The objective of this article is to examine differences in the effect of ambient particulate matter on first, second, and overall hospital admissions for respiratory disease among the elderly. We studied 8989 adults 65 yr of age or older living in the greater Vancouver area who were admitted to hospital for any acute respiratory disease (ICD-9 codes 460-519) between June 1, 1995, and March 31, 1999. Time-series analysis was used to evaluate the association between respiratory admissions and daily measures of particulate matter (PM10, PM2.5, and PM10 - 2.5) in urban air, after adjustment for gaseous copollutants (CO, O3, NO2, and SO2) and meteorological variables. Repeated admissions for respiratory disease were common among the elderly. Approximately 30% of the subjects were readmitted to hospital after the first admission; 9% had more than 2 admissions for respiratory disease during the 4-yr study period. PM10 - 2.5 was significantly associated with the second and overall admissions for respiratory disease, but not with the first admission. The adjusted relative risks for an increment of 4.2 microg/m(3) in -day average PM10 - 2.5 concentrations were 1.03 (95% confidence interval: 0.98-1.09) for the first admission, 1.22 (1.10-1.36) for the second admission, and 1.06 (1.02, 1.11) for overall admissions. There was no significant association between PM2.5 and hospital admissions for respiratory disease among the elderly. Our data suggest that (1) people with a history of respiratory admissions are at a higher risk of respiratory disease in relation to particulate air pollution in urban areas, (2) analyses based on overall rather than repeated hospital admissions lead to lower estimates of the risk of respiratory disease associated with particulate air pollution, and (3) PM10 - 2.5 has a larger effect on respiratory admissions than PM2.5.