Air parcel trajectory data are used in two ways to elucidate the temporal trend analysis of 5 years of peroxyacetyl nitrate (PAN) and O3 observations at a rural site in eastern Canada. In the first method, ''probability of residence'' contours are constructed to determine the most probable origin of air parcels containing the highest and lowest 10% of PAN and O3 mixing ratios. High PAN is found to emanate always from areas of high anthropogenic activity, except when the transport path is to a large extent over the ocean, especially in the summer. High O3 originates from the same regions except in the winter when because of low photochemical activity, O3 is actually titrated and air from less populated areas is richer in O3. The trajectories indicate that transport at higher altitude leads to higher mixing ratios; this is especially the case for O3 in winter and spring. The complementary method of clustering the trajectories has allowed qualitative derivation of seasonal cycles for background and polluted air masses of different origin. Background air from the north shows a distinct PAN maximum in March; it is discussed that this could be either due to enhanced photochemistry or to the import of polluted air from the Arctic. Polluted air masses show the same March peak, but a second peak in late summer/early fall. Oceanic air from the south has a January maximum in PAN but otherwise is consistently low in PAN. O3 also has a spring maximum, but in polluted air it is broader stretching into the summer. It is postulated that this is due to additional O3 formation in the summer, while in the winter, actual O3 loss is indicated for polluted air. By inference it is deduced that the second PAN peak in polluted air is also due to additional formation in comparison with background air, while in the summer, extra PAN loss mechanisms operate that are less important for O3. Indications for both dry deposition and thermal decomposition have been found. ¿ American Geophysical Union 1995