Fine and ultrafine particles in ambient air are more consistently associated with severe adverse health effects than coarse particles. We assessed whether the effects of PM(2.5) on peak expiratory flow (PEF) and respiratory symptoms in asthma patients differ by the source or the chemical properties of particles. A panel of 57 adult asthmatics was followed for 181 days from November 1996 to April 1997 with 3 daily PEF measurements and diaries. Air quality, including elemental analyses of PM(2.5) filters every 2 days (n= 83), was monitored at a central site. Daily concentrations of PM(2.5) from different sources were estimated using principal component analysis and multiple linear regression. Associations of PM(2.5) from different sources with respiratory endpoints were examined using a generalized least squares autoregressive model after adjustment for covariates. PM(2.5) attributable to local combustion was consistently negatively associated with all measurements of PEF. One interquartile increase (1.3 microg/m(3)) in 5-day average concentrations of PM(2.5) attributable to local combustion was associated with an average 1.14 L/min decline in evening PEF (95% CI: -1.95 to -0.33 L/min). We also observed that PM(2.5) attributable to long-range transport was positively, and soil-derived PM(2.5) negatively, associated with PEF. No consistent associations were observed between source-specific PM(2.5) and respiratory symptoms or between individual chemical elements and any respiratory endpoints. Our results suggest that the negative effects of PM(2.5) on PEF in adult asthmatics are mainly mediated by particles related to local combustion sources.