Long-term acclimation of leaf production, development, longevity and quality following 3 yr exposure to free-air CO2 enrichment during canopy closure in Populus


Author for correspondence: Gail Taylor Tel: +44 2380 592335 Fax: +44 2380 594269 Email: g.taylor@soton.ac.uk


  • • The effects of elevated CO2 on leaf development in three genotypes of Populus were investigated during canopy closure, following exposure to elevated CO2 over 3 yr using free-air enrichment.
  • • Leaf quality was altered such that nitrogen concentration per unit d. wt (Nmass) declined on average by 22 and 13% for sun and shade leaves, respectively, in elevated CO2. There was little evidence that this was the result of ‘dilution’ following accumulation of nonstructural carbohydrates. Most likely, this was the result of increased leaf thickness. Specific leaf area declined in elevated CO2 on average by 29 and 5% for sun and shade leaves, respectively.
  • • Autumnal senescence was delayed in elevated CO2 with a 10% increase in the number of days at which 50% leaf loss occurred in elevated as compared with ambient CO2.
  • • These data suggest that changes in leaf quality may be predicted following long-term acclimation of fast-growing forest trees to elevated CO2, and that canopy longevity may increase, with important implications for forest productivity.


Given the importance of forests for global productivity, the consequences of increased atmospheric carbon dioxide partial pressure(s) (pCO2) for the global carbon cycle are potentially extremely large (Malhi et al., 1999). Despite this, there are still relatively few large-scale, long-term experiments from which predictions about likely forest responses to future pCO2 can be made. Few studies have been completed where trees are allowed to develop to canopy closure and where a ‘stable’ response to pCO2 is likely. Increase in C availability drives changes in plant structure at multiple levels contributing to altered plant, community and ecosystem level function (Pritchard et al., 1999). Given that European forests are considered an important sink for atmospheric CO2 (Janssens & Pilegaard, 2003), an understanding of leaf-level responses to increased C is still warranted. Leaves are crucial to plant function as they regulate light interception, photosynthesis, water use and therefore total plant productivity (Murthy & Dougherty, 1997). These aspects are particularly important for poplars as they are fast-growing species, characterized by high productivity.

Leaf growth is highly responsive to the environment, including the supply of atmospheric CO2. Elevated pCO2 is known, in many instances, to cause stimulation in the rate of leaf expansion (Ferris & Taylor, 1994), whole-plant leaf-area development (Taylor et al., 1994; Pritchard et al., 1999) and, to a lesser extent, leaf initiation (Gregory, 1996). Maintenance of increased biomass productivity in elevated pCO2 will depend on continued stimulation of leaf area as the canopy closes, or on any photosynthetic acclimation, i.e. change in the biochemical efficiency of CO2 assimilation (Norby et al., 1999).

Acclimation of leaf biochemistry, morphology and anatomy under elevated pCO2 has been widely reported (Allen, 1990; Bazzaz, 1990; Rogers et al., 1992; Ceulemans et al., 1997). Altered stomatal physiology and density (Penuelas & Matamala, 1990), greater numbers of mesophyll cells and chloroplasts (Masle, 2000), and increased leaf area, leaf thickness and leaf longevity (reviewed by Pritchard et al., 1999) may increase gas-exchange rates, light harvesting and, ultimately, C assimilation potential, but few studies have been conducted over the longer time scale. In field conditions, reduced light intensities, soil water stress, suboptimal soil fertility and intra- and interspecific competition may all limit the magnitude of a response to elevated pCO2 (Curtis et al., 1995). Physiological acclimation to elevated pCO2 has been reported in many species including the trees Pinus radiata (Griffin et al., 2000) and Quercus ilex (Blaschke et al., 2001). An acclimatory decrease in ribulose bisphosphate carboxylase/oxygenase (Rubisco) activity or quantity in elevated pCO2 may limit photosynthetic efficiency where nitrogen availability is also limited and negate any stimulation of biomass derived from increased leaf growth (Hymus et al., 2001).

Changes in leaf tissue morphology and chemistry, including reduced leaf N concentrations and increased starch and soluble sugar concentrations, have been proposed as mechanisms for photosynthetic acclimation based on N reallocation and feedback-driven downregulation (Drake et al., 1997). Studies with trees unrestricted in root growth that are exposed to elevated pCO2, rarely describe any acclimation in photosynthetic C assimilation, but often report changes in leaf biochemistry, in particular N content (Curtis & Wang, 1998). Only from the results of long-term experiments will it be possible to conclude whether changes in leaf morphology and biochemistry are a common response to elevated pCO2 and whether they preface increases or decreases in photosynthetic capacity.

Mechanistic studies of increased leaf-area growth in poplars under elevated pCO2 have shown that this increase is determined by a stimulation in growth through increased cell expansion (Taylor et al., 2001; Ferris et al., 2001), with or without increased cell numbers through increased cell production (Taylor et al., 2003). How cell expansion and cell production are thought to interact in the control of leaf growth rate and final leaf size is the subject of ongoing speculation. The long-term responses and interactions between these mechanisms of growth in leaves exposed to elevated pCO2 are unknown. A fundamental understanding of the influence of C supply on cellular growth processes is required in order to predict plant responses in future CO2 atmospheric concentrations.

Here, the long-term effects of elevated pCO2 on leaf development were investigated at the POPFACE plantation (Miglietta et al., 2001) where three selected species of Populus were exposed to elevated CO2 at 550 µmol mol−1 over 3 yr. Leaf growth, anatomy, cell size and production, and basic chemistry were assessed during three growing seasons as the canopy closed. The influences were observed of species, shading as the canopy closed, and different growing seasons, characterized by weaker or stronger competition (Gielen et al., 2001, 2003). The effects of elevated pCO2 on spatial and temporal patterns of leaf development were also quantified to determine the relative importance of cell production and cell expansion for final leaf size and shape, and broad implications for long-term acclimation of leaf function in managed forest canopies of the future are discussed here. In addition, the impact of elevated pCO2 on Populus nigra leaves was followed at different leaf ages to determine the relevance of developmental stage.

Materials and Methods

Site and growth conditions

The experimental plantation was laid out at a 9 ha site in central Italy (Tuscania, Province of Viterbo, latitude 42°22′ N, longitude 11°48′ E, altitude 150 m) as described by Ferris et al. (2001). During the spring of 1999 experimental plots were planted in equal numbers with cuttings of three Populus species: P. alba (clone 2AS-11), P. nigra (clone Jean Pourtet) and P. × euramericana (P. deltoides × P. nigra, clone I-214) within a plantation of P. × euramericana. The entire plantation was managed as a short-rotation coppice system, with single stems for the first 3 yr, precoppice. Six octagonal areas of 314 m2 within the plantation were designed as experimental plots, each containing two sectors per species of 52 plants. Cuttings were spaced at 1 × 1 m within the plots (planting density 10 000 trees ha−1) and 2 × 1 m within the remaining plantation. The six plots were categorized into three blocks; within each block one plot received CO2 enrichment at a target concentration of 550 µmol mol−1 (FACE) from bud burst to leaf fall (March–November), while the other was left as a control. The mean pCO2 within the enriched plots was 544 ± 48 µmol mol−1 during 1999 (year 1); 532 ± 83 µmol mol−1 during 2000 (year 2); and 554 ± 95 µmol mol−1 during 2001 (year 3). Throughout the fumigation, pCO2 measured at 1 min intervals was within 20% of the target at least 80% of the time.

Throughout the experiment the entire area and all plants were drip-irrigated. Irrigation was monitored to ensure equal application in each experimental plot. Irrigation was applied to match transpiration, reaching a maximum of 10 mm d−1 during year 3. No additional fertilizers were used during the experiment.

During year 1 the canopy was open and there were no shade leaves. Average wind speed was 1.8 m s−1 during the sampling campaign, reaching a maximum of 5.8 m s−1. Average air temperature was 24°C reaching a maximum of 36°C during this period. Precipitation was 23 mm and relative humidity averaged 68%. During year 2 canopy closure had been achieved but main stem leaves were not shaded (Gielen et al., 2001). Average air temperature was 20°C reaching a maximum of 31°C during the sampling campaign. Precipitation was 14 mm with 10 mm falling on one day. Mean relative humidity was 73% and average wind speed was 2.5 m s−1 reaching a maximum of 8.1 m s−1. During year 3 the canopy was completely closed (Gielen et al., 2003), although main stem leaves were not shaded. This period was characterized by strong competition between trees. Average air temperature during the sampling campaign was 16°C, mean relative humidity was 76% and average wind speed was 1.13 m s−1. Days were overcast with mean global radiation of 266 W m−2 and 13 mm rain during the period.

Further information on the site characteristics and FACE facility is given by Calfapietra et al. (2003); Miglietta et al. (2001).

Yearly assessments of leaf length, area and specific leaf area

In each of the 3 yr of the experiment a campaign to assess leaf development, morphology and quality was undertaken in all three species. In each year, leaves of equivalent, measured leaf plastochron index (LPI) from the main stem (sun leaves) and from the second or third primary branch (shade leaves) were sampled.

During year 1,100 unfurled, sun leaves of all ages from each species were collected and the relationship between leaf length and area was established. Leaf length, width and area were measured on a DELTA-T image analyser (DELTA-T Devices, Cambridge, UK). Coefficients of determination of 0.99, 0.98 and 0.99 for P. × euramericana, P. nigra and P. alba, respectively, were determined for the regression of leaf length against area. Subsequently, 12 young but unfurled leaves were selected from each of the three species in each experimental plot, and length measured at intervals of 2–3 d. Leaf extension rate (LER) was calculated as the daily increment in length (mm d−1). The regression equation was used to convert leaf length measurements to area. Main stem leaf production was observed, and specific leaf area (SLA) was determined from five recently matured leaves harvested from each plot and measured, dried and weighed to calculate SLA (mm2 g−1).

In years 2 and 3 measurements of leaf length, area and SLA were made as before, except that the coefficient of determination did not satisfactorily demonstrate a relationship between length and area in P. × euramericana (0.42). Therefore P. × euramericana leaf area was followed by digital photography in the selected leaves (Nikon Coolpix 950) every 2 d and measured using the image analysis freeware computer programme Scion Image (http://www.scioncorp.com). In addition, mature sun and shade leaves of each species were harvested, images made and incorporated into Scion Image for analysis of mature leaf area.

A detailed investigation of leaf area and SLA in relation to developmental stage: P. nigra, year 3

During year 3 leaf quality and function were assessed in detail in P. nigra at several contrasting stages of development as leaves expanded and matured. A number of young leaves, just emerged from the bud and still in the leaf-expansion phase, were labelled. Measurements of these leaves were carried out the day after labelling, day 0, and repeated after 5, 10 and 20 d, respectively (called day 5; day 10; and day 20 measurements). On each measurement day three leaves per plot were collected, area was determined using a Li-Cor 3000 leaf-area meter (Li-Cor Inc., NE, USA) and these leaves were utilized for SLA (mm2 g−1) measurement as previously described. Leaves 20 d old were considered mature, confirmed by measurements of chlorophyll content and fluorescence (data not shown).

Yearly assessments of leaf anatomy

In year 1 transverse sections of five fully expanded leaves of each species were prepared as described by Ferris et al. (2001). In year 2, 10 mm2 sections from the base of nine P. × euramericana young and mature sun leaves were prepared as described by Taylor et al. (2003). Briefly leaves were placed in fixative [formalin : glacial acetic : 70% ethanol (1 : 1 : 18, v/v)]. For light microscopy the discs were cut into 1–2 mm squares and fixed in buffered osmium tetroxide. The samples were rinsed in 0.1 m PIPES buffer, dehydrated in an ethanol series, cut into 1 mm squares and embedded in TAAB resin (TAAB Laboratories, Aldermaston, UK). Sections (0.5 µm) were cut on a Leica OMU 3 Ultramicrotome (Reichert, Austria) and stained with 1% toluidine blue in 1% borax. Images were captured at ×400 magnification using a digital camera attached to a Zeiss microscope. In both years the leaf thickness and transverse length of mesophyll and palisade cells were measured from sectioned photographs using the Scion Image imaging programme.

Yearly measurements of foliar nitrogen

In each year five mature leaves from each plot and of each species were sampled and the area measured as before using the Scion Image analysis programme. These leaves were dried, weighed and ground in a rotor-speed mill (0.5 mm sieve perforations). Ground samples were prepared by sulphuric digest, liberating C as CO2, converting all metals to their sulphates and converting most forms of N into ammonium ions (ammonium sulphate). The sulphuric digest solution was then analysed using a continuous flow analysis system (System 4, ChemLab, Great Dunmow, UK and Series 2000, Burkard Scientific, Uxbridge, UK, implemented by MB Scientific Services Ltd, Longstanton, UK). This method utilizes the reaction of ammonia with salicylate and dichloroisocyanurate (DIC), in alkaline solution. The DIC decomposes in alkaline solution to release hypochlorite ions which react with the salicylate to give a substituted indophenol which was measured at 650 nm. Nitroprusside was added as a catalyst. Leaf total N was calculated as a percentage g−1 d. wt (Nmass) and also on a leaf-area basis as g m−2 (Narea).

A detailed investigation of leaf C and N content in relation to developmental stage: P. nigra, year 3

Populus nigra leaves were selected and labelled as described for the detailed investigation of area and SLA, at contrasting developmental stages. Three sample leaves from each plot were collected on day 0 and after 5, 10 and 20 d, respectively (called day 5, day 10 and day 20 measurements) and area was determined using a Li-Cor 3000 leaf-area meter as before. These leaves were powdered, weighed and C and N content estimated by combustion using a CHN-Carlo Erba instrument (Milan, Italy). Carbon and N contents were expressed as percentages g−1 d. wt (Cmass and Nmass) and also on a leaf-area basis as g m−2 (Carea and Narea).

Yearly measurements of epidermal cell area and number

During each of the field campaigns a single mature leaf was harvested from sun and shade branches of five trees of each species in each plot, and the area was measured using the Scion Image programme as before. A small area of adaxial tissue at the base of the leaf was painted with nail varnish, left to dry for ≈ 20 min and an imprint obtained by applying sticky tape to the dried varnish, then mounting on a glass microscope slide (Gardner et al., 1995). Images of imprints were examined under a light microscope and the epidermal cell size of 10 randomly chosen cells on each slide measured using Scion Image as before. Epidermal cell numbers per leaf were calculated by dividing leaf area by mean epidermal cell area.

Yearly measurements of cell wall extensibility

Cell-wall properties of plasticity and elasticity were measured in years 1 and 2 as described by Ferris et al. (2001); Taylor et al. (2003). Briefly, five expanding leaves from each species and experimental plot, previously stored in 70% methanol, were rehydrated and a strip of tissue (3 × 5 mm) from the basal area cut from each. Leaf strips were attached by brass clamps to a custom Instron apparatus for assessment of biophysical characteristics and stretched to give irreversible and reversible extension per 10 g load. Results were expressed as percentage plasticity (percentage irreversible extension per 10 g load) and percentage elasticity (percentage reversible extension per 10 g load).

Yearly measurements of leaf production and longevity

During years 2 and 3, for all three species a detailed investigation was undertaken of leaf production, assessed using the plastochron index (PI), and leaf longevity, assessed from leaf retention measurements. The uppermost 10 leaves of five selected main stems and branches in each plot were labelled in accordance with the PI system (Erickson & Michelini, 1957) so that the most recently emerged leaf just < 20 mm long (30 mm in P. × euramericana) was Ln+1 and the leaf immediately below was Ln, then Ln−1, and so on down the branch. Leaf lengths were measured weekly throughout the growing season and PI was calculated using the equation (Erickson & Michelini, 1957):

PI = n + [(log Ln − log 20) / (log Ln − log Ln+1)]

where n is the serial number of the leaf.

In year 3, leaf loss was followed on the same branches. Percentage leaf fall was estimated by counting the leaf scars along a branch at each measurement date and the life span of every fifth leaf was recorded throughout the year 3 growing season.

Statistical analysis

Measured and calculated values were analysed for statistical significance with the software package minitab 13.0 for Windows (Minitab Inc., Philadelphia, PA, USA). Independent data were analysed using a fully randomized block, custom factorial design ANOVA with block an untested random source of variation and pCO2 and LPI fixed factors in the model. For each year, two-way ANOVA for pCO2 and species were conducted for each characteristic assessed. Bartlett's and Levene's tests were conducted for homogeneity of variance and post hoc Dunnett's tests were applied where appropriate. Significant P values are annotated as ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001.


Leaf growth

Final leaf length (mm) increased in each successive growing season in the sun for all three species (Fig. 1). Leaf length also increased from year 2 to year 3 in shade leaves of P. × euramericana and P. nigra. Overall, leaf length was not significantly enhanced by pCO2 enrichment after year 1, and increases in area, not reflected in increases in length, demonstrated changing leaf shape in P. × euramericana in response to elevated pCO2. In each year LER (mm d−1) varied significantly with species (Table 1). In the sun leaves of P. × euramericana extended rapidly, but there was little difference between sun and shade LERs of P. nigra and P. alba except that, unusually, in year 2 the extension of shade leaves of P. alba was stimulated in comparison with other species and with sun leaves. The pCO2 enrichment significantly stimulated LER in full sun; this trend persisted but was no longer significant in the second or third growing seasons.

Figure 1.

(a) Mature leaf area and (b) mature leaf length for leaves of three Populus spp. exposed to either ambient or elevated pCO2 (550 µmol mol−1) at the POPFACE experiment over 3 yr and in full sun. Each value represents average (± SE) of 36 measurements per treatment. A summary of two-way ANOVA is shown, and significant effect of treatment at each data point where significance was found is also indicated: ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001.

Table 1.  Summary of two-way ANOVA
ParameterYear 1Year 2Year 3
CO2 treatmentSpeciesSpecies ×  CO2 interactionCO2 treatmentSpeciesSpecies ×  CO2 interactionCO2 treatmentSpeciesSpecies ×  CO2 interaction
  • *

    , **, *** Significant at < 0.5, < 0.01 and < 0.001 probability levels, respectively; ns, not significant.

  • ¶Study on all three species undertaken August 1999.

  • †Study undertaken on both sun and shade leaves for all three species July 2000.

  • ‡Study undertaken on both sun and shade leaves for all three species May 2001.

Leaf extension rate (mm d−1) *** (sun)*** (sun)ns (sun)ns (sun)*** (sun)ns (sun)ns (sun)*** (sun)ns (sun)
   ns (shade)** (shade)ns (shade)ns (shade)** (shade)ns (shade)
Mature leaf area (mm2) *** (sun)*** (sun)*** (sun)ns (sun)** (sun)ns (sun)ns (sun)* (sun)ns (sun)
   ns (shade)ns (shade)ns (shade)ns (shade)** (shade)ns (shade)
Plastochron index   ns (sun)*** (sun)*** (sun)   
   ** (shade)* (shade)* (shade)   
Leaf thickness (µm)ns (sun)ns (sun)ns (sun)ns (sun)     
   ns (shade)     
Mesophyll thickness (µm)** (sun)** (sun)** (sun)* (sun)
   ns (shade)     
Palisade thickness (µm)** (sun)** (sun)ns (sun)** (sun)     
   ns (shade)     
Epidermal cell area (µm2)*** (sun)** (sun)*** (sun)ns (sun)ns (sun)* (sun)ns (sun)** (sun)ns (sun)
   ns (shade)* (shade)ns (shade)ns (shade)* (shade)ns (shade)
Epidermal cell number (105)*** (sun)*** (sun)*** (sun)ns (sun)** (sun)** (sun)ns (sun)ns (sun)ns (sun)
   ns (shade)* (shade)* (shade)ns (shade)** (shade)ns (shade)
Cell wall extensibility (%)ns (sun)* (sun)ns (sun)ns (sun)** (sun)ns (sun)   
   ns (shade)ns (shade)ns (shade)   
Nitrogen (% d. wt)ns (sun)* (sun)ns (sun)*** (sun)*** (sun)*** (sun)*** (sun)* (sun)ns (sun)
      * (shade)ns (shade)ns (shade)
Specific leaf area (mm2 g−1)ns (sun)*** (sun)ns (sun)* (sun)ns (sun)ns (sun)* (sun)ns (sun)ns (sun)
   ns (shade)ns (shade)ns (shade)ns (shade)ns (shade)ns (shade)

Populus × euramericana sun leaf area (mm2) was significantly stimulated by pCO2 enrichment in each year with a mean increase of 32% over the course of the experiment (Table 2). Final leaf areas of P. nigra and P. alba leaves were stimulated by pCO2 enrichment in year 1, statistically equal in sun leaves in year 2, but were reduced under elevated pCO2 in year 3 in both species. Over the course of the experiment differences in final leaf area were not significant in these two species. Final leaf area was larger in P. × euramericana in year 3 than it had been in either preceding growing season and there was little difference in the final leaf area of P. nigra in any year. Final leaf area of P. alba was largest in its second growing season in shade leaves.

Table 2.  Summary of percentage increases in growth and developmental characteristics of leaves exposed to elevated pCO2 (550 µmol mol−1) during the POPFACE experiment from year 1 (open canopy with newly planted material), year 2 (closing canopy) and year 3 (entirely closed canopy)
CharacteristicLeavesYear 1Year 2Year 3
Populus × euramericana
Mature leaf area (mm2)Sun+27+27+41
Shade −27+20
Epidermal cell area (µm2)Sun+48 +2−19
Shade   0+28
Epidermal cell numberSun+42+21 +3
Shade +20 −7
Leaf nitrogen (% d. wt)Sun+27−17−15
Shade  −17
Specific leaf area (mm2 g−1)Sun +4 −6−30
Shade  −2 −4
Populus alba
Mature leaf area (mm2)Sun+22 +1−21
Shade −41−20
Epidermal cell area (µm2)Sun −2−13−41
Shade −17−25
Epidermal cell numberSun+25 +7+19
Shade −14  0
Leaf nitrogen (% d. wt)Sun−21−11−22
Shade   −6
Specific leaf area (mm2 g−1)Sun −4 −7−35
Shade −19 +5
Populus nigra
Mature leaf area (mm2)Sun+24+11−34
Shade +43 −2
Epidermal cell area (µm2)Sun+19+10  0
Shade +15+19
Epidermal cell numberSun +3−25−18
Shade −25 −7
Leaf nitrogen (% d. wt)Sun +8−16−29
Shade  −17
Specific leaf area (mm2 g−1)Sun −4 −6−23
Shade −16−15

Leaf morphology

Figure 2(a–c) shows that SLA (mm2 g−1) did not differ significantly with pCO2 treatment in any species in year 1. In years 2 and 3, however, SLA was greater in ambient than elevated pCO2 in all three species in sun leaves. This was largely the result of an increase in SLA with successive years in ambient pCO2. Spongy mesophyll and palisade cell thicknesses were significantly increased in both year 1 (P ≤ 0.01) and year 2 (P  0.05) in sun leaves exposed to elevated pCO2, suggesting both the pattern and the means for the trend towards increased leaf thickness with time (Table 1). During single leaf development of P. nigra (year 3), SLA showed significant differences between pCO2 treatments on days 5 and 10, i.e. during most of the maturation period of the leaf (Fig. 2c inset). In contrast, very similar values were reached at the young (day 0) and mature (day 20) stages. Final values for mature leaves were 13 145 and 13 073 mm2 g−1 for elevated and ambient pCO2, respectively. It is interesting that values increased from days 0 and 5 corresponding with the phase of leaf expansion, whereas they decreased from day 5 onwards because of the predominance of accumulation of photosynthates in the leaf.

Figure 2.

Effects of ambient or elevated pCO2 (550 µmol mol−1) on: (a–c) specific leaf area (SLA, mm2 g−1); (d–f) foliar nitrogen content (% d. wt) at the POPFACE experiment over 3 yr in three species, (a,d) Populus × euramericana; (b,e) P. alba; (c,f) P. nigra[insets, effects on (c) SLA, (f) leaf N content during leaf development of P. nigra in the third growing season at the POPFACE experiment]. Each value represents average (± SE) of 15 measurements per treatment. Closed symbols, elevated pCO2; open symbols, control (ambient pCO2). Summary results of two-way ANOVA are given: ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001.

Leaf nitrogen content

The general pattern of change in SLA was reflected in the pattern of change in foliar N content. Leaf N content as a percentage of d. wt (Nmass) varied significantly with pCO2 treatment in all 3 yr. However, in year 1 leaf Nmass was increased in P. × euramericana and P. nigra in elevated pCO2 (Fig. 2d–f), whereas thereafter leaf Nmass was significantly reduced under elevated pCO2 in all three species. When N content was calculated on a leaf-area basis (Narea) no consistent trend of either increase or decrease was observed in any species, and differences between responses to elevated and ambient pCO2 were insignificant in all years (data not shown).

During single leaf development of P. nigra (year 3) a decrease in leaf Nmass with leaf development was observed, from values of 4.86 and 5.15% for young leaves to values of 2.89 and 3.33% for mature leaves from elevated and ambient pCO2, respectively (Fig. 2f inset). No significant differences were observed between treatments for C content as % d. wt (Cmass, data not shown). The values ranged from 44 to 46% throughout the maturation period, showing a decrease from day 0 to day 5 and an increase from day 5 to days 10 and 20. Given the decrease in Nmass and the small but insignificant increase in Cmass in elevated pCO2, the C : N ratio was significantly influenced by elevated pCO2, particularly during the later developmental stages. No significant differences were observed between treatments for Narea or Carea during P. nigra leaf development (Fig. 3). Narea was reduced and Carea increased in elevated pCO2 during the majority of leaf development, but these differences were small and insignificant.

Figure 3.

(a) Leaf nitrogen and (b) carbon contents of Populus nigra leaves, on a leaf-area basis, exposed to ambient (open symbols) and elevated pCO2 (closed symbols, 550 µmol mol−1) during 20 d of leaf development. Each value represents average (± SE) of nine measurements per treatment. No statistically significant differences were found at any data point.


Table 2 shows the percentage change in mature epidermal cell areas (µm2) in leaves exposed to elevated pCO2 in sun and shade for each species in each year. Elevated pCO2 stimulated sun leaf epidermal cell areas of P. × euramericana in the first growing season but not thereafter at advanced LPI, and mature cell areas were larger in ambient rather than elevated pCO2 (although not significantly so) by the third growing season. Epidermal cells of young P. × euramericana leaves remained significantly larger and more extensible under elevated pCO2. In P. nigra, epidermal cells were larger under pCO2 enrichment in year 1 although this difference was not statistically significant in sun leaves by year 2. The epidermal cells of P. alba were smaller when exposed to elevated pCO2 in both sun and shade leaves in all 3 yr. Extensibility analysis revealed that, in years 1 and 2, epidermal cell walls were more extensible under elevated pCO2 in P. alba although P. alba cells were not larger.

In sun leaves epidermal cell numbers (105) were significantly stimulated by elevated pCO2 in P. × euramericana in all 3 yr (Table 2). Numbers of epidermal cells were very similar in both pCO2 in P. nigra in year 1 when leaf area was stimulated, but significantly greater in ambient pCO2 in years 2 and 3 when there was no significant increase in leaf area. In P. alba sun leaves the number of epidermal cells was increased significantly by elevated pCO2 in all years. Populus alba shade leaves in year 2 were, again, unusual in that epidermal cell numbers were greater in ambient than elevated pCO2. The number of epidermal cells in P. × euramericana increased with each successive growing season while the size of those cells varied little from year to year.

Leaf production and longevity

Plastochron index (PI) in sun and shade leaves increased with time for all three species, but the effects of pCO2 enrichment varied depending on species and shade (Fig. 4). In years 2 and 3 the PI was increased under elevated pCO2 in sun leaves of P. alba, suggesting that the initiation of leaves was slowed by the elevated pCO2. In direct contrast, for P. nigra a reduction in PI was observed, while there was no significant effect of elevated pCO2 for sun leaves in P. × euramericana. Contrasting responses were also documented for shade leaves but for both P. alba and P. × euramericana, where a significant effect of elevated pCO2 on PI was seen, a reduction was observed. This suggests that leaves in the shade subjected to elevated pCO2 were initiated more quickly.

Figure 4.

Plastochron index of leaves exposed to either ambient (open symbols) or elevated pCO2 (closed symbols, 550 µmol mol−1) in sun (circle) and shade (square) leaves in (a) Populus ×euramericana; (b) P. alba; (c) P. nigra. Each value represents average of 15 measurements per treatment. Summary results of one-way ANOVA at the end of growth are shown for each species: ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001.

Leaf longevity was marginally increased by the elevated CO2 treatment in P. nigra with 1 d delay in measured 50% leaf fall at the branch apex and 10% more leaves remaining attached during the winter (Table 3). In year 3 leaf longevity was increased under elevated pCO2 in both P. alba and P. × euramericana with significantly fewer leaves falling during September and more during November. Fifty per cent leaf fall at the branch apex was delayed by 17 and 13 d, respectively, and the number of leaves remaining attached during the winter period was 10% higher in elevated than in ambient pCO2 in P. alba.

Table 3.  Estimated leaf fall during the growing season in three Populus species exposed to elevated pCO2 (550 µmol mol−1) and ambient pCO2 (≈ 360 µmol mol−1) during the POPFACE experiment in year 3 (entirely closed canopy)
LeavesP. × euramericanaP. albaP. nigra
Elevated [CO2]Ambient [CO2]Elevated [CO2]Ambient [CO2]Elevated [CO2]Ambient [CO2]
 Percentage leaf fall
May 2 3 1 2 2 1
June1415 6 5 4 5
July2524121311 9
September1619 9141113
November 6 122123841
December 0 0 7 9 0 0
May 9 710 7 2 0
June4431 717 5 9
September 4 9    
Attached leaves
over winter (%) 0 037275545
Number of days delay in 50% leaf fall at apex1317 1
 Leaf longevity (d)
Leaf no.      
ln-10  156134  
ln-15113 92165149151152
ln-40    130138
ln-45    129130
ln-50    123121
ln-55    122112


After three growing seasons of exposure to elevated pCO2 in POPFACE, an intensively managed forest plantation of poplar trees, there remained significant effects of elevated pCO2 on leaf development, biochemistry and growth. Some of these changes occurred consistently in all three species, others were highly species-specific.

Leaf growth

Where leaves remained unshaded, leaf length (mm) increased year on year in all three species whether or not pCO2 had been enriched (Fig. 1). This could reflect greater irrigation water application; however this seemed unlikely because leaf area was greater in year 2 than in year 3. The stimulation of leaf growth by elevated pCO2 was diminished over three growing seasons so that the significant increases in leaf area and length observed in all species in the first season were no longer significant by the second in P. nigra and P. alba. Populus × euramericana, the species with the largest leaves, continued to display significantly enhanced sun-leaf size through to the end of year 3. The effects of elevated pCO2 on shade leaves were less clear, with no statistically significant stimulation of leaf area recorded in either year 2 or year 3 (no true shade leaves existed in year 1). For sun leaves there were few effects on developmental leaf age expressed as LPI, and it appeared that stimulation of individual leaf size was an important mechanism for leaf-area development. In poplars, leaves of the main stem are reported to contribute to the height growth of the tree (Scarascia-Mugnozza et al., 1999), and in year 3 height growth was still stimulated in elevated pCO2 (Calfapietra et al., 2003). In contrast to leaf-area development, the production of shade leaves was influenced and suggested that leaf production was stimulated by elevated pCO2 as PI was reduced in P. × euramericana and P. alba. Bosac et al. (1995) considered that both poplar leaf production and expansion were sensitive to C supply, depending on species.

The relationship of these leaf-level measurements to those of the canopy LAI can also be determined. In general, LAI in this experiment was stimulated in an open canopy (Gielen et al., 2001) and increases in leaf size almost certainly contributed to this effect. As the canopy closed, LAI was no longer significantly stimulated and any enhancement depended on species and time of season (Gielen et al., 2003).

Leaf morphology

The most consistent difference across species in leaves grown under elevated pCO2 with the progression through the growing seasons was in SLA (mm2 g−1). Both decreases and increases in SLA have been reported in the literature for plants exposed to elevated pCO2 (Gardner et al., 1995). The mesophyll tissues of sun leaves were thicker in enriched pCO2 in years 1 and 2 and, by the third growing season, whole leaves were probably thicker as indicated by SLA. This suggested that accumulation of photosynthetic machinery was being progressively increased in elevated pCO2, possibly with tree carbohydrate status acting as a signal from year to year (Bernacchi et al., 2003). Above-ground harvested biomass increased significantly in all three species in both year 2 and year 3 (Calfapietra et al., 2003), as did photosynthetic assimilation per unit leaf area (Bernacchi et al., 2003).

During single leaf development of P. nigra, SLA was reduced by elevated pCO2 during intermediate developmental stages but not when leaves became mature after 20 d, suggesting a different speed in the expansion of the leaves and/or in photosynthate accumulation between treatments. Elevated pCO2 consistently increased LER (mm d−1) in P. nigra and there was no significant difference in carbohydrate (starch) accumulation in the leaves at this time (P. Davey, personal communication). The effect of elevated pCO2 seemed to be an acceleration of mesophyll expansion during growth. From this point of view, the similarity in mature epidermal cell size and the contrasting expansion in the cells of mesophyll tissues after exposure to elevated pCO2 is puzzling. However, Donnelly et al. (1999) confirmed that, in Arabidopsis, cell production and complementary patterns of cell expansion were organized differently within different tissues. Differential expansion by tissue is explained by the prolonged duration of cell cycling in the mesophyll tissue compared with the earlier cessation of cell cycling in the epidermis, and this experiment has already confirmed that cell expansion and cycling within tissues are coupled in this way in P. × euramericana (Taylor et al., 2003). We hypothesize that the effect of elevated pCO2 during development of P. nigra leaves was to increase the rate of accumulation of mesophyll cells during the intermediate stages of development and to decrease cell production earlier in elevated than in ambient pCO2.

Leaf nitrogen

Foliar N content (% d. wt) also changed in both elevated and ambient pCO2 over the 3 yr as the canopy closed. An initial and unusual increase in Nmass in elevated pCO2 in P. ×euramericana and P. nigra was followed by significantly lower leaf Nmass under elevated compared with ambient pCO2 for all three species. Leaf Nmass of trees grown under elevated pCO2 varies considerably (Norby et al., 1999) but declines on average. There was no significant response of Narea to elevated pCO2 in any year. In all three species Nmass increased over 3 yr in ambient pCO2, although no N fertilizer was added to the soil. Reduced Nmass in elevated pCO2 often indicates a comparatively lower Rubisco content in leaves (Stitt & Krapp, 1999), a symptom of acclimation (Makino et al., 2000). However, measurements of C assimilation revealed no significant photosynthetic acclimation in any of the three species, although, interestingly, there was some indication of a reduction in maximum carboxylation capacity (Bernacchi et al., 2003). Drake et al. (1997) noted that at elevated pCO2 a leaf may lose a substantial content of Rubisco with no effect on assimilation rate, and previous studies on trees planted in the ground have shown consistently enhanced photosynthetic C assimilation in response to elevated pCO2 despite regular changes in leaf biochemistry (Norby et al., 1999). Here we report a similar finding, coupled to the maintenance of enhanced biomass productivity in elevated pCO2 (Calfapietra et al., 2003).

The single-leaf experiment in P. nigra provided the opportunity to analyse further the decline in Nmass. During P. nigra leaf development, differences in Nmass with elevated pCO2 had already begun at the first developmental stage measured, and increased throughout the maturation period. Differences in Nmass in response to elevated pCO2 were evident in young leaves and became more pronounced in 5-d-old leaves, and were similar in magnitude to those reported on a yearly basis for all three species. In several experiments decrease in leaf N under elevated pCO2 has been explained by increased structural or nonstructural C content in CO2-enriched leaves. In some cases, however, N content decreased even when correction was made for starch and soluble sugar accumulation (Stitt, 1991), suggesting that ‘C dilution’ did not account entirely for this effect. This agrees with our results, where neither the Carea nor the Cmass of P. nigra leaves was significantly increased in response to elevated pCO2, and there was no significant accumulation of carbohydrates. On a leaf-area basis, however, Narea was not reduced under elevated pCO2, which strongly suggests that the decrease in Nmass is related to denser leaves as also indicated by measurements of leaf thickness (in previous years) and SLA. This has important implications for continuing photosynthetic efficiency, as acclimation may be the result feedback-driven downregulation of photosynthetic capacity or limitations to internal CO2 diffusion following an increased accumulation of leaf carbohydrate in elevated pCO2 (Sage, 1994; Moore et al., 1999). Here, the foliar N findings were indicative of improved N-use efficiency in leaves exposed to elevated pCO2 over the long term. As Makino et al. (2000) have pointed out, increased N-use efficiency under elevated pCO2 does not necessarily lead to a reallocation of N to limiting processes or, ultimately, greater biomass production. In long-term exposure to elevated pCO2, low N availability has limited tree responses so that total leaf area and stem biomass were not stimulated in low-N conditions (Curtis et al., 2000). In elevated pCO2 and high-N soil at POPFACE, however, photosynthetic N-use efficiency was improved.

Leaf longevity

An important finding of this research is the increase in leaf longevity apparent from assessments of leaf retention and leaf fall throughout the season. Few studies have considered the effects of elevated pCO2 on leaf longevity, and conclusions to date are contradictory. For example, Craine & Reich (2001) showed that longevity in response to elevated pCO2 appeared to be increased in C3 but not C4 species. For forest trees in a FACE experiment Thomas & Herrick (2003) reported no effect, while Sigurdsson (2001) showed that elevated pCO2 resulted in accelerated bud set irrespective of N treatment, although the effect was most pronounced with a low N supply. In this fast-growing forest plantation clear and significant effects were apparent for P. alba and P. ×euramericana. For poplar it is known that canopy longevity is an important indicator of biomass productivity (Rae et al., 2003). Leaf senescence processes are complex and have rarely been investigated for closed-canopy forests in elevated pCO2 but, for temperate trees, senescence is linked to the developmental events resulting in bud set and dormancy, also reported as delayed by elevated pCO2 following exposure of Populus tremuloides in the Aspen FACE experiment (Karnosky et al., 2003). Any effects of elevated pCO2 on tree phenology are important as they not only help determine productivity, but also provide the synchronization between tree and environment, with enhanced sensitivity to low temperature likely if bud set is delayed or altered by elevated pCO2. These responses remain to be further elucidated in the long-term ongoing forest FACE experiment, but our data are in accord with those of Reich et al. (1992, 1999) who showed that leaf quality and longevity relationships can be generalized among and within species, across distantly related taxa and across diverse biomes where SLA (mm2 g−1) and Nmass decrease with increasing leaf life span, suggesting that the long-term changes in leaf morphology and N status found here are important aspects of increased biomass potential under elevated pCO2.

Cellular growth

Epidermal cell size was increased by elevated pCO2 in line with stimulated leaf expansion in P. × euramericana and P. nigra in year 1. Thereafter, although phases were uncovered of increased cell expansion relating to increased leaf growth in young P. × euramericana leaves (Taylor et al., 2003), increased cell number was the more important contributor to any stimulation of leaf expansion under elevated pCO2 and to patterns of leaf growth, regardless of pCO2 in all species. This evidence supports previous work showing that cell production is able to contribute to plant growth (Beemster & Baskin, 1998; Donnelly et al., 1999) and changing leaf shape (Wang et al., 2000; Wyrzykowska et al., 2002). Cell production responded to pCO2 over the long term, and the basis for an adaptation in leaf morphology was suggested by increases in epidermal cell numbers in older trees and leaves.


In the realistic conditions provided by the FACE technology and over long-term exposure, an initial stimulation of individual leaf-area development by elevated pCO2 was reduced in magnitude as canopy shading and competition increased, but there were consistent, species-specific increases in leaf-area development where leaves remained in full sun. After an initial stimulation of cell-expansion processes, increases in leaf area under elevated pCO2 were provided by increased cell production and leaf thickness. Over the longer term, basic leaf anatomy and chemistry were changed in a consistent way by exposure to enriched pCO2.

Leaf quality was altered consistently in all three species such that N concentration (% d. wt) declined as the experiment progressed with an average decline of 22 and 13% for sun and shade leaves, respectively, in elevated pCO2. Similarly, SLA declined in elevated pCO2 in all three species, with an average decline of 29 and 5% for sun and shade leaves, respectively. Leaf longevity was increased in elevated pCO2, an effect that may be related to altered leaf quality. The findings for leaf development, quality and longevity here suggest that a continued enhancement of biomass productivity will be likely when intensively managed forest plantations are exposed to a future rising concentration of atmospheric CO2.


We thank Dr A. Page (Biomedical Imaging Unit, University of Southampton) for help in the preparation of leaf sections and Dr P. Davey (University of Essex) for providing us with unpublished data. This work was supported by the EC through its Environment R&D programme within the Fourth Framework as a research contract ENV4-CT97-0657 (POPFACE) coordinated by the University of Viterbo and by the Natural Environment Research Council and Department of Environment, Food and Rural Affairs (grant nos GR9/04077 and NFO410 to G.T.). P.J.T. was awarded a research studentship from the Natural Environment Research Council (no. GT04/99/TS250). This study also contributes to the Global Change and Terrestrial Ecosystems elevated CO2 consortium of the International Geosphere–Biosphere Programme.