The development and usefulness of biomarkers in the evaluation of relationships between occupational or environmental exposures and disease were discussed in this paper. The development of biomarkers was described. The need for evidence of their usefulness in resolving uncertainties about exposure and disease relationships was considered. A biomarker would be considered to be valid if it resolves uncertainty when epidemiological data is definitive, equivocal, and lacking. Examples of definitive biomarkers were blood lead (7439921) levels measured in bridge workers and purified protein derivatives in hospital workers to monitor control procedures. When equivocal evidence occurs, biomarkers can help inform about the potential existence of an exposure/disease relationship. The carcinogenicity of ethylene-oxide (75218) in humans is inconclusive, but evidence from studies points to genetic and cytogenetic effects. Biomarkers can be used to assist in the interpretation of inconclusive epidemiological evidence. Even without epidemiological data of an exposure disease relationship, biomarkers can be of use by contributing risk information about exposed groups and serving as early warning signals. A matrix was developed to help organize biomarker types and study designs. The horizontal axis contained a continuum of biomarkers between exposure and disease and biomarkers of susceptibility. The vertical axis contained a continuum of study designs and uses that reflect the development and validity of markers. The author concludes that the selection and use of biomarkers to resolve uncertainties or explain epidemiologic findings should be considered part of a process in which biomarkers are seen as part of a dynamic process of scientific exploration.