Effects of elevated CO2 on the protein concentration of food crops: a meta-analysis

Authors

  • DANIEL R. TAUB,

    1. Biology Department, Southwestern University, 1001 East University Avenue, Georgetown, TX 78626, USA,
    2. Environmental Studies Program, Southwestern University, 1001 East University Avenue, Georgetown, TX 78626, USA
    Search for more papers by this author
  • BRIAN MILLER,

    1. Biology Department, Southwestern University, 1001 East University Avenue, Georgetown, TX 78626, USA,
    Search for more papers by this author
  • HOLLY ALLEN

    1. Environmental Studies Program, Southwestern University, 1001 East University Avenue, Georgetown, TX 78626, USA
    Search for more papers by this author

Daniel R. Taub, 1001 East University Ave., Georgetown TX 78626, USA, fax +1 512 863 1696, e-mail: taubd@southwestern.edu

Abstract

Meta-analysis techniques were used to examine the effect of elevated atmospheric carbon dioxide [CO2] on the protein concentrations of major food crops, incorporating 228 experimental observations on barley, rice, wheat, soybean and potato. Each crop had lower protein concentrations when grown at elevated (540–958  μmol mol−1) compared with ambient (315–400  μmol mol−1) CO2. For wheat, barley and rice, the reduction in grain protein concentration was ∼10–15% of the value at ambient CO2. For potato, the reduction in tuber protein concentration was 14%. For soybean, there was a much smaller, although statistically significant reduction of protein concentration of 1.4%. The magnitude of the CO2 effect on wheat grains was smaller under high soil N conditions than under low soil N. Protein concentrations in potato tubers were reduced more for plants grown at high than at low concentrations of ozone. For soybean, the ozone effect was the reverse, as elevated CO2 increased the protein concentration of soybean grown at high ozone concentrations. The magnitude of the CO2 effect also varied depending on experimental methodology. For both wheat and soybean, studies performed in open-top chambers produced a larger CO2 effect than those performed using other types of experimental facilities. There was also indication of a possible pot artifact as, for both wheat and soybean, studies performed in open-top chambers showed a significantly greater CO2 effect when plants were rooted in pots rather than in the ground. Studies on wheat also showed a greater CO2 effect when protein concentration was measured in whole grains rather than flour. While the magnitude of the effect of elevated CO2 varied depending on the experimental procedures, a reduction in protein concentration was consistently found for most crops. These findings suggest that the increasing CO2 concentrations of the 21st century are likely to decrease the protein concentration of many human plant foods.

Introduction

Atmospheric concentrations of carbon dioxide (CO2) have been steadily rising from preindustrial values of approximately 280 μmol mol−1 to a current global mean of approximately 380 μmol mol−1 (Keeling & Whorf, 2005; IPCC, 2007). Concentrations are projected to increase to approximately 540–958 μmol mol−1 by the year 2100 (IPCC, 2001).

Numerous effects of elevated atmospheric CO2 concentrations on plants have been documented, including changes in plant elemental composition. As growth CO2 concentrations increase, plants typically show increased concentrations of carbon in their tissues, with correspondingly reduced concentrations of other elements, including nitrogen (Cotrufo et al., 1998; Gifford et al., 2000), phosphorus (Gifford et al., 2000) and several trace elements (Loladze, 2002). Along with changes in elemental composition, changes have frequently been noted in macromolecular composition, with proteins (which contain substantial amounts of nitrogen and sulfur) decreasing and relatively carbon-rich molecules such as carbohydrates increasing in concentration at higher concentrations of atmospheric CO2 (e.g. Poorter et al., 1997).

Such changes in plant composition might be expected to have important implications for the growth and nutrition of animals that consume plant material. In a recent meta-analysis, Zvereva & Kozlov (2006) found that insect herbivore performance was diminished when feeding on plants grown at elevated vs. ambient concentrations of CO2. Several authors have also considered the possible implications of altered chemical composition of plants in elevated CO2 for human nutrition. Loladze (2002) argued that elevated CO2 may lead to ‘globally imbalanced plant stoichiometry’ and negatively impact human nutrition, particularly with regard to micronutrients such as zinc and iodine. Idso & Idso (2001) in a narrative review, examined a number of studies on the effects of elevated CO2 on food composition. They found that, for a given nutrient, the results of CO2 enrichment varied: for example various studies have shown that CO2 may increase, decrease or have no effect on the protein concentration of crops.

In spite of the potential for elevated CO2 to affect the nutritional composition of foods, there have been few attempts at meta-analysis or quantitative synthesis of the available data. Loladze (2002) synthesized data from five published studies on wheat grains and found reductions in the concentration of eight elements (including nitrogen) when plants were grown at elevated CO2. Jablonski et al. (2002) performed a meta-analysis of studies on the effects of elevated CO2 on plant reproductive characteristics, including seed N concentration, for several seed/grain crops. They found that growth at elevated CO2 resulted in significant decreases in seed N for wheat and barley, but not for soybean or rice.

Neither of these studies focused exclusively on the effect of elevated CO2 on crop nutrient composition and both surveyed only a limited number of crops and a limited selection of the available literature on those crops. In order to rigorously address the question of how elevated CO2 affects the protein composition of food crops, we performed a comprehensive meta-analysis, attempting to include all available data for all food crop species.

Materials and methods

In summer 2006, we performed a variety of keyword searches in the broad scientific bibliographic databases Agricola and Web of Science (ISI), and in one database exclusively focused on the effects of CO2 on plants (Jones & Curtis, 2000). We also exhaustively examined the entries in a second database exclusively focused on the effects of CO2 on plants (Strain & Cure, 1994) and examined the lists of cited references for a number of relevant review and meta-analysis papers (including Rogers & Dahlman, 1993; Conroy et al., 1994; Cotrufo et al., 1998; Pleijel et al., 1999; Gifford et al., 2000; Nakagawa & Horie, 2000; Amthor, 2001; Idso & Idso, 2001; Ainsworth et al., 2002; Jablonski et al., 2002; Loladze, 2002; Kimball et al., 2003; Poorter & Navas, 2003; Davis et al., 2004; Ainsworth & Long, 2005).

We included in our analyses studies that compared crop plants grown at ambient CO2 (315–400 μmolmol−1) with those grown at elevated CO2 (540–958 μmol mol−1). This ambient CO2 range was intended to represent CO2 concentrations of the present and recent past. The low end of this range corresponds to the ambient atmospheric CO2 concentration for the late 1950s while the upper end is somewhat higher than current global means (Keeling & Whorf, 2005) but well within the variation currently seen across terrestrial locations (Ziska et al., 2001). The elevated CO2 range was chosen to reflect the likely atmospheric concentrations for the year 2100. This range represents the central 80% of the projected CO2 concentrations for the year 2100 generated by the Bern-CC model for a variety of SRES emissions scenarios (IPCC, 2001, Appendix II.2.1).

We included data from all studies that reported either total protein or total N for the portion of a crop plant commonly used as a human food item. Total N was regarded as the equivalent of total protein as the most common analytical methods for determination of total protein in plant tissues are based on measurement of N followed by multiplication by a conversion factor (e.g. AOAC, 1990). We did not include studies that reported data for only a fraction of total protein, such as soluble protein, concentrations of individual proteins, or a subset of the amino acids. We included multiple data points from an individual publication if the data were obtained from separate experiments or from substantially different treatments within a single experiment (e.g. different species, cultivars, soils or temperatures). Data were taken directly from tables or text whenever possible. When necessary, data presented in graphical form were digitized using datathief III software (Bas Tummers, Eindhoven, the Netherlands).

We analyzed each crop species separately to obtain an estimate of the mean effect of CO2 for that species. This was accomplished using standard meta-analytic methods (Cooper & Hedges, 1994) with the natural logarithm of the response ratio (protein concentration in elevated CO2/protein concentration in ambient CO2) as the experimental effect metric (Hedges et al., 1999). As few studies reported measures of variation in the data (standard deviation, standard error or variance), we were unable to analyze the data using parametric meta-analysis techniques, and instead calculated 95% confidence intervals for effect size by resampling the data using 9999 bootstrap replicates for each analysis (Adams et al., 1997). These bootstrap confidence intervals were corrected for asymmetry bias (Rosenberg et al., 2002).

We found appropriate studies for fifteen crop species, although most of these species were represented by only one to three experimental observations. We report here data only for the five species for which we found four or more observations in the literature (barley, Hordeum vulgare L.; rice, Oryza sativa L.; wheat, Triticum aestivum L.; soybean, Glycine max (L.) Merr.; potato, Solanum tuberosum L.). Appendix A lists the publications from which data were obtained that were included in the analyses.

We also assessed potential patterns of variation in the effects of CO2 by including categorical variables in meta-analysis models, including CO2-enrichment technology (FACE, open-top chamber, closed-top field chamber, glasshouse or growth chamber), rooting environment (‘pot’– rooted within a container or ‘ground’– rooted directly in the earth) and whether the chemical analyses were performed on flour or on whole grains (for wheat). Heterogeneity among levels of these categorical variables was assessed by resampling following Adams et al. (1997). For multiple comparisons, groups were considered to differ significantly if their 95% confidence intervals did not overlap, providing a conservative test of group differences (Grantz et al., 2006). Analyses comparing enrichment technologies could not be performed for potato as all available data were from a single experimental type (open-top chambers).

We also considered the effect of several environmental variables (temperature, water supply, soil nitrogen and atmospheric ozone) on the size of the CO2 effect. To do so, we examined studies that had compared the effect of elevated CO2 on crop protein concentration at two or more levels of another environmental variable. We report results for each environmental variable for which there were at least seven experimental observations on an individual species. When there were more than two levels of the relevant environmental variable, we used the data from the highest and lowest levels.

All meta-analyses were performed using the software meta-win 2.0 (Rosenberg et al., 2002). While all analyses were performed on log-transformed data, values have been back transformed for presentation.

Results

All of the crops included in the analysis had significantly lower protein concentrations when grown at elevated vs. ambient CO2 (Fig. 1). For potato, the mean reduction in protein was 13.9% and for the grain crops (barley, rice and wheat) the reduction in protein was 15.3%, 9.9% and 9.8%, respectively. For soybean the reduction was a much smaller 1.4%.

Figure 1.

 Response of crop protein concentrations to growth at elevated CO2 for five major crops. Means and 95% confidence limits are depicted. Numbers of experimental observations for each species are in parentheses.

For wheat and soybean, the two species with the largest sample size of studies, there were significant differences among CO2 enrichment technologies in the effect of CO2 on protein concentration (Fig. 2; P<0.001 for each species). For both species, the largest effects of elevated CO2 were seen in open-top chamber studies. For barley and for rice there were no significant differences among CO2 enrichment technologies.

Figure 2.

 Response of crop protein concentrations to growth at elevated CO2 in studies using various CO2 enrichment technologies. Means and 95% confidence limits are depicted. Numbers of experimental observations are in parentheses. FACE, free-air CO2 enrichment; OTC, open-top chamber; CTC, closed-top field chamber; GH, glasshouse; GC, growth chamber.

There was some suggestion that CO2 had a greater effect in studies performed in pots than in studies in which plants were rooted in the ground (Fig. 3). Comparing all studies performed in pots with those involving plants rooted in the ground, no species showed a significant rooting environment effect, although for rice there was a near-significant trend toward a greater CO2 effect in pot studies vs. ground studies (P=0.051; Fig. 3a). However, for both soybean and wheat it was additionally possible to make the pot vs. ground comparison focusing solely on studies performed in open-top-chambers (OTC; Fig. 3b). For both species, in OTC studies there was a significantly greater effect of CO2 when plants were grown in pots (Fig. 3b; for soybean P=0.005, for wheat P=0.003).

Figure 3.

 Response of crop protein concentrations to growth at elevated CO2 in studies with plants rooted in pots vs. rooted in the ground. (a) All studies, (b) studies in open-top chambers. Means and 95% confidence limits are depicted. Numbers of experimental observations are in parentheses.

The effect of elevated CO2 on protein concentrations was affected by environmental variables in several instances. Protein concentrations in wheat grains were reduced more when elevated CO2 was applied to plants at low N supply than at high N supply (Fig. 4a). Across this group of studies, grain protein concentrations were decreased by 16.4% in low nitrogen treatments compared with 9.8% in high nitrogen treatments, with this difference statistically significant (P=0.038).

Figure 4.

 Response of crop protein concentrations to growth at elevated CO2 in studies that varied (a) nitrogen, (b) temperature or (c) ozone. Each point represents one study. Percent change is the percent change in protein concentration under elevated [CO2]. The diagonal lines represent the 1 : 1 relationship.

There was no significant difference in the effect of elevated CO2 on wheat grain protein concentration between plants grown at high vs. low temperatures (Fig. 4b), although the trend was for elevated CO2 to have a greater effect on protein concentrations at high than at low temperatures.

The effect of ozone differed greatly between species (Fig. 4c). In potato, tuber protein concentrations were decreased by 19.3% under high ozone compared with 7.7% under low ozone, with this difference statistically significant (P=0.013). For soybean, the effect of ozone was the reverse of that seen for potato. Elevated CO2 increased protein concentrations by 3.0% under high ozone, and decreased protein concentrations by 1.3% under low ozone, with this difference statistically significant (P=0.005).

For wheat, the effect of CO2 on protein concentration was nearly twice as large when protein was measured in grain rather than flour. (Table 1; P=0.004).

Table 1.   Response of crop protein concentrations to growth at elevated CO2 for studies on wheat in which protein concentration was measured in grains or flour
Item
measured
Number of
observations
Percent decrease in
protein concentration
under elevated CO2
95%
confidence
interval
Grain8711.08.7–13.3
Flour286.04.1–8.2

Discussion

The effects of elevated CO2 on protein concentrations seen in these analyses differ somewhat from those reported in previous quantitative syntheses. Both Jablonski et al. (2002) and Loladze (2002) found mean decreases in protein in wheat of more than 20%, compared with the mean 9.8% decrease seen here. Jablonski et al. (2002) also reported a decrease in protein concentration for barley of nearly 20%, but did not find a significant effect of CO2 concentration on protein for either rice or soybean. The differences between the results of the present study and those of previous studies are likely attributable to the much less extensive survey of the literature carried out in the previous studies, neither of which exclusively focused on this particular question. Jablonski et al. (2002) included 11, 8, 3 and 4 observations for wheat, barley, rice and soybean, respectively, compared with 120, 20, 14 and 56, respectively, in the present study. Loladze (2002) included data for wheat drawn from five publications, compared with 22 in the present study. It is, therefore, likely that the present study provides a more representative estimate than previous studies of the effects of elevated CO2 on the protein concentrations of crops.

The mechanism(s) by which elevated CO2 decreases tissue concentrations of N and protein are not thoroughly understood, as growth at elevated CO2 can affect multiple processes involved in nitrogen uptake and metabolism (Gifford et al., 2000; BassiriRad et al., 2001; Bloom et al., 2002; McDonald et al., 2002; Lynch & St.Clair, 2004; Reich et al., 2006). One mechanism that has often been suggested to account for decreased N and protein concentrations is dilution by increased concentrations of nonstructural carbohydrates (Gifford et al., 2000). The few studies that have measured both nonstructural carbohydrates and protein (or N) in the edible portions of crop plants suggest that dilution can account for at best a small proportion of the observed decreases in protein concentrations. Increased starch concentrations explained less than one-third of the 15.1% decrease in wheat grain protein concentrations under elevated [CO2] observed by Wu et al. (2004). Donnelly et al. (2001) observed a decrease in the N concentration of potato tubers of 24% under elevated [CO2] unaccompanied by changes in either tuber starch or sugar concentrations.

Fangmeier et al. (1999, 2000) have suggested that decreased protein concentrations in cereal grains under elevated [CO2] might be a consequence of the decreased protein concentrations in photosynthetic tissues commonly seen under elevated CO2. These decreases are largely the result of decreased Rubisco concentrations (Ainsworth & Long, 2005), due most likely to a carbohydrate-dependent decrease in the expression of photosynthetic genes (Moore et al., 1999), although other mechanisms may operate as well (Stitt & Krapp, 1999). These decreases in leaf protein can lead to decreased seed protein concentration as the N supply to seeds during filling is largely from translocation from catabolized proteins in senescing photosynthetic tissues (Fangmeier et al., 1999, 2000; Salon et al., 2001). This same mechanism may also contribute to the decreased protein concentrations in potato (Fangmeier et al., 2002) as approximately 50% of the nitrogen supply to tubers is from translocation from leaves and stems (Kolbe & Stephan-Beckmann, 1997).

This hypothesis is also consistent with the fact that soybean seeds show much smaller changes in protein concentration than grains under elevated [CO2]. Nodulated soybeans show little if any decrease in leaf N concentrations under elevated CO2 (Ainsworth et al., 2002), perhaps because root nodules provide a large sink for photosynthate, preventing the increase in leaf hexose flux responsible for downregulation of Rubisco (Moore et al., 1999). Elevated CO2, therefore, can be expected not to affect the substantial proportion (up to 100%) of soybean seed N obtained through translocation from vegetative tissues (Zeiher et al., 1982). The relatively small magnitude of the effect of elevated [CO2] on soybean protein concentrations seen in this meta-analysis is also consistent with the results of Jablonski et al. (2002) who found no significant effect of elevated CO2 on seed N across a wide range of wild and domestic legumes, compared with a 15% decrease in nonleguminous C3 species.

Differences in N metabolism between legumes and nonlegumes may also explain the very different interactions of elevated [CO2] with elevated ozone seen in potato vs. soybean. The primary effect of ozone exposure is damage to leaf mesophyll, leading to decreased photosynthetic carbon assimilation and growth (Long & Naidu, 2002). Elevated [CO2] has frequently been found to ameliorate the effects of ozone, principally by decreasing the flux of ozone into leaves through decreasing stomatal conductance, although other mechanisms may operate as well (Booker & Fiscus, 2005).

In potatoes, chronic ozone exposures increase tuber N concentrations (Fangmeier et al., 2002; Heagle et al., 2003). The greater effect of elevated [CO2] on tuber protein concentrations at high than low [O3] observed in the current meta-analysis may result because elevated CO2 negates this effect of ozone in addition to exerting its own direct effect on tuber protein concentrations.

The fact that elevated [CO2] increased seed protein in soybeans under high O3 conditions may also be due to CO2 amelioration of the effects of O3. Ozone can decrease nitrogen fixation in legumes, presumably by decreasing the translocation of photosynthate to nodules (Pausch et al., 1996), while elevated CO2, by increasing carbon assimilation, may increase symbiotic N fixation (De Graaff et al., 2006; Rogers et al., 2006). In the experiments included in the current meta-analysis, it may be that the amelioration of the negative effects of ozone on nitrogen fixation had a larger effect on seed protein than the direct effects of elevated CO2, leading to a slight increase rather than a slight decrease in seed protein concentrations with elevated CO2.

Global tropospheric ozone concentrations are steadily rising, even in relatively unpolluted areas (Vingarzan, 2004). If potato and soybean are representative of other crop species, the effects of elevated CO2 on protein concentrations in the 21st century are likely to be larger for nonleguminous crops and smaller (or reversed in sign) for leguminous crops than indicated by Table 1.

In addition, the values in Table 1, while accurately reflecting the available experimental data, may be influenced by several types of experimental artifacts. One possible artifact is suggested by the fact that larger CO2 effects were generally seen with open-top chambers than with other CO2 enrichment technologies. Open-top chambers typically create conditions that are hotter and drier than in adjacent unenclosed fields, and the plants within them are subject to substantial edge effects (Long et al., 2004). It is not clear whether this explains the findings of the current meta-analysis that, for some species, open-top chamber studies showed larger effects of elevated CO2 than studies performed in glasshouses or growth chambers. Plants in both of these environments should be subject to edge effects similar to those in open-top chambers, and there is no obvious reason to expect open-top chambers to have environments that are consistently hotter or drier than those in such facilities.

The technology for manipulating CO2 that most closely mimics realistic future agricultural conditions is free-air carbon dioxide enrichment (FACE), which elevates CO2 under field conditions without using enclosures. Comparisons of crop growth between FACE and chamber experiments have shown a number of important differences, including a typically smaller yield response to CO2 in FACE than in chamber experiments (Ainsworth & Long, 2005; Long et al., 2005, 2006). However, in a recent meta-analysis, the decease in leaf N concentration across FACE experiments was 13.2% (Ainsworth & Long, 2005), similar to values found in previous meta-analyses of data largely obtained from chamber studies (16.4% for tree leaves; Curtis & Wang, 1998; 14% for all plant tissues; Cotrufo et al., 1998). In the current study, the magnitude of the CO2 effect in FACE experiments was similar to the mean for all experimental technologies for all crops except wheat. FACE technology may therefore yield estimates of the effect of CO2 on crop protein concentrations which are similar to other technologies.

While artifacts introduced by the CO2 enrichment technologies may be of some importance, those caused by unrealistic rooting environments are perhaps of greater consequence (Körner, 2006). Ainsworth et al. (2002), e.g., reported that the CO2 effect on soybean yield was three times greater for plants grown in the ground compared with those grown in large pots. In the present study, the larger CO2 effect in OTC studies performed in pots vs. the ground suggests that the magnitude of elevated CO2 effects on crop protein might be less under realistic agricultural conditions than the effects depicted in Fig. 1.

The impact of elevated CO2 on human protein nutrition may also be affected by postharvest processing and food preparation, as well as by the protein composition of edible plant tissues. The greater effect of elevated CO2 on whole wheat grains than on wheat flour suggests that the impact of elevated CO2 on human nutrition may differ between those who consume refined (white) flour and those who consume whole-grain wheat products. This finding also suggests that elevated CO2 may affect the protein concentration in endosperm less than in other portions of the wheat grain. We know of no studies that have directly examined this possibility. Differences in the effects of elevated [CO2] on different portions of the wheat grain might occur due to the very different accumulation patterns of the structural and metabolic proteins of the aleurone and embryo compared with the major storage proteins of the endosperm. In wheat grains structural and metabolic proteins accumulate earlier in developing wheat grains than the major endosperm storage proteins (Triboïet al., 2003). Accumulation of structural and metabolic proteins in wheat grains also appears to be sink-regulated, while endosperm storage protein accumulation is largely constrained by N sources to the developing grain (Martre et al., 2003).

The effect of elevated CO2 on human protein nutrition might be mitigated by agricultural practices that increase the protein content of crops. Idso & Idso (2001) argued that any effects of elevated CO2 on crop protein content could be ameliorated by increased use of N fertilizer, and the current results demonstrate that increasing soil N minimizes the effect of elevated CO2 on crop protein concentration, at least for wheat. However, even under high nitrogen treatments, wheat still experienced a mean reduction in protein concentration of nearly 10%. It is, therefore, possible that nitrogen fertilizer can minimize, but not eliminate, the reduction in protein concentration associated with increased atmospheric CO2. However, if fertilizer can be used to mitigate the effects of elevated CO2 on protein concentration, it cannot be regarded as a panacea as smallholder farmers in many countries have little access to chemical fertilizers (United Nations Millennium Project, 2005) and may be unable to increase fertilizer use in response to increases in atmospheric [CO2]. In addition, runoff from agriculturally applied N fertilizer is associated with potentially undesirable ecological effects across a wide variety of ecosystems (Galloway et al., 2003).

While our analyses have focused on changes in crop protein as atmospheric [CO2] rises over the 21st century, it is also possible that crop protein concentrations have already been affected by recent increases in atmospheric [CO2]. Conroy & Hocking (1993) reported that for wheat grown in New South Wales, protein concentration had declined by ∼16% between 1968 and 1990. Davis et al. (2004) found a median decrease in protein concentration of 6% across 43 garden crops from 1950 to 1999. The authors of both papers suggested that changes in the cultivars used probably explained most of this reduction, but that increasing CO2 may have played an additional role. Experimentally, both Ziska et al. (2004) and Rogers et al. (1998) found that the protein concentration of wheat flour was lower in plants grown at current atmospheric CO2 concentrations than in those grown at concentrations representative of atmospheric concentrations before the industrial revolution. It, therefore, appears likely that rising atmospheric [CO2] has already affected the protein content of some plant foods.

Conclusions

Rising atmospheric [CO2] is likely to reduce the protein concentration for many plant crops. The magnitude of this effect is difficult to estimate, due to the sensitivity of this effect to experimental conditions. Nonetheless, decreases in protein are seen consistently for several species across a wide range of experimental techniques and environmental conditions. This effect may be partially mitigated by increased use of nitrogen fertilizers, but this seems likely to be only a partial solution to the effect of elevated CO2 on the protein concentration of human foods. The effect of atmospheric CO2 on crop protein therefore seems likely to be of genuine importance for human nutrition in and beyond the 21st century.

Acknowledgements

This work was supported by the Fleming Fund for Collaborative Research of Southwestern University. We thank Eli Taub for assistance in typing the manuscript, Lisa Anderson for help in obtaining literature and Xianzhong Wang for valuable comments on the research and manuscript.

Appendix

Appendix A. Publications with data included in the analyses.

Allen LH, Vu JCV, Valle RR, Boote KJ & Jones PH (1988) Nonstructural carbohydrates and nitrogen of soybean grown under carbon dioxide enrichment. Crop Science, 28, 84–94.

Amthor JS, Mitchell RJ, Runion BR, Rogers HH, Prior SA & Wood CW (1994) Energy content, construction cost and phytomass accumulation of Glycine max (L.) Merr. and Sorghum bicolor (L.) Moench grown in elevated CO2 in the field. New Phytologist, 128, 443–150.

Bai Y, Tischler CR, Booth DT & Taylor EM (2003) Variations in germination and grain quality within a rust resistant common wheat germplasm as affected by parental CO2 conditions. Environmental and Experimental Botany, 50, 159–168.

Bencze S, Veisz O & Bedö Z (2004) Effects of high atmospheric CO2 and heat stress on phytomass, yield and grain quality of winter wheat. Cereal Research Communications, 32, 75–82.

Blumenthal C, Rawson HM, McKenzie E, Gras PW, Barlow EWR & Wrigley CW (1996) Changes in wheat grain quality due to doubling the level of atmospheric CO2. Cereal Chemistry, 73, 762–766.

Chadhuri UN, Burnett RB, Kanemaso ET & Kinkham MB (1986) Effect of Elevated Levels of CO2on Winter Wheat under two Moisture Regimes (Response of Vegetation to Carbon Dioxide No. 29). United States Department of Energy, Carbon Dioxide Research Division, Washington DC.

Conroy JP (1992) Influence of elevated atmospheric CO2 concentrations on plant nutrition. Australian Journal of Botany, 40, 445–456.

Conroy JP, Seneweera S, Basra AS, Rogers G & Nissen-Wooler B (1994) Influence of rising atmospheric CO2 concentrations and temperature on growth, yield and grain quality of cereal crops. Australian Journal of Plant Physiology, 21, 741–758.

Cure JD, Israel DW & Rufty TW (1988) Nitrogen stress effects on growth and seed yield of nonnodulated soybean exposed to elevated carbon dioxide. Crop Science, 28, 671–677.

Donnelly A, Lawson T, Craigon J, Black CR, Colls JJ & Landon G (2001) Effects of elevated CO2 and O3 on tuber quality in potato (Solanum tuberosum L.). Agriculture, Ecosystems and Environment, 87, 273–285.

Fangmeier A, Grüters U, Vermehren B & Jäger H-J (1996) Responses of some cereal cultivars to CO2 enrichment and tropospheric ozone at different levels of nitrogen supply. Angewandte Botanik, 70, 12–18.

Fangmeier A, Grüters U, Högy P, Vermehren B & Jäger H-J (1997) Effects of elevated CO2, nitrogen supply and tropospheric ozone on spring wheat- II. Nutrients (N, P, K, S, Ca, Fe, Mg, Zn). Environmental Pollution, 96, 43–59.

Fangmeier A, De Temmerman L, Mortensen L, Kemp K, Burke J, Mitchell R, van Oijen M & Weigel HJ (1999) Effects of nutrients on grain quality in spring wheat crops grown under elevated CO2 concentrations and stress conditions in the European multiple-site experiment ‘ESPACE-wheat’. European Journal of Agronomy, 10, 215–229.

Fangmeier A, Chrost B, Högy P & Krupinska K (2000) CO2 enrichment enhances flag leaf senescence in barley due to greater grain nitrogen sink capacity. Environmental and Experimental Botany, 44, 151–164.

Fangmeier A, De Temmerman L, Black C, Persson K & Vorne V (2002) Effects of elevated CO2 and/or ozone on nutrient concentrations and nutrient uptake of potatoes. European Journal of Agronomy, 17, 353–368.

Hakala K (1998) Growth and yield potential of spring wheat in a simulated changed climate with increased CO2 and higher temperature. European Journal of Agronomy, 9, 41–52.

Heagle AS, Miller JE & Pursley WA (1998) Influence of ozone stress on soybean response to carbon dioxide enrichment: III. Yield and seed quality. Crop Science, 38, 128–134.

Heagle AS, Miller JE & Pursley WA (2003) Atmospheric pollutants and trace gases

Growth and yield responses of potato to mixtures of carbon dioxide and ozone. Journal of Environmental Quality, 32, 1603–1610.

Israel DW & Rogers HH (1982) The effect of N2-fixing ability of Rhizobium strain on response of nodulated soybeans to atmospheric CO2 enrichment. In Field Studies of Plant Responses to Elevated Carbon Dioxide Levels (eds. H.H. Rogers and G.E. Bingham), pp. 122–161. United States Department of Energy, Carbon Dioxide Research Division, Washington, DC.

Kimball BA, Morris CF, Pinter PJ, Wall GW, Hunsaker DJ, Adamsen FJ, LaMorte RL, Leavitt SW, Thompson TL, Matthias AD & Brooks TJ (2001) Elevated CO2, drought and soil nitrogen effects on wheat grain quality. New Phytologist, 150, 295–303.

Kleemola J, Peltonen J & Peltonen-Sainio P (1994) Apical development and growth of barley under different CO2 and nitrogen regimes. Journal of Agronomy and Crop Science, 173, 79–92.

Lieffering M, Kim H-Y, Kobayashi K & Okada M (2004) The impact of elevated CO2 on the elemental concentrations of field-grown rice grains. Field Crops Research, 88, 279–286.

Manderscheid R, Bender J, Jäger H-J & Weigel HJ (1995) Effects of season long CO2 enrichment on cereals. II. Nutrient concentrations and grain quality. Agriculture, Ecosystems and Environment, 54, 175–185.

Pleijel H, Gelang J, Sild E, Danielsson H, Younis S, Karlsson P-E, Wallin G, Skärby L & Selldén G (2000) Effects of elevated carbon dioxide, ozone and water availability on spring wheat growth and yield. Physiologia Plantarum, 108, 67–70.

Rogers GS, Gras PW, Batey IL, Milham PJ, Payne L & Conroy JP (1998) The influence of atmospheric CO2 concentration on the protein, starch and mixing properties of wheat flour. Australian Journal of Plant Physiology, 25, 387–393.

Rogers GS, Milham PJ, Gillings M & Conroy JP (1996) Sink strength may be the key to growth and nitrogen responses in N-deficient wheat at elevated CO2. Australian Journal of Plant Physiology, 23, 253–264.

Rogers HH, Bingham GE, Cure JD, Smith JM & Surano KA (1983) Responses of selected plant species to elevated carbon dioxide in the field. Journal of Environmental Quality, 12, 569–574.

Rogers HH, Cure JD & Smith JM (1986) Soybean growth and yield response to elevated carbon dioxide. Agriculture, Ecosystems and Environment, 16, 113–128.

Rogers HH, Cure JD, Thomas JF & Smith JM (1984) Influence of elevated CO2 on growth of soybean plants. Crop Science, 24, 361–366.

Rudorff BFT, Mulchi CL, Fenny P, Lee EH & Rowland R (1996) Wheat grain quality under enhanced tropospheric CO2 and O3 concentrations. Journal of Environmental Quality, 25, 1384–1388.

Sæbø A & Mortensen LM (1996) Growth, morphology and yield of wheat, barley, and oats grown at elevated atmospheric CO2 concentration in a cool, maritime climate. Agriculture, Ecosystems and Environment, 57, 9–15.

Sánchez de la Puente L, Pérez Pérez P, Martínez-Carrasco R, Morcuende Morcuende R & Martín del Molino IM (2000) Action of elevated CO2 and high temperatures on the mineral composition of two varieties of wheat. Agrochimica, XLIV, 5–6.

Seneweera SP & Conroy JP (1997) Growth, grain yield and quality of rice (Oryza sativa L.) in response to elevated CO2 and phosphorus nutrition. Soil Science and Plant Nutrition, 43, 1131–1136.

Terao T, Miura S, Yanagihara T, Hirose T, Nagat K, Tabuchi H, Kim H, Lieffering M, Okada M & Kobayashi K (2005) Influence of free-air CO2 enrichment (FACE) on the eating quality of rice. Journal of the Science of Food and Agriculture, 85, 1861–1868.

Thomas JMG, Boote KJ, Allen LH, Gallo-Meagher M & Davis JM (2003) Elevated temperature and carbon dioxide effects on soybean seed composition and transcript abundance. Crop Science, 43, 1548–1557.

Thompson GB & Woodward FI (1994) Some influences of CO2 enrichment, nitrogen nutrition and competition on grain yield and quality in spring wheat and barley. Journal of Experimental Botany, 45, 937–942.

Veisz O, Bencze S & Bedö Z (2005) Effect of elevated CO2 on wheat at various nutrient supply levels. Cereal Research Communications, 33, 333–336.

Weigel HJ & Manderscheid R (2005) CO2 enrichment effects on forage and grain nitrogen content of pasture and cereal plants. Journal of Crop Improvement, 13, 73–89.

Wolf J (1996) Effects of nutrient supply (NPK) on spring wheat response to elevated atmosperic CO2. Plant and Soil, 185, 113–123.

Wu D-X, Wang G-X, Bai Y-F & Liao J-X (2004) Effects of elevated CO2 concentration on growth, water use, yield and grain quality of wheat under two water soil levels. Agriculture, Ecosystems and Environment, 104, 493–507.

Ziska LH, Bunce JA & Caulfield F (1998) Intraspecific variation in seed yield of soybean (Glycine max) in response to increased atmospheric carbon dioxide. Australian Journal of Plant Physiology, 25, 801–807.

Ziska LH, Namuco O, Moya T & Quilang J (1997) Growth and yield response of field-grown tropical rice to increasing carbon dioxide and air temperature. Agronomy Journal, 89, 45–53.

Ziska LH, Morris CF & Goins EW (2004) Quantitative and qualitative evaluation of selected wheat varieties released since 1903 to increasing atmospheric carbon dioxide: can yield sensitivity to carbon dioxide be a factor in wheat performance? Global Change Biology, 10, 1810–1819.

Ancillary