Isotopically labelled ozone ((18)O(3)) is an ideal tool to study the deposition of O(3) to plants and soil, but no studies have made use of it due to the technical difficulties in producing isotopically enriched ozone. For (18)O(3) to be used in fumigation experiments, it has to be purified and stored safely prior to fumigations, to ensure that the label is present predominantly in the form of O(3), and to make efficient use of isotopically highly enriched oxygen. We present a simple apparatus that allows for the safe generation, purification, storage, and release of (18)O(3). Following the purification and release of O(3), about half (by volume) of the (18)O is present in the form of O(3). This means that for a given release of (18)O(3) into the fumigation system, a roughly identical volume of (18)O(2) is released. However, the small volume of this concurrent (18)O(2) release (100 nmol mol(-1) in our experiment) results in only a minor shift of the much larger atmospheric oxygen pool, with no detectable consequence for the isotopic enrichment of either soil or plant materials. We demonstrate here the feasibility of using (18)O as an isotopic tracer in O(3) fumigations by exposing dry soil to 100 nmol mol(-1) (18)O(3) for periods ranging from 1 to 11 h. The (18)O tracer accumulation in soil samples is measured using gas chromatography/isotope ratio mass spectrometry (GC/IRMS), and the results show a linear increase in (18)O/(16)O isotope ratio over time, with significant differences detectable after 1 h of exposure. The apparatus is adapted for use with fumigation chambers sustaining flow rates of 1 m(3) min(-1) for up to 12 h, but simple modifications now allow larger quantities of O(3) to be stored and continuously released (e.g. for use with open-top chambers or FACE facilities).