This paper summarises the main results, conclusions and recommendations from nine related projects on Phosphorus (P) loss from agriculture to water during 2001 to 2003. The final reports from studies are available from the EPA (www.epa.ie, from mid 2006). Irish soils were naturally P deficient and most P now in top soils on intensive farms was added in chemical fertiliser and purchased feedstuffs, during the past forty years. This reservoir of P has a large influence on P losses. Up to 1960's Irish soils had a low Soil Test P (STP) and eutrophication was not a major problem. Since then there has been an input of about 3 million tonnes P in fertiliser and purchased feed to the 4.0 m ha of agricultural land; equivalent to doubling total P in agricultural top soils, on average. Field plot experiments show the build-up of soil P increases P loss from grassland fields to water. The highest loads and concentrations of P in overland flow from grassland occurred during autumn after the relatively dry summer period. Exports of P from nested sub-catchments in three catchments (with similar STP) were studied. The Dripsey (Co. Cork) and Oona (Co. Tyrone) catchments were consistently high (greater than 2.0 kg/ha/yr total P in 2002) but losses from the calcareous Clarianna (Co. Tipperary) catchment were only 10% of the above two catchments. The chemical and hydrological characteristics of soils in catchments are important; the free draining Clarianna soils ensure that surface runoff rarely occurs and the inherent capacity of the soil to absorb P prevents significant losses of P to the river. Laboratory studies indicated that the chemical characteristics of soils can influence land as a source of mobile P, e.g. soils that have high amounts of aluminium will hold P more strongly and mineral soils with high organic matter will bind P more loosely. Dung from grazing animals and soil microbial biomass plays an important role in P cycling in soils. Calculations showed that from 18 to 27 weeks storage is required for different parts of Ireland with less safe slurry spreading days in the wetter North West region. Recommendations based on the studies and proposals for future research are included. It is recommended that STP for grassland should be maintained at the lowest level compatible with agronomic production, this will mean an STP in Index 2 (3.1 to 6.0 mg/l Morgan P in soil) in most cases and fertiliser P should not be applied to soils in Index 4 (>10). The Morgan STP and soil sampling depth of 100mm should continue to be used as it provides a good indication of agronomic response and the relative loss potential of P in overland flow. A new initiative, including: a) an awareness campaign for farmers on how to reduce P loss to water and b) a national field soil testing programme is recommended. It is concluded that a reduction in P loss to water will be achieved when P inputs are approximately balanced with outputs on fields with optimum STP for grassland production and when soils with higher P levels adjust to suitable levels by non application of P fertiliser or excess P inputs in animal feed. This may take several years or decades for soils with high STP and high P buffering capacity. Most effort should be directed at reducing P loss from these soils in catchments sensitive to eutrophication due to P loss from agriculture. The P loss from farmyards needs to be controlled as a priority. Sustainable agriculture and water quality is possible if farmers and other stakeholders understand and adopt best management practices.