Trade-offs in low-light CO2 exchange: a component of variation in shade tolerance among cold temperate tree seedlings



1. Does enhanced whole-plant CO2 exchange in moderately low to high light occur at the cost of greater CO2 loss rates at very-low light levels? We examined this question for first-year seedlings of intolerant Populus tremuloides and Betula papyrifera, intermediate Betula alleghaniensis, and tolerant Ostrya virginiana and Acer saccharum grown in moderately low (7·3% of open-sky) and low (2·8%) light. We predicted that, compared with shade-tolerant species, intolerant species would have characteristics leading to greater whole-plant CO2 exchange rates in moderately low to high light levels, and to higher CO2 loss rates at very-low light levels.

2. Compared with shade-tolerant A. saccharum, less-tolerant species grown in both light treatments had greater mass-based photosynthetic rates, leaf, stem and root respiration rates, leaf mass:plant mass ratios and leaf area:leaf mass ratios, and similar whole-plant light compensation points and leaf-based quantum yields.

3. Whole-plant CO2 exchange responses to light (0·3–600 µmol quanta m−2 s−1) indicated that intolerant species had more positive CO2 exchange rates at all but very-low light (< 15 µmol quanta m−2 s−1). In contrast, although tolerant A. saccharum had a net CO2 exchange disadvantage at light > 15 µmol quanta m−2 s−1, its lower respiration resulted in lower CO2 losses than other species at light < 15 µmol quanta m−2 s−1.

4. Growth scaled closely with whole-plant CO2 exchange characteristics and especially with integrated whole-plant photosynthesis (i.e. leaf mass ratio × in situ leaf photosynthesis). In contrast, growth scaled poorly with leaf-level quantum yield, light compensation point, and light-saturated photosynthetic rate.

5. Collectively these patterns indicated that: (a) no species was able to both minimize CO2 loss at very-low light (i.e. < 15 µmol quanta m−2 s−1) and maximize CO2 gain at higher light (i.e. > 15 µmol quanta m−2 s−1), because whole-plant respiration rates were positively associated with whole-plant photosynthesis at higher light; (b) shade-intolerant species possess traits that maximize whole-plant CO2 exchange (and thus growth) in moderately low to high light levels, but these traits may lead to long-term growth and survival disadvantages in very-low light (< 2·8%) owing, in part, to high respiration. In contrast, shade-tolerant species may minimize CO2 losses in very-low light at the expense of maximizing CO2 gain potential at higher light levels, but to the possible benefit of long-term survival in low light.


A general hypothesis underlying many studies of shade adaptation in plants is that: shade-adapted species should express morphological and physiological traits that enhance energy capture (i.e. growth) in low light relative to sun-adapted species (Björkmann 1981; Givnish 1988). However, despite the intuitive appeal of the enhanced low-light growth hypothesis, tests to date either do not support it or are insufficient. For example: (1) the few tests of the enhanced low-light growth hypothesis that were conducted at low light levels (e.g. < 4% of open sky light) have not consistently supported it and growth comparisons conducted at slightly higher, moderately low light levels (e.g. 4–12% light) have consistently shown the opposite trend, i.e. greater growth rate for intolerant than tolerant species (see reviews by Venenklaas & Poorter 1998; Walters & Reich 1999a); (2) although there are many studies of leaf-level CO2 exchange and morphology in low light (see reviews by Givnish 1988; Chazdon et al. 1996), few studies have examined the relationships of morphological and CO2 exchange characteristics to low-light growth in a whole-plant context, and how these characteristics vary with species differences in shade tolerance (e.g. Loach 1967, 1970; Walters, Kruger & Reich 1993a,b; Kitajima 1994; Reich, Walters et al. 1998).

One possible reason why shade-tolerant species have not shown clear growth advantages over intolerant species in low light is because traits enhancing growth in all but the lowest light levels may occur at the cost of high respiration losses at very-low light levels. High respiration could lead to negative growth rates and death at the very-low light levels that typify closed forest understories (often < 2% of open-sky light, Canham et al. 1990), thus selection in these habitats may be for low respiration rates and not for high low-light growth rates. Although an enhanced growth (or photosynthesis) vs low respiration trade-off has long been hypothesized and explored at the leaf level for sun and shade phenotypes and tolerant and intolerant species (see reviews by Bazzaz 1979; Björkman 1981; Givnish 1988), it has not been adequately investigated on a whole-plant basis. That is the goal of this study.

Specifically we ask: (1) which leaf and whole-plant CO2 exchange and morphological characteristics are associated with variation in shade tolerance and growth rates for cold-temperate, broad-leaved species grown in moderately low and low light; (2) what are the inter-relationships among CO2 exchange, morphological and growth characteristics; (3) do these inter-relationships suggest a trade-off between enhancing growth at moderately low to high light levels and minimizing respiration at very-low light levels? We investigated these questions for first-year seedlings of shade-intolerant Populus tremuloides and Betula papyrifera, mid-tolerant Betula alleghaniensis, and tolerant Ostrya virginiana and Acer saccharum grown in two low-light environments (2·8 and 7·3% of open-sky light). Data are lacking for O. virginiana and P. tremuloides in 2·8% light owing to low germination success and complete mortality, respectively. The characteristics we quantified on leaf and/or whole-plant bases included: relative growth rate; initial slope of the net photosynthesis vs light relationship (i.e. apparent quantum yield); light level at which net photosynthesis equals zero (i.e. light compensation point); integrated in situ photosynthesis; fraction of total plant mass in leaves (i.e. leaf mass ratio, leaf area ratio); leaf, stem and root respiration rates.

Materials and methods

Study species, growing conditions and experimental design

Populus tremuloides, B. papyrifera, B. alleghaniensis, O. virginiana and A. saccharum are all common, deciduous, broad-leaved trees of the cold-temperate forests of central and eastern North America. They vary widely in reported shade tolerance (Baker 1949; Burns & Honkala 1990), low-light growth (Pacala et al. 1994; Walters & Reich 1996, 1999; Reich, Tjoelker et al. 1998), survival (Lorimer 1981; Kobe et al. 1995; Walters & Reich 1996, 1999) and seed size (Hewitt 1998). Seeds used in this experiment were from wild central Minnesota and northern Wisconsin sources (~ latitude 47°N).

Except for O. virginiana seedlings, the seedlings and growth conditions used in this experiment were a subset of those used for another experiment that investigated the effects of seed size and growth rate on survival for 10 species grown at four light and three nitrogen levels (Walters & Reich 2000). Ostrya virginiana germinants were added to the experimental treatments used by Walters & Reich (2000) but germinant numbers were too low to be used for survival and growth determinations. Seed mass, low-light mortality, growth and morphological data for all of our study species are included in Walters & Reich (1996, 2000).

Seeds were germinated in sand or wet paper towels beneath 50% neutral-density shade cloth in a temperature-controlled greenhouse at the University of Minnesota, St Paul, MN, USA, in mid-February 1994. Single germinants (< 4 days following radicle emergence) were transferred to 2 cm top width × 15 cm depth plastic pots filled with washed silica sand and placed into (1·6 m × 1·6 m × 0·8 m high) shade houses in the same greenhouse. Shade houses were covered with either one or two layers of neutral-density woven polypropylene shade cloth with different light transmission characteristics (Carlin Co., Milwaukee, WI, USA). Because the experiment started in mid-February, we extended day length with sodium halide lamps positioned above each shade house. Lamps were on from 15.00 to 22.00 h local time for the duration of the experiment. We monitored photosynthetic photon flux density (PPFD) nearly continuously over the experiment with calibrated galium arsenide photodiodes (G1118, Hamamatsu Corp., NJ, USA) or quantum sensors (LI-COR, Inc., Lincoln, NE, USA). Mean total daily PPFD outside the greenhouse was 37·75 mol m−2 day−1. Including light from the sodium halide lamps, the two shade cloth treatments resulted in (mean of the three houses over the experiment ± 1 SE) 2·77 ± 0·15 and 7·34 ± 0·48% of open-sky PPFD outside the greenhouse. These levels were similar to those found in closed-forest understories (1–3%) and small tree-fall gaps (5–10%) in the cold-temperate forests of North America (Canham et al. 1990).

Trays with seedlings were rotated within shade houses several times during the experiment (mid-February to late June). Except for O. virginiana, 18–20 seedlings of each species were allocated to each block (3) × light treatment (2). Owing to low germination success, only two O. virginiana germinants were allocated to each block and then only in 7·3% light. For P. tremuloides we have no data for 2·8% light because complete mortality occurred prior to our measurements (Walters & Reich 2000). Nutrients were supplied daily as a 1/2 strength modified Hoagland's solution with 3·4 mm N (NH4+ and NO3 in equal amounts). Pots were flushed with distilled water twice a week to prevent the accumulation of nutrient salts.


Leaf net CO2 exchange was measured 21–29 June from 09.00 to 16.00 h local time with a CO2 analyser operating as a closed system in the absolute mode (LI-6200, LI-COR, Inc., Lincoln, NE, USA). During measurements, greenhouse air temperature was 27·4°± 1·7° (mean ± SD), relative humidity was 54·9 ± 7·0% and ambient CO2 concentration was 355·3 ± 8·6 µl l−1. Chamber CO2 concentrations and relative humidities during photosynthesis measurements were nearly identical to ambient conditions. Of the three or more seedlings used in each light × species treatment, approximately half were moved each day to a side-walled, open-topped enclosure in an adjacent greenhouse. Photosynthetic photon flux density (PPFD) from natural sky light was varied from 0·3 to 600 µmol quanta m−2 s−1 at leaf surface by covering the top of the enclosure with one to several layers of neutral-density shade cloth. Over this range, PPFD was both increased and decreased between measurements over the day. After enclosing a leaf in the chamber, CO2 exchange measurements were taken after PPFD was stable for more than 30 s. One typical, fully expanded leaf was used for photosynthesis measurements per plant. Following the first such measurement, the area of the leaf in the chamber was marked so that the same leaf, and leaf area could be used for measurements later in the day and later in the week. Each day following measurements, plants were returned to their light treatments.

Prior to sunrise on 28–30 June plants were moved to a dark closet in a laboratory close to the greenhouse and separated into leaves, stems and roots immediately prior to measuring respiration. Dark respiration was measured at 21·0 °C ± 0·5 SD and 365 µl l−1 CO2 with two portable, open gas-exchange systems (LCA-3, Analytical Development Corporation, Hoddesdon, UK). To eliminate possible cross sensitivity of the IRGA to water vapour, water was scrubbed from the line returning from the chamber with a magnesium perchlorate column. Complete root, stem and leaf systems for single plants were placed in the chamber and given c.5–10 min for readings to stabilize before measurements were recorded. Following respiration measurements, total leaf area and leaf area for photosynthesis was measured with an image analysis system (Decagon Devices, Pullman, WA, USA) and plant organs were dried in a forced-air oven at 70 °C prior to mass measurements.

At least six seedlings per species × light treatment not used for CO2-exchange measurements were harvested mid-April (8–9 weeks since germination) and again late June (19–20 weeks since germination) and dried at 70 °C. These plants were used to estimate mean relative growth rate between harvest dates. Leaf and plant morphology (e.g. LMR) values of plants harvested for relative growth rate determinations were similar to those for plants used for CO2 exchange and are not reported here (see Walters & Reich 2000 for additional data on relative growth rate and morphology at other seedling ages).

Parameter calculation and analysis

Individual plants were considered experimental units (≥ 3 per species × light treatment), thus parameters were determined for each plant measured and these values were averaged for each light × species treatment. From mass and leaf area data we calculated: leaf mass ratio (LMR, g leaves/g plant); stem mass ratio (g stems/g plant); root mass ratio (g roots/g plant); specific leaf area (SLA, cm2 fresh leaf area/g dry leaf mass); leaf area ratio (LAR, cm2 fresh leaf area/g plant). Average relative growth rate (mg g−1 day−1) was estimated for harvested seedlings as: {ln[mass harvest (a + 1)] − ln(mass harvest a)}/days (Evans 1972).

Leaf-level net CO2 exchange vs PPFD relationships for each plant (eight to 37 measurements per plant) were fitted using the non-linear platform of JMP (SAS Institute, Cary, NC, USA). For these fittings, dark respiration data were not included. In general, Hanson et al.'s (1988) non-linear model resulted in similar or lower residual sums of squares error compared to other non-linear models (data not shown). The resulting fits (P < 0·0001 for all individuals) from this model were used to estimate light compensation point for leaf CO2 exchange and maximum light-saturated CO2 exchange rate (Amax). For presentation, we pooled plants within light × species treatments prior to fitting the model. Quantum yield (QY) was determined for each plant as the slope of the least-squares linear regression for area-based net CO2 exchange vs PPFD relationships between 0·3 and 40 µmol quanta m−2 s−1 (R2 = 0·77–0·997, average R2 = 0·98, n = 5–12).

In addition to presenting leaf (RL), stem (RS) and root (RR) respiration (per unit mass), we estimated net average 24 h whole-plant respiration rates as: RW = [(LMR × RL × 9/24) + (stem mass ratio × RS) + (root mass ratio × RR)]. For this calculation we did not include RL during the light period (c. 15 h per day) and we assumed that the respiration rates measured for stems and roots during the dark period also occurred during the light period. For all CO2 exchange and morphological parameters described above we tested for significant (P < 0·05) differences among species means within light treatments with a Tukey–Kramer Honestly Significant Difference test (JMP, SAS Institute, Cary, NC, USA).

Leaf-level net CO2 exchange vs PPFD relationships were modified to whole-plant net CO2 exchange vs PPFD curves by incorporating respiration rates, photoperiod length and allocation to leaves, roots and stems as: AMASS × LMR × 15/24 − RW. AMASS for the canopy was extrapolated from single leaf measurements, which was justified given the small seedlings simple canopies. Whole-plant quantum yield was estimated from whole-plant net CO2 exchange vs PPFD curves as the average slope of the function from 0 to 40 µmol quanta m−2 s−1 PPFD. Whole-plant light-compensation points were estimated from whole-plant net CO2 exchange vs PPFD curves as the PPFD at which net CO2 exchange equals zero. We calculated integrated net in situ CO2 exchange for a 13 day period (10–22 June) using modelled whole-plant CO2 exchange responses to PPFD (explained above) and measurements of PPFD recorded continuously in the experimental treatments. Integrated net in situ CO2 exchange was calculated as: daytime leaf CO2 exchange per leaf mass (∫AMASS); daytime leaf CO2 exchange per plant mass (∫AMASS × LMR); 24 h whole-plant CO2 exchange (∫AMASS × LMR × 15/24 − RW).


Tissue-level CO2 exchange and morphology: differences among species

Within each low-light environment, mass-based, light-saturated leaf CO2 exchange rates (AmaxMASS) varied two- to threefold among species and were highest for the most shade-intolerant species (i.e. P. tremuloides in 7·3% light and B. papyrifera in 2·8% light). In contrast, area-based rates (AmaxAREA) varied less than 50% among species and showed no trend with shade tolerance classifications (Figs 1 and 2; Table 1). Differences in the patterns of AmaxAREA and AmaxMASS among species were owing to differences in specific leaf area (SLA) which was two- to threefold higher for the least shade-tolerant than the most shade-tolerant (i.e. A. saccharum) species. Patterns of mass-based leaf dark respiration (RL) among species and light environments strongly paralleled those for AmaxMASS (AmaxMASS = 33·86 + 14·86 × RL, R2 = 0·90, P = 0·0004, n = 8), such that RL was threefold greater for shade-intolerant P. tremuloides than tolerant A. saccharum. Like RL, stem (RS) and root (RR) respiration were also generally highest for shade-intolerant P. tremuloides and lowest for shade-tolerant A. saccharum (Table 1).

Figure 1.

Net leaf CO2 exchange per unit leaf area responses to photosynthetic photon flux density (PPFD). Hatched lines and filled squares are for leaves from seedlings grown in 2·8% light. Empty circles and solid lines are for leaves from seedlings grown in 7·3% light. Each symbol is a single CO2 exchange measurement. Lines are best fits (P < 0·0001 in all cases) using Hanson et al.'s (1988) non-linear model. All plants and individual measurements within a treatment were pooled prior to curve fitting.

Figure 2.

Net CO2 exchange per unit leaf area (a) and mass (b) responses to photosynthetic photon flux density (PPFD). The curves in Fig. 2a are the same as in Fig. 1a and are presented here together to facilitate visual comparisons.

Table 1.  Mean values (± SE) of leaf morphology and leaf, stem and root CO2 exchange parameters for each species × light treatment (n ≥ 3 seedlings for all values). Letters indicate significant differences (P < 0·05, Tukey–Kramer HSD) among means for species within a light treatment. Parameter abbreviations and units are: AmaxM and AmaxA, light-saturated leaf CO2 exchange rates (nmol CO2 g−1 s−1 and µmol CO2 m−2 s−1, respectively); SLA, specific leaf area [cm2 leaf (g leaf)−1]; LCP, leaf light compensation point (µmol quanta m−2 s−1); QY, apparent quantum yield (µmol CO2 (mmol quanta)−1), QY × SLA mass-based apparent quantum yield (µmol CO2 (µmol quanta)−1 g−1). RL, RS and RR are respiration rates (nmol CO2 g−1 s−1) for leaves, stems and roots, respectively
P. tremuloides290 ± 483·9 ± 0·2662 ± 717·9 ± 1·134 ± 22·6 ± 0·219·1 ± 3·110·7 ± 2·418·6 ± 3·0
B. papyrifera141 ± 123·1 ± 0·3497 ± 245·3 ± 0·634 ± 21·7 ± 0·18·5 ± 0·86·8 ± 0·518·7 ± 1·3
B. alleghaniensis153 + 183·1 ± 0·3535 ± 224·3 ± 0·133 ± 21·5 ± 0·17·9 ± 0·28·0 ± 0·914·8 ± 0·9
O. virginiana156 ± 143·0 ± 0·3552 ± 64·7 ± 0·734 ± 11·9 ± 0·18·8 ± 1·45·9 ± 1·215·2 ± 1·8
A. saccharum144 ± 104·3 ± 0·3331 ± 95·1 ± 0·438 ± 11·3 ± 0·15·8 ± 0·41·8 ± 0·44·8 ± 0·2
B. papyrifera306 ± 193·3 ± 0·1913 ± 275·2 ± 1·031 ± 33·0 ± 0·415·9 ± 1·611·0 ± 1·525·5 ± 2·4
B. alleghaniensis228 ± 83·0 ± 0·1729 ± 275·1 ± 0·537 ± 22·8 ± 0·311·9 ± 1·711·3 ± 1·023·2 ± 2·4
A. saccharum113 + 162·7 ± 0·3387 ± 323·1 ± 1·238 ± 41·5 ± 0·26·9 ± 2·02·4 ± 0·35·4 ± 1·7

Increased apparent quantum yield (QY), and decreased light compensation point (LCP) can both result in increased leaf CO2 exchange at low PPFD. Compared to intolerant species, shade-tolerant species generally had greater QY and lower LCP, but trends were weak (Table 1). However, because SLA was generally much greater for the least-than the most-tolerant species and because SLA × QY equals the rate that AMASS increases with increasing light at low light levels (i.e. mass-based QY), greater SLA × QY for the most intolerant species (Table 1) indicate that these species had greater AMASS than more tolerant species at PPFD above LCP. Despite generally large differences in AmaxMASS, SLA and R between the most- (A. saccharum) and least-tolerant (P. tremuloides) species, the three Betulaceae species, although varying greatly in shade tolerance, had similar tissue-level CO2 exchange and leaf morphology (Table 1, Figs 1 and 2).

Tissue-level CO2 exchange and morphology: differences among light environments

In 2·8 compared to 7·3% light, A. saccharum had lower AmaxAREA which occurred via modest increases in SLA and decreases in AmaxMASS, whilst for B. alleghaniensis and B. papyrifera, AmaxAREA was similar despite much greater SLA, because AmaxMASS also increased appreciably (Table 1, Figs 1 and 2). Neither light compensation point nor QY were affected by light treatments. However, AMASS at low PPFD (as indicated by SLA × QY) was greater for all species in 2·8 than 7·3% light with B. papyrifera showing the largest difference and A. saccharum the least difference between light treatments. RL, RS and RR were greater in 2·8 than 7·3% light for B. papyrifera and B. alleghaniensis but were only marginally greater for A. saccharum.

Whole-plant CO2 exchange and morphology: differences among species

In 7·3% light, shade-intolerant P. tremuloides and B. papyrifera had twice the leaf mass ratio (LMR) and half the root mass ratio of tolerant O. virginiana and A. saccharum. Because intolerant species had twofold greater SLA and twofold greater LMR, they had fourfold greater leaf area ratio (LAR) than tolerant species. Patterns among the three species grown in 2·8% light were similar to those in 7·3% light but differences among species common to both treatments were even greater. For example, LAR was fivefold greater for B. papyrifera than A. saccharum in 2·8% light but only threefold greater in 7·3% light. Walters et al. (1993a) reported an inverse LMR vs plant mass relationship within species for some of the same species used in this study, thus raising the possibility that differences in LMR among species could be confounded by differences in mass among species. We tested this by analysing LMR vs plant mass relationships for seedlings harvested (> 10 per light × species treatment, except O. virginiana unavailable) within a few days of our CO2 exchange seedling harvest. Of seven tests, relationships for B. alleghaniensis in 7·3% light and B. papyrifera and B. alleghaniensis in 2·8% light were significant (P < 0·05). Using regression equations from these relationships to estimate LMR at the mass of A. saccharum in each light treatment, we found that differences in biomass allocation among species were primarily owing to species differences and much less so to differences in biomass among species (Table 2). For P. tremuloides, LMR was not estimated because its mass range did not extend to A. saccharum's mean mass.

Table 2.  Mean values (± SE) of whole-plant mass, CO2 exchange and morphology parameters for each species × light treatment (n ≥ 3 seedlings for all values except LMRY). LMRY are for treatment means adjusted to the mass of A. saccharum by using plant mass vs LMR relationships. Significant (P < 0·05) plant mass vs LMR relationships are indicated (*). Within light treatments, species are arranged, top to bottom, from least to most shade tolerant (Baker 1949; Burns & Honkala 1990). Parameter abbreviations and units are: mass (mg), LMR, leaf mass ratio [g leaf (g plant)−1]; SMR, stem mass ratio [g stem (g plant)−1]; RMR, root mass ratio [g root (g plant)−1]; LAR, leaf area ratio [cm2 leaf area (g plant)−1]; RW, average 24 h whole-plant respiration rate [nmol CO2 (g plant)−1 s−1]; QYW, whole-plant quantum yield [µmol CO2 (µmol quanta)−1]; LCPW, whole-plant light compensation point (µmol quanta m−2 s−1). RW, QYW and LCPW were calculated from treatment means, thus, there are no SEs and tests of significance for these parameters
P. tremuloides196 ± 570·51 ± 0·02 0·25 ± 0·020·23 ± 0·02340 ± 4110·60·6821·2
 cab aca   
B. papyrifera686 ± 610·54 ± 0·010·530·24 ± 0·010·22 ± 0·01270 ± 147·50·5317·7
 aa acab   
B. alleghaniensis603 + 570·52 ± 0·030·48*0·23 ± 0·010·25 ± 0·02288 ± 227·10·5717·0
 abab acab   
O. virginiana273 ± 990·43 ± 0·02 0·23 ± 0·020·34 ± 0·01233 ± 118·10·5219·6
 bcb abb   
A. saccharum792 ± 830·28 ± 0·040·280·27 ± 0·020·45 ± 0·0392 ± 113·20·2318·1
 ac aac   
B. papyrifera76 ± 260·60 ± 0·010·54*0·27 ± 0·020·14 ± 0·01546 ± 2210·11·3013·0
 ba aba   
B. alleghaniensis134 ± 320·60 ± 0·030·58*0·23 ± 0·010·18 ± 0·01436 ± 259·51·0713·2
 aba abb   
A. saccharum245 + 560·29 ± 0·030·290·27 ± 0·020·44 ± 0·03111 ± 133·80·2415·5
 ab aac   

Whole-plant respiration rates (RW) for A. saccharum were equal to 45% that of other species (Table 2), owing to lower RL, RS and RR (Table 1). Whole-plant net CO2 exchange vs PPFD curves (Fig. 3) indicate that: (1) shade-tolerant A. saccharum generally had the lowest net CO2 loss rates at PPFD < 13 µmol quanta m−2 s−1 in 2·8% light and < 18 µmol quanta m−2 s−1 in 7·3% light. However, the other species always had higher (and positive) net CO2 exchange rates at PPFD above these (i.e. above the roughly common LCP for all species); (2) least tolerant P. tremuloides had the greatest net CO2 loss rates at PPFDs < 15 µmol quanta m−2 s−1 in low light, but because it had the highest whole-plant QY of any species it had the highest CO2 exchange rates of any species at PPFD > 40 µmol quanta m−2 s−1.

Figure 3.

Twenty-four hour whole-plant net CO2 exchange vs photosynthetic photon flux density (PPFD). These curves were constructed by modifying predicted values of AMASS from best-fit curves for leaf-level net CO2 exchange vs PPFD responses (i.e. Fig. 2) as follows: whole-plant net CO2 exchange = AMASS × LMR × 15/24 − RW. See Materials and methods for additional details. Lines are as in legend to Fig. 2.

LMR multiplied by AmaxMASS is whole-plant Amax (i.e. Amax per unit plant mass). Species with greater whole-plant AmaxMASS and whole-plant QY (e.g. P. tremuloides in 7·3% light) had both greater leaf-level AmaxMASS and greater LMR than other species (Table 2). Conversely, lower whole-plant CO2 loss rates at very-low PPFD (e.g. A. saccharum) were primarily owing to lower RW (i.e. lower CO2 loss rates) (Table 2).

Among species, whole-plant net CO2 exchange vs PPFD responses (Fig. 3) and generally strong positive correlations of tissue and whole-plant respiration rates (i.e. RL, RS, RR and RW) with AmaxMASS, SLA, LMR, LAR and AMASS × LMR (Table 3) indicate that high respiration rates are associated with characteristics that enhance whole-plant photosynthetic function. This implies a high-light vs low-light CO2 exchange trade-off, because no species had both high whole-plant photosynthetic rates at PPFD > 15 µmol quanta m−2 s−1 and low CO2 loss rates at PPFD < 15 µmol quanta m−2 s−1. Like patterns in leaf and whole-plant morphology and Amax, patterns in relative growth rate and whole-plant CO2 exchange revealed three distinct groups among our study species, because values were greatest for P. tremuloides, lower and similar among the three Betulaceae species, and lowest for A. saccharum (Table 4).

Table 3.  Pearson product-moment correlation coefficients and P-values of tissue and whole-plant respiration rates with leaf and whole-plant morphology and light-saturated photosynthetic rates for species × light environment means (n = 8). *P < 0·05, **P < 0·01, ***P < 0·001. Acronyms are defined in the text
Table 4. In situ leaf and whole-plant CO2 exchange parameters and relative growth rate (RGR, mg g−1 day−1). CO2 exchange parameters are integrated totals for 13 days (10–22 June) that were estimated from CO2 exchange responses to PPFD and continuous measurements of PPFD taken in the seedlings light treatments: RGR is the average from mid-April to late June. Parameter abbreviations and units are: ∫AMASS, total daytime leaf CO2 exchange per gram leaf mass, mmol CO2 (g leaf)−1; ∫AMASS × LMR, total daytime leaf CO2 exchange per gram plant mass, mmol CO2 (g plant)−1, ∫AMASS × LMR × 15/24 − RW, total 24 h whole-plant CO2 exchange per gram plant mass, mmol CO2 (g plant)−1
P. tremuloides44·452·426·79·3
B. papyrifera41·537·520·27·4
B. alleghaniensis39·339·820·77·9
O. virginiana 45·519·66·5
A. saccharum14·132·1 9·03·4
B. papyrifera34·330·018·04·2
B. alleghaniensis33·426·315·83·2
A. saccharum 4·516·3 4·80·3

CO2 exchange and morphology: relationships with growth

An underlying assumption of many CO2 exchange studies is that CO2 exchange parameters are sensitive indices of growth (Givnish 1988). We tested this assumption by relating leaf and whole-plant morphology and CO2 exchange parameters to relative growth rate determined over a 2 month interval that ended during our net CO2 exchange measurements. Given the caveat that we determined morphology and CO2 exchange at the end of the growth interval and ontogenetic drift in these characteristics may have occurred (e.g. Walters et al. 1993a; Walters & Reich 1999a), patterns among the 18 CO2 exchange, morphology and combined parameters we tested indicated that relationships of growth with photosynthesis were (not surprisingly) stronger for in situ than light-saturated photosynthetic rates and were strongest when leaf fraction (i.e. LMR) was combined with leaf CO2 exchange rates (Table 5). Interestingly, several parameters assumed to be important for low-light growth correlated poorly (P > 0·10) with growth. These parameters included leaf-level QY and leaf and whole-plant LCP (Table 5).

Table 5.  Pearson product-moment correlation coefficients and P-values of leaf and whole-plant morphological and photosynthetic parameters with relative growth rate in 7·3 and 2·8% light combined (RGR-all) and in 7·3% light (RGR-7·3%). *P < 0·05, **P < 0·01, ***P < 0·001. Acronyms are defined in Tables 1, 2 and 4
ParameterRGR-all (n = 7)RGR-7·3% (n = 4)
AmaxAREA  0·01− 0·69
AmaxMASS  0·55  0·46
AMASS  0·70  0·72
SLA  0·58  0·91
QY− 0·72− 0·94
QY × SLA  0·51  0·66
LCPLEAF  0·65  0·37
RL  0·60  0·61
RS  0·80*  0·94
RR  0·78*  0·98*
RW  0·83*  0·92
LMR  0·85*  0·98*
LAR  0·66  0·98*
AmaxMASS × LMR  0·65  0·76
AMASS× LMR  0·93**  0·95*
AMASS× LMR − RW  0·77*  0·97*
LCPW  0·22  0·28
QYW  0·54  0·97*

Whole-plant CO2 exchange and morphology: differences between light environments

Compared to 7·3% light, seedlings from 2·8% light had greater LMR, lower root mass ratios, and similar stem mass ratios. Differences between light environments were greater for less-tolerant B. papyrifera and B. alleghaniensis (e.g. c. 11 and 15% higher LMR in 2·8% light, respectively) than for tolerant A. saccharum (3% greater LMR in 2·8% light). Compared to 7·3% light, RW in 2·8% light ranged from 35% higher for B. papyrifera to 19% higher for A. saccharum. However, if RW and LMR could have been compared between light treatments at a common plant size, they may have differed less, because they both decrease with size and/or age (Walters et al. 1993b), and seedlings were smaller in 2·8 than 7·3% light. Yet, even ignoring the effects of plant size on these parameters, differences between light environments in LMR and RW were small compared to differences in SLA and leaf-level AmaxMASS. Therefore, differences in 24 h whole-plant CO2 exchange vs PPFD responses between light environments (Fig. 3) were primarily owing to linked changes in leaf level-morphology (SLA) and AMASS (Walters et al. 1993a; Reich & Walters 1994) and much less to differences in biomass distribution.


Species and light environment differences in CO2 exchange and morphology

Among seedlings of five, broad-leaved cold-temperate deciduous species grown in low light, species with high whole-plant CO2 exchange at PPFD > 15 µmol quanta m−2 s−1 also had high CO2 loss rates at PPFD < 15 µmol quanta m−2 s−1. This trade-off in high vs low PPFD CO2 exchange corresponded broadly to species’ shade tolerance classifications because intolerant P. tremuloides had the greatest CO2 loss at very low PPFD and the greatest CO2 gain at low to moderately high PPFD, whereas shade-tolerant A. saccharum had the lowest CO2 loss rates at very-low PPFD and the lowest CO2 gain at low to moderately high PPFD. In cold-temperate closed-canopy forest understories, where A. saccharum seedlings are common, mean daytime light levels for most of the growing season are often < 10 µmol quanta m−2 s−1 (Ellsworth & Reich 1992; Tobin & Reich, unpublished data) which is lower than A. saccharum’s, and the other species, whole-plant CO2 compensation point in this study. Given these light environments, minimizing CO2 loss at very low PPFD could be more important for A. saccharum's whole-season carbon balance than maximizing CO2 gain at higher PPFD. In addition, A. saccharum's low LMR may benefit low-light carbon balance over the long-term because autumn leaf turnover is a lower fraction of total plant mass compared to high LMR plants (Walters & Reich 2000). In contrast to shade-tolerant A. saccharum, less tolerant species express traits (high LMR, LAR, Amax, SLA, and subsequently high RW) that suggest a predisposition to maximize growth potential even in low light, despite the possibility that light levels may be too low for these growth potential advantages to be realized. It is important to note that shade-tolerant A. saccharum had the lowest growth rates and integrated whole-plant CO2 exchange of any species in both 2·8 and 7·3% light. Thus, the light levels at which CO2 exchange would be more positive for A. saccharum than the other study species, at least for seedlings acclimated to 2·8% light, would be less than the 2·8% light treatments that the seedlings were grown in. In Walters & Reich (2000) we found that growth rates in 1·5% light were similar and near zero for A. saccharum, B. papyrifera, B. alleghaniensis and P. tremuloides, suggesting that whole-plant CO2 exchange was similar among species at that light level. However, Walters & Reich (2000) also found that seedling survival for A. saccharum was much greater than for the other species at light levels ranging from 0·6 to 7·3% light.

In contrast to the large differences in mass-based R and photosynthesis, and in biomass fractions in roots and leaves among species, leaf-area-based measures of QY and RL, and light compensation points were similar for tolerant and intolerant species, indicating that these characteristics were not important bases for differences in shade adaptation among our study species.

Although we identified potentially adaptive differences in whole-plant CO2 exchange characteristics over the entire range of reported species shade tolerances we examined (i.e. P. tremuloides to A. saccharum), the three Betulaceae species had similar CO2 exchange characteristics despite large differences in reported shade tolerance. In part, these similarities could be owing to differences in plant mass. Walters et al. (1993a) reported that respiration declined with plant mass for the same three Betulaceae species and it was lowest at a common mass for O. virginiana. In this study, mass for O. virginiana was less than half that of B. papyrifera and B. alleghaniensis. Thus, at a common mass, O. virginiana may have had a lower RW and possibly lower whole-plant CO2 loss rates at very-low PPFD than B. papyrifera and B. alleghaniensis.

Between light environments, our data support those of other studies of cold-temperate tree seedlings in indicating that: (1) plasticity in whole plant biomass distribution (i.e. LMR) is, at most, a minor component of light acclimation compared to plasticity in leaf morphology; (2) shade-intolerant species showed greater plasticity, especially in leaf characteristics, than tolerant species (see also Walters et al. 1993b;Reich, Tjoelker et al. 1998). It is important to note that, although strong and similar light effects on SLA have been reported for these species in both greenhouse/growth chamber and field experiments (Ellsworth & Reich 1992;Walters et al. 1993a,b;Walters & Reich 1996, 1997;Reich, Tjoelker et al. 1998) increasing AmaxMASS with decreasing light occurs primarily in greenhouse/growth chamber experiments (Walters et al. 1993a,b;Reich, Walters et al. 1998; this study).

Necessary caveats of our whole-plant CO2 exchange analysis include: (1) we measured respiration on washed roots at 365 µmol CO2 mol−1 and 21 °C. However, wounding, temperature (Tjoelker & Reich 1999) and CO2 (Burton et al. 1997; but not Tjoelker et al. 1999) can affect respiration rates and both temperature and soil CO2 concentration likely varied temporally in the greenhouse. Despite these and other factors affecting differences between measured and in situ rates, our strong RWvs growth relations suggest that the impact of measurement protocol on respiration was similar for all treatments; (2) our data are only relevant for first-year seedlings because morphology and CO2 exchange characteristics can change with ontogeny (Shukla & Ramakrishnan 1984; Walters et al. 1993b; Niinemets 1998); (3) greenhouse conditions do not mimic the combination of multiple resource limitations, environmental heterogeneity and biotic stresses that seedlings experience in the field (Wayne & Bazzaz 1993; Ackerly 1997). However, for O. virginiana, A. saccharum, P. tremuloides and B. alleghaniensis our greenhouse data and those for seedlings grown in forest understories (Lusk & Reich, in press) were similar in terms of rankings among species (i.e. respiration was lower in shade-tolerant species).

Low-light grown shade-tolerant A. saccharum had lower relative growth rates, LMR, SLA and whole-plant CO2 exchange at PPFD > 15 µmol quanta m−2 s−1 than intolerant P. tremuloides. These results are inconsistent with the enhanced low-light growth hypothesis of shade tolerance. Conversely, compared to P. tremuloides, A. saccharum had lower whole-plant respiration rates, which account for its lower whole-plant CO2 loss rates at PPFD < 15 µmol quanta m−2 s−1 than P. tremuloides. Three birch species (B. papyrifera, B. alleghaniensis and O. virginiana) despite large differences in reported shade tolerance, had similar characteristics and were intermediate to A. saccharum and P. tremuloides in their responses. Thus lower whole-plant respiration rates may be one of many traits underlying shade-adaptation, and they may be part of a larger resource conservation syndrome that includes traits that enhance storage and protection from herbivores, pathogens, and mechanical damage. Collectively these resource conservation traits and their possible trade-offs with enhanced low-light growth potential may account for both the higher survival and the equal to lower growth rates documented for shade-tolerant tree seedlings and saplings in low light (Lorimer 1981; Kitajima 1994; Kobe et al. 1995; Kobe & Coates 1997; Venenklaas & Poorter 1998; Walters & Reich 1996, 1999, 2000).