Drained and undrained grassland lysimeter plots were established in 1982 on a clay loam of the Hallsworth series at a long-term experimental site in south-west England. The plots were continuously grazed by beef cattle, and received fertilizer at either 200 or 400 kg N ha -I per annum to the existing permanent sward, or at 400 kg N ha-' to a new sward, reseeded to perennial ryegrass following cultivation. Drainage water was monitored at V-notch weirs and sampled daily for the analysis of nitrate-N. Seven years of data are presented (five years for the reseeded swards). On the drained plots a large proportion of the rainfall was routed preferentially down large pores to the mole drains, whilst on the undrained plots, drainage was mainly by surface runoff. The average quantities of nitrate N leached per year were 38.5,133.8 and 55.7 kg ha-' from the old sward that received 200 and 400 kg N ha-', and from the reseed that received 400 kg N ha-' fertilizer, respectively. Ploughing and reseeding resulted in a two-fold reduction in leaching, except during the first winter after ploughing, and twice as much leaching occurred after a hot, dry summer as after a cool, wet one. Nitrate concentrations in drainage from either drained or undrained plots were rather insensitive to rainfall intensity, such that concentration was a good predictor of nitrate load for a given drainage volume. The drainage volume determined the proportion of the leachable N that remained in the soil after the winter drainage period. Initial (peak) concentrations of nitrate N ranged, on average, from 55 mg dm-3 for the drained old sward that received 400 kg N ha-' fertilizer, to 12 mg dm-3 for the undrained sward at 200 kg N ha-' fertilizer input. Concentrations of nitrate N in drainage from similar, unfertilized plots rarely exceeded 1 mg dm-3. The results suggest that manipulating the nitrate supply can lessen leaching and that the route of water through soil to the watercourse determines the maximum nitrate concentration for a given load.

The model simulates the cycling of N in grassland systems grazed by beef cattle and predicts the annual amount of N in liveweight gain, and the amounts lost through ammonia volatilization, denitrification and leaching, on the basis of fertilizer application and soil and site characteristics. It aims to provide a better understanding of the way in which these various factors interact in their influence on N transformations. The model has been programmed to run on IBM-compatible personal computers and responds rapidly to changes in input parameters.The model has been constructed from the average annual amounts of N passing through various components of the N cycle in ten field systems grazed by beef cattle. The amounts were either measured directly or were calculated from empirical sub-models, assuming a balance between inputs to, and outputs from the soil inorganic N pool. The model is given wide applicability through the inclusion of a mineralization sub-model which is sensitive to soil texture, sward age, previous cropping history, and climatic zone. Another important sub-model determines the partitioning of soil inorganic N to either plant uptake or the processes of loss: the proportion partitioned to plant uptake decreases as the total amount of soil inorganic N increases.Outputs from the model indicate that fertilizer N has a strong influence on ammonia volatilization, denitrification and leaching at a given site but that, over a range of sites with a given rate of fertilizer N, total loss and the proportions lost by the three processes are greatly influenced by the amount of N mineralized by the soil. The model indicates how fertilizer N should be matched with mineralization to limit gaseous and leaching losses and to achieve optimum efficiency of N use in grazing systems.

Annual liveweight gain of beef cattle (steers) grazing grass pasture fertilized with 200 kg N\ha was compared over a period of 7 years with that of steers grazing grass\white clover pasture given no artificial N fertilizer at North Wyke, Devon, UK. Nitrogen lost by leaching over the ensuing winter drainage periods was monitored from both pastures. Nitrogen leaching loss from the fertilized pasture over an extended period of 13 years (1983-95) is also reported.The average annual liveweight gain of the steers grazing the grass\clover pasture (0n81 t\ha) was 19 % lower than that of the steers grazing the N-fertilized grass pasture (1n00 t\ha). The average annual loss of nitrate-N by leaching in winter drainage from the grass\clover pasture (13 kg\ha) was only 26 % of that recorded from the fertilized grass (50 kg\ha). A possible reason for this difference may arise from the previous history of the grass\clover pasture which had been ploughed in 1982, causing a flush of N mineralization and consequently greater immobilization of N in the soil in subsequent years.Losses of N each winter by leaching measured over a 13-year period from the fertilized grass were highly correlated (P 0n001) with the preceding summer's soil moisture deficit, with the highest losses following dry summers. The nitrate-N concentration in the drainage water exceeded the European Union limit in drinking water (11n3 mg\l) in the initial 25 mm of drainage during 11 of the 13 autumns. The average loss of N each winter (53 kg\ha) was equivalent to 26 % of the fertilizer-N applied annually. Immediate losses of N by leaching of fertilizer applied in early spring and throughout one very wet summer (1993) were minimal.

The yield and quality of herbage produced by six grasses (perennial ryegrass, cocksfoot, timothy, rough-stalked meadow grass, tall fescue and Italian ryegrass) were examined both without irrigation and under two irrigation regimes. Water was applied according to the potential soil water deficit (potential SWD): the soil was either partially returned to field capacity (FC) after each cut or fully returned to FC whenever the potential SWD reached 25 mm. The swards were cut either at 3 (C 3 ) or 6 (C 6 ) week intervals over a 2 year period. Partial irrigation increased yields by 12-14 % in the first year and by 36-58 % in the second. Full irrigation produced little more growth than partial irrigation in the first year (maximum SWD, 188 mm) but increased yield by 78-93 % in the second, very dry, year (maximum SWD, 311 mm). Under treatment C 3 response per unit of water applied was similar with both partial and full irrigation, but under C 6 the response was greater with partial (2-86 kg D.M./m 3 ) than with full irrigation (1-79 kg D.M./m 3 ).There were marked differences between the species in their ability to grow under drought conditions in the second year of the experiment. Without irrigation, roughstalked meadow grass and Italian ryegrass did not survive the drought. The performance of tall fescue was markedly superior to both perennial ryegrass and cocksfoot in these conditions. Of the surviving grasses timothy made least growth.

A long-term field grazing experiment was begun in 1982 to examine the impact of efficient field drainage on herbage and animal production from swards on an impermeable clay loam in the south-west of England. Drained and undrained lysimeter plots (each of 1 ha) were established on the existing permanent sward and received annual applications of fertilizer N of 200 or 400 kg ha~'. Similar plots were initially ploughed and reseeded with Lotium perenne (cv. Melle), and received fertilizer N at an annual rate of 400 kg ha " '. All plots were continuously stocked by beef cattle and stock numbers were adjusted to maintain a constant sward height and to avoid poaching damage. Results for the first 5 years show that the benefits from drainage were modest and, for beef production, unlikely to pay for its costs over the shorter term. The main benefit was in spring when herbage dry matter yield was 1 l'^o greater on the drained plots, but with no significant interaction with fertilizer N level or sward type. This benefit was reduced to 3*70 on an annual basis, due to the effect of the larger soil water deficits sustained by the drained swards in mid-season. Drainage increased the

Crop yields and soil properties of grazed grass-clover pasture and arable plots at the Grasslands Research Institute, Hurley, UK. were measured over 30 years . Yields of herbage declined from 9-6 t DM/ha in the first harvest year to 4-6 t in years 5 to 8 but subsequently increased (after liming) to give a mean yield of 8-2 t for 1963-84. Yields of barley in plots receiving a seed-bed application of 45 kg N/ha averaged (1962-80) 31 t/ha and, in plots top dressed with an additional 100 kg N/ha, 41 t/ha but after 1969, response to additional N and yield under both N treatments showed a marked decline.Soil carbon increased during the first 10 years under pasture from 1-2 to 1-8% and soil N from 0-13 to 0-17%. These increases, equivalent to 1 t C and 75 kg N/ha per year, occurred at the same time as the initial decline in herbage yield noted above. During the subsequent 20 years, C and N remained relatively constant and herbage yields increased. Soil C and N decreased under 30 years' arable management to 0-8 % C and 0-1 % N, equivalent to a loss of 290 kg C and 25 kg N/ha per year.Major factors that appear to have influenced herbage yields were (i) soil moisture, (ii) accumulation of N in the soil organic matter during the first 8 years and the subsequent increase in the rate of mineralization, and (iii) the action of the earthworm population in creating a stone-free, fertile layer of top soil. As for barley, it was concluded that yields could not be maintained indefinitely on this soil series, as a result of the loss of soil structure and the development of a compact layer of soil at plough depth.

SummaryDrainage lysimeters were used to estimate actual transpiration of a perennial ryegrass sward; this is compared with the calculated values for potential transpiration. Losses to drainage of nitrate nitrogen (NO3–-N) under differing levels of application of nitrogen fertilizer were examined. When 250 kg N/ha were applied annually losses of NO3-N were negligible; when 500 kg N/ha were applied up to 142·9 kg N/ha were lost to drainage. Concentration of NO3-N in the leachate with an application of 250 kg N/ha was low (1–5 ppm); with 500 kg N/ha the concentration ranged from 10 to 109 ppm. Losses of NO3–-N were particularly severe during winter following a dry autumn. When drainage occurred during the growing season losses were minimal.Concentrations and quantities of K, Ca, Mg, Na, P, Cl and S (as sulphate) leached wore also examined.

When high rates of fertilizer N are applied to grass in an area of low rainfall, considerable amounts may be lost to drainage in the subsequent winter if the N is not fully utilized owing to dry soil conditions during the growing season (Garwood & Tyson, 1973). The data now presented were obtained from the lysimeters described in that paper. The soil is a sandy loam, overlying chalk. The swards of perennial ryegrass (Lolium perenne cv. S. 23) were cut four times each year. Two rates of N were applied as ammonium nitrate (Nitrana), 250 kg (LN) and 500 kg N/ha per year (HN). Two-fifths of this N were applied in March and one-fifth after each of the first three cuts.