In 1990, the sale, marketing and distribution of bituminous coal, primarily used for domestic heating, were banned across Dublin. This study exploits the potential of a 'natural experiment' to assess a temporal association between adjusted annual lung cancer death rates and the changing annual mean urban air-pollution concentrations in Dublin from 1981 to 2000. Annual mean 'black smoke' (BS) concentration was used as an indicator variable for the urban air-pollution mixture. Log-linear Poisson regression model (with an offset) was used to estimate adjusted rate ratios of lung cancer death rates between two periods (1981-1990 and 1991-2000) relative to the year 1990. A significant (p<0.0001) two-third decline in BS concentration (28.2 microg/m(3)) was seen between the two periods [pre-ban (46.4 microg/m(3)) vs. post-ban (18.2 microg/m(3))]. Relative to 1990 (rate ratio= 1 ), a slightly greater decline (2%) in death rates was achieved in the pre-ban period (1981-1990) when mean annual BS concentrations were very high, but a lower decline (1%) was seen in the post-ban period (1991-2000) corresponding to very low mean annual BS concentrations. In other words, a further fall in adjusted rates in lung cancer was achievable both in the pre-ban and the post-ban periods when simultaneously controlling for BS and smoking. A temporal association thus observed between lung cancer death rates and the changing BS concentrations suggests that control of particulate air-pollution could further reduce lung cancer rates, irrespective of smoking patterns.