Light-acquisition and use of individuals as influenced by elevated CO2 in even-aged monospecific stands of Chenopodium album

Authors


*Author to whom correspondence should be addressed. E-mail: hikosaka@mail.tains.tohoku.ac.jp

Summary

  • 1We studied the effect of CO2 elevation on photosynthesis and growth of individuals in even-aged monospecific stands of an annual, Chenopodium album, established at ambient and doubled CO2 concentrations in open-top chambers.
  • 2The whole-plant photosynthesis of every individual in the stand was calculated from (1) the distribution of light and leaf nitrogen and (2) the relationships between photosynthetic parameters and leaf nitrogen content per area. Elevated CO2 increased light-saturated rates of photosynthesis by 10–15% and the initial slope of the light-response curve by 11%, but had no effect on the convexity of the light-response curve and dark respiration.
  • 3The relative rate of photosynthesis (RPR, the rate of photosynthesis per unit above-ground mass) was analysed as the product of light capture (Φmass, the photon flux captured per unit above-ground mass) and light-use efficiency (LUE, plant photosynthesis per unit photon capture).
  • 4At an early stage of stand development (33 days after germination), RPR was nearly constant and no difference was found between ambient and elevated CO2. However, CO2 elevation influenced the components of RPR such that the reduction in Φmass at elevated CO2 offsets the effect of the higher LUE. Later (47 days), RPR was positively correlated with plant mass at both CO2 concentrations. When compared at an equal plant mass, RPR was lower at elevated CO2, which was caused by a reduction in Φmass despite some compensation by higher LUE.
  • 5We conclude that elevated CO2 increases size inequality of a stand through enhanced photosynthesis and growth of dominants, which reduce the light availability for subordinates and consequently increase size inequality in the stand.

Introduction

Although many studies have been carried out on the effect of elevated CO2 on photosynthesis and growth of plants (Bazzaz 1990; Poorter, Roumet & Campbell 1996; Curtis & Wang 1998), studies on the stand-level response to elevated CO2 are still limited (Körner, Bazzaz & Field 1996). Particularly, individuals growing in a stand have rarely been studied (but see Morse & Bazzaz 1994; Wayne & Bazzaz 1997). Individuals in the stand are subject to different availability of resources (Hikosaka, Sudoh & Hirose 1999; Hikosaka & Hirose 2001), which may result in different responses to elevated CO2 among individuals.

Studying the effect of elevated CO2 on the size structure in a stand of Chenopodium album, we found that the effect of elevated CO2 on individuals in the stand changed depending on the nutrient availability and stage of stand development (Nagashima et al. 2003). At an early stage of plant growth, CO2 elevation benefited all individuals and shifted the whole size distribution to larger size classes in the stand. At later stages, dominant individuals were still larger at elevated than at ambient CO2, but the difference in the size of subordinate individuals between CO2 treatments became smaller. These tendencies were more prominent at high than at low nutrient availability. Morse & Bazzaz (1994) found increasing size inequalities with elevated CO2 in the stand of Abutilon theophrasti and Amaranthus retroflexus, whereas Wayne & Bazzaz (1997) found a size inequality decreasing with elevated CO2 in the stand of Betula alleghariensis. These inconsistencies in the effect of elevated CO2 suggest the necessity for a more mechanistic approach to the development of size inequality.

Difference in size structure results from different size-dependent growth rates of individuals in the stand (Westoby 1982). In a dense stand, plant growth is strongly influenced by light availabilities. Because the light availability decreases with increasing depth in the canopy (Monsi & Saeki 1953), smaller subordinate individuals are disadvantaged in capturing light. However, maintaining leaves at lower positions needs less investment of biomass for support tissues (Givnish 1982). Smaller individuals can thus increase biomass allocation for leaf expansion, and partly ameliorate the limited light availability (Anten & Hirose 1998). To indicate the efficiency of biomass use to capture light, Hirose & Werger (1995) introduced the parameter Φmass, defined as photon flux captured per unit above-ground mass. They suggested that Φmass might not differ between dominant and subordinate species in multispecies systems. However, plant growth is determined not only by the amount of acquired resources, but also by the efficiency of resource use (growth per unit amount of resource acquired). Hikosaka et al. (1999) defined light-use efficiency of photosynthesis (LUE) as photosynthesis per unit photon interception, and described the photosynthesis of individuals as the product of Φmass and LUE:

RPR = Φmass × LUE(eqn 1)

where RPR is the relative photosynthetic rate, i.e. photosynthetic rate per unit above-ground mass. Provided plant growth is proportional to leaf photosynthesis, RPR is closely related to the relative growth rate (RGR). Hirose et al. (1997) found a positive correlation between photosynthesis and growth observed across ambient and elevated CO2. With a modification of the canopy photosynthesis model of Hirose & Werger (1987), Hikosaka et al. (1999) estimated the photosynthetic rate of individuals in a natural monospecific stand of an annual, Xanthium canadense. They found that Φmass was higher in larger individuals, while LUE was highest in intermediate individuals. As a consequence, RPR was high in intermediate and larger individuals, and lowest in smaller individuals. Anten & Hirose (2001) showed how Φmass and LUE of individuals are affected by light and nitrogen availability.

Elevated CO2 is expected to increase LUE because it enhances photosynthetic rates (Pinter et al. 1994; Dewar 1996; Hirose & Bazzaz 1998). However, long-term growth at elevated CO2 often reduces photosynthetic capacity via reduction in leaf N content (CO2 acclimation; Stitt 1991; Sage 1994; Makino & Mae 1999). If CO2 acclimation is caused by CO2 sink limitation of photosynthesis, it is probable that the reduction in photosynthetic capacity is greater in larger individuals whose photosynthetic rates are higher due to exposure to strong light. Φmass may also be influenced by elevated CO2. Φmass is the product of Φarea (photon flux density captured per unit leaf area) and LAR (leaf area ratio, leaf area per unit above-ground mass) (Hirose & Werger 1995):

Φmass = Φarea × LAR(eqn 2)

Φarea measures light availability per unit leaf area, while LAR is a measure of biomass allocation for leaf area expansion. Many studies have shown that elevated CO2 reduces LAR (Pettersson, McDonald & Stadenberg 1993; Hirose et al. 1996; Makino et al. 1997; Lutze & Gifford 1998; Harmens et al. 2000; Ishizaki, Hikosaka & Hirose 2003). Wayne & Bazzaz (1997) suggested that reduction in LAR of larger dominant plants might increase light availability in lower layers of the canopy at elevated CO2.

In the present paper, following Nagashima et al. (2003), we studied monospecific stands of C. album established at ambient and elevated CO2. To elucidate the mechanism of size structure development in the stand, we analysed how elevated CO2 influenced Φarea, LAR, Φmass and LUE, and consequently RPR and growth of individuals in the stand.

Materials and methods

We established Chenopodium album L. stands in two open-top chambers (OTC) of Tohoku University, Sendai, Japan. The species, OTC system and growth conditions were basically the same as described by Nagashima et al. (2003), except for the size of the stand. In the present study, we prepared stands in two wooden boxes (130 × 130 × 30 cm each) which were filled with washed river sand. Of the two, one was exposed to ambient CO2 (320 µmol mol−1) and the other to elevated CO2 (700 µmol mol−1). In our previous study we observed no significant difference in stand growth, size structure and growth rates of individuals between chambers with the same CO2 concentration (Nagashima et al. 2003). Actual CO2 concentration changed 320–400 µmol mol−1 in the ambient, and 670–720 µmol mol−1 in the elevated CO2 chamber. Daily mean air temperature changed from 22·2 °C early in July to 23·9 °C late in August. Photosynthetic photon flux density (PPFD) above the stand at noon ranged between 180 (on rainy days) and 2000 µmol m−2 s−1.

Seeds of C. album were sown on the sand surface on 5 July 1999 and watered every day until the end of the experiment. Emergence was completed by 12 July, after which nutrient solution diluted with tap water was added every week: 12 l of 1/100 times solution of Hyponex (5 : 10 : 5, Hyponex-Japan, Osaka, Japan) per box per week (N, 3·6; P, 7·1; K, 3·6 g m−2 week−1). Grids of 5 × 5 cm were laid on the stands, and seedlings were thinned to one plant per grid (equivalent to 400 plants m−2) by the first measurement on 28 July. To reduce edge effects, the stands were surrounded with a 90% shade cloth, the top of which was adjusted to the height of the canopy.

Two quadrats (each 20 × 40 cm) were established in each stand and harvested in the vegetative growth period: one on 13 August (33 days after emergence) and the other on 27 August (47 days after emergence). Each quadrat was divided into four subquadrats (10 × 20 cm). In the quadrat for harvest on day 47, stem height (H) and diameter (D) of all individuals were measured every week until harvest. Before harvest, vertical distribution of PPFD in the stand was determined every 2·5 cm (day 33) or 5 cm (day 47) using a line sensor SF-80 (Decagon Devices, Pullman) with a point sensor (LI-190SA, LiCor, Lincoln, NE, USA) as reference PPFD measured at top of the canopy. Plants were harvested at the stem base and immediately brought to the laboratory, where plant height, and height and area of every leaf, were measured for each individual. Leaf area was determined with a leaf area meter (LI-3100, LiCor). Stems (petioles were included in the stem fraction) and individual leaves were dried separately at 70 °C for more than 3 days, and weighed. After weighing, leaves were milled and dried again. The total N content of all harvested leaves was measured with an NC analyser (NC-80, Shimadzu, Kyoto, Japan).

Photosynthetic rates of leaves of individuals remaining in the stand were determined on 14 and 29 August using a portable open gas-exchange measurement system (LI-6400, LiCor). In total, 15 leaves each from an ambient and an elevated CO2 stand were collected from different depths in the canopy to yield a range of leaf N contents per area. Light-response curves of photosynthesis were obtained at growth CO2 concentration (either 360 or 700 µmol mol−1). The light source was a mixture of blue and red LEDs (6400–02B, LiCor) and leaf temperature was regulated at 25 °C. Total N content of leaves was determined after drying at 70 °C for more than 3 days.

model

Calculation of photosynthetic rates of individuals followed Hikosaka et al. (1999). PPFD on the horizontal surface at F (leaf area per ground area cumulated from top of the stand) is described as:

I = I0 exp(−KF)(eqn 3)

where I and I0 are PPFD at F and above the canopy, respectively, and K is the coefficient of light extinction (Monsi & Saeki 1953). Photon flux (Φ) absorbed by an individual i at layer j was described as:

Φij = −ΔIjfiji Δfij)(eqn 4)

where Δfij is the cumulative leaf area per ground area of an individual i at layer j (Hirose & Werger 1995). Photon flux absorbed by an individual i was calculated:

Φi = ΣjΦij(eqn 5)

Photon flux density (Ij′) intercepted by leaves in layer j was described as:

image(eqn 6)

(Saeki 1960), where m is a transmittance of a leaf and was assumed to be 0·1. Daily change of PPFD above the canopy was described as a sine square curve:

I0 = I00 sin2[π(t − 5)/14] (5 < t < 19)(eqn 7a)
I0 = 0 (0 < t < 5, 19 < t < 24)(eqn 7b)

where I00 is the noon irradiance above the canopy and is assumed to be 2000 µmol m−2 s−1, and t is the solar time.

Non-rectangular hyperbola was used for the instantaneous rate of leaf photosynthesis as a function of PPFD intercepted by a leaf:

image(eqn 8)

where P is the net photosynthetic rate; Pmax the light-saturated rate of gross photosynthesis; φ the initial slope of the light-response curve; θ the convexity of the light-response curve; and R the rate of dark respiration. Pmax, φ, θ and R are expressed as a function of N content per unit leaf area (Hirose & Werger 1987). After substituting equations 3, 6 and 7 into equation 8, daily photosynthesis (Pday) of a leaf at F was calculated by integrating the instantaneous rate with respect to t.

image(eqn 9)

The sum of the Pday values multiplied by their leaf area gives whole-plant photosynthesis. RPR was calculated from Pday divided by the total above-ground mass of a plant (Hikosaka et al. 1999).

Results

Large differences were observed in plant height in the stand (Fig. 1, monitored non-destructively in the quadrat for harvest 2), and the differences increased with stand development at both CO2 concentrations. At early stages to day 33 (harvest 1), elevated CO2 increased plant height in all size classes, whereas at later stages CO2 elevation benefited large individuals more than smaller ones. This was confirmed by Spearman's rank correlation coefficients between each census and the final harvest (day 47) which increased from 0·35 and 0·13 (day 18) to 0·70 and 0·64 (day 25), 0·86 and 0·91 (day 33), and 0·89 and 0·99 (day 47) at ambient and elevated CO2, respectively. Similar trends were also observed in other size measures, such as diameter and biomass (estimated as D2H, squared basal diameter multiplied by stem height), and in other experiments carried out in different years (1997, unpublished results; 1998, Nagashima et al. 2003). Between the two harvests (day 33 and 47), stand height (defined as the stem height of the tallest plant) increased from 8·9 to 43·2 cm in ambient, and from 12·5 to 49·5 cm at elevated CO2 (Table 1; anova, F = 33·4, P < 0·001); above-ground mass increased from 23 to 141 g m−2 at ambient and 36 to 194 g m−2 at elevated CO2. CO2 elevation increased above-ground mass by 55% on day 33 and 38% on day 47 (F = 4·6, P < 0·05). Between the two harvests, LAI increased by 3·5 times at ambient and by 2·8 times at elevated CO2. CO2 elevation gave 41 and 13% larger LAI on day 33 and 47, respectively (Table 1; F = 3·3, P < 0·1). Thus the effect of CO2 elevation on above-ground mass and leaf area development was larger on day 33 than on day 47, although the effects were only marginally significant.

Figure 1.

Height of individuals in the stand of Chenopodium album grown at ambient (○) and elevated (•) CO2. Plants are arranged according to stem height from smallest to largest. (a) 18; (b) 25; (c) 33; (d) 39; (e) 47 days after emergence.

Table 1.  Stand height, above-ground dry mass, and leaf area index of the stand of Chenopodium album grown at ambient (370 µmol mol−1) and elevated (700 µmol mol−1) CO2, harvested at 33 and 47 days after emergence
DayCO2Stand height* (cm)Above-ground mass (g m−2)Leaf area index (m2 m−2)
  • Mean ± SD (n = 4).

  • *

    Stem height of highest plant in subquadrat.

33Ambient 8·9 ± 0·923·0 ± 7·01·07 ± 0·31
Elevated12·5 ± 1·535·6 ± 5·61·51 ± 0·17
47Ambient  43 ± 2 141 ± 143·71 ± 0·35
Elevated  50 ± 2 194 ± 594·21 ± 0·96

Vertical distribution of leaf area is shown in Fig. 2, where individuals are categorized according to plant height. As we defined plant height as a distance between ground surface and stem apex, some plants had leaves at a layer higher than plant height. Reflecting differences in plant height (Fig. 1), stands at elevated CO2 developed their leaves to higher positions. Individuals tended to have the highest leaf area densities at the upper layers. In each height category, upper leaves contain more N per unit area (Narea; Fig. 3) than lower leaves. When compared at the same layer in the canopy, smaller individuals tended to have larger Narea, as observed in previous studies (Anten et al. 1998; Hikosaka et al. 1999). There was no significant difference in the distribution of Narea between ambient and elevated CO2.

Figure 2.

Vertical distribution of leaf area per unit ground area (Narea) in the stand of Chenopodium album grown at ambient (a,c) and elevated (b,d) CO2 at 33 (a,b) and 47 (c,d) days after emergence. Plants are categorized according to stem height.

Figure 3.

Changes in leaf nitrogen content per leaf area of the Chenopodium album stand grown at ambient (a,c) and elevated (b,d) CO2 at 33 (a,b) and 47 (c,d) days after emergence. Plants are categorized according to stem height. Mean values for leaves at each height class are shown. Height classes: (a,b) •, 2·5–5; ○, 5–7·5; ▪, 7·5–10; □, 10–12·5; ◆, 12·5–15 cm; (c,d) •, 10–20; ○, 20–30; ▪, 30–40; □, 40–50; ◆, 50–60 cm.

PPFD decreased with increasing depth in the canopy. Equation 3 fitted well the relationship between PPFD and cumulative LAI. Difference in this relationship between the two harvests was significant (slope of the regression, P < 0·05), whereas the difference between ambient and elevated CO2 was not significant (ancova, P > 0·05). We thus calculated K values separately for day 33 and 47 using equation 3, combining data for different CO2:

Day 33: I/I0 = exp(−0·70F) (r2 = 0·56)
Day 47: I/I0 = exp(−0·99F) (r2 = 0·97)

For the calculation of canopy photosynthesis below, we used 0·70 and 0·99 for K on days 33 and 47, respectively.

There was no difference in CO2 dependence of the photosynthetic rate between CO2 treatments, indicating no CO2 acclimation in plants growing at elevated CO2 (data not shown). The photosynthetic rate increased with intercellular CO2 concentration up to 400 µmol mol−1, above which it became independent of CO2 concentration. There was no significant difference in the response between plants grown at different CO2 levels.

Relationships between Narea and photosynthetic parameters (Pmax, φ, θ and R) showed no significant difference between the two measurement dates (day 34 and 48; P > 0·05). Except for the initial slope (φ), parameters were significantly correlated with Narea (P < 0·05; Fig. 4). Difference between ambient and elevated CO2 was significant in Pmax (ancova, F = 4·67, P < 0·05) and φ (P < 0·05), and was not significant in θ (F = 1·74, P > 0·10) and R (F = 2·70, P > 0·10). Thus the following equations were used to characterize the photosynthesis of C. album growing at ambient and at elevated CO2:

Figure 4.

Relationships between photosynthetic parameters and leaf nitrogen content per area (Narea) in plants grown at ambient (○) and elevated (•) CO2. (a) Light-saturated photosynthesis (Pmax); (b) initial slope of the light-response curve (φ); (c) convexity (θ); (d) dark respiration (R). Parameters were determined at growth CO2 concentrations. See text for fitted regression equations.

Ambient CO2

Pmax = −4·06 + 16·5Narea (r2 = 0·75)

φ = 0·0519

θ = 0·77 − 0·13Narea (r2 = 0·16)

R = 0·0047 + 0·90Narea (r2 = 0·68)

Elevated CO2

Pmax = −3·79 + 18·0Narea (r2 = 0·85)

φ = 0·0575

θ and R, same as above.

The φ value above was calculated by multiplying the mean of the measured values (Fig. 4) by a factor 0·9. This correction was done as the LED light source led to a higher quantum yield than that for white light due to difference in spectral composition (S. Oikawa, personal communication).

Photosynthetic rates of a leaf in a given layer in the canopy were calculated as a function of PPFD and Narea, and whole-plant photosynthesis was calculated as the sum of leaf photosynthesis. The rate of whole-plant photosynthesis increased with increasing above-ground mass of individual plants (Fig. 5). The photosynthetic rate on day 33 was nearly proportional to above-ground mass, and there was no difference between ambient and elevated CO2. On day 47, the relationships were curvilinear, and the difference between the two CO2 treatments was significant (P < 0·05). Photosynthetic rates of the smallest individuals were close to zero at both CO2 levels. When compared at an equal above-ground mass, daily photosynthesis of individuals at ambient CO2 exceeded that at elevated CO2.

Figure 5.

Relationship between whole-plant photosynthetic rate and above-ground dry mass of individuals in the stand of Chenopodium album grown at ambient (○) and elevated (•) CO2 at (a) 33 and (b) 47 days after emergence.

The relative photosynthetic rate (RPR, photosynthesis per unit above-ground mass) on day 33 slightly increased with above-ground mass, and this relationship was not different between ambient and elevated CO2 (Fig. 6a). However, Φmass and LUE, two components of RPR, were differentially dependent on above-ground mass and the CO2 concentrations (Fig. 6c,e). Φmass was positively correlated to above-ground mass, though the correlation was weak at ambient CO2. At a given above-ground mass, ambient CO2 gave a higher Φmass than elevated CO2. On the other hand, LUE was nearly constant (ambient) or slightly decreased (elevated) with increasing above-ground mass. Elevated CO2 gave a higher LUE when compared at an equal above-ground mass. On day 47, on the other hand, RPR showed a curvilinear, positive relationship with above-ground mass, with higher RPR at ambient than elevated CO2 for a given above-ground mass (Fig. 6b). Φmass was also positively correlated with above-ground mass. For a given above-ground mass, individuals at ambient CO2 had higher Φmass (Fig. 6d). LUE was very low in the smallest individuals (around zero at elevated CO2), but increased sharply to the maximum in medium-sized individuals and then decreased in individuals with larger above-ground mass (Fig. 6f).

Figure 6.

(a,b) Relative photosynthetic rates (RPR, whole-plant photosynthetic rate per unit above-ground mass); (c,d) photon flux captured per unit above-ground mass (Φmass); and (e,f) light-use efficiency (LUE, photosynthesis per unit captured photon) as a function of above-ground dry mass at 33 (a,c,e) and 47 (b,d,f) days after emergence. RPR = Φmass × LUE. ○, ambient; •, elevated CO2.

Φmass was further analysed as the product of Φarea and LAR (Fig. 7). Φarea was correlated positively, and LAR negatively, to above-ground mass. At a given above-ground mass, Φarea was higher in ambient CO2 at both harvests due to lower LAI (Table 1). LAR decreased with increasing above-ground mass at both harvests at both CO2 concentrations. Analysis of covariance after logarithmic transformation indicated that the relationship between LAR and above-ground mass was not significantly different on day 33 (F = 0·71, P > 0·05), but on day 47 (F = 5·67, P < 0·05), where LAR was larger at ambient CO2 when compared at the same above-ground mass.

Figure 7.

(a,b) Photon flux density captured per unit leaf area (Φarea); (c,d) leaf area ratio (LAR, leaf area per unit above-ground mass) as a function of above-ground mass at 33 (a,c) and 47 (b,d) days after emergence. ○, ambient; •, elevated CO2.

There was a parabolic relationship between LUE and Φarea across two harvest dates (Fig. 8). LUE was maximized at an intermediate Φarea, and lowest at the lowest end of Φarea. Data points on day 47 covered the whole range of Φarea, while data on day 33 were concentrated at the highest end of Φarea.

Figure 8.

Light-use efficiency (LUE) as a function of Φarea. Circles, 33 days; squares, 47 days after emergence. Open symbols, ambient; closed symbols, elevated CO2.

Discussion

We estimated whole-plant photosynthesis by applying the canopy photosynthesis model to individuals (Hikosaka et al. 1999) in the stand established under ambient and elevated CO2. To validate the model, growth rates estimated from the two harvests were plotted against mean photosynthetic rates calculated from the model (Fig. 9). There was a strong correlation between the two variables (r2 = 0·90). However, when photosynthesis was converted into dry mass with an assumption that 1 mol CO2 is equivalent to 30 g dry mass (Hikosaka et al. 1999), dry mass growth was only 40% of photosynthesis. This may be explained by several factors. First, root mass was excluded from our harvests. In the early growth of X. canadense, dry mass allocation for root growth was about 30% (Shitaka & Hirose 1993). Second, some photosynthates were lost through respiration of non-photosynthetic organs, which amounts up to 50% of total photosynthesis (Lambers 1985). Third, actual PPFD at noon varied from 180 to 2000 µmol m−2 s−1 and was mostly less than the value (2000 µmol m−2 s−1) that was assumed to calculate photosynthesis. Fourth, dominance of diffuse light implicitly assumed with equation 3 overestimates PPFD in the canopy on sunny days (Anten 1997). There was no difference in the growth–photosynthesis relationship between CO2 concentrations (Fig. 9), which may disprove the importance of structural materials and secondary compounds in C. album (Körner & Arnone 1992; Poorter et al. 1997). It also indicates that the photosynthesis model gives good qualitative predictions of the effect of CO2 on growth.

Figure 9.

Relationship between estimates of dry mass growth (g day−1) and photosynthetic rates (mmol day−1). Both were estimated for individuals in the stand for harvest 2. Dry mass growth was estimated using an allometry between D2H (squared stem diameter multiplied by stem height) and dry mass at harvests 1 and 2. Photosynthetic rates are a mean of calculated values for harvests 1 and 2. ○, ambient; •, elevated CO2. Dotted line indicates equivalence between CO2 uptake and dry mass growth.

Elevated CO2 enhanced stand-level growth (Table 1) and altered the size distribution of individuals such that, at early stages, all individuals were larger at elevated than at ambient CO2, while at later stages the difference between CO2 concentrations in the size of subordinate individuals became smaller and thus size inequality became greater at elevated CO2 (Fig. 1; Nagashima et al. 2003). We used individual plant photosynthesis calculated at two stages in their vegetative growth to analyse the mechanism involved in the change in size structure. Size inequality developed after canopy closure (Nagashima 1999; Nagashima et al. 2003) and CO2 elevation accelerated leaf area growth (Nagashima et al. 2003; Table 1). Accordingly, Nagashima et al. (2003) suggested that faster development of size inequality might be due to ‘putting the clock forward’ at elevated CO2. Coleman, McConnaughay & Bazzaz (1993) suggested that the effect of elevated CO2 on plant growth resulted simply from plants at high CO2 becoming large earlier. The present study, however, suggests that these developmental effects are exerted through increasing LUE, despite reductions in Φmass at elevated CO2.

On day 33, the rate of whole-plant photosynthesis was nearly proportional to the above-ground mass irrespective of CO2 concentrations (Fig. 5a), and the photosynthetic rate per unit above-ground mass (RPR) was only slightly correlated to above-ground mass with no difference between CO2 levels (Fig. 6a). Weak, positive correlations between Φmass and above-ground mass on day 33 indicate that competition for light was slightly asymmetric at both CO2 levels: larger individuals tended to capture a disproportionately large amount of light relative to their size (Weiner 1990). On day 47, whole-plant photosynthesis was curvilinearly related to above-ground mass (Fig. 5b). A stronger dependence of Φmass on above-ground mass on day 47 than on day 33 indicates that competition for light became more asymmetrical later in stand development, consistent with earlier observations (Weiner & Thomas 1986; Nagashima 1999; Nagashima et al. 2003). Lower Φmass at elevated CO2 may indicate that light became more limiting for growth with CO2 elevation, particularly in smaller subordinates. In a natural stand of X. canadense, Anten & Hirose (1998) showed that Φmass was larger in dominant than in subordinate individuals, and that Φmass was larger in individuals in an open than in a dense stand.

Large dominants having larger LUE at elevated CO2 partly compensated for their lower Φmass, while subordinate individuals with their smaller LUE did not. These differences in response to CO2 and light conditions between dominant and subordinate individuals may account for increased size inequality at elevated CO2 (Fig. 1; Nagashima et al. 2003). At an early stage, when light was not limiting, all individuals were benefited by elevated CO2 and increased their growth in proportion to their size (symmetric competition). With stand development, large dominants captured light more than proportionate to size (asymmetric competition) and preempted light from subordinates. Thus subordinates slowed their growth more at elevated than at ambient CO2.

Different Φmass between CO2 treatments (Fig. 6) was attributable mainly to the difference in Φarea rather than in LAR (Fig. 7). Φarea represents the photon flux captured by individuals (Hirose & Werger 1995; Anten & Hirose 1998). The gradient of PPFD developed in the canopy reduced Φarea for subordinate individuals, but the smaller Φarea was partly compensated for by a larger LAR. As photon flux density per unit ground area is limited, increasing LAI necessarily reduces stand-level Φarea. In the present study, elevated CO2 increased LAI (Table 1) and dominants deployed their leaves at higher positions at elevated than at ambient CO2 (Fig. 2). Thus light availability to subordinates was reduced at elevated CO2. Significant reduction in LAR by elevated CO2 has been reported (Pettersson et al. 1993; Hirose et al. 1996; Makino et al. 1997; Wayne & Bazzaz 1997; Lutze & Gifford 1998; Harmens et al. 2000). In the present study, however, CO2 elevation only slightly (but significantly) decreased LAR on day 47, but not on day 33 (Fig. 7c,d).

Elevated CO2 increased LUE, except for small subordinate individuals on day 47 (Fig. 6f). Because photosynthesis is a saturating function of irradiance, LUE changes depending on light availability (Niinemets & Tenhunen 1997; Hirose & Bazzaz 1998). To separate the effect of light availability from that of CO2 elevation, LUE was plotted against Φarea in Fig. 8. This indicates that the smaller LUE of the smallest individuals at elevated CO2 on day 47 (Fig. 6f) is attributable to their small Φarea, and that a greater LUE in larger individuals (Fig. 6c,f) was a direct response to elevated CO2. Several authors (Pinter et al. 1994; Hirose & Bazzaz 1998; Luo et al. 2000; Rodriguez et al. 2001) have also reported an increase in LUE with CO2 elevation.

As Narea did not change with CO2 elevation (Fig. 3), enhanced LUE at elevated CO2 was accrued to increase in photosynthesis as represented by Pmax and the initial slope (φ) (Fig. 4). CO2 elevation increases Pmax and φ in C3 species through enhanced carboxylation with suppression of oxygenation (Farquhar, von Caemmerer & Berry 1980). When measured at the same CO2 concentration, no difference was found in the PmaxN relationship between ambient and elevated grown plants (data not shown), suggesting that the photosynthetic activity was not downregulated in our experiment. However, doubling CO2 concentration enhanced Pmax only by 10–15% (Fig. 4), which was less than the expected enhancement at 25 °C without downregulation (40–50%; Sage 1994; Hikosaka & Hirose 1998; Makino & Mae 1999). Large enhancement, however, is possible in C3 species only when either RuBP (ribulose-1,5-bisphosphate) regeneration or RuBP carboxylation (Sage 1994) limits photosynthetic rates. When photosynthesis is limited by triose-phosphate utilization, photosynthetic rates are independent of CO2 concentration (Sharkey 1985). In our experiment, photosynthetic rates were not increased above 400 µmol mol−1 CO2, suggesting that photosynthesis might have been limited by triose-phosphate utilization.

In conclusion, elevated CO2 enhanced photosynthesis and growth rates in all individuals at an early stage, and shifted the plant size distribution upward. Increased leaf area in the stand reduced light availability for smaller subordinates earlier at elevated than at ambient CO2. Reduction in light capture was compensated for by enhanced LUE at elevated CO2 at an early stage. Later, the reduction exceeded the enhancement of LUE and reduced the photosynthesis and growth rate of subordinates. Elevated CO2 thus increased size inequality in the stand (1) by enhancing photosynthesis and growth of dominants because of increased LUE; and later (2) by dominants reducing the light available to subordinates.

Acknowledgements

We thank Niels Anten for comments, and Ken-ichi Sato, Masaharu Kato, Hiroshi Kimura, Kou Kijima and Shimpei Oikawa for advice and help in the experiment. This work was supported in part by grants-in-aid from the Japan Ministry of Education, Science and Culture.

Ancillary