A passive sampler (the polar organic chemical integrative sampler; POCIS) was assessed for its ability to sample natural estrogens (17 beta-estradiol, E2; estrone, E1 and estriol, E3) and the synthetic estrogen (17 alpha-ethynylestradiol, EE2) in the outlet of a sewage treatment works over several weeks. The performance of the POCIS was investigated and optimised in the laboratory before field deployment with high recoveries (66-99%) were achieved for all estrogens. Moreover, it was shown that POCIS does not exhibit any preferential selectivity towards any of the target compounds. The sampling rates of E1, E2 and E3 were 0.018 +/- 0.009, 0.025 +/- 0.014 and 0.033 +/- 0.019 L d(-1), respectively. Following field deployments of 28 days in the discharge of a sewage works, POCIS was shown to enhance the sensitivity of estrogen detection, especially for E3, and provide time-weighted average (TWA) concentrations of E1, E2 and E3, ranging from undetectable to 12 ng L-1 upstream of the outflow of a sewage treatment works, 13 to 91 ng L-1 at the outflow and 8 to 39 ng L-1 downstream of the outflow. This revealed that E1, E2 and E3 are not completely removed during sewage treatment, with concentrations most likely being maintained by contributions from conjugated estrogen analogues. Grab water samples showed considerable variation in the concentrations of estrogens over a longer period (6 months). The results confirm that POCIS is an effective and non-discriminatory method for the detection of low concentrations of estrogens in the aquatic environment.