Seattle firefighters participate in a voluntary annual medical surveillance program including measurements of ventilatory capacity (FVC and FEV1) and single-breath diffusing capacity of carbon monoxide (DLCO). From 1989 to 1996, average % predicted DLCO (Crapo) for all participating firefighters declined from 94.4% (95% confidence interval [CI]: 93.4% to 95.5%) to 87.3% (95% CI: 86.2% to 88.3%), with no significant change in average FVC or FEV1. A random-effects regression model based on data from 812 firefighters with at least two annual sets of DLCO measurements showed the expected associations between DLCO and age, height, gender, race, ventilatory capacity, and smoking. In addition, two important temporal changes were observed, including, for an average firefighter, a large mean decline in DLCO of -1.02 ml/min/mm Hg associated with year of measurement, and a relatively smaller decline of -0.006 ml/min/mm Hg associated with number of fires fought. Although the stability of ventilatory capacity over time is reassuring, the marked temporal decline in diffusing capacity among this population of firefighters raises issues of concern. Interpretation of the observed decline poses a dilemma in terms of the reliability and efficacy of diffusing capacity as a screening tool, in whether DLCO is subject to unacceptable technical variability or whether it might provide more sensitive detection of early adverse respiratory effects of smoke inhalation.