Two clones of 5-yr-old Norway spruce (Picea abies [L.] Karst.) were exposed to two atmospheric concentrations of CO2 (350 and 750μmol mol−1) and of O3 (20 and 75 nmol mol−1) in a phytotron at the GSF-Forschungszentrum (Munich) over the course of a single season (April-October). The phytotron was programmed to recreate an artificial climate similar to that at a high elevation site in the Inner Bavarian forest, and trees were grown in 40 1 containers of soil (pH 3.5) fertilized to achieve two levels of potassium nutrition; well fertilized and K-deficient. Foliar nutrient analyses performed at the beginning of the exposure indicated that the fertilization programmes achieved their goal without significantly altering the levels of other nutrients or the soil pH. At the beginning of the fumigation, foliar K concentrations were 7–9 mg g−1d. wt for well fertilized trees and 4–5 mg g−1 d. wt for trees receiving no supplemental K. Over the course of the season, differences between K treatments intensified so that by the end of the experiment there was a five to sixfold difference between foliar K concentrations. This was associated with slight, but significant (P < 0.05), decreases in S and Zn (and of Cu in the 1989 needle year age class) and higher levels of C, N and Mg in K-deficient trees. Foliar N concentrations were low for all trees (9–15 mg g−1needle d. wt) but were similar to levels found in the field. Elevated O3 was found to decrease significantly the C (P < 0.05) and N (P < 0.001) content of both current-year (1989) and previous-year (1988) needles independent of CO2 concentration, but apart from some minor changes in the concentrations of Cu and Mn in the current-year needles no other effects of the pollutant on plant nutrient status were found. In contrast, CO2 enrichment resulted in significantly (P < 0.01) lower concentrations of K and P (effects on Mg were also on the borderlines of statistical significance) in current-year needles, but there was no influence on the nutrient composition of the previous-year needles (although effects on N were on the borderlines of statistical significance). CO2 enrichment also increased (P < 0.05) the C:N ratio of both current-year and previous-year needles. One factor contributing to the decline in foliar K at elevated CO2 appeared to be a marked increase (25–30%) in the rate at which cations were leached from the canopy by repeated simulated acid mist (pH 4.0) events, and this effect occurred independently of the O3 concentration. The information presented will aid the interpretation of parallel studies examining the effects of elevated CO2 and/or O3 on seasonal changes in photosynthesis, non-structural carbohydrate content, antioxidants, tree growth and water use efficiency, and sheds further light on the growing scepticism concerning the role of O3 in the development of Mg and K-deficiency symptoms characteristic of certain types of forest decline in central Europe.