Attention is given to two different approaches to determine the water quality in relation to toxic stress.The first approach is based on direct observations on the state of health of the biota naturally occurring in the environment to be judged. In the original concept of biological indicators of water pollution the existence of differences in stress-susceptibility of different species is assumed. However, toxicity studies indicated that there are no species whose absence or presence gives information on the degree of toxic pollution. When attention is solely directed to health aspects of one species, a higher specificity and response-rate is obtained, but water quality indexing is not possible. Examples of effects of chemical pollution on the health of fish from the river Rhine are presented.The second approach is based on indirect observations, determining the water quality by examination of the state of health of organisms experimentally exposed to the water under controlled conditions. To save testing time it is useful to concentrate the toxic compounds prior to testing. An example of these methods is given, describing the water quality of the rivers Rhine and Meuse in terms of toxicity and mutagenicity.As both approaches are complementary with respect to ecological significance and specificity, it is recommended to apply them simultaneously to obtain appropriate information on environmental quality and stress factors.