Human biomonitoring (HBM) of environmental contaminants plays an important role in estimating exposure and evaluating risk, and thus it has been increasingly applied in the environmental field. The results of HBM must be compared with reference values (RV). The term 'reference values' has always been related to the interpretation of clinical laboratory tests. For physicians, RV indicate 'normal values' or 'limits of normal'; in turn, toxicologists prefer the terms 'background values' or 'baseline values' to refer to the presence of contaminants in biological fluids. This discrepancy leads to the discussion concerning which should be the population selected to determine RV. Whereas clinical chemistry employs an altered health state as the main exclusion criterion to select a reference population (that is, a 'healthy' population would be selected), in environmental toxicology the exclusion criterion is the abnormal exposure to xenobiotics. Therefore, the choice of population to determine RV is based on the very purpose of the RV to be determined. The present paper discusses the concepts and methodology used to determine RV for biomarkers of chemical environmental contaminants.