『Abstract
　Phosphorus (P) losses from intensively farmed dairy pastures can impair surface water quality. One of the first steps in mitigating this loss is to determine where in a field the potential for P loss is greatest. This study compared P export in overland flow from grazed pasture with areas that receive elevated P inputs and stock traffic (e.g. gateway, water trough, stream crossing and cattle lane). Intact soil blocks were removed, simulated rainfall applied and overland flow analysed for P fractions and suspended sediment (SS). Soil bulk density, hydraulic conductivity, porosity, Olsen P and water soluble P were also measured. P loss from the sites was in the order: trough＞crossing＞gateway＞pasture. Total P losses from the trough averaged 4.20 mg P/m² while the pasture exported 0.78 mg P/m². In addition, runoff from lane soil was measured with total P averaging 5.98 mg P/m², however the method used as different from the other soils. Using stepwise linear regression, Olsen P or H2O-P, ％ bare ground and ％ saturation were the most commonly occurring variables to predict P loss among the sites. This suggests that locating and minimizing the size of these areas in fields has the potential to significantly decrease P loss to surface waters.

Keywords: source areas; dairy; pasture; phosphorus; export; water quality』

Introduction
Materials and methods
　Study area
　Rainfall simulation
　Surface runoff and soil analysis
Results and discussion
　Soil properties
　Loss of P and SS in surface runoff
　Predicting P loss for sites
Conclusions
Acknowledgements
References