Trifolium repens L. and Lolium perenne L. were grown in monocultures and bi‐species mixture in a Free Air Carbon Dioxide Enrichment (FACE) experiment at elevated (60 Pa) and ambient (35 Pa) CO2 partial pressure (pCO2) for three years. The effects of defoliation frequencies (4 and 7 cuts in 1993; 4 and 8 cuts in 1994/95) and nitrogen fertilization (10 and 42 g m–2 y–1 N in 1993; 14 and 56 g m–2 y–1 N in 1994/95) on the growth response to pCO2 were investigated. There were significant interspecific differences in the CO2 responses during the first two years, while in the third growing season, these interspecific differences disappeared. Yield of T. repens in monocultures increased in the first two years by 20% when grown at elevated pCO2. This CO2 response was independent of defoliation frequency and nitrogen fertilization. In the third year, the CO2 response of T. repens declined to 11%. In contrast, yield of L. perenne monocultures increased by only 7% on average over three years at elevated pCO2. The yield response of L. perenne to CO2 changed according to defoliation frequency and nitrogen fertilization, mainly in the second and third year. The ratio of root/yield of L. perenne increased under elevated pCO2, low N fertilizer rate, and frequent defoliation, but it remained unchanged in T. repens. We suggest that the more abundant root growth of L. perenne was related to increased N limitation under elevated pCO2. The consequence of these interspecific differences in the CO2 response was a higher proportion of T. repens in the mixed swards at elevated pCO2. This was evident in all combinations of defoliation and nitrogen treatments. However, the proportion of the species was more strongly affected by N fertilization and defoliation frequency than by elevated pCO2. Based on these results, we conclude that the species proportion in managed grassland may change as the CO2 concentration increases. However, an adapted management could, at least partially, counteract such CO2 induced changes in the proportion of the species. Since the availability of mineral N in the soil may be important for the species’ responses to elevated pCO2, more long‐term studies, particularly of processes in the soil, are required to predict the entire ecosystem response.

Acclimation of photosynthesis to elevated CO 2 has previously been shown to be more pronounced when N supply is poor. Is this a direct effect of N or an indirect effect of N by limiting the development of sinks for photoassimilate? This question was tested by growing a perennial ryegrass (Lolium perenne) in the field under elevated (60 Pa) and current (36 Pa) partial pressures of CO 2 (pCO2) at low and high levels of N fertilization. Cutting of this herbage crop at 4-to 8-week intervals removed about 80% of the canopy, therefore decreasing the ratio of photosynthetic area to sinks for photoassimilate. Leaf photosynthesis, in vivo carboxylation capacity, carbohydrate, N, ribulose-1,5-bisphosphate carboxylase/ oxygenase, sedoheptulose-1,7-bisphosphatase, and chloroplastic fructose-1,6-bisphosphatase levels were determined for mature lamina during two consecutive summers. Just before the cut, when the canopy was relatively large, growth at elevated pCO 2 and low N resulted in significant decreases in carboxylation capacity and the amount of ribulose-1,5-bisphosphate carboxylase/oxygenase protein. In high N there were no significant decreases in carboxylation capacity or proteins, but chloroplastic fructose-1,6-bisphosphatase protein levels increased significantly. Elevated pCO 2 resulted in a marked and significant increase in leaf carbohydrate content at low N, but had no effect at high N. This acclimation at low N was absent after the harvest, when the canopy size was small. These results suggest that acclimation under low N is caused by limitation of sink development rather than being a direct effect of N supply on photosynthesis.Acclimation of photosynthesis to growth in elevated pCO 2 has frequently been shown to be more marked under suboptimal N supply (Drake et al., 1997). Growth in low N limits the development of the shoot and root, and therefore the capacity for utilization of the additional photoassimilate formed under elevated pCO 2 . Low N may therefore exacerbate the accumulation of carbohydrate observed under elevated pCO 2 (Webber et al., 1994; Drake et al., 1997). Alternatively, nitrate accumulation within the plant can alter gene expression (Paul and Driscoll, 1997; Scheible et al., 1997), and could lead to different patterns of acclimation to elevated pCO 2 depending on the N supply. Wheat grown under limiting N supply showed a greater loss of Rubisco in response to elevated pCO 2 than plants grown with free access to N (Rogers et al., 1996). This appeared to result from an accumulation of soluble carbohydrates in leaves, resulting in sugar repression of the expression of the genes encoding the LSU and the small subunit of Rubisco (rbcL and rbcS, respectively) (Stitt, 1991; Sheen, 1994; Krapp and Stitt, 1995; Koch, 1996).Most studies of acclimation to elevated pCO 2 under different levels of N nutrition have been conducted in containers in the laboratory. However, Arp (1991) demonstrated that such restriction of rooting volume might accentuate acclimation to elevated pCO 2 . In addition to the p...

~Symbiotic N, fixation is one of the main processes that introduces N into terrestrial ecosystems. As such, it may be crucial for the sequestration of the extra C available in a world of continuously increasing atmospheric CO, partial pressure (pC0,). The effect of elevated pC0, (60 Pa) on symbiotic N, fixation ('5N-isotope dilution method) was investigated using Free-Air-C0,-Enrichment technology over a period of 3 years. Trifolium repens was cultivated either alone or together with Lolium perenne (a nonfixing reference crop) in mixed swards. Two different N fertilization levels and defoliation frequencies were applied. The total N yield increased consistently and the percentage of plant N derived from symbiotic N, fixation increased significantly in T. repens under elevated pC0,. All additionally assimilated N was derived from symbiotic N, fixation, not from the soil. In the mixtures exposed to elevated pCO,, an increased amount of symbiotically fixed N (+7.8,8.2, and 6.2 g m-' a-' in 1993, 1994, and 1995, respectively) was introduced into the system. lncreased N, fixation is a competitive advantage for T. repens in mixed swards with pasture grasses and may be a crucial factor in maintaining the C:N ratio in the ecosystem as a whole.

The extent of the response of plant growth to atmospheric CO enrichment depends on the availability of resources other than CO. An important growth-limiting resource under field conditions is nitrogen (N). N may, therefore, influence the CO response of plants. The effect of elevated CO (60 Pa) partial pressure (pCO) on the N nutrition of field-grown Lolium perenne swards, cultivated alone or in association with Trifolium repens, was investigated using free air carbon dioxide enrichment (FACE) technology over 3 years. The established grassland ecosystems were treated with two N fertilization levels and were defoliated at two frequencies. Under elevated pCO, the above-ground plant material of the L. perenne monoculture showed a consistent and significant decline in N concentration which, in general, led to a lower total annual N yield. Despite the decline in the critical N concentration (minimum N concentration required for non-N-limited biomass production) under elevated pCO, the index of N nutrition (ratio of actual N concentration and critical N concentration) was lower under elevated pCO than under ambient pCO in frequently defoliated L. perenne monocultures. Thus, we suggest that reduced N yield under elevated pCO was evoked indirectly by a reduction of plant-available N. For L. perenne grown in association with T. repens and exposed to elevated pCO, there was an increase in the contribution of symbiotically fixed N to the total N yield of the grass. This can be explained by an increased apparent transfer of N from the associated N-fixing legume species to the non-fixing grass. The total annual N yield of the mixed grass/legume swards increased under elevated pCO. All the additional N yielded was due to symbiotically fixed N. Through the presence of an N-fixing plant species more symbiotically fixed N was introduced into the system and consequently helped to overcome N limitation under elevated pCO.

The class I chitinases are vacuolar proteins implicated in the defense of plants against pathogens. Leaves of transgenic Nicotiana sylvestris plants homozygous for a chimeric tobacco (Nicotiana tabacum) chitinase gene with Cauliflower Mosaic Virus (CaMV) 35S RNA expression signals usually accumulate high levels of chitinase relative to comparable leaves of non-transformed plants. Unexpectedly, some transgenic plants accumulated lower levels of chitinase than nontransformed plants. We call this phenomenon silencing. The incidence of silencing depends on the early rearing conditions of the plants. When grown to maturity in a greenhouse, approximately 25% of plants raised as seedlings in closed culture vessels were of the silent type; none of the plants raised from seed in a greenhouse showed this phenotype. Silencing is also developmentally regulated. Plants showed three patterns of chitinase expression: uniformly high levels of expression in different leaves, uniformly low levels of expression in different leaves, and position-dependent silencing in which expression was uniform within individual leaves but varied in different leaves on the same plant. Heritability of the silent phenotype was examined in plants homozygous for the transgene. Some direct descendants exhibited a high-silent-high sequence of activity phenotypes in successive sexual generations, which cannot be explained by simple Mendelian inheritance. Taken together, the results indicate that silencing results from stable but potentially reversible states of gene expression that are not meiotically transmitted. Gene-specific measurements of chitinase and chitinase mRNA showed that silencing results from co-suppression, i.e. the inactivation of both host and transgene expression in trans. The silent state was not correlated with cytosine methylation of the transgene at the five restriction sites investigated.