The phosphorus fertilizer requirements and long term productivity of nitrogen-fertilized Gatton panic (Panicum maximum cv. Gatton) pastures, grazed by lactating dairy cows, were evaluated in a 7-year experiment. Cows grazed at 2·6 cows/ha on pastures that received annually 100 or 300 kg N/ha and each of 0, 22·5 or 45 kg P/ha. Cows received no energy supplements in years 1–3 and were offered molasses at 3·5 kg/day from year 4 to year 7. Cows grazed their experimental paddocks from the start of the wet season until they started to lose weight in the dry season.

In years 6 and 7 there was significantly less green pasture and leaf on offer in 300N pastures at 0P than with 22·5P and 45P. This was reflected in a reduced milk yield by cows at 300N/0P in these two years. There was no influence of rate of P fertilizer at 100N on milk yield in any year. Lactation milk yields at 300N in years 6 and 7 averaged 3930, 4310 and 4610 kg/cow (P<0·05) for 0P, 22·5P and 45P, respectively. Nitrogen fertilizer increased milk yield in each year (P<0·01) except the first. Milk yields at 100N and 300N averaged 2860 and 3320 kg/cow respectively in years 1–3 and 3720 and 4290 kg/cow in years 4–7.

The milk yield responses to P fertilizer were related to the greater amounts of pasture and green leaf on offer, which led to a higher proportion of leaf in the diet, and the response to P fertilizer was dependent on the rate of N fertilizer applied. Phosphorus intakes were estimated to be below that of published requirements for cows producing this quantity of milk. An annual model of P flow between plant, animal and soil pools demonstrated that at 100N/22·5P more P was returned to the soil as excreta (15·7 kg P/ha) than with 300N/22·5P (7·1 kg P/ha). The major pathway of return of P to the soil at 300N was through plant litter. Soil organic P was the largest, but least exploited, pool of phosphorus.

This study has illustrated how the demand for phosphorus by the plant in grazed pastures is modified by the input of N fertilizer, is poorly predicted from plant analysis and published standards for animal requirements, and indicates that a response in milk production may be mediated through the effects of P on leaf growth and not on dietary P content.