Molecular epidemiology is a promising new tool in the study of environmental carcinogenesis and, particularly, in cancer prevention. Genetic damage and mutation are believed to play a critical role in chemical carcinogenesis. By incorporating biologic markers of dose or response to carcinogens (such as mutagenicity of body fluids, carcinogen-DNA adducts, chromosomal abnormalities, and somatic cell mutation) into human bio-monitoring or molecular epidemiologic studies, one can detect potential hazards early and increase the power of studies to determine causal relationships. Such markers can also improve extrapolation of risks from experimental animals to humans or from one human population to another. During the past 5 years, there has been considerable progress in developing markers and applying them in human (largely pilot) studies. A review of this experience--with particular emphasis on carcinogen-DNA adducts--affords a better awareness both of the significance of biologic markers and the research needed to fill gaps in understanding. Criteria for marker validation and sound study design are presented that should greatly enhance future research.