A field study was carried out in the high rainfall zone (HRZ, >600 mm p.a.) of southern Australia from March 1994 to August 1997 to test the hypothesis that sown perennial grasses and liming could make the existing pastures more sustainable through better use of water and nitrogen. The site, on an acid duplex soil at Book Book near Wagga Wagga in southern New South Wales, was typical of much of the HRZ grazing country in southern New South Wales and north-east Victoria. The experiment consisted of 4 replicate paddocks (each 0.135 ha) of 4 treatments: annual pasture (mainly ryegrass Lolium rigidum, silver grass Vulpia spp., subterranean clover Trifolium subterraneum and broadleaf weeds) without lime, annual pasture with lime, perennial pasture (phalaris Phalaris aquatica, cocksfoot Dactylis glomerata and subterranean clover T. subterraneum) without lime, and perennial pasture with lime. Soil pH (0–10 cm) in the limed treatments was maintained at 5.5 (0.01 mol/L CaCl2), compared to 4.1 in the unlimed treatments. The pastures were rotationally grazed with Merino ewe or wether hoggets at a stocking rate which varied with the season, but was 10–25% higher on the limed pastures [14.8–17.3 dry sheep equivalent (dse)/ha] than the unlimed pastures. One replicate set of pasture treatments was intensively monitored for surface runoff, subsurface flow (at the top of the B horizon), water potential gradients and ammonium volatilisation. Other measurements of nitrogen inputs, transformations and losses were made on all paddocks. In a normal to wet year, surface runoff, subsurface flow and deep drainage (>180 cm depth) were about 40 mm less from the perennial than the annual pastures. The reduction in deep drainage under the perennials was about one-third to one-half (20–29 mm/year). The smaller loss of solution NO3– from the perennial pastures (up to 12 kg N/ha.year) suggested soil acidification under perennials was reduced by about 1 kmol H+/ha.year. Denitrification and volatilisation losses of N were small (1–12 kg N/ha.year). Nitrogen fixed by subterranean clover (above ground parts) ranged from 2–8 kg N/ha in the drought of 1994–95 to 128 kg N/ha in a normal year (1996). The soil-pasture nitrogen balance was positive for all treatments and averaged 76 kg N/ha.year over 2 years. The abundance of introduced and native earthworms increased from 85 to 250/m2 in the limed pastures between 1994 and 1997. Introduced species, such as Aporrectodea trapezoides, were especially responsive to lime. Animal production per hectare was 10–25% higher on pastures with lime. Critical gross margins per dse were lowest ($16/ha) for a long-lived perennial pasture (>15 years), and highest ($20/ha) for a short-lived perennial (5 years). Overall, there were substantial benefits in animal production, improved soil quality and water use from establishing perennial grass pastures with lime on these strongly acid soils.

Summary Mineral N accumulates in autumn under pastures in southeastern Australia and is at risk of leaching as nitrate during winter. Nitrate leaching loss and soil mineral N concentrations were measured under pastures grazed by sheep on a duplex (texture contrast) soil in southern New South Wales from 1994 to 1996. Legume (Trifolium subterraneum)‐based pastures contained either annual grass (Lolium rigidum) or perennial grasses (Phalaris aquatica and Dactylis glomerata), and had a control (soil pH 4.1 in 0.01 m CaCl2) or lime treatment (pH 5.5). One of the four replicates was monitored for surface runoff and subsurface flow (the top of the B horizon), and solution NO3– concentrations. The soil contained more mineral N in autumn (64–133 kg N ha−1 to 120 cm) than in spring (51–96 kg N ha−1), with NO3– comprising 70–77%. No NO3– leached in 1994 (475 mm rainfall). In 1995 (697 mm rainfall) and 1996 (666 mm rainfall), the solution at 20 cm depth and subsurface flow contained 20–50 mg N l−1 as NO3– initially but < 1 mg N l−1 by spring. Nitrate‐N concentrations at 120 cm ranged between 2 and 22 mg N l−1 during winter. Losses of NO3– were small in surface runoff (0–2 kg N ha−1 year−1). In 1995, 9–19 kg N ha−1 was lost in subsurface flow. Deep drainage losses were 3–12 kg N ha−1 in 1995 and 4–10 kg N ha−1 in 1996, with the most loss occurring under limed annual pasture. Averaged over 3 years, N losses were 9 and 15 kg N ha−1 year−1 under control and limed annual pastures, respectively, and 6 and 8 kg N ha−1 year−1 under control and limed perennial pastures. Nitrate losses in the wet year of 1995 were 22, 33, 13 and 19 kg N ha−1 under the four respective pastures. The increased loss of N caused by liming was of a similar amount to the decreased N loss by maintaining perennial pasture as distinct from an annual pasture.

The effects of lime additions on exchangeable and soil solution cations of four soil types in north-eastern Victoria are discussed. Liming significantly (P < 0.05) reduced the concentration of exchangeable (1 M KCl), extractable (0.01 M CaCl2), soil solution total and monomeric aluminium. Raising the soil pHCa to 4.8 decreased Ale, concentrations below 1 mg kg-1, Al saturation % of the effective cation exchange capacity below 5 and AlTot below 5 PM; and raising the soil pHCa to 5.8 decreased MnCa concentrations below 10 mg kg-1 and AlTot below 2�m on the four soil types used in this study. Grain yield responses were best described by the sum of the activities of the Al monomers. Where organic C was present, responses could also be attributed to the complexing of monomeric Al. Grain yield responses could not always be reliably predicted by the Al saturation % of the effective cation exchange capacity. Liming significantly (P < 0.05) increased the concentration of Ca in the ECEC, but the Ca activity was not well correlated with lime response for all sites. The In ratio of aCa2+/�aAl- mono shows promise in predicting negative responses to lime applications (with values > 6) where soil pHCa is less than 5. The combination of Ca activity and the sum of the activity of the Al monomers, together with organic C content, may provide a better description of the responsiveness of acid soils to lime applications.

Liming experiments were conducted at 13 sites (soil pH range 4.99-6.27, 0-10 cm depth) in the dryland cropping region of north-eastern Victoria with wheat grown at all sites and barley at 3 sites. Lime increased wheat yields at 9 of the 13 sites with the acid sensitive cultivar Oxley, but the yield increase was not correlated (r2=0.07) with exchangeable Al. Exchangeable A1 was closely related to pH (in 0.01 mmol/L CaCl2). The acid-tolerant wheat cultivars (Matong and Millewa) out-yielded Oxley at a soil pH (CaCl2) of 4.7 and the acid-tolerant cultivars were less responsive to liming. The barley responded to the lime treatment at each of the 3 sites. The use of acid tolerant crop species is recommended on these soils, but an improvement in the predictability of a lime response is required before liming is widely recommended.

The nutrient requirements of wheat grown on a soil ameliorated with lime and deep ripping were studied in a field experiment over 5 seasons in northeastern Victoria. Phosphorus (P) fertiliser was required when this soil was limed, although the lime treatment may have increased P uptake (P x lime interaction). The nitrogen (N) concentration in wheat foliage was increased with both lime and deep ripping. It is likely that the availability of N will be the most important factor for sustaining high yields following soil treatment. Molybdenum (Mo) fertiliser (62 g Mo ha-1) increased grain yield in 2 seasons. Molybdenum fertiliser increased grain weight by an average of 4.4% over 3 years, and also the N content of grain in the absence of lime, indicating that a deficiency of available Mo in the soil was affecting the assimilation of N in the wheat. The magnesium (Mg) concentration in the wheat foliage was very low (0.07-0.10%) without soil treatment, but there was no grain yield response when Mg fertiliser (10.5 kg Mg ha-1) was applied in the absence of lime. Lime and ripping both resulted in big increases in manganese (Mn) concentrations in the foliage. Calcium (Ca) and copper (Cu) concentrations were little affected and zinc (Zn) was not affected by lime (Ca, 0.35-0 55%; Cu, 7-20 �g g-1; Zn, 50-60 �g g-1 at 50 days after emergence), and each nutrient appeared to be adequate for wheat. The concentration of Mn in the foliage was high without lime, but did not approach published figures for critical toxicity concentrations. Liming the soil greatly reduced the manganese concentration.