Search

In Bambara groundnut [Vigna subterranea (L) Verdc], long photoperiods can cause a decline in pod and seed number. While negative photoperiod effects on pod yield have been reported, positive effects and the existence of genotypes less sensitive for photoperiod have not been reported. Ten geographically diverse genotypes were evaluated over 3 years under short and long photoperiod for yield‐related traits (pod number, pod weight, seed number, seed weight, and 100 seed weight), pod growth habit (geocarpic or ageocarpic pods), peduncle elongation and soil penetration, and pod set/abortion. Anatomical sections of gynophores and embryonic pod development of a highly photoperiod sensitive genotype “Ankpa‐4” from Nigeria were examined. A strong photoperiodic effect on yield component traits (pod number, pod weight, seed number, seed weight, and 100 seed weight) was observed and with considerable variation between genotypes. Three classes of photoperiod effect on pod yield were identified, (a) qualitative short‐day types; (b) quantitative short‐day types; and (c) quantitative long‐day types. In long photoperiods, above‐ground vegetative biomass and the length of lateral branches in some genotypes increased by at least twofold. Morphological and anatomical characterization of gynophores and developing pods of the most sensitive line shows continuous geocarpic growth, but with healthy embryonic pods failing to enlarge after soil penetration. Results from the yield patterns of the three phenotypic classes confirm that pods and seeds are high priority sinks, and long photoperiod alters the balance in assimilate distribution between competing yield forming processes and vegetative sinks. By exploiting the genetic variation characterized here for photoperiod requirement for pod filling, the geographical range of this crop could be extended beyond current growing regions. In the near future, we anticipate that Bambara groundnut will become a significant contributor to global food, nutritional, and environmental security once these photoperiod issues are resolved.

Crops For the Future Research Centre (CFFRC) is currently developing CropBASE—a cross-cutting web-based platform for decision support and knowledge sharing on underutilised crops and their end uses. It provides a quantitative basis to compare underutilised crop productivity and resource-use-efficiency with that of major crops and cropping systems under current and future climate scenarios, and will include interactive tools to integrate underutilised crop data with geo-referenced information in existing and new databases. Specific crop application models form the analysis engine of CropBASE, including crop-climate, crop-quality, crop-economic models to assess agroecological suitability, nutritional value, livelihood or food security potential of underutilised crops and products at field, farm and regional scales. A geospatial information system framework will link currently available climate databases as input data for crop models. CFFRC, already using major models for a range of crops around the world under the AgMIP program, will spearhead the inclusion of specific underutilised crops into recognised crop models. After briefly introducing CropBASE, this paper presents use cases as examples comparing yield and water productivity predictions for bambara groundnut (an underutilised African legume) and pearl millet (an alternative major cereal) under current and future climate scenarios in contrasting sub-Saharan African locations.

In Mediterranean regions drought is the major factor limiting spring barley and durum wheat grain yields. This study aimed to compare spring barley and durum wheat root and shoot responses to drought and quantify relationships between root traits and water uptake under terminal drought. One spring barley (Hordeum vulgare L. cv. Rum) and two durum wheat Mediterranean cultivars (Triticum turgidum L. var durum cvs Hourani and Karim) were examined in soil-column experiments under well watered and drought conditions. Root system architecture traits, water uptake, and plant growth were measured. Barley aerial biomass and grain yields were higher than for durum wheat cultivars in well watered conditions. Drought decreased grain yield more for barley (47%) than durum wheat (30%, Hourani). Root-to-shoot dry matter ratio increased for durum wheat under drought but not for barley, and root weight increased for wheat in response to drought but decreased for barley. The critical root length density (RLD) and root volume density (RVD) for 90% available water capture for wheat were similar to (cv. Hourani) or lower than (cv. Karim) for barley depending on wheat cultivar. For both species, RVD accounted for a slightly higher proportion of phenotypic variation in water uptake under drought than RLD.

SUMMARYSimulation models for many crops are often simple, empirical equations that lack generality across different locations or different levels of abiotic stress making them difficult to apply by independent researchers. The objective of this paper is to describe and demonstrate a user-oriented model ‘BAMGRO’ for an underutilized crop, bambara groundnut with enough detail to be general across locations and genotypes but not so complex that independent users cannot apply it to their own situations. The model consists of different sub-modules that deal specifically with weather (thermal time), crop growth (canopy development, biomass production and yield), temperature (heat and cold stress), photoperiod (daylength control of phenology and reproductive development) and soil water (drought stress). The model predictions are achieved by defining genotype (i.e. cultivar) input files, daily weather parameters and soil characteristics. The model achieves a good fit between observed and predicted data for leaf area index (Nash and Sutcliffe (N-S), 0.80–0.84; mean absolute error (MAE) with maximum less than ± 0.50) for tested genotypes. Pod yield simulation correlated well with measured values (N-S 0.73–0.87; MAE ± 16 g m−2). Available soil moisture content correlated well with the observed data for a field site in Botswana indicating successful performance of the soil water module.

SUMMARYSimulation of yield response to water plays an increasingly important role in optimization of crop water productivity (WP) especially in prevalent drought in Africa. The present study is focused on a representative crop: bambara groundnut (Vigna subterranea), an ancient grain legume grown, cooked, processed and traded mainly by subsistence women farmers in sub-Saharan Africa. Over four years (2002, 2006–2008), glasshouse experiments were conducted at the Tropical Crops Research Unit, University of Nottingham, UK under controlled environments with different landraces, temperatures (23 ± 5 °C, 28 ± 5 °C, 33 ± 5 °C) and soil moisture regimes (irrigated, early drought, late drought). Parallel to this, field experiments were conducted in Swaziland (2002/2003) and Botswana (2007/2008). Crop measurements of canopy cover (CC), biomass (B) and pod yield (Y) of selected experiments from glasshouse (2006 and 2007) and field (Botswana) were used to calibrate the FAO AquaCrop model. Subsequently, the model was validated against independent data sets from glasshouse (2002 and 2008) and field (Swaziland) for different landraces. AquaCrop simulations for CC, B and Y of different bambara groundnut landraces are in good agreement with observed data with R2 (CC-0.88; B-0.78; Y-0.72), but with significant underestimation for some landraces.

SUMMARYThe association of carbon isotope discrimination of grain (Δ13C) with yield performance under rain-fed and well-watered conditions was analysed using a doubled-haploid (DH) winter wheat population, derived from the cross between cvars Beaver×Soissons, within field experiments at two site-seasons. The aim of this work was to quantify associations between Δ13C and yield responses to drought and to test effects of major genes (the semi-dwarf genes, Rht-B1b, Rht-D1b, an awn suppressor gene, B1 and the 1BL.1RS wheat–rye chromosome translocation) segregating in the population for associations with Δ13C and drought performance. Carbon isotope discrimination, through its negative relationship with transpiration efficiency, may be used as a surrogate for this trait. Grain Δ13C was positively associated with grain yield under both irrigated and unirrigated conditions in each site-season and, overall, explained 0·34 of the phenotypic variation in grain yield amongst DH lines under drought and 0·14 under well-watered conditions. There was a positive association between specific leaf lamina N content (SLN) at anthesis and Δ13C under drought amongst DH lines in one site-season, suggesting higher SLN may confer increased stomatal conductance via higher photosynthetic capacity, hence increased grain Δ13C. Overall the Rht-D1b (semi-dwarf) lines had slightly higher Δ13C of grain (20·0‰) than the Rht-B1a/Rht-D1a (tall) group of lines (19·8‰). There were no significant differences between the Rht-B1b (semi-dwarf) or the Rht-B1b/Rht-D1b (dwarf) lines and the tall lines. Comparing their performance under irrigated and unirrigated conditions, the Rht groups of lines (Rht-B1b semi-dwarf, Rht-D1b semidwarf and dwarf and tall groups) responded no differently to drought for Δ13C. The Rht-D1b semi-dwarf lines had higher grain yield (9·50 t/ha) than the tall lines (8·76 t/ha), while the yield of the Rht-B1b semi-dwarf and dwarf lines did not differ significantly from the tall lines. In each site-season, the presence of the 1BL.1RS chromosome increased grain Δ13C (PP13C was not statistically significant in either site-season. Overall, the present results show that Rht-D1b confers higher Δ13C and grain yield, and the 1BL.1RS translocation confers higher Δ13C. This implies that modern UK wheat cultivars may have lower water-use efficiency during the grain filling period than their predecessors, and therefore may require more water to fulfil their yield potential.

Canopy development of bambara groundnut (Vigna subterranea (L.) Verdc) is affected by temperature stress, drought stress and photoperiod. The quantification of these documented effects by means of a suitable crop model, BAMGRO is presented in this paper. Data on canopy development from five growth chamber, four glasshouse and three field experiments were analyzed to calibrate and validate the BAMGRO model to produce simulations for temperature stress, drought stress and photoperiodic effect on two contrasting landraces; Uniswa Red (Swaziland) and S19-3 (Namibia). The daily initiation rate of new leaves is calculated by means of a Gaussian function and is altered by temperature stress, drought stress, photoperiod and plant density. The rate in dead leaf number is dependent upon the maximum senescence fraction which can be explained by physiological maturity, mutual shading, temperature stress and drought stress. The leaf area expansion is calculated as a function of leaf number and individual leaf size. The canopy development of bambara groundnut is modelled (and tested) in BAMGRO model by means of leaf initiation and leaf area expansion and branching and stem production are not considered. Nash and Sutcliffe value (N–S) is used as the measure of the mean square error to the observed variance. The simulated leaf number per plant correlated well with the glasshouse measurements for Uniswa Red for the two considered years (2007 and 2008) with high N–S value (Nash and Sutcliffe; from 0.76 to 0.92), low mean absolute error (MAE; from ±2.91 to 3.98) and Botswana field grown in January 18 and February 1, 2007 (N–S, 0.88 and 0.74; MAE, ±0.59 and ±6.18). Overall, simulated LAI correlated well with measured values for the two tested landraces; S19-3 (N–S value, 0.84) and Uniswa Red (N–S value, 0.80) with maximum MAE less than ±0.50 thus explaining canopy development under temperature stress, drought stress and variable photoperiods.

This study involves an assessment of the potential effects of greenhouse gas climate change, as well as the direct fertilization effect of CO2 on crop yields in Cameroon. The methodology involves coupling the transient diagnostics of 2 atmosphere-ocean general circulation models, namely NASA/Goddard Institute GISS and the Hadley Centre's HadCM3, to the CropSyst crop model to simulate current and future (2020, 2080) crop yields (bambara nut, groundnut, maize, sorghum and soybean) in 8 agricultural regions of Cameroon. For the future we estimate substantial yield increases for bambara groundnut, soybean and groundnut, and little or no change and even decreases of maize and sorghum yields, varying according to the climate scenario and the agricul- tural region. Maize and sorghum (both C4 crops) yields are expected to decrease by 14.6 and 39.9%, respectively, across the whole country under GISS 2080 scenarios. The results also show that the effect of temperature patterns on climate change is much more important than that of precipitation. Findings call for monitoring of climate change/variability and dissemination of information to farmers, to encourage adaptation to climate change.

The effect of drought on the growth and development of bambara groundnut (Vigna subterranea (L.) Verdc.) was studied in controlled-environment glasshouses in the UK. There were three landraces (S19-3, DipC and UN from Namibia, Botswana and Swaziland, respectively) and two watering regimes; a control that was irrigated weekly to 90% field capacity and a drought treatment with no irrigation from 49 days after sowing (DAS) until final harvest (147 DAS). Bambara groundnut responded to drought by reducing the rate of leaf area expansion, final canopy size and total dry matter (TDM) during vegetative growth. Drought also caused significant reductions in pod dry matter (PDM), pod number, seed weight and harvest index (HI), leading to a decrease in final pod yield that was different between landraces. Across landraces, drought reduced mean pod yield from 298 g m−2 to 165 g m−2, representing 45% yield loss. Despite the reduction in all landraces, the mean pod yield across the droughted treatments that had received no water for almost 100 days indicated the resilience of the species to drought. The three landraces differed in their phenology; S19-3 exhibited a reduced phenology while UN maintained the longest life cycle. The different responses of the landraces reflect their adaptation to their local climates where mean annual rainfall ranges between 365 mm (Namibia) and 1390 mm (Swaziland). We discuss the significance of these results for future breeding programmes on bambara groundnut.

Controlled crosses in bambara groundnut were attempted between a range of thirty-six bambara groundnut landraces (thirty domesticated (V. subterranea var. subterranea) and six wild (V. subterranea var. spontanea)). Ten F-1 seed were produced. Of these, eight germinated producing F-2 populations. On seed set, four populations could be unambiguously confirmed as true crosses by F-3 seed coat colour. A single F-2 population, derived from a domesticated landrace from Botswana (DipC; female parent) crossed with a wild accession collected in Cameroon (VSSP11; male parent) was used to study a range of agronomic and domestication traits. These included; days to emergence, days to flowering, internode (fourth) length at harvest, number of stems per plant, leaf area, Specific Leaf Area (SLA), Carbon Isotope Discrimination (CID), 100 seed weight, testa colour and eye pattern around the hilum. On the basis of variation for internode length and stems per plant, 14 small F-3 families were selected and grown under field conditions to further investigate the genetic basis of the 'spreading' versus 'bunched' plant character, a major difference between wild and cultivated bambara groundnut. Results presented suggest that traits including leaf area, SLA, CID and 100 seed weight are controlled by several genes. In contrast, the variation for traits such as internode length, stems per plant, days to emergence and seed eye pattern around the hilum are likely to be under largely monogenic control. The results of this work are discussed in relation to the domestication of bambara groundnut.

Ten bambara groundnut [Vigna subterranea (L.) Verdc] landraces of diverse origin were used to examine the effect of temperature on the rate of leaf appearance (RLA) and to determine the base temperature (T b ) among landraces. Knowledge of leaf development is required in breeding programs where crop morphological development is considered an important selection criterion. The ability to predict leaf appearance would also facilitate modeling of plant development. Plants were grown in controlled-environment growth rooms under different temperatures (20, 25, 30, 32, and 35°C ± 1). The study to examine the effect of temperature on leaf appearance in bambara groundnut landraces indicated that the base temperature (T b ) for leaf appearance ranged from 8.1 to 12.0°C. Within the range of temperatures evaluated in this study, the observation of differences in T b values among the landraces suggests that T b values for leaf appearance are landrace dependent. The phyllochron was also variable among landraces ranging from 40.9 to 53.0°C. The RLA ranged from 0.19 to 0.63 leaves d -1 depending on temperature and landrace. The landrace variations in relation to leaf appearance may allow selection of landraces for different environments.

Three landraces of bambara groundnut (Vigna subterranea (L.) Verdc.) were grown as crop stands in controlled environment glasshouses at the Tropical Crops Research Unit, University of Nottingham, in 1995. Two soil moisture treatments were imposed: irrigated to 90% field capacity each week and irrigated to 60% field capacity until establishment (27 days after sowing) with no further irrigation. Seasonal mean fractional interception varied between 0·20–0·37 for the droughted treatments and 0·62–0·74 for the irrigated treatments, resulting in cumulative intercepted radiation of 228–350 MJ/m2 and 662–794 MJ/m2, respectively. The maximum total dry matter (DM) produced was 5·8 t/ha at final harvest (145 days after sowing) with a pod yield of 2·7 t/ha. Under moisture stress there was little difference in DM production between landraces, with the highest total DM of 1·1 t/ha and a pod yield of 0·05 t/ha, representing a harvest index of 0·05 compared with an average of 0·46 for the irrigated treatments. The conversion coefficient was reduced from 1·00 under irrigation to 0·51 g DM/MJ radiation intercepted by soil moisture stress. Two of the landraces showed adaptive mechanisms to avoid drought; these are discussed in relation to maximizing seasonal radiation interception.

Bambara groundnut is an indigenous African legume crop plant. It is largely grown by subsistence farmers, but can also be used as a cash crop to supplement family income. The fact that it is highly drought tolerant means that it has considerably potential to provide part of food security in regions of the world where water availability is a serious issue. As one part of the international effort to understand and improve this species, we report here the first 10 microsatellite markers derived from bambara groundnut.

SUMMARYStands of bambara groundnut (Vigna subterranea (L.) Verde.) were grown in five controlledenvironment glasshouses at the Tropical Crops Research Unit, University of Nottingham, Sutton Bonington Campus, in 1990. Five soil moisture regimes were imposed (one per house), from fully irrigated each week (treatment A), to no irrigation after crop establishment at 35 days after sowing (DAS) (treatment E). Decreasing the amount of water applied resulted in a decline in total dry matter production and harvest index, and a reduction in pod yield from 412 (treatment B) to 0·041 ha-1 (treatment E) at 125 DAS. A maximum leaf area index of 5–4 was achieved by treatments B and C at 90 DAS, resulting in a fractional interception of c. 0·8 of incoming radiation. Total accumulated radiation interception values were 749, 693, 688, 618 and 554 MJ m-2 for treatments A, B, C, D and E, respectively. The efficiency of conversion of the radiation intercepted into dry matter was reduced from 1·41 to 0·50 g MJ-1 by drought.

The effects of elevated atmospheric CO 2 , alone or in combination with water stress, on stomatal frequency in groundnut (Arachis hypogaea (L.) cv. Kadiri-3) were investigated. CO 2 exerted significant effects on stomatal frequency only in irrigated plants. The effects of drought on leaf development outweighed the smaller effects of CO 2 concentration, although reductions in stomatal frequency induced by elevated atmospheric CO 2 were still observed. When stands of groundnut were grown under irrigated conditions with unrestricted root systems, an increase in atmospheric CO 2 from 375 to 700 ppmv decreased stomatal frequency on both leaf surfaces by up to 16%; in droughted plants, stomatal frequency was reduced by 8% on the adaxial leaf surface only. Elevated atmospheric CO 2 promoted larger reductions in leaf conductance than the changes in stomatal frequency, indicating partial stomatal closure. As a result, the groundnut stands grown at elevated CO 2 utilized the available soil moisture more slowly than those grown under ambient CO 2 , thereby extending the growing period. Despite the large variations in cell frequencies induced by drought, there was no treatment effect on either stomatal index or the adaxial/abaxial stomatal frequency ratio. The data suggest that the effects of future increases in atmospheric CO 2 concentration on stomatal frequency in groundnut are likely to be small, especially under conditions of water stress, but that the combination of associated reductions in leaf conductance and enhanced assimilation at elevated CO 2 will be important in semi-arid regions.

SUMMARYBambara groundnut (Vigna subterranea (L.) Verde.) is grown throughout southern and western Africa, primarily as a subsistence crop, but agronomic information is scarce. The rates of emergence, flowering and pod production were assessed for 20 weekly sowings between 16 October 1990 and 26 February 1991 at Sebele, Gaborone, Botswana. Daylength during this period changed from 12·7 h to 13·7 h and back to 12·6 h. Mean time to 50% emergence was 111 days while mean time to 50% flowering was 47·5 days. The thermal time to first pod production varied with daylength, decreasing as daylength decreased, to an approximately constant value when daylength after flowering was less than about 12 h. Consequently, the plant size at which pods were produced also varied. The practical implications for date of sowing and plant spacing are discussed.

Stands of groundnut (Arachis hypogaea L.), a C 3 legume, were grown in controlled-environment glasshouses at 28 o C (±5 o C) under two levels of atmospheric CO 2 (350 ppmv or 700 ppmv) and two levels of soil moisture (irrigated weekly or no water from 35 d after sowing). Elevated CO 2 increased the maximum rate of net photosynthesis by up to 40%, with an increase in conversion coefficient for intercepted radiation of 30% (from 1.66 to 2.16 g MJ -1 ) in well-irrigated conditions, and 94% (from 0.64 to 1.24 g MJ -1 ) on a drying soil profile

SUMMARYAt a field site in central India, four populations of groundnut (Arachis hypogaea L.) were grown on stored water to investigate how the production of shoot and root dry matter is related to transpired water and intercepted radiation. Throughout the season, total dry matter was closely related to transpiration (slope = 3.0 mg dry matter g−1 water) and the amount of radiation intercepted by foliage (slope = 0.74 g dry matter MJ−1 radiation intercepted). Accumulated transpiration increased linearly with intercepted radiation at 0.37 kg water MJ−1 in the sparser stands. In the densest spacing, the initial slope of the relation at 0.28 kg MJ−1 decreased later in the season because water deficits curtailed growth without a concomitant reduction in the interception of radiation.

In central India, four populations of groundnut were grown to assess the interaction between population and water stress. Transpiration was calculated from measurements or estimates of stomatal resistance, r s , boundary layer resistance, r a , vapour concentration difference between leaf and air, δ χ , and leaf area index, L. The frequency distributions of r s , r a , δ χ and seasonal changes in L were plotted to analyse the dependence of transpiration rate on each variable, both per unit area of leaf surface, E l , and per unit land surface, E e . For estimates of E l , both r s and δ χ were of similar importance, exerting a far greater influence than changes in r a . However, in terms of E e , changes in L were far more important than in any other variable, particularly late in the season when water was scarce. This study provides further experimental support for estimates of evaporation based on stomatal resistance and allied measurements and confirms similar estimates obtained earlier for crops of pearl millet maintained on stored water. The ability of the technique to describe temporal and spatial variations as well as the dominant environmental and physiological influences on transpiration may outweigh any small loss in accuracy of estimates thus obtained.

SUMMARYThe objective of this paper is to identify the major factors responsible for the differences in transpiration rate (T) between stands of groundnut grown at three row spacings on stored water in central India. A method of analysis is developed to distinguish between the impact on transpiration of ground cover (which varied threefold between spacings) and of root:shoot ratio, which was substantially greater in the wider row spacings. When the soil was wet, both T and the canopy conductance, gc, were approximately proportional to the fraction (f) of incident radiation intercepted by foliage. But when the soil water content decreased below a threshold value, T/f and gc/f decreased because of an increase in stomatal resistance. Stomatal closure in response to soil water stress occurred sooner in the denser stands, partly because of more rapid depletion of soil water. A second contributory factor was that the sparser stands (which had a relatively large root:shoot ratio) had a greater capacity to keep stomata open as the soil water deficit increased.

SUMMARYPearl millet (Pennisetum typhoides) was grown on stored water at Niamey, Niger, using three row spacings (38, 75 and 150 cm), to determine the physiological basis of exploitation and conservation of water by crops during drought Between 18 and 32 days after sowing, roots grew rapidly beneath all crops reaching 140 cm in the narrow spacing, but there were differences between crops in the pattern of growth. Soil cores and trench profiles indicated that plants in wider rows had fewer, deeper roots. Substantial differences in both the amount and pattern of shoot growth were recorded in the different populations. Initially growth was fastest at the narrow spacing but stopped by day 45 and eventually the wide spacing produced most dry matter due mainly to greater survival of tillers. The partitioning of above ground dry matter into vegetative and reproductive fractions was similar at all three spacings and was consistent with figures for comparable crops elsewhere. The important role of tillers is discussed in relation to the development and maintenance of a canopy.

SUMMARYThe growth of roots and shoots was measured in stands of groundnut grown at a number of populations on stored water in central India. Total weight and length of roots per unit land area increased with population density, but the proportional increases were much less than for shoot weight. Consequently the root:total weight ratio increased from 0.3 in the densest stand to almost 0.5 in the widely spaced crop. The denser stands produced a greater proportion of their roots at depth. In wide rows there was little change in rooting density across the inter-row space.Total dry matter per unit land area increased with population, although the weight per plant was less in denser stands. Although the crops were harvested prematurely, pod yield per unit land area, unlike total dry matter, was no greater in dense stands than in more widely spaced crops. The greatest number of pods per unit land area was recorded at an intermediate population density.

SUMMARYPearl millet was grown on stored water at Niamey, Niger, using three row spacings. Water extraction based on neutron probe readings was compared with crop transpiration using a porometer and allied measurements. Between 23 and 52 days after sowing, plants at the narrow and medium spacings used about 77 and 100 mm of water, respectively, and those at the wide spacing used between 59 and 75 mm. Estimates of seasonal crop evaporation from leaf resistances and from the green leaf area index (GLAI) of the crops were 103, 130 and 123 mm for the narrow, medium and wide spacings, respectively. The water use per unit of dry weight produced was similar for both narrow and medium spacings but water was used more efficiently in the wide spacing. Dry weight increased in proportion to intercepted radiation with the same efficiency (1·3 g MJ−1) irrespective of spacing.

This paper reports an analysis of the yield of sugar beet crops grown under experimental conditions between 1993 and 1995 in the UK. Crops were either healthy (unstressed) or subjected to drought, infection with Beet yellows virus (BYV) or a combination of both. The study investigated whether the large differences in yield between the crops grown in different seasons and subjected to different stresses could be accounted for by simple relationships between total biomass and radiation interception (εs), transpiration (εw) or εs and εw adjusted for mean saturation deficit (Ωs and Ωw respectively). Mean values of εs, εw, Ωs and Ωw in healthy crops were 1·42 g/MJ, 0·89 g/kg, 6·76 g/kPa/MJ and 4·29 g/kPa/kg respectively. Variations in the dry matter yield between seasons were best accounted for by Ωw and less well by εw or εs. Ωs accounted for least variation in yield between seasons. None of these relationships remained constant in stressed plants; both drought and BYV-infection decreased εs (and Ωs) but Ωw was increased by drought and decreased by BYV-infection. However, in common with healthy crops, seasonal variation in yields was best accounted for by Ωw. Mean values of εs, Ωs, εw and Ωw for all healthy, infected and droughted crops accounted for 61, 50, 88 and 97% of the variation in dry matter yield between experiments respectively. Accurate prediction of the yield of stressed plants requires a knowledge of their infection and drought status. If this information is unavailable then the mean value of Ωw for healthy, infected and droughted crops will provide a reasonable prediction of the yield of such crops.