Four cultivars of winter barley (Hordeum vulgare L.) and two cultivars of combining pea (Pisum sativum L.) were grown in the field in the UK (52°N) and irradiated under banks of UV-B lamps in 1994/95 (barley) and 1996 (pea). Supplementary UV-B radiation was applied to treated plots as a proportional addition to the UV-B dose received under a control plot. Treated plants received a UV-B enhancement simulating the consequence of a 15% reduction in the amount of stratospheric ozone. No significant effect on yield and few significant effects on growth, pigment composition or chlorophyll fluorescence variables were detected. However, interplot variability was such that yield differences of <8·5% (pea) and <21·6% (barley) had less than a 95% probability of being detected as significant at the 5% level. The results indicate that yields of pea, and probably barley, would not be markedly affected by the increase in UV-B associated with a 15% reduction in stratospheric ozone. However, given uncertainties, such as the possible interactions between the effects of UV-B and those of other environmental factors, the possibility of significant crop responses to stratospheric ozone depletion cannot be excluded.

The recovery of nitrogen ‘retained’ through cover crop uptake, delayed ploughing and immobilization by straw was assessed in a spring cropping rotation on a chalk loam in Eastern England (1989–96). The effect of annual cover cropping on yield of the subsequent spring crops and on the soil N balance was also investigated. The recovery of retained N was in part dependent upon cover crop management. Late August-sown cover crops which were incorporated in February/March tended to reduce spring crop yields and crop N offtake. Adverse effects on soil N supply, seedbed conditions and soil water reserves were not in evidence and so an allelopathic effect from the decomposition of the rye cover crop, previously reported by others, may be responsible for the reduction in yield of spring crops. When the cover crops were drilled later and their early destruction was followed by a short fallow period, spring crop yields and N offtake were increased. The soil N balance indicated that over the course of the experiment there was a positive N input to the system due to continuous cover cropping. This input may be held as immobilized organic N, in which case it could be made available to subsequent crops over a number of years or lost via other routes. Nitrate concentrations in drainage water increased with the number of years under cover cropping.

A field experiment was carried out on soyabean (Glycine max (L.) Merr.) to measure the effect of planting date (soil temperature) on seedling emergence. Seeds were sown at weekly intervals on seven planting dates from April until the end of May in SE England in 1997. Planting date had no significant effect on final percentage emergence but had a highly significant effect on time to emergence. The coefficients of variation (c.v.) for the number of days to emergence (calendar days) were high (43–45%), and therefore not a reliable method for predicting emergence. Three accumulated heat unit (AHU) methods based on air and soil temperatures were compared with the calendar day method to determine the most reliable system for predicting seedling emergence. Accumulated soil temperatures above a base of 9·0 °C had the lowest c.v.s (8–15%) and were shown to be the most reliable predictor of emergence.

Soil sampling in autumn gives important information on the soil N dynamic. In the growing seasons 1991/92 to 1995/96, the effects of different crop management systems on soil mineral N (NO3-N plus NH4-N:Nmin) were investigated in a factorial field experiment at Hohenschulen Experimental Station near Kiel in NW Germany. The crop rotation was oilseed rape – winter wheat – winter barley, and soil tillage (conservation tillage without ploughing, conventional tillage), application of pig slurry (none, autumn, autumn+spring), mineral N fertilization (0, 120 and 240 kg N ha−1) and fungicide application (none, three applications) were all varied. Each year, the treatments occurred in all three crops of the rotation and were located on the same plots. Nmin was determined on four dates (‘After drilling’, ‘End of autumn growth’ before winter, ‘Beginning of spring growth’ before N fertilizer application, and ‘After harvest’) to 90 cm in 30 cm horizons.

Under all crops, Nmin showed a large year to year variation. Highest values of 132 kg N ha−1 were observed ‘After drilling’, which decreased until ‘End of growth’. The increase of autumn Nmin (‘After drilling’, ‘End of autumn growth’) was mainly due to autumn slurry, whereas mineral N fertilizer mainly affected Nmin ‘After harvest’. Soil tillage and fungicide application only slightly modified Nmin at all dates.

The relationship between N leaching and Nmin measured either ‘After drilling’ or at the ‘End of autumn growth’ in 1991/92–1994/95 remained too poor to be used to estimate N leaching. N net mineralization during autumn and winter varied with crops, as estimated by the Nmin changes between ‘After drilling’ minus ‘Start of spring growth’ plus N uptake by the crop at ‘Start of spring growth’ plus N leaching during winter. On average over the years, 39 kg N ha−1 were mineralized under oilseed rape and 42 kg N ha−1 under wheat compared with 31 kg N ha−1 under barley. However, a large year-to-year variation occurred. In addition, the ranking of the years differed with the crops. Slurry application led to different amounts of mineralized N. Under barley only 25 kg N ha−1 were calculated for the autumn slurry, but 42 kg N ha−1 for the autumn plus spring slurry treatment. In contrast, under oilseed rape the highest value of 41 kg N ha−1 occurred in the autumn slurry plots. Under wheat, slurry application only slightly affected N mineralization. Increased mineral N fertilization decreased N release under oilseed rape, but significantly increased it under cereals. Application of fungicides did not affect N mineralization during winter.

A comprehensive study was conducted over a 4-year period (1984–87) to evaluate the water use, growth and yield responses of pearl millet (Pennisetum glaucum (L.) R. Br.) cv. CIVT grown with and without fertilizer (30 kg P2O5 and 45 kg N ha−1) at the ICRISAT Sahelian Centre, Sadoré, Niger. Our study showed significant year and fertilizer effects on the growth and yield of millet at the study site. Observed year effects were primarily due to the variations in the amount and distribution of rainfall in relation to the potential demand for water. During 1984, 1985 and 1987, total rainfall was below the long term average, while in 1986 it was above average. While the onset of rains (relative to the average date of onset) was early from 1984 to 1986, in 1987 the sowings were delayed by as much as 33 days. Of all the four years, the separation between the treatments in the cumulative evaporation is most evident for 1984, which was a drought year with below-average rainfall in all the months from June to September. Cumulative evaporation patterns in 1985 and 1986 were similar because of regular rains and high average rainfall per rainy day from June to October. In 1987, sowings were delayed until 15 July and only 6·9 mm of rainfall was received per rainy day in July. Hence cumulative evaporation was initially low and showed a significant increase only after two significant rain events in early August. There was a large response to fertilizer in all the years as small additions of fertilizer phosphate increased the soluble phosphate in the soil. Fertilizer application resulted in a small increase in water use (7–14%) in all years except 1987. Increased yield due to the application of fertilizer was accompanied by an increase in the water-use efficiency (WUE) in all the four years with the largest increase in 1985. The beneficial effect of fertilizers could be attributed to the rapid early growth of leaves which can contribute to reduction of soil evaporative losses and increased WUE. Over the four seasons, average increase in the WUE due to the addition of fertilizer was 84%.

A simple model of canopy expansion and senescence with accumulated thermal time from sowing was used to describe differences in canopy development and potential size across a range of sites in the UK between 1992 and 1995. The principal model inputs were nitrogen, temperature and sowing date. The model was calibrated across six sites within the major wheat growing areas using canopy data collected during the 1992/93 and 1993/94 growing seasons and was validated using data collected in the 1994/95 season. The model was able to predict the time course of green area index (GAI) over a season with an r2[ges ]0·87 for the five English sites, and an r2=0·68 for the single Scottish site. This model may prove to be a useful approach to forecasting the potential canopy size, based upon an estimate or measure of the total nitrogen available to the crop (both from applied fertilizer and the soil).

The re-introduction of forages into the farming systems of central and southern Italy has been advocated by EU policies in order to reduce high inputs and favour more sustainable agricultural systems. Since legumes could certainly play a major role, landraces and natural populations of Trifolium alexandrinum (2), T. incarnatum (5) and T. squarrosum (3) were investigated. Two experiments were conducted during the 1991/92, 1992/93 and 1993/94 seasons in Perugia, central Italy. The objective of the first was to obtain the morphological and physiological characteristics of the entries, while the second assessed the biomass production, forage quality (stem: leaf: inflorescence ratio, protein, fibre, O.M. digestibility, milk and meat Forage Units) and seed yields. In T. incarnatum, flowering time, leaf area, stem length and diameter were important characteristics for discriminating between populations. Norcia and Rocca di Corno, two natural populations of T. incarnatum, flowered early, produced the most biomass and had high forage quality. Their potential as forage/pasture crops and/or cover crops in Mediterranean farming systems is discussed.

The effects of weather and nitrogen application rate on bread volume, protein concentration and gluten strength were investigated in spring and winter wheats in Sweden during 1990–96. The investigation was carried out in wheats containing the high-molecular-weight (HMW) glutenin subunit combinations 2+12 or 5+10.

The results showed that a warm (c. 20°C mean dry temperature) and dry grain-filling period, as during 1994 and 1995, led to high gluten strength. As a relatively gentle dough mixing is normally applied in Sweden when baking bread, excessively high gluten strength results in low bread volumes. Wheat cultivars containing HMW glutenin subunits 5+10 generally had higher gluten strength than those containing subunits 2+12. However, not all the cultivars containing subunits 5+10 showed overstrong gluten properties, whereas some cultivars containing subunits 2+12 were overstrong during 1994 and 1995. Different cultivars produced the highest gluten strength in different years. Generally, cultivars containing HMW subunits 2+12 gave higher bread volumes than those containing subunits 5+10. This is probably due to the Swedish baking method, where the dough is given a relatively gentle mixing for a fixed period of time.

Increased fertilizer rates led to increased protein concentration and decreased gluten strength. However, during certain years, for example 1994, the gluten strength was only decreased slightly by increased fertilizer rates, particularly in cultivars containing HMW glutenin subunits 5+10.

Three pot experiments were performed at the University of Wales, Bangor, Gwynedd, UK during 1993 to determine if differences in the competitive ability of three annual weeds (Chenopodium album L., Sinapis arvensis L. and Phalaris minor Retz.) were related to their relative effects on leaf growth, gas exchange and nitrogen uptake of spring wheat (cv. Alexandria). In all experiments, wheat density (316 plants m−2) was similar to that in a commercial crop and five weed density treatments (between 0 and 600 plants m−2) were tested. Measurements of gas exchange were made on fully expanded, attached wheat flag leaves on four occasions between emergence and complete senescence in the control and highest weed density treatments. High weed density resulted in a lowering of net photosynthetic rate due to stomatal and non-stomatal factors. Lamina area and stomatal density of wheat flag leaves were decreased, and specific leaf area was increased by weed competition, but the effects on these variables were smaller than on net photosynthesis. Weed density did not affect wheat plant height, but dry weight, grain yield and total N-uptake were decreased with an increase in density of all weed species. The rank order of competitive ability of the species (C. album>P. minor>S. arvensis) was unaffected by weed density and was the same irrespective of whether it was based on the % decreases in wheat grain yield or in total plant dry weight. Averaged over the four measurements made during the grain-filling period there were only small differences between the weed species in their effects on net photosynthetic rate. However, when these were combined with effects on flag leaf area, there were larger differences in calculated net photosynthetic productivity, which were related to differences in the effects of weeds on grain yield. Differences in the competitive ability of weeds were not related to differences in their effects on wheat flag leaf lamina area, specific leaf area, stomatal density or total nitrogen uptake. Differences in competitive ability between weed species were not related to differences in weed plant height, dry weight or nitrogen uptake. It was concluded that the observed effects of weeds on wheat were due either to shading, or to competition for a nutrient other than nitrogen.

The potential for using crop wastes as a source of nutrients on the yields of rice grown under submerged conditions was studied in both field and pot experiments at the Experimental Farm of the Agricultural and Food Engineering Department, Indian Institute of Technology, Kharagpur during 1992/93 and 1993/94. The application of groundnut haulm at 3·0 t/ha, maize stover at 5·0 t/ha, mustard stover at 3·0 t/ha and green manure at 1·5 t/ha contributed considerable amounts of nutrients, which improved components of yield, yield and nutrient uptake by rice and also improved the organic carbon (OC) content and available N, P and K in the soil. Yields and soil fertility were both further improved when these organic materials were combined with inorganic fertilizer to supply the recommended amounts of nutrients, except that inorganic fertilizer alone did not cause an increase in the OC content of the soil. In the pot experiment, the production of NH4-N and NO3-N was both larger and more consistent when organic and inorganic nutrient sources were added together than when inorganic sources alone were used.

Autumn-sown sugarbeet (Beta vulgaris L.) responses (sugar yield, plant N-uptake and juice quality) were studied in relation to the residual NO3−-N in a soil of southwestern Spain which, for the previous five years (1989–93), had received high N rates, in accordance with conventional fertilization schedules used by farmers in the area. Three different combinations of fertilizers, supplying equal amounts of N, were used during the fertilization period (1989–93): a mineral fertilization treatment (MF, a complex 15N-15P2O5-15K2O) and two organo-mineral fertilization treatments (an olive mill wastewater sludge compost, AC, and a depotassified concentrated beet vinasse, V). All these treatments also received a top-dressing with urea (46% N). A control treatment (C), without fertilization was included for comparison.

During the major part of the beet growing season, the presence of almost four times as much mineral N in the 0·100 cm soil layer of previously fertilized plots (AC, V and MF) than in the unfertilized one (C), led to a significant increase (P<0·05) in total fresh weight yield and N-uptake, but also to a significant decrease (P<0·05) in sugar content and beet processing quality. The time course of NO3−-N concentration in sugarbeet petioles and the evolution of the nutritional state of leaf-blades gave advance information about the final response of the crop to the different fertilization treatments. Besides N, Na was the element which, due to the repeated and high fertilization rates applied, had a major effect in reducing the technological quality of the sugarbeet.

The effects of residual phosphate (P) fertilizer were monitored for six seasons on Mediterranean grassland. The phosphate fertilizer was originally applied annually at three rates (0, 25 and 60 kg P2O5/ha) for 7 years (1984–1990) to phosphate-deficient grassland at Tel Hadya, northern Syria. The herbage biomass productivity, botanical composition and the seed bank in the soil were monitored for six seasons (1991/92–1996/97). The experiment was grazed at two annual stocking rates (1·1 sheep/ha (low) and 2·3 sheep/ha (high)). The experimental site was typical of native grassland within the cereal zone of west Asia, where cropping is not possible because of shallow, stony soils and steep slopes.

Available soil phosphorus in May 1991 was 6·5, 20·8 and 40·1 mg P/kg under the 0, 25 and 60 kg P2O5/ha treatments and 6·6, 13·4 and 14·8 mg P/kg respectively, in May 1997. Yields of both legumes and total herbage (legume+grass+other species) were significantly influenced by the residual phosphate. Legume yields were between 6- and 7-times the control yield in the first two years of the study. This decreased with time but was still in the range of 1·5 to 1·9 times the yield of the control in 1997, six years later. Total herbage yield was consistently higher on the plots previously fertilized with P, which ranged between 1·5- and 2·5-times the control. Both legume seed and grass seeds were significantly larger with residual P, which ranged between 5·4- and 2·0-times the control for the legume and 2·5- and 1·4-times for the grasses. All these factors have practical implications for the use of P fertilizer on grassland which could help control soil erosion and improve livestock production on marginal lands on which farming communities largely depend.

Swards on wet hill land in mid-Wales, which had been converted from semi-natural rough grazing (dominated by Molinia caerulea, Nardus stricta or Calluna vulgaris) by surface-sowing in the period 1965–75, were examined in 1982, 1984 and 1996/97, using a point quadrat technique, in order to record the extent to which the sown species had persisted and the ingress of unsown species.

Lolium perenne had persisted to 1996/97 to the extent of being 20% of the sward on sites c. 300 m above sea level and 3% on sites at c. 500 m. Trifolium repens had persisted to c. 7% of the sward at both elevations. Festuca rubra was 6% of the sward on the higher sites and close to zero on the lower ones. The other main sown species, Phleum pratense, was present in only trace amounts on all sites in 1982, 1984 and 1996/97. The failure of T. repens to persist in larger amounts was attributed in part to below optimum soil pH and K and to the annual application of fertilizer N. The decline in L. perenne was attributed in part to severe winters on the higher sites and to below optimum soil K and pH. Agrostis spp. were the most prominent of the unsown species. Poa spp. appeared to stabilize at 8–15% of the sward. Stellaria media and Holcus lanatus became fairly prominent on the lower sites and Anthoxanthum odoratum on the higher sites. A supplementary study of the lower sites showed that the lime had been effective to at least the 20–25 cm layer of soil, that the distribution of plant roots in the soil was not abnormal, that T. repens plants had a satisfactory number of root nodules and that the soil contained a satisfactory number of earthworms.

The productivity of eight forage maize cultivars (Zea mays L.) in response to N was investigated in a 3-year field experiment located in Northern Germany. Nitrogen fertilizer applications were zero, 50 and 150 kg N/ha given each year shortly after sowing. Each cultivar was grown on the same plot, beginning in 1993, following the ploughing up of a 2-year old grass sward, to which slurry had been added. Plants were sampled regularly for dry matter (DM) production and N uptake. A non-linear regression equation was used to compare the data. Growth analysis and N uptake characteristics (maximum crop growth rate, duration of maximum crop growth rate, period until maximum crop growth rate, maximum N uptake rate, duration of maximum N uptake rate, period until maximum N uptake rate) which derived from the function were used to compare the cultivars.

The cultivars DM yield and N uptake were highest in 1993 and declined in the next two years partly due to a decrease in soil N mineralization following the ploughing of the grass sward and partly due to the drier weather conditions during the summers of 1994 and 1995. Duration of the maximum crop growth rate was greater during the cool year of 1993. In contrast, maximum crop growth rate was at its highest in the dry vegetation period of 1995. A significant N×cultivar interaction for growth analysis characteristics (P<0·05) was found in 1995. Cultivars with a high maximum crop growth rate (above the average value of the eight cultivars tested) and a short duration of maximum crop growth rate (below the average) accumulated more DM than those genotypes which showed an inverse relationship.

Increasing N yield was determined by increased maximum N uptake rather than by a greater duration of maximum N uptake. A significant N×genotype interaction for N uptake parameters (P<0·05) was found in 1994 and 1995. With some exceptions, cultivars with a high maximum N uptake rate (above average) accumulated more N per unit area compared to those genotypes which had low uptake rates. The exceptions had a longer duration of uptake, which could not, however, compensate for the lower rate. Maximum N uptake rate occurred earlier and duration of maximum N uptake rate increased compared to the start and duration of maximum crop growth rate. Especially in 1995, the amount of N taken up before the day of maximum crop growth rate accounted for 71% of total N uptake. The N uptake rate and the amount of accumulated N until the day of maximum crop growth rate were highly correlated with DM yield.

This result indicates the availability of genotypic variability in crop growth and N uptake rate to assist the improvement of DM yield by selection.

The response of spring field beans (Vicia faba) (cv. Victor) to irrigation applied before, during and after flowering and in combinations was studied on sandy soils, overlying sand, in Nottinghamshire, England in three years, 1993, 1994 and 1995. Irrigation was only needed pre-flowering in 1995, thus the opportunity to evaluate the effect at this timing was limited. Seasonal rainfall was higher than the long-term 30-year mean in 1993, but lower in 1994 and much lower in 1995.

Irrigation increased spring field bean yield in all three years, even in 1993 when a period of drought post-flowering caused a yield depression of 2 t/ha on unirrigated plots. The yield response for fully irrigated, compared with unirrigated field beans per 25 mm water was 0·34, 0·28 and 0·36 t/ha in 1993, 1994 and 1995 respectively. Irrigation applied during or post-flowering gave statistically significant yield increases and there were indications of greater efficiency of water use from post-flowering applications.

The effect of late season insect infestation on seed yield, yield components, oil content and oil quality of two canola species (Brassica napus L. and B. rapa L.) and two mustard species (B. juncea L. and Sinapis alba L.) was examined over 2 years. In each year, ten genotypes from each species were evaluated with late season insects controlled with either methyl parathion or endosulfan insecticides, and without insecticides. Major late season insect damage in 1992 was caused by cabbage seedpod weevil (Ceutorhynchus assimilis Paykull), while diamondback moth (Plutella xylostella L.) and aphids (primarily cabbage aphids, Brevicoryne brassicae L.) were major insect pests in 1993. Insecticide application was very effective in controlling diamondback moth larvae and adult cabbage seedpod weevils, but only partially effective in controlling aphids. Higher numbers of diamondback moth larvae were observed on mustard species compared to canola species. S. alba was completely resistant to cabbage seedpod weevil and there was no damage due to this pest observed. Aphid colonization was observed on plants from all species, but infestation on S. alba and B. rapa occurred too late to have a major effect on seed yield. Seed oil content of canola species was significantly reduced by insect damage although oil quality (indicated by fatty acid profile) was not affected by insect attack. Uncontrolled insect infestation reduced seed yield of canola species by 37 and 32% in B. napus and B. rapa, respectively. Least yield reduction occurred in S. alba, where average yield reduction from plants in untreated control plots was <10% of insecticide treated plants. S. alba, therefore, has good potential as an alternative crop suitable for northern Idaho because it can be grown with reduced late season insecticide application.

A natural rock phosphate, ‘Kodjari” (RP), sparingly soluble, indigenous to Burkina Faso, and a commercial single superphosphate (SSP) fertilizer were studied at Farako-Bâ in the Northern Guinea Savannah (NGS) and at Oipasse in the Sudan Savannah (SS) regions of Burkina Faso. A 3-year crop sequence either of cowpea, maize and sorghum or of cowpea, sorghum and sorghum was used, respectively, in the NGS and SS regions. The sorghum crop in the third year was not fertilized with any phosphorus (P) source. The objectives of the research were to study the direct and residual effects of the P fertilizers on soil fertility improvement in order to boost agricultural productivity in both regions.

Cowpea daylength-insensitive cultivars, in both regions, and maize and sorghum in the NGS and SS regions, respectively, responded more strongly to SSP than to RP fertilizer treatments. The optimum rate of SSP and RP source was 21·8 kg P/ha and 43·6 kg P/ha, respectively, for cowpea in both regions. The optimum rates of phosphorus fertilizer applied in the second year to maize in the NGS and to sorghum in the SS region, in addition to the optimum rate of P applied to cowpea the previous year, was 10·9 kg P/ha of SSP or 43·6 kg P/ha of RP, and 21·8 kg P/ha of SSP or 43·6 kg P/ha of RP, respectively. Both P sources had significant residual effects for up to 2 years. The agronomic effectiveness of RP relative to SSP, in the year of application of both fertilizers, was greater for cowpea than for maize in the NGS region and similar for cowpea and sorghum in the SS region; it increased markedly for the two subsequent cereal crops in both regions. Cowpea fertilization with both P sources proved, therefore, to be effective in improving the soil fertility and boosting the productivity of cereal crops in the 3-year crop sequence.

Canola (Brassica napus L.), yellow mustard (Sinapis alba L.) and intergeneric crosses of S. alba×B. napus were assessed for resistance (antixenosis) to the cabbage seedpod weevil (Ceutorhynchus assimilis Paykull). Pod trichomes did not appear to be a major factor in the resistance of S. alba to weevils. The number of feeding punctures and eggs per pod in S. alba was not significantly different in pods with trichomes than in those where the trichomes had been removed. Choice and no-choice laboratory tests examining feeding punctures and eggs laid per pod suggested that resistance in S. alba is not conferred in the intergeneric cross, S. alba×B. napus. Similar data on feeding and weevil oviposition were found in field test plots. However, despite many eggs being laid in S. alba×B. napus hybrid plants, fewer cabbage seedpod weevil larvae developed to exit the intergeneric hybrid pods. Glucosinolate analyses of leaves, pods and seeds showed that S. alba plants have a high concentration of p-hydroxybenzyl glucosinolate in all three plant parts, but B. napus has no p-hydroxybenzyl. Interestingly the intergeneric hybrid examined in this study had 62% and 60% of p-hydroxybenzyl concentration in the leaves and seeds, respectively, than was found in the S. alba parent. However, pod tissues contained very little (3%) compared with the S. alba parent. It is possible, therefore, that the adult cabbage seedpod weevil feeds on the pods of the intergeneric hybrid and lays eggs in the pod, because of the low concentration of p-hydroxybenzyl glucosinolate, but the larvae then fail to develop as they feed on the seeds containing high concentrations of p-hydroxybenzyl glucosinolate. It should be noted also that this hybrid produced pods that were more similar in physical shape to canola pods and that this may also be a factor determining cabbage seedpod weevil feeding and subsequent egg laying. In addition, both B. napus and the intergeneric hybrid produced 3-butenyl and 4-pentenyl glucosinolates in their pods, and degradation products (3-butenyl, and 4-pentenyl isothiocyanates) from these glucosinolate types, are known to be stimulatory kairomones that attract cabbage seedpod weevil. Further studies are being conducted to examine these factors in more detail.

The paper reports a study of the chemical composition of 25400 grass silages produced in England and Wales, of which 1040 were from the period 1961–63, 17900 (15480 clamped and 2420 baled) were from 1982–88 and 6460 (5640 clamped and 820 baled) were from 1992–93. A sample of each silage had been analysed for dry matter (DM), pH, crude protein (CP) and volatile N and in most cases for ash.

The baled silages were drier on average than the clamped silages, with a wider spread of DM contents. The clamped silages with the lowest pH on average were those with a DM content of c. 240 g/kg; clamped silages wetter than this were progressively higher in pH with reduction in DM content; clamped and baled silages drier than 240 g/kg were progressively higher in pH with increase in DM content, up to pH 5·9 in baled silages of 580 g DM/kg. At a given DM content pH was on average 0·6 units higher in baled than in clamped silages. The ash content of baled silages was similar to that of clamped silages. There was an increase in the mean CP content of clamped silages between 1961–63 and 1992–93 and a reduction in the proportion of CP which was volatile. The relationship between DM and volatile N as a proportion of total N was a curve rather than a straight line, with the steepest part of the curve at the lower end of the DM range. The wet silages harvested in 1961–63 were particularly high in volatile N. At a given DM content in the range 170–320 g/kg, volatile N was rather higher in baled than in clamped silages in both 1982–88 and 1992–93. Clamped silages with a DM content of at least 250 g/kg and baled silages with a DM content of at least 310 g/kg were typically well preserved in the sense of having <100 g volatile N/kg total N.

The response of two cultivars of dry harvest field peas (Pisum sativum), Solara and Bohatyr, to irrigation at different growth stages was studied on light soils overlying sand in Nottinghamshire, England in 1990, when the spring was particularly dry, in 1991 which had a dry spring and summer and in contrast, 1992, when rainfall was greater compared with the long-term (40 year) mean.

Solara, short haulmed and semi-leafless was more sensitive to drought than the tall conventional-leaved cultivar Bohatyr and gave a greater yield response to irrigation, particularly at the vegetative growth stage in the first two dry years 1990 and 1991, of 108% and 55% respectively, compared with unirrigated plots. Bohatyr was less sensitive to the timing of single applications.

In all years, peas irrigated throughout on several occasions produced the highest yields, but this was the least efficient use of water.