The author provides an overview of the problems and progress to date in studies of the respiratory health effects of multiple air pollutants. The problem arises because the preference for single (or even multiple) pollutants in lab studies does not replicate the conditions of actual exposure to complex pollutant mixtures for the general population. The limits of interpretation possible in more realistic studies is stressed. The examples of SO2-particulate matter and lead-nutritional factors are presented as two cases where laboratory-generated data lagged behind epidemiologic studies in their ability to corroborate multiple exposure-disease associations identified in field studies of humans. An extensive review of laboratory and field studies of multipollutant interactions is provided. The paper ends with a call for more field studies with enhanced measurements of exposures, and outcomes.