Exogenic endocrine-active substances are also called 'Endocrine Disrupting Chemicals' (EDC). They imitate or hinder the function of natural endogenic hormones or disturb the synthesis or the metabolism of hormones or of hormone receptors. The Enzyme-Linked Receptor Assay (ELRA) can detect estrogenic and anti-estrogenic effects at the level of receptor binding and is a useful tool for the integrative detection of contaminant effects. Although the test system has been used repeatedly in sediment assessments, the questions have remained concerning how it responds to variations in the physico-chemical matrix. For some bioassays, the salinity of the sample is a critical factor. This is especially relevant when testing wastewater samples or when sediment-associated samples in the tidal reaches of rivers are tested. Sediments in the tidal reaches of rivers change their salinity several times a day. Against this background, it would be beneficial to have a test procedure of known salinity tolerance. On account of this, the salinity tolerance of the ELRA was tested, assessed with reference substances at several salinity levels, and compared with the E-Screen method and a Yeast Estrogen Screen (YES), which are also frequently applied in environmental testing. The aim of this paper was to explore when the salinity limits within these test procedures are applicable. The trials should reveal the working range to be expected, characterize the salinity-dependent variations in sensitivity of the test, and provide options for methodological adjustments to improve the stability against increased salinity. The ELRA was carried out with the human Estrogen Receptor alpha. (ER) using the same principle like a competitive immunoassay based on ligand-protein interaction. However, an essential difference is the use of a physiologically relevant receptor instead of an antibody as a linking protein. The ELRA measures the competition of sample estrogens and anti-estrogens against estradiol supplied as a BSA-coating conjugate for the binding site of dissolved ER. Estradiol or xeno-estrogen binding is quantified by a biotynilated anti-ER antibody and the subsequent measurement of peroxidase activity by a streptavidin-POD-biotin complex. The E-Screen was performed with the human breast cancer cell line MCF-7, which expresses the estrogen receptor constitutively. Cell proliferation depends on binding of estrogens or xeno-estrogens with the receptor. After incubation, estrogen-dependent cell growth was measured by sulforhodamin B staining. The YES was performed with a recombinant yeast strain, transfected with a receptor and a reporter plasmid bearing the estrogen receptor and a vitellogenin gene fused with the reporter gene lacZ. Estrogen or xeno-estrogen-dependent gene induction was measured indirectly by LacZ activity. The salinity levels were simulated in varying concentrations with NaCl from 0 to 40 per thousand or Artificial Sea Water (ASW) from 0 to 32 per thousand. The study characterized the factor 'salinity' for the prospective application fields of the ELRA. With reference substances such as 17-beta-estradiol, the ELRA showed classical sigmoidal concentration-effect relations in a range from 0.05 to 100 microg/l under physiological conditions. After a methodological adjustment to compensate decreasing receptor-binding affinity of estrogens and xeno-estrogens at higher salinity levels, the ELRA became applicable under salinity conditions up to concentrations of 20.5 per thousand. In tests, the ELRA reached under the influence of salinity a mean limit of detection of 0.062 microg/l 17-beta-estradiol. The mean relative inter-test error was around 11%. Above concentrations of 20.5 per thousand there is a risk of false negative assessment. Compared with the E-Screen method using the MCF7 cell line and the yeast estrogen test system (YES), the ELRA shows a lower sensitivity to 17-beta-estradiol. In the E-Screen, the cell proliferation was strongly reduced by sodium chloride induced cytotoxicity. In comparison with the E-Screen, the salinity tolerance of the YES and YAS methods is significantly higher. Despite adaption, total salinity tolerance could not be achieved with the ELRA. Freshwater samples were generally appraisable. Higher salinity levels above 20.5 per thousand would tend towards false negative results. The low inter-test error of 11% makes the ELRA suitable for the detection of estrogenic and anti-estrogenic potentials of single substances, substance mixtures, and of environmental samples. The ELRA is very fast and reproducible, it can be used for high-throughput screening in a microplate format at low cost, it is robust to microbial contamination, and is less susceptible to cytotoxic interferences than cell culture methods. In their established form, the YES and the E-Screen methods are not applicable for liquid phase testing at higher salinity conditions. The salinity-adapted test version of the ELRA described here shows a broader working range for samples. Native water samples of more or less brackish origin or high-salinity effluent samples are testable. Results of tests with sediment associated samples of different salinity will be subject of a forthcoming publication.