Drainage water loss of dissolved phosphorus (DP) and particulate phosphorus (PP) from four artificially drained, loamy soil, arable catchments was studied over a 1-yr period comparing intensive (hourly samples pooled weekly) and discrete (weekly) sampling strategies. Annual total phosphorus (TP) loss determined by intensive sampling averaged 0.241 kg P ha(-1) yr(-1), and was particularly high (0.627 kg P ha(-1) yr(-1)) at a catchment partly comprised of low-lying riparian areas used for grazing, where animal manure input was consequently high. At the other three catchments annual loss was lower (0.071-0.169 kg P ha(-1) yr(-1)). The major part of TP loss was accounted for by DP in the case of the catchment with the highest loss (71%), but by PP at the other three catchments (55-71%). Annual loss of PP ranged from 0.043-0.182 kg P ha(-1) yr(-1), being episodic and mainly occurring during storm events. Significant linear relationships were established between the concentrations of PP and particulate matter. Evidence was found indicating exhaustion of fine particulate matter availability during storm events, either in the soil macropores or in the drain pipe itself. As compared with intensive sampling, the discrete sampling strategy normally used for monitoring purposes underestimated P loss by more than 50% because most P loss occurred during storm events. It is therefore concluded that reliable estimates of P losses from artificially drained catchments can only be obtained using a sampling strategy that incorporates intensive sampling during storm events. Copyright (C) 1996 Elsevier Science Ltd