Interactive effects of low atmospheric CO2 and elevated temperature on growth, photosynthesis and respiration in Phaseolus vulgaris


  • S. A. Cowling,

    1. Department of Botany, University of Toronto, 25 Willcocks Street, Toronto, Ontario, Canada, M5S 3B2
    Search for more papers by this author
    • Present address: Climate Impacts Group, Department of Plant Ecology, Lund University, S-22362 Lund, Sweden

  • R. F. Sage

    1. Department of Botany, University of Toronto, 25 Willcocks Street, Toronto, Ontario, Canada, M5S 3B2
    Search for more papers by this author

R. F. Sage Fax: 416 978 5878; e-mail:


For most of the past 250 000 years, atmospheric CO2 has been 30–50% lower than the current level of 360 μmol CO2 mol–1 air. Although the effects of CO2 on plant performance are well recognized, the effects of low CO2 in combination with abiotic stress remain poorly understood. In this study, a growth chamber experiment using a two-by-two factorial design of CO2 (380 μmol mol–1, 200 μmol mol–1) and temperature (25/20 °C day/night, 36/29 °C) was conducted to evaluate the interactive effects of CO2 and temperature variation on growth, tissue chemistry and leaf gas exchange of Phaseolus vulgaris. Relative to plants grown at 380 μmol mol–1 and 25/20 °C, whole plant biomass was 36% less at 380 μmol mol–1× 36/29 °C, and 37% less at 200 μmol mol–1× 25/20 °C. Most significantly, growth at 200 μmol mol–1× 36/29 °C resulted in 77% less biomass relative to plants grown at 380 μmol mol–1× 25/20 °C. The net CO2 assimilation rate of leaves grown in 200 μmol mol–1× 25/20 °C was 40% lower than in leaves from 380 μmol mol–1× 25/20 °C, but similar to leaves in 200 μmol mol–1× 36/29 °C. The leaves produced in low CO2 and high temperature respired at a rate that was double that of leaves from the 380μmol mol–1× 25/20 °C treatment. Despite this, there was little evidence that leaves at low CO2 and high temperature were carbohydrate deficient, because soluble sugars, starch and total non-structural carbohydrates of leaves from the 200μmol mol–1× 36/29 °C treatment were not significantly different in leaves from the 380μmol mol–1× 25/20 °C treatment. Similarly, there was no significant difference in percentage root carbon, leaf chlorophyll and leaf/root nitrogen between the low CO2× high temperature treatment and ambient CO2 controls. Decreased plant growth was correlated with neither leaf gas exchange nor tissue chemistry. Rather, leaf and root growth were the most affected responses, declining in equivalent proportions as total biomass production. Because of this close association, the mechanisms controlling leaf and root growth appear to have the greatest control over the response to heat stress and CO2 reduction in P. vulgaris.


Much research has addressed plant responses to projected increases in atmospheric CO2. However, plants have experienced much lower CO2 levels than today for most of their recent evolutionary history. Atmospheric CO2 concentrations were ≈270μmol mol–1 between 100 and 12 000 years ago, and averaged 215μmol mol–1 during the late Pleistocene epoch between 12 000 and 110 000 years ago (Barnola et al. 1987; Jouzel et al. 1993). During the peak of the last two glaciations at 21 000 and 150 000 years ago, atmospheric CO2 concentration was as low as 180 μmol mol–1 (Jouzel et al. 1993). Despite the recent prevalence of low atmospheric CO2, relatively few studies have addressed its physiological, ecological, and morphological consequences (Sage 1995). Theoretical models and physiological experiments consistently show that plant processes, particularly photosynthesis and transpiration, are highly sensitive to variation in CO2 below 360μmol mol–1. Photosynthesis and productivity of C3 plants at the last glacial maximum (21 000 years ago) are predicted to have been 25–50% lower than at present, assuming moderate temperatures near 25 °C (Sage 1995). Studies of plant responses to a range of subambient CO2 levels support the theoretical prediction (Allen et al. 1991; Sage & Reid 1992; Polley et al. 1993a, b; Dippery et al. 1995; Tissue et al. 1995).

Research on the effects of low atmospheric CO2 on plants has not focused on interactive effects of low CO2 and environmental stress. Whereas an increase in atmospheric CO2 can ameliorate effects of stress (Kriedemann, Sward & Downton 1976; Rogers et al. 1984; Hsaio 1993; Baker et al. 1997), low atmospheric CO2 may aggravate stress frequency because of combined negative effects on plant carbon balance. Atmospheric CO2 depletion alone promotes carbohydrate deficiency through reduced photosynthesis. CO2 depletion limits ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) activity, while simultaneously increasing the rate of photorespiration (Farquhar & von Caemmerer 1982; Sage & Reid 1994). Abiotic stress can also depress carbon reserves by inhibiting stomatal conductance and the biochemical reactions of photosynthesis, and by stimulating respiration (Osmond et al. 1987). In the case of high temperature, for example, both photorespiration and mitochondrial respiration are stimulated, while the electron transport capacity of photosynthesis is often impaired (Brooks & Farquhar 1985; Sage, Santrucek & Grise 1995).

How CO2 depletion interacts with mild stress is uncertain, but has great significance, as most of the modern flora persisted during recent episodes of low atmospheric CO2. Here, we combined analyses of growth, leaf gas exchange, and the carbon and nitrogen content of leaves and roots to investigate interactions between CO2 depletion and an elevated temperature that is moderately inhibitory at current CO2 levels. The C3 crop Phaseolus vulgaris was selected because previous work has shown that 32–36 °C produces modest symptoms of heat stress in this species, including reduced photosynthetic rate, growth and duration of leaf expansion (Lin & Markhart 1996). During the Pleistocene, exposure to moderate heat was probably common in tropical and subtropical latitudes (where P. vulgaris originated), as these regions were not much colder than today (Kutzbach et al. 1997).


Plant material and treatments

Phaseolus vulgaris (cv. Black Turtle) plants were grown in either 3 dm3 pots containing sand for growth assessments, or 8 dm3 pots containing 50% loam, 17% sand, and 33% Pro-Mix (Premier Brands Inc., Redhill, PA, USA) for all other assessments. Sand was used for the growth analysis because it is easily washed from roots, and the smaller pots allowed an adequate sample size to be placed in the growth chamber. Soil was used for all other experiments because its higher capacitance for nutrients and water better supported the growth of the plants beyond 30 d after planting (DAP). No differences in plant characteristics were observed between sand-grown plants and their soil-grown counterparts prior to 30 DAP. Seeds were germinated in greenhouses, and the seedlings were transferred to treatments 7 DAP. The plants were watered daily with a 0·5 strength Hoagland's solution (Hoagland & Arnon 1938). The plants were grown in Conviron E-15 growth chambers (Winnipeg, Manitoba, Canada) under fluorescent and incandescent lamps. The photosynthetic photon flux density (PPFD) averaged 520 ± 20 mol m–2 s–1 for 12 h photoperiods.

A two-by-two factorial design involving atmospheric CO2 concentration and chamber air temperature was employed to test independent and interactive effects of CO2 and temperature on growth and leaf gas exchange. The design resulted in the following four treatments: 380 μmol CO2 mol–1 air × 25/20 °C day/night temperature, 200 μmol mol–1× 25/20 °C, 380 μmol mol–1× 36/29 °C, and 200 μmol mol–1× 36/29 °C. Low CO2 was maintained 24 h a day by a CO2-controlling infrared gas analyser (PP Systems, Haverhill, MA, USA) which regulated a solenoid valve on a CO2-free air source. CO2 was removed from the air source by passing compressed air through soda-lime (WR Grace & Co., Atlanta, GA, USA) filters. The CO2 concentration in the ambient and subambient treatments was 383 ± 8 and 196 ± 2 μmol mol–1, respectively. The actual day/night air temperature in moderate and high temperature treatments was 24·9 ± 0·3/19·9 ± 0·2 °C and 35·7 ± 0·4/29·1 ± 0·1 °C, respectively. The humidity in the chambers was 50 ± 7% (mean ± range) for the chambers with 380 μmol CO2 mol–1 air, and 78 ± 8% for the low CO2 chambers.

Plant growth, leaf area and tissue chemistry

Growth and biomass distribution were determined on equal-aged cohorts grown in sand and harvested at 19, 21 and 23 DAP. The harvest intervals were chosen to avoid harvest periods which resulted in more than a doubling of leaf area (Beadle 1993). The leaf area was determined with a Li-Cor area meter (Licor 3000, Lincoln, NE, USA). To assess leaf morphology and leaf and root chemistry, plants grown on soil in a second trial were harvested 22 ± 2 DAP. For carbon, carbohydrate and nitrogen determinations, leaves and roots were sampled during a 2 h window beginning 5 h into the photoperiod. The leaves and roots were oven dried at 65 °C and analysed for carbon and nitrogen using a LECO gas analyser (LECO Corp., St. Joseph, MI, USA). Carbohydrates and leaf chlorophyll content were analysed on 4–6 cm2 leaf discs which were frozen in liquid nitrogen immediately upon harvest and stored at – 80 °C until assay.

The soluble sugar (sucrose, glucose, fructose) content of leaves was measured using the phenol–sulphuric acid method following extraction in 95% ethanol (Sturgeon 1990; Farrar 1993). All samples were run in duplicate and the measurements were compared against a glucose standard. Leaf starch was measured as glucose equivalents. Starch was extracted with a methanol:chloroform:water solvent [12:5:3; Dickson (1979)], digested with 35% perchloric acid, and assayed using the phenol–sulphuric acid technique (Tissue & Wright 1995). The chlorophyll content was determined by extracting leaves in 80% dimethylformamide and measuring absorbance at 647 and 664 nm (Porra, Thompson & Kriedemann 1989).

Gas exchange measurements

The temperature response of photosynthesis (A), stomatal conductance (gs), and the ratio of intercellular to ambient CO2 partial pressure (Ci/Ca) were determined at 200 and 380 μmol mol–1 for single, attached leaves using a null balance gas exchange system, similar to that described by Sharkey (1985). At 27–40 DAP, recent fully expanded leaves were placed in an environmentally controlled cuvette and equilibrated to the plant's growth temperature and CO2 concentration. The leaf temperature was increased to 40 °C, and after equilibration with the environment in the leaf cuvette, CO2 and water vapour exchange were measured. The leaf temperature was decreased in steps of ≈ 5 °C from 40 to 10 °C, with gas exchange measurements taken at each step. Using this procedure, A at the growth temperature was not depressed by more than 10% following exposure to the 40 °C measurement condition. Measurement PPFD was equivalent to the growth PPFD of 520 μmol m–2 s–1. The vapour pressure deficit between leaf and air was maintained between 0·8 kPa (at cooler temperature) and 1·7 kPa (at the warmer temperatures). Gas exchange parameters were calculated according to von Caemmerer & Farquhar (1981).

Night respiration of attached leaves was determined at 39–59 DAP using the gas exchange system described above. Near the end of a 12 h light period, attached leaves were equilibrated in the gas exchange chamber at their daytime growth conditions. The leaves were then exposed to their night growth conditions and the respiration rates were monitored for the next 4 h, a period which was found to allow the dark respiration rate to decay to stable levels in all treatments.

Statistical analysis

Experiments were repeated three times, occurring sequentially through time. Because of the fixed nature of the technology for controlling low CO2, treatments could not be rotated between chambers. However, close monitoring ensured conditions were maintained within treatments between replication. Treatment differences were tested with one-way ANOVA for parametric data, and with ANOVA on ranks for non-parametric data. Multiple pairwise comparisons, Dunn's test and the Bonferroni test, were used for parametric and non-parametric data, respectively.


Plant dry weight and total leaf area

At the last harvest (23 DAP), biomass at 380μmol mol–1 was 35% lower in plants grown at 36/29 °C compared with 25/20 °C (Fig. 1a). At 25/20 °C, the plant mass was 37% less in plants grown at 200 μmol mol–1 relative to those grown at 380μmol mol–1. Biomass at low CO2 and elevated temperature was 77% less relative to that at 380μmol mol–1× 25/20 °C at 23 DAP. An increase in temperature at 380μmol mol–1 reduced leaf growth such that plants had 42% less area at 36/29 °C than at 25/20 °C (Fig. 1b). At 25/20 °C, plants grown at low CO2 produced 28% less area than their counterparts at 380μmol mol–1. Growth at high temperature and low CO2 produced additive effects, such that leaf area at 23 DAP was 77% less in the combined treatment relative to plants grown at 380μmol mol–1× 25/20 °C.

Figure 1.

. The effects of low CO2 and elevated temperature on (a) plant dry weight as a function of time, and (b) leaf area as a function of time. Growth conditions: 380 μmol mol–1× 25/20 °C (○), 200 × 25/20 °C (○), 380 × 36/29 °C (□), 200 × 36/29 °C (□). Mean ± SE, n = 15.

Treatments produced inconsistent, and when significant, minor shifts in patterns of biomass distribution. At 36/29°C, root:shoot was 15% lower in low CO2 than in ambient CO2 (0·39 ± 0·20 vs 0·46 ± 0·01). At both 25/20°C and 36/29°C, the growth in low CO2 reduced root mass by 10% relative to 380 μmol mol–1, from ≈ 31 to 28% of total biomass. This reduction in fractional root mass was not associated with increased biomass distribution to leaves, because all treatments had 56 ± 1% of biomass in leaves. Instead, the proportion of stem biomass was 10–15% higher in the low CO2 treatments (not shown).

Leaf characteristics

Neither temperature nor CO2 concentration independently had a significant effect on the number of leaves per plant, but a significant interaction occurred at low CO2 and elevated temperature (Table 1). Leaf number was reduced by more than 30% in 200 μmol mol–1× 36/29 °C relative to other treatments (Table 1). The area of the middle leaflet of the first trifoliate was 77% less in 200μmol mol–1× 36/29 °C than in 380μmol mol–1× 25/20°C. Associated with this loss of leaflet area was a large reduction of leaf length in the low CO2× high temperature treatment. At 200μmol mol–1, leaflets produced at a high temperature were narrower than those produced at 25/20°C, as shown by 20% greater length to width ratios in the high temperature treatments. Narrowing of the leaflets indicates proportionally greater treatment effects on lateral leaf expansion. Reduction of atmospheric CO2 at 25/20°C reduced leaf weight per area by 25%, but had no significant effect at 36/29°C (Table 1). Consequently, at 200 μmol mol–1, the weight per area of leaves produced at 36/29 °C was 35% greater than for leaves formed at 25/20 °C.

Table 1.  . Treatment effects on leaf characteristics including the number of leaves per plant (n = 15), the area of the middle leaflet of the first trifoliate (n = 8), the maximum leaf length (n = 20), the ratio of leaflet length to width (L/W) (n = 20), the leaf dry weight per area (LWA) (n = 20) and leaf chlorophyll content (n = 6). Mean ± SE. Dissimilar letters denote significant differences (ANOVA, P < 0·05) Thumbnail image of

Leaf and root chemistry

Throughout the growth and gas exchange experiments, leaves appeared healthy, despite differences in size, exhibiting no obvious signs of senescence, such as chlorosis. In support of our visual observations, the leaf chlorophyll content at 22 DAP was unaffected by CO2 treatment, while increased growth temperature increased leaf chlorophyll at both CO2 levels by over 10% (Table 1).

At 25/20 °C, the plants grown at low CO2 exhibited lower values in all categories of carbon content, with soluble sugars and starch in particular being 40–60% less (Table 2). Similarly, at 25/20 °C, the leaves produced in low CO2 contained 16% less nitrogen. Increasing temperature at 380 μmol mol–1 reduced total carbon levels ≈10%, but had no effect on starch, sugars, total non-structural carbohydrates or leaf N content. At 200 μmol mol–1, all leaf carbohydrate parameters (except starch) and leaf nitrogen content were significantly higher at 36/29°C than at 25/20°C.

Table 2.  . Treatment effects on leaf carbon and nitrogen at 22 d after planting. Mean ± SE, n = 20. Dissimilar superscripts denote significant differences (ANOVA, P < 0·05) Thumbnail image of

Root carbon was minimally affected by atmospheric CO2, decreasing from 36 to 35% of dry weight at 200μmol mol–1× 36/29°C relative to both 380μmol mol–1 treatments (data not shown). The root nitrogen concentration was not affected by atmospheric CO2 or temperature (mean = 4% of dry weight). Similarly, C:N was ≈ 8 in all treatments.

CO2 exchange

When measured at the growth temperature and CO2 concentration, the net CO2 assimilation rate per unit area (A) was 37–42% less at 200μmol mol–1 than at 380μmol mol–1 (Table 3). There was no significant difference in A between low CO2 treatments at 25 °C and at 36 °C. The thermal optimum of A was 4–6 °C greater in plants measured at 380 μmol mol–1 than at 200μmol mol–1 (Table 3 & Fig. 2). No difference in the photosynthetic thermal optimum was observed between the four treatments when measured at equivalent CO2 levels (Fig. 2).

Table 3.  . The effects of low atmospheric CO2 and elevated temperature on net photosynthesis (An) and dark respiration (Rd) at the growth conditions, photosynthesis at the thermal optimum (Aopt) for growth and reciprocal CO2 concentrations, and the thermal optima of photosynthesis (Topt). Thermal optima were determined for growth conditions. Rates of photosynthesis and respiration were determined on plants 27–40 and 39–59 d after planting, respectively. Photosynthetic photon flux density = 537 ± 10 μmol m–2 s–1. Mean ± SE, n = 3. Dissimilar superscripts denote significant differences (ANOVA, P < 0·05) Thumbnail image of
Figure 2.

. Treatment effects on net CO2 assimilation rate as a function of leaf temperature. Measurement CO2 levels: 380 μmol mol–1 (○), 200 μmol mol–1(○). Mean ± SE, n = 3. Gas exchange was measured at both growth and reciprocal CO2 concentrations. Photosynthetic photon flux density = 537 ± 10 μmol photons m–2 s–1. Numbers represent intercellular CO2 concentration (μmol mol–1).

There were two important effects of growth conditions on the thermal response of photosynthesis. First, A at the thermal optimum (Aopt) was enhanced by higher growth temperature. Plants grown at 36/29 °C had identical Aopt when measured at the same CO2 level (Table 3). Similarly, plants grown at 25/20 °C exhibited similar Aopt when measured at the same CO2. However, this rate was ≈ 20% less than in plants grown at 36/29 °C and measured at equivalent CO2 levels. The second growth effect was the greater degree of CO2 sensitivity of photosynthesis at measurement temperatures below 16 °C in plants grown at 380 versus 200μmol mol–1. Below 16 °C measurement temperature, A in plants grown at 380 μmol mol–1 was stimulated by a rise in ambient CO2 from 200 to 380μmol mol–1 (Fig. 2a & b), while plants grown at 200μmol mol–1 showed little stimulation of A by CO2 increase at these cooler temperatures (Fig. 2c & d).

At 380 μmol mol–1, the growth temperature did not affect the rate of dark respiration. However, growth CO2 did (Fig. 3). In plants grown at 25/20 °C, dark respiration was 33% lower at 200μmol mol–1 than at 380 μmol mol–1 when measured at the night temperature. In contrast, at 29 °C, respiration was double at 200μmol mol–1 relative to 380μmol mol–1 (Table 3 & Fig. 3). This rise in respiration at low CO2 and high temperature contributed to a large reduction in the ratio of A to dark respiration. Only the combined treatment of low CO2 and elevated temperature exhibited a pronounced change in respiration after the imposition of darkness, declining from 2·6μmol m–2 s–1 immediately after darkness to 1·9μmol m–2 s–1 4 h later (Fig. 3).

Figure 3.

. Leaf respiration as a function of time after light period for plants grown at (a) 25/20 °C and (b) 36/29 °C. Growth CO2 level: 380 μmol mol–1(○), 200 μmol mol–1(○). Mean ± SE, n = 3. Respiration was measured at growth night temperatures. The numbers in brackets represent respiration in mmol m–2 integrated over the 4 h measurement period.

H2O exchange

At similar measurement leaf-to-air vapour pressure difference (VPD), stomatal conductance at the growth temperature and CO2 was greater as a result of exposure to high temperature alone, and at 25 °C, of reduction in CO2 alone (Table 4). The combination of low CO2 and high temperature produced conductance values three times those of plants grown at 380μmol mol–1× 25/20 °C. Associated with this increase in conductance was an increase in the ratio of intercellular to ambient CO2 partial pressure (Ci/Ca; Table 4). Growth CO2 also modified the thermal response of stomatal conductance relative to photosynthesis, as indicated by Ci/Ca (Fig. 4). Plants grown at 380 μmol mol–1 showed a rise of Ci/Ca with increasing temperature, while Ci/Ca was little affected by temperature in plants grown at 200μmol mol–1. Water use efficiency under gas exchange conditions was reduced by ≈ 60% by either increasing temperature at 380 μmol mol–1, or by reducing CO2 at 25 °C (Table 4). The combined treatment exhibited a 76% reduction in water use efficiency relative to 380μmol mol–1× 25°/20 °C. Differences in conductance did not affect leaf water status, as the relative water content was 92–93% in all treatments.

Table 4.  . The effects of low atmospheric CO2 and elevated temperature on stomatal conductance to water vapour (gs, mol m–2 s–1); Ci/Ca, the ratio of intercellular to ambient CO2 (μmol mol–1); and water use efficiency (WUE, mmol CO2 per mol H2O) at a leaf-to-air vapour pressure difference of 1·0–1·2 kPa. Photosynthetic photon flux density = 537 ± 10 μmol photons m–2 s–1. Mean ± SE, n = 3. Dissimilar superscripts denote significant difference (ANOVA, P < 0·05) Thumbnail image of
Figure 4.

. Treatment effects on the ratio of intercellular to ambient CO2 concentration as a function of leaf temperature. Measurement CO2 level: 380 μmol mol–1(○), 200 μmol mol–1(○). Photosynthetic photon flux density = 537 ± 10 μmol photons m–2 s–1. Mean ± SE, n = 3. Regression equations for significant trends (P < 0·05; TL = leaf temperature in °C): (a) Ci/Ca = 0·60 + 0·003TL at 380 μmol mol–1; Ci/Ca = 0·58 + 0·005TL at 200 μmol mol–1; (b) Ci/Ca = 0·67 + 0·005TL at 380 μmol mol–1; Ci/Ca = 0·70 + 0·003TL at 200 μmol mol–1; (c) Ci/Ca = 0·73 + 0·003TL at 200 μmol mol–1.


Increasing growth temperature to 36/29 °C at ambient CO2 concentration produced mild symptoms of stress [sensuOsmond et al. (1987)], in P. vulgaris including decreased leaf expansion, reduced biomass and a narrowing of the leaves. Such responses are common in plants exposed to elevated temperature (Ong & Baker 1985; Lin & Markhart 1996). Lowering CO2 to 200μmol mol–1 at 25/20°C reduced leaf expansion and plant growth by approximately one third, which is consistent with CO2 depletion effects on C3 plants under non-stressful conditions [reviewed in Sage (1995)]. Significantly, moderate heat inhibition of growth in P. vulgaris at 380μmol mol–1 became severe at 200μmol mol–1, demonstrating strong interactions between subambient CO2 and carbon-depleting stresses, such as elevated temperature. Net CO2 assimilation per unit leaf area was also influenced by CO2 reduction, but this was not highly correlated with growth responses. For example, plants at 200μmol mol–1 had similar A at different growth temperatures, yet markedly different growth rates. Consequently, the mechanisms controlling leaf expansion appear more important than A in determining the growth response to heat and CO2 in P. vulgaris. The reduction in leaf production at low CO2 and 36/29 °C occurred because of a reduction in the initiation of new leaves, reduced expansion of existing leaves and a narrowing of the leaves. This decline in leaf production reduced the potential for light interception, which is a major control over plant productivity (Potter & Jones 1977; Ong & Monteith 1992). In P. vulgaris, biomass production is closely correlated with cumulative light interception (Hsiao 1993). In addition, leaves from low CO2 and high temperature were shorter, thus producing narrower canopies with exponentially reduced potential for light interception. Assuming a circular canopy with radius equal to leaf length, the canopy of the 200μmol mol–1× 36/29°C treatment at 23 DAP would cover a ground surface area of 150 cm2 while the 380μmol mol–1×25/20°C treatment would cover 660 cm2.

Severe carbon depletion promotes leaf senescence (Pell & Dann 1991; Irigoyen, Emerich & Sanchez-Diaz 1992; Berchtold, Besson & Feller 1993), which can aggravate stress intensity through premature loss of photosynthetic surface area. Because the combination of high temperature and low CO2 potentially cause severe carbon deficiency, we hypothesized that carbohydrate starvation was a primary mode of stress intensification in the combined treatment. This hypothesis was not supported, because no symptoms of premature senescence or carbohydrate depletion were observed in the 200μmol mol–1× 36/29°C treatment. Compared with plants from 380μmol mol–1× 25/20°C, the 200μmol mol–1× 36/29 °C plants had more chlorophyll, similar levels of leaf and root carbohydrate and nitrogen, and no signs of accelerated chlorosis. If carbohydrate starvation did occur, it may have been localized in meristematic regions, and either directly through reduced substrate for growth, or indirectly through interaction with growth regulators, leaf growth was impaired. Numerous groups have recently discussed possible interactions between temperature, sucrose supply and phytohormone signalling on root and shoot growth rate (Beck 1994; Dodd & Davies 1994; Jacqmard, Houssa & Bernier 1994).

Photosynthesis and respiration

The CO2 assimilation rate was decreased in low CO2 treatments, consistent with theoretical predictions and previous measurements (Johnson, Polley & Mayeux 1993; Sage 1995; Tissue et al. 1995). In contrast to theory, the combination of low CO2 and elevated temperature did not result in a further reduction in photosynthesis relative to plants in low CO2 alone. At 200 μmol mol–1, increasing leaf temperature from 25 to 36 °C should have theoretically resulted in ≈ 40% reduction in A[modelled according to Farquhar & von Caemmerer (1982), assuming an intercellular CO2 of 160 μmol m–2 s–1 and a RuBP regeneration limitation]. Differences between theoretical expectations and observed values of A appear to occur because high temperature enhanced stomatal conductance and leaf nitrogen content. The reduction in leaf growth apparently reduced competition for nitrogen within plants so that more nitrogen was invested in A per unit area. While the warmer treatments had up to 20% higher A per unit area at a given CO2 level, it was small compensation for much larger effects on leaf area production.

There has been extensive work on acclimation of photosynthesis to elevated CO2 levels. However, few low CO2 studies have examined acclimation responses of photosynthesis. Those that have report little acclimation to subambient CO2 levels (Thomas & Strain 1991; Tissue et al. 1995). Sage & Reid (1992) reported identical Rubisco, chlorophyll and A/CO2 responses in P. vulgaris plants grown at 200 and 350 μmol mol–1 when measured at CO2 concentrations below 350 μmol mol–1. At higher CO2 levels, Sage & Reid (1992) observed reduced assimilation capacity in low CO2-grown plants that they interpreted as evidence for reduced capacity of starch and sucrose to consume triose phosphates and regenerate Pi for photophosphorylation. Here, we also saw indications that bean plants grown at low CO2 had lower capacity to utilize triose phosphates in that they showed little short-term CO2 enhancement of A below 16 °C. Limitations in the capacity for triose phosphate use cause a loss of CO2 responsiveness of photosynthesis, typically at cooler temperatures (Sage & Sharkey 1987). In contrast, at a measurement temperature below 16 °C, plants grown at 380 μmol m–2 s–1 exhibited an increase in A with increasing measurement CO2 concentration. These results indicate that the enzymatic processes contributing to triose phosphate use are major controllers of acclimation of photosynthesis to CO2-depleted atmospheres.

Interactive effects of temperature and CO2 on respiration have received little attention. In one of the few interactive studies conducted, Ziska & Bunce (1994) observed that increases in growth temperature eliminated inhibitory effects of high CO2 on respiration in Dactylus glomerata. Our results also showed an interaction between CO2 and temperature on respiration, but it was variable in direction and magnitude. At 25/20 °C, a reduction in growth CO2 reduced respiration by 30%, while at 36/29 °C, CO2 reduction tripled the respiration rate. The mechanisms for these responses are unknown, but may be related to differences in leaf nitrogen and carbohydrate status. High leaf nitrogen and carbohydrate levels can stimulate respiration (Azcon-Bieto, Lambers & Day 1983; Ryan 1991), and we observed the greatest respiration in the low CO2× high temperature treatment which had relatively high nitrogen and carbohydrate content. The significance of respiratory carbon loss to the plant can be large, because a large fraction of photosynthate (as much as 50%) can be lost over a 24 h period, and respiratory flux is often inversely correlated to yield (Lambers 1985). Here, we estimated that leaf respiratory flux was relatively modest, being ≈ 20% of daily carbon uptake in 200 μmol mol–1× 36/29 °C, and less than 10% in the other treatments (estimates were based on integrated flux values over 12 h, using data in Table 3 and Fig. 3, and assuming respiration and A did not change between the separate measurement times). However, if respiration of whole plants responded in a similar manner to respiration in leaves [as can occur in other species; Ziska & Bunce (1994)], then respiratory drain in beans at low CO2 and high temperature may have been substantial, and could have led to carbon deficiency in developing tissues.

Stomatal acclimation

Acclimation of stomata to low CO2 may involve partial closure to control water loss or further opening to enhance carbon gain. Alternatively, there may be no acclimation, with the short-term response predominating through time. Using Ci/Ca as an index of stomatal acclimation (Ball & Berry 1982; Sage 1994), we observed that the stomata of plants grown at low CO2 lost the ability to respond to short-term increases in temperature, because they exhibited high Ci/Ca (above 0·75) at lower temperature. In contrast, the stomata of plants grown at 380 μmol mol–1 maintained Ci/Ca below 0·75 at lower temperatures. We interpret these results to indicate that the plants grown at low CO2 acclimated to CO2 reduction by opening stomata relative to photosynthesis at cooler measurement temperatures. CO2, rather than temperature, appeared to drive stomatal acclimation, primarily by uncoupling stomatal responses to temperature.

Significance for Pleistocene environments

If responses of modern genotypes to low CO2 reflect those of genotypes from the Pleistocene, then two implications are evident from our data. First, in low CO2 atmospheres of the Pleistocene, the set of thermal conditions considered stressful would have been greater in frequency and intensity, with the degree of inhibition being compounded by reduced canopy size and light interception. As a result, primary production may have been more haphazard, with major consequences for faunal dynamics (including human foraging patterns) and vegetation–climate interactions. Second, because a wider range of conditions would cause stress in low CO2 atmospheres, the selection value of traits conferring stress tolerance might have been greater in Pleistocene environments. Such traits may have included high allocation to storage, and a stronger reaction to a given level of stress than might be needed in current atmospheric conditions. If modern plants still express Pleistocene adaptations, the possibility exists that adaptations conferring fitness in low CO2 environments may be important constraints affecting plant responses to future, high CO2 conditions.


We thank Debbie Tam for technical assistance, and Bob Jefferies, John Coleman and Jarmila Pittermann for comments and advise. This research was supported by Natural Science and Engineering Research Council of Canada Grant #OGP0154273 to R. F. Sage, and an Ontario Graduate scholarship to S. A. Cowling.


  1. Present address: Climate Impacts Group, Department of Plant Ecology, Lund University, S-22362 Lund, Sweden