Much progress has been made in abating the impacts on aquatic ecosystems of industrial wastewaters, intensive agriculture, and large urban centres. Nowadays the short term consequence of stress from such sources is less frequently the death and destruction of fish populations or entire communities of organisms. Large scale fish kills are now less common. Scientific attention has shifted to the effects on ecosystems of long term exposures to sublethal stressors. Although use of the term 'ecosystem health' is a topic of debate, the metaphor can usefully reflect a state of well being or absence of impaired survival, growth, reproduction, and recruitment problems in an ecosystem's key organisms. The present article explores the use of the physiological and biochemical responses of organisms to stressors, the so-called 'biomarkers, ' to assess and study the sublethal effects of chemical stressors in fish. Fish were chosen as the organism of example because they are key components of practically all aquatic ecosystems. Most biomarkers can be used as an early warning that fish have been exposed to putative stressors, and can often be used to help identify the stressor(s). Biomarkers, however, tell us little about the eventual ecological outcome of such exposures. Some biomarkers are mechanistically linked to toxic modes of action, and can thus be classed as 'biomarkers of effect' at the level of the individual organism. None, however, have been fully validated and calibrated as predictive indicators of adverse ecological effects at either the population or community levels. Biomarkers are valueable as part of a broader strategy for monitoring the effects of stressors on aquatic ecosystems.