The potential of seven septic tank installations in the Perth (Western Australia) metropolitan area to contribute phosphate to the groundwater was investigated. The phosphate concentration in the soil solution below the soak wells and leach drains was measured using immiscible displacement and compared with the phosphate concentration of the water flowing into the systems. The phosphate sorbing properties of the subsoils were measured, and these were found to vary up to 100-fold within the same profile. A very strong correlation was established between a laboratory measure of the ability of the soil to sorb phosphate and the phosphate sorbed in the soil profile below leach drains and soak wells. The correlation held only for those systems for which little further phosphate was removed by reaction with the soil, and the phosphate in the soil solution was at or near the same concentration as the phosphate in the effluent. For a system receiving water from the bathroom, laundry and kitchen the phosphate concentration was as low as 8 p.g P/ml. For a system receiving water from a toilet only, the concentration was as high as 29 ng P/ml. In systems receiving water from both sources the values were intermediate. For systems that had been installed for more than a few years, the concentration of phosphorus in the soil water down to 6 m below the soak well and leach drain was similar to that in the effluent being discharged into the soil.