Legume-based pastures, particularly those containing a large proportion of lucerne (alfalfa, Medicago sativa), have a prodigious capacity to fix atmospheric N2. Budgets of N show that permanent pastures in south-eastern Australia, when growing with no management limitations, can supply more N than is removed in animal products and can eventually lead to excess soil N. For a mixed crop-livestock farm, legume-dominant ley pastures occupying ?40% of the land area can maintain a stable N balance. The actual performance of pastures on farms normally falls below the potential. Pastures are being replaced by crops in the wheat-sheep zone and, to a lesser extent, in the high-rainfall zone. Pasture productivity, as indicated by the area topdressed, the mean stocking rate, input of superphosphate and sale of pasture legume seed has decreased in the period 1990-2010. It is therefore likely that N2 fixation by pastures is falling sharply in the wheat-sheep zone and is static or falling slightly in the high-rainfall zone. Reversing the decrease in N2 fixation by pastures will become important if the real price of N fertilisers increases, as seems likely because the efficiency of fertiliser synthesis is approaching a maximum and the reserves of natural gas feedstock will eventually be depleted. Increased N2 fixation by pastures will depend on more profitable grazing industries, improved management methods and genotypes, and re-adoption of ley pastures by farmers. There is evidence that profitability of grazing enterprises is approaching that of crops so investment in pasture science is likely to lead to increased N2 fixation.