When soils are extracted with sodium bicarbonate to measure soil test phosphorus (P) by the standard Colwell procedure, the concentration of P in the extract solution is measured using the Murphy and Riley procedure. This measures P as orthophosphate (PO4) and most of the organic P extracted is not measured. The total (inorganic plus organic) P extracted can be measured by digesting an aliquot of the extract solution with perchloric acid to convert all the extracted P to PO4before measuring the P concentration by the Murphy and Riley procedure. This is called total Colwell P. Data from one crop and two pasture field experiments in Western Australia are presented in which soil test P measured in 1991 and 1992 by the standard and total Colwell procedures were compared. Fertiliser P residues, including organic P, accumulate in the topsoil because, (i) the fertiliser is applied to the surface of pastures, and (ii) crops are being increasingly sown by minimum (conservation) tillage when P is concentrated in the top approximate 5 cm of the soil compared with about 10 cm when crops are sown conventionally. Consequently, to measure soil test P in the present study, soil samples were collected in February 1991 to either 5 or 10 cm depth. For the two Colwell procedures and both sampling depths, soil test P was related to: (i) the level of P applied in previous years and to (ii) plant yield measured later on that year (the soil test P calibration). Soil test P was larger for the total Colwell procedure and for soil samples collected to 5 cm depth. Total Colwell P as the dependent variable was closely correlated with standard Colwell P as the independent variable. The relationship between soil test P and the level of P applied, and the calibration relating yield to soil test P, were different for standard and total Colwell P. Neither method can be claimed to be significantly better than the other. It is concluded that use of the total Colwell procedure instead of the standard procedure is not warranted. Likewise, there was no evidence that collecting soil samples to 5 cm to measure soil test P provided better estimates of the current P status of soils than collecting soils to the standard 10 cm depth. There is therefore no need to change the sampling depth.