. This approach is discussed in the context of farming systems research in the eastern hills of Nepal.

Data recently published in this journal on grain yields of new rice varieties, grown under rainfed conditions on-farm in two villages in Uttar Pradesh, were analysed. The new varieties were spectacularly better than the landraces with which they were compared and showed a high regression slope (>2) against the landraces. If supported by further studies, such experiments could provide an appropriate, but hitherto unexplored, approach to plant breeding for poor, small farmers operating with low inputs.

This paper uses survey results to show how one group of small scale farmers in Santa Cruz, Bolivia, were able to mechanize annual cropping despite a farming systems diagnosis which indicated that this was not a viable development path. It attempts to identify the underlying causes of farmer success and diagnosis failure.

Rapid initiation and growth of nodal roots (NR) are critical for the establishment of post-rainy season (rabi) sorghums grown in drying soil. The growth of NR was studied in sorghum in the field when sown at different depths in order to vary the crown depth, when different levels of irrigation water were applied to create different moisture contents in the topsoil, and in a number of cultivars to assess genotypic variability. Variation in sowing depth did not result in differences in NR growth, partly because of the capacity of sorghum to vary the mesocotyl length, which brought the crown to approximately the same depth. Different soil moisture levels in the field affected the elongation rate of NR, but not their initiation. Genotypic variation existed in the thermal time required for NR initiation and rate of elongation. The variation in NR growth was independent of the variation in shoot growth. The genotypic variability in NR growth offers an opportunity for combining root-related traits with useful agronomic characters.

The relation between current yield and future yield components in the oil palm (Elaeis guineensis Jacq.) was investigated by removing a proportion of inflorescences before anthesis. This disbudding caused mean fruit bunch weight to increase by over 60%, with corresponding increases in the components of bunch weight. The total number of inflorescences produced increased by 10%, the sex ratio increased from 72 to 91% females, and inflorescence abortion and bunch failure rates were halved. As a result, the number of bunches harvested was reduced by less than the level of disbudding, and removal of 75% of inflorescences only reduced yield by 42%. Disbudded palms also grew more vigorously. The time intervals between a change in the level of fruiting activity and the resulting changes in yield components were determined by recording the yield components during the period after disbudding was stopped.

The diversity of cropping patterns on smallholder farms in southern Malawi was analysed using a framework that explicitly incorporates the extent of intercropping in each field. Diversity is defined as the relative abundance of each crop in the overall cropping pattern. Six indices of diversity were constructed for 208 farms and used in a model of a welfare-maximizing farm household to examine the reasons for diversity in cropping patterns; multiple regression techniques were used to determine the effect of different household characteristics on diversity. The results suggest that an increase in labour availability over the production period is associated with a more diverse cropping pattern. Landholding size also influences diversity, which rises to a maximum and then falls as the area cultivated per capita increases. Farmers who grow a non-food cash crop (tobacco) have more diverse cropping patterns than those who do not. As diversity increases, farmers use intercrop patterns that are more substitutive than additive.

Grassland productivity was studied for four years near Terbol, Lebanon. In the first year pasture availability and plant numbers were monitored along transects, and in the following three years the effects of sowing four Mediterranean annual legumes (three cultivars of subterranean clover and a local ecotype of Medicago rigidula) and top-dressing with super-phosphate were studied. The results indicated that natural grasslands were dominated by annual grasses (Aegilops, Hordeum, Bromur, Lolium and Poa). Legume density was low, which resulted in poor legume productivity, especially in winter. Exotic legumes only resulted in a slight improvement in pasture productivity even when phosphate fertilizer was added. However, partial protection from grazing (for one or two months in late winter and spring) more than doubled the number of legume seeds in the seed bank compared with full protection and open grazing. The build up of seeds in the soil is an essential step towards the improvement of productivity in these degraded pastures, which form a large part of the land surface in Lebanon and on which small ruminant production largely depends.

The extent to which the number and weight components of oil palm fruit yield are affected by the total weight of fruit bunches on the palm (fruiting activity) was investigated by multiple regression analysis, with plant available water and sunshine hours as additional covariates. Fruiting activity significantly increased the percentage of aborted inflorescences six months after leaf opening, and reduced the ratio of female to total inflorescences (sex ratio) seven to eight months prior to leaf opening. Mean weight of bunches was negatively affected by fruiting activity ten to eleven months before harvest, the number of flowers per spikelet and the mean single fruit weight being the only components to show a clear response. Fruiting activity significantly reduced the petiole cross sectional area of leaves six months before opening.

Simulation models for oil palm yield forecasting must take account of the endogenous yield cycles resulting from these effects of fruiting activity.

The poor quality of the natural vegetation in the dry season affects both animal productivity and the lives of the pastoral community in the Adamawa Plateau of Cameroon but may be improved by introduction of the forage legume Stylosanthes guianensis. When used in a ‘fodder bank’ system, Stylosanthes grew to an average plant height of 1.41 m, had a dry matter productivity of 5.5 t ha−1 and comprised 93.4% of the pasture's botanical composition at the end of the growing season, indicating that fodder banks have considerable potential to enhance animal and crop production in Cameroon. Complementary studies conducted in Nigeria on the yield benefits of cropping leguminous pastures support the feasibility of using fodder banks as a means of improving the livelihoods of small scale farmers.

The best available varieties of barley, durum wheat, rapeseed, safflower and sunflower were compared for their yield of grain and oil and other traits under rainfed conditions and with supplementary irrigation in the relatively dry Mediterranean climate of Cyprus. The grain yields of the oilseed crops safflower and rapeseed when grown under conditions of plentiful rainfall or with supplementary irrigation were 3–5 t ha−l. Considering the high prices of oilseeds on the world market, the income from oil crops compares favourably with that from barley and wheat. However, under conditions of limited rainfall (300 mm) and without supplementary irrigation the yields of oilseed crops were extremely poor, at the most 0.6 t ha−1. Grain and oil yields of irrigated safflower and rapeseed were larger than those of sunflower. Among the oilseed crops, sunflower received most irrigation water and rapeseed the least. Under rainfed conditions, with precipitation less than 300 mm, cereals were the only successful crops. It is concluded that in the high rainfall areas of West Asia and North Africa, or with supplementary irrigation, including treated municipal wastewater, oilseed crops can be successfully grown.