Chemical exposures vary tremendously within and between persons in different groups. This variability creates difficulties in characterizing levels of exposure and ultimately in relating exposures to health effects. In this lecture, Prof. Rappaport considers various alternatives for assessing exposures, namely environmental measurements, biomarkers, and kinetic or statistical models. Points are illustrated with recent studies conducted by his group, including occupational and environmental exposures to benzene and other VOCs and air levels of polycyclic aromatic hydrocarbons following the World Trade Center disaster. Although it is shown that each approach for assessing exposure has its advantages, some appear to be more appropriate than others in particular situations.