The development of air pollution standards ideally involves the integration of data from the disciplines of epidemiology, controlled clinical studies, and animal toxicology. Epidemiological studies show statistical associations between health outcomes and exposure; they cannot establish a definite cause-effect relationship. The utility of toxicological studies is to establish this relationship. Recently, there was simultaneous promulgation of a new National Ambient Air Quality Standard (NAAQS) for particulate matter < 2.5 microns in aerodynamic diameter (PM2.5) and a revised NAAQS for ozone (O3). The O3 NAAQS was based, in part, on a sound foundation of toxicological data from controlled exposure studies in humans and animals. It also relied on epidemiological studies of hospital admissions for respiratory diseases. Such studies also served as important bases for the new PM2.5 NAAQS. However, the most influential bases for the PM NAAQS were the numerous and generally consistent epidemiological studies that associated exposure with premature mortality in susceptible subpopulations and the inability of numerous hypothesized confounding factors to negate the associations. Using ozone and PM as examples, this paper discusses the scientific basis for NAAQS promulgations in situations in which the underlying database differed greatly in the extent of toxicological support.