Plant species typical of semi-natural grasslands were tested for their sensitivity to long-term ozone effects on growth under two irrigation regimes, and the relationship between ozone sensitivity, relative growth rate, and plant strategy was analysed. Individuals of 24 grasses, herbs and legumes were grown in pots and exposed for two seasons to four levels of ozone in open-top chambers: (a) CF, charcoal-filtered air, (b) CFAA, charcoal-filtered air plus ozone to match ambient levels, (c) CF+, charcoal-filtered air plus ozone added to match 1.5 times ambient levels and (d) CF++, charcoal-filtered air with ozone added to match twice ambient levels during selected episodes of 7–13 d. During the ozone episodes, one half of the plants received reduced irrigation (dry treatment) while the rest was kept well watered (wet treatment). The effect of ozone on the relative growth rate was determined during the initial growth period during the first season. Above-ground plant dry weight was determined three times during each season, and stubble dry weight at the end of the experiment. In most species, the effect of ozone on relative growth rate was not significant, except in Centaurea jacea and Lychnis flos-cuculi for which a significant stimulation of relative growth rate was found. There was a significant negative relationship between percentage relative growth rate and relative growth rate measured in the CF treatment, indicating that in species growing faster in clean air relative growth rate tended to be more sensitive to ozone than in slower-growing species. After the first re-growth period during the first season, a significant effect of ozone on above-ground biomass was found in five species, i.e. a decrease in Chrysanthemum leucanthemum, Trifolium pratense and T. repens, and an increase in Silene dioïca. During the second season no significant ozone effect was detected, except for an increase in stubble mass in C. leucanthemum. In 10 species, the dry treatment caused a significant decrease in biomass, but in Plantago lanceolata during the first year, and in Knautia arvensis and T. repens during the second year, an ozone×irrigation interaction was observed. Exposure–response patterns in the wet treatment varied between species. In some species growth tended to be increased in charcoal-filtered air plus ozone to ambient levels, in others either a continuous increase or decrease in biomass was observed with increasing ozone. Trifolium repens showed the lowest AOT40 (ozone accumulated above a threshold of 40 ppb) corresponding to a 10% change in biomass. The comparison between exposure–response pattern between the wet and dry treatments revealed reduced, increased, or unchanged ozone sensitivity. The shift was related to plant strategy and not to the specific sensitivity to the dry treatment. The specific response to both ozone and dry treatment tended to be related to Grime's triangular classification of plant strategies, i.e. the C-S-R model. In species with a large component of C (competitor strategy), and/or R (ruderal strategy), biomass was more sensitive to both stresses, compared with species with a high component of S (stress tolerators), and ozone×drought interactions were more pronounced. The results indicate that specific relative growth rate is an important determinant for the potential sensitivity of grassland species to long-term effects of ozone, and that reduced irrigation causes no general protection from the effects of ozone on growth.