Background : multicentric studies in Europe are required to gain knowledge on the short-term impacts of PM2.5 and PM10 2.5. We present an analysis of the short-term associations between particulate matters (PM10, PM10 2.5 and PM2.5) and mortality by causes, age-groups and seasons in nine French cities. Methods : the associations between PM and daily mortality were investigated in each city using a generalized additive Poisson regression model for the 2000 2006 period. The percent increases in the mortality rate were estimated for a 10 Œg/m3 increase and for an interquartile range increase in PM levels in each city, for the whole year and by season. The models also compared the PM effect observed on "non-warm" days and on "warm" days. Results : a significant effect of PM10 (+0.8% CI 95% [0.2; 1.5] for a 10 Œg/m3 increase) and PM2.5 (+0.7% [0.1; 1.6]) on all-ages non-accidental mortality whole year was observed. The largest impacts were observed on all-ages cardiovascular mortality during summer for PM2.5 (+5.1% [1.8; 8.4]) and PM10 2.5 (+7.2% [2.8; 11.7]). These estimates were lowered when the model included PM2.5 and PM10 2.5. We also report a significant interaction between warm days and PM. Adjusting PM on ozone did not modify the results for the whole year, but decreased the estimates for summer, when a high correlation is observed between these pollutants. Conclusions : our results confirm the short-term impacts of PM10 on mortality, even at concentrations complying with the European annual regulation. They underline the short-term impacts of PM2.5 and PM10 2.5 and call for the setting of regulation values for these PM indicators. (R.A.)