Effects of different concentrations of ozone on grain filling,leaf ozone flux (CFO 3 0 ). The flag leaves senesced earlier and flag leaf senescence and final grain yield in field-grown spring the grain-filling duration was significantly shorter at higher ozone exposure compared to F (−5, −13 and −18% in wheat (Triticum aesti7um L. cv. Dragon) were studied using NF1+, NF2+ and NF3+ , respectively). The relative grain-open-top chambers. The hypothesis tested was that an ozoneinduced reduction in grain yield is mainly related to an filling rate did not differ between the treatments. The 1000grain weights were 10, 28 and 37% lower, and the grain yields enhanced senescence and a shortening of the grain-filling were 15, 29 and 46% lower than F in NF1+ , NF2+ and period. The plants were exposed to filtered air (F), non-filtered NF3+, respectively. Ozone exposure had no significant effect air without extra ozone (NF) or non-filtered air with 3 different levels of ozone added (NF1 +, NF2+ and NF3+ ). The on the number of grains per unit ground area or on straw mean daytime (08.00-20.00 h) ozone concentrations during yield, but significantly reduced the harvest index and increased the exposure period (31 days) were 7, 20, 34, 48 and 62 nmol the grain protein concentration in NF2 + and NF3 + commol − 1 in F, NF, NF1+ , NF2+ and NF3+, respectively. pared to F. The grain yield was negatively correlated with the The corresponding ozone doses, expressed as the accumulated ozone dose, expressed either as AOT40 or as CFO 3 with or exposure over a concentration threshold of 40 nmol mol − 1 without an ozone flux threshold. The 1000-grain weight was (AOT40), were 0, 12, 1989, 5881 and 10375 nmol mol − 1 h, positively correlated with the grain-filling duration (R 2 = respectively, and 884, 2594, 4557, 6188 and 7900 mmol m − 2 , 0.998), which in turn was positively correlated with the leaf respectively, expressed as the calculated cumulative flag area duration (R 2 = 0.989). (Tottman and Broad 1987). After anthesis, both the number of ears per unit area, apart from late tillers which normally only contribute little to the yield, and the maximum number of grains per ear are already fixed. Then, the rate and the duration of the grain-filling processes are the dominant factors influencing the final grain yield, within the limits set by the preceding life history of the plant. The maximum grain size is related to the number of endosperm cells in the developing grains, which is determined before starch storage Abbre6iations -AOT40, accumulated ozone exposure over a concentration threshold of 40 nmol mol − 1 ; CFO 3 , cumulative ozone flux; CFO 3 0 , cumulative ozone flux using no ozone flux threshold; CFO 3 2 , cumulative ozone flux over an ozone flux threshold of 2 nmol m − 2 s − 1 ; F, filtered air; NF, non-filtered air; (NF1 + , NF2 + and NF3 + ), non-filtered air with addition of ozone; OTC, open-top chamber; PPFD, photosynthetic photon flux density; RGFR, relative grain-filling rate; VPD, vapour pressure deficit.