The risk assessment process must encompass all available toxicological data and scientific evidence on the plausible toxicities of a chemical or chemical mixture. As an extension to the approaches used to conduct risk assessments on chemical mixtures, a preliminary scheme, analogous to the IARC classification of carcinogens, is proposed to express the weight of evidence for the interactions in binary mixtures. This scheme is based on composite representation of all the toxicological evidence from animal bioassays and human data, pharmacokinetics studies, metabolism studies, and structure activity relationships. In addition, factors such as the relevance of route, duration and sequence of exposure, toxicological significance of interactions and the quality of in vivo and in vitro data are taken into consideration. The scheme yields an alphanumeric classification that can be used for qualitative risk assessment, and has the potential, as demonstrated in this paper, for quantitative application to site-specific risk assessments. Furthermore, the scheme can be used to estimate interactions or form hypotheses concerning binary interactions. It is flexible and allows all pertinent information to be incorporated in a methodical and consistent manner. Research is needed to identify interaction patterns for simultaneous and sequential exposure scenarios of chemical pollutants in order that this scheme may be developed further and its usefulness and limitations may be tested.