The association between daily variations in all-cause mortality from 1983-1991 in Rotterdam, the Netherlands, and ambient air pollution was investigated. Twenty-four-hour average concentrations of total suspended particulates, Black Smoke, ozone, sulfur dioxide, and carbon monoxide were available on a daily basis. Every other day, total iron content in total suspended particulates samples was available. Poisson regression analysis was used to study associations between air pollution and mortality; generalized additive models were used to adjust for confounders (e.g., seasonal trends, weather). Daily mortality was associated most consistently with previous-day concentrations of total suspended particulates (relative risk = 1.05 for a change of 91 microg/m3) and ozone (relative risk = 1.06 for a change of 67 microg/m3). Total iron was associated less consistently with mortality than total suspended particulate mass was. The associations of mortality with ozone and total suspended particulates were independent of sulfur dioxide and carbon monoxide. The relative risks of total suspended particulates and particularly ozone were higher for subjects older than 78 y. The relationship between mortality and ozone did not deviate significantly from linear. The relationship between mortality and total suspended particulates was linear below 100 microg/m3 and leveled off at higher concentrations. If a threshold exists for the effects on mortality of these components, it exists at very low levels.