To examine the effect of automobile exhaust on respiratory symptoms and pulmonary function, the authors studied 175 tunnel and turnpike workers employed by the Massachusetts Turnpike Authority on two occasions 3 yr apart beginning in 1972. A standard respiratory symptom and illness questionnaire was administered, spirometry was performed, and proximal hair lead and blood lead content were measured as biologic indices of automobile exhaust exposure. One hundred nine (63%) workers were current cigarette smokers, 41 (23%) were exsmokers, and 24 (14%) had never smoked. Smoking was strongly related to respiratory symptoms of cough (P less than .001) and phlegm production (P less than .001), but not to wheezing (P = .41), breathlessness (P = .14), bronchial asthma (P = .13), or frequent chest colds (P = .14). When workers were stratified by smoking status, no effect could be seen between high automobile exhaust exposure as measured by a variety of parameters and all of the above respiratory symptoms and illnesses. The level of pulmonary function [forced expiratory volume in 1 sec (FEV1.0) and forced vital capacity (FVC)] was not related to past or current exhaust exposure in a cross-sectional analysis when we controlled for age, height, and cigarette consumption. In a prospective analysis of 84 of these workers, the observed changes in FEV1.0 and FVC over 3 yr were unrelated to exhaust exposure after controlling for age, height, cigarette consumption, and initial level of pulmonary function.