Dynamic light use and protection from excess light in upper canopy and coppice leaves of Nothofagus cunninghamii in an old growth, cool temperate rainforest in Victoria, Australia


  • Michael Tausz,

    Corresponding author
    1. School of Forest and Ecosystem Science, University of Melbourne, Water Street, Creswick, Victoria 3363, Australia;
      Author for correspondence:Michael Tausz Tel: +61 35321 4310 Fax: +61 35321 4166 Email: michael.tausz@unimelb.au or michael.tausz@uni-graz.at
    Search for more papers by this author
  • Charles R. Warren,

    1. School of Forest and Ecosystem Science, University of Melbourne, Water Street, Creswick, Victoria 3363, Australia;
    Search for more papers by this author
  • Mark A. Adams

    1. Centre for Excellence in Natural Resource Management, Faculty of Natural and Agricultural Sciences, University of Western Australia, Nedlands, WA 6008, Australia
    Search for more papers by this author

Author for correspondence:Michael Tausz Tel: +61 35321 4310 Fax: +61 35321 4166 Email: michael.tausz@unimelb.au or michael.tausz@uni-graz.at


  • • Responses to simulated sunflecks were examined in upper canopy and coppice leaves of Nothofagus cunninghamii growing in an old-growth rainforest gully in Victoria, Australia.
  • • Shaded leaves were exposed to a sudden increase in irradiance from 20 to 1500 µmol m−2 s−1. Gas exchange and chlorophyll fluorescence were measured during a 10 min simulated sunfleck and, in the ensuing dark treatment, we examined the recovery of PS II efficiency and the conversion state of xanthophyll cycle pigments.
  • • Photosynthetic induction was rapid compared with tropical and northern hemisphere species. Stomatal conductance was relatively high in the shade and stomata did not directly control photosynthetic induction under these conditions. During simulated sunflecks, zeaxanthin was formed rapidly and photochemical efficiency was reduced. These processes were reversed within 30 min in coppice leaves, but this took longer in upper canopy leaves.
  • • Poor drought tolerance and achieving a positive carbon balance in a shaded canopy may be functionally related to high stomatal conductance in the shade in N. cunninghamii. The more persistent reduction in photochemical efficiency of upper canopy leaves, which means less efficient light use in subsequent shade periods, but stronger protection from high light, may be related to the generally higher irradiance and longer duration of sunflecks in the upper canopy, but potentially reduces carbon gain during shade periods by 30%.


Closed forests are often spatially complex and the light environment can also be spatially and temporally variable. Irradiance generally decreases vertically from the top of the canopy to forest floor. Average irradiance on the floor of a closed forest may be as little as 0.5% of full sunlight whilst leaves in the upper canopy receive full sunlight (Björkman & Ludlow, 1972). The understorey of some evergreen closed forests is amongst the darkest environments habitable by higher plants (Allen & Pearcy, 2000a). In this layer, the light environment is characterised by diffuse irradiance of low intensity interrupted by intense sunflecks that might last from a second or less to 15 min or more. Depending on the ecosystem in question, sunflecks may account for between 20 and 80% of the total irradiance (Watling et al., 1997). Pearcy (1987) estimated that sunflecks resulted in 30–60% of the total daily carbon gain, and noted that the utilisation of sun flecks is of paramount importance for the overall assimilation balance of understorey plants.

For leaves acclimated to shade, the photosynthetic response to a sudden increase in irradiance (via sunflecks) is not instantaneous owing to biochemical and stomatal limitations. Biochemistry is initially limiting because after prolonged shading, ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) and other photosynthetic enzymes are deactivated, and because of the small size of the pools of Calvin cycle intermediates. Appreciable periods of time are required for enzyme re-activation and to build the pools of intermediates. The stomatal limitation arises because stomata typically close at least partially in the shade and, again, it takes some time for the stomata to re-open and thereby reduce the limitation on supply of CO2 for photosynthesis. The efficient use of sunflecks is thus a function of the rapidity with which biochemical and stomatal limitations are removed due to, inter alia, light activation of enzymes and stomatal opening, and/or the maintenance of high enzyme activation and wide-open stomata during low-light periods (Pepin & Livingston, 1997; Allen & Pearcy, 2000a,b; Naumberg & Ellsworth, 2000; Timm et al., 2002).

Previous studies in diverse ecosystems (tropical rainforests, Allen & Pearcy, 2000a; temperate pine plantations, Naumberg & Ellsworth, 2000) have shown considerable variation in the relative contribution of stomatal control to sunfleck utilisation. In a tropical ecosystem, stomatal response times determined the photosynthetic response times to sunflecks when water was limiting, but stomata exhibited little or no control when water was not limiting (Allen & Pearcy, 2000a). In a temperate pine plantation, stomatal responses generally limited sunfleck responses, but to different extents in different species (Naumberg & Ellsworth, 2000).

Sunflecks are also potentially harmful to plants because of the time-lag between the beginning of a sunfleck and the realisation of maximum rates of photosynthesis. This time-lag leads to the quantum of absorbed irradiance being, at least temporarily, in excess of what is used in photochemistry (Logan et al., 1997; Watling et al., 1997; Thiele et al., 1998) and can cause ‘photoinhibition’ (defined as a persistent decrease in quantum use efficiency). Photoprotective mechanisms that function to dissipate excess energy are thus likely to be active in shade-adapted species subject to sunflecks. Foremost among these are: (1) the energy (pH) dependent dissipation that reduces the quantum efficiency of photosystem II; (2) a dissipation mechanism related to the conversion (de-epoxidation) of the xanthophyll cycle pigments violaxanthin via antheraxanthin to zeaxanthin; and (3) the quenching of excited chlorophylls and scavenging of active oxygen species in thylakoids (Niyogi et al., 2001; Demmig-Adams, 2003; Morosinotto et al., 2003).

Only a handful of studies have examined photoinhibition in relation to sunflecks. Investigations on (sub)tropical plants showed that fast responses of photo-protective systems accounted for flexible protection, but that photoinhibition (a persistent decline in quantum efficiency) was not completely avoided (Logan et al., 1997; Watling et al., 1997; Thiele et al., 1998). A reasonable primary hypothesis is that shade-adapted leaves are more susceptible to photoinhibition during sunflecks because they contain fewer protective pigments (xanthophylls) than sun foliage (Logan et al., 1997). Previous studies on sunfleck responses have focused on emerging saplings or understorey shrubs in comparison with gap plants, but did not compare differently light-adapted foliage within the same species (i.e. sun leaves vs shade leaves). Similarly, while photoinhibition and sunflecks have been studied in detail in tropical and subtropical forests (Logan et al., 1997; Watling et al., 1997; Stegemann et al., 1999; Allen & Pearcy, 2000a,b) and in northern hemisphere temperate ecosystems (Pepin & Livingston, 1997; Naumberg & Ellsworth, 2000), the same is not true for cool-temperate rainforests in the southern hemisphere.

In mainland Australia these Nothofagus-Eucalyptus forests are confined to relatively small areas and are therefore of high conservational value. One view is that once closed-canopy Nothofagus rainforest is established, it may become self-maintaining and less fire-prone than mixed stands with Eucalyptus spp. (Howard, 1973b). Such a view makes the maintenance and regeneration of these forests an important aim of local forest and park management. Myrtle beech (Nothofagus cunninghamii) – the dominant overstorey species of these ecosystems – can regenerate in the light-limited understorey of a closed canopy rainforest (Howard, 1973b) – a phenomenon supported by previous studies of steady state light compensation points (Howard, 1973b; Ashton & Turner, 1979). In Victoria, coppice shoots are of paramount importance for this regeneration (Howard, 1973b). The response of N. cunninghamii to sunflecks has not yet been investigated, despite the putative importance of sunflecks for carbon gain of coppice leaves and regeneration of this species.

Here we report the responses of gas exchange and photoprotection to simulated sunflecks in N. cunninghamii. The study was carried out in situ in a natural old-growth rainforest gully in Victoria. Measurements were made on more light-adapted leaves from the upper canopy and shade leaves from coppice shoots near the forest floor. We tested the following hypotheses: (1) as a rainforest species that can regenerate under low irradiances –N. cunninghamii displays adaptive responses to sunflecks such as fast induction of photosynthesis or maintenance of open stomata under low irradiances; (2) because coppice leaves receive a larger proportion of irradiance as sunflecks and these are more important for their daily carbon gain their response to sunflecks is faster compared with upper canopy leaves; and (3) because upper canopy leaves experience longer periods of direct sunlight they have stronger photoprotection (measurable as xanthophyll related decrease in quantum use efficiency) at the potential expense of carbon gain compared with coppice leaves.

Materials and Methods

Site conditions

The field site is located in a sheltered gully at c. 600 m asl on the southern slopes of Mt Donna Buang in the Yarra Ranges National Park in Victoria, Australia. The vegetation type is cool-temperate rainforest with a mid-storey canopy dominated by Nothofagus cunninghamii and Atherosperma moschatum at 15–20 m above the forest floor under a sparse over-storey of high (c. 60 m tall) Eucalyptus regnans. Detailed descriptions of this vegetation type are found in Howard (1973a). A bi-level boardwalk established by Parks Victoria allowed access to upper canopy leaves (15 m above ground level) and coppice leaves (2 m above ground level).

The study was carried out from 7 to 23 of October 2003 (early spring). All measurements were made on overcast days. This ensured that the leaves were uniformly acclimated to low irradiances prior to all measurements. The ambient photosynthetic photon flux density (Q) measured adjacent to coppice leaves was 10–20 µmol m−2 s−1 (sometimes less), while Q measured beside upper canopy leaves was up to 50 µmol m−2 s−1. Air temperatures varied from 8 to 17°C.

Sampled trees

Measurements were made on coppice leaves of four Nothofagus cunninghamii Oerst. trees c. 2.0 m above the forest floor, and on light-exposed upper-canopy leaves of four trees c. 15 m above the forest floor. Sampled trees were mature individuals that were estimated to be between 300 and 400 yr old. All measurements were made on the youngest fully expanded leaves; these were between the fifth and eighth leaf from the shoot apex.

Light environment of the measured leaves

Cool-temperate rainforest canopies are structurally complex and we made no attempt to determine the exact interception of irradiance by individual leaves. Instead, we measured broad gradients in relative irradiance at canopy locations close to the measured branches and estimated the frequency of sunflecks – future studies might include a more detailed examination of canopy structure and interception of irradiance by individual leaves. Variation in the light environment between the upper canopy leaves and coppice leaves was quantified by hemispherical photographs that were taken adjacent (within 10 cm) to four measured branches at each canopy level (ter Steege, 1997). This technique accurately estimates the fraction of irradiance transmitted to foliage (Easter & Spies, 1994; ter Steege, 1997). Hemispherical photographs (Nikon 601, Nikon, Tokyo, Japan) were taken at dusk, black and white negatives were scanned and evaluated using WinPhot 5.0 software (ter Steege, 1997). For 20 October we estimated the diurnal course of direct and diffuse Q received by leaves and the total Q per day.

Leaf morphology

The projected leaf area of leaves was measured with a leaf area meter (LI-3000 A + LI-3050 A, Li-Cor, Lincoln, NE, USA). Dry mass was determined after 72 h in an oven at 60°C. Specific leaf area (S) was calculated as projected area divided by dry mass.

Cuticular conductance

To evaluate the relative importance of stomatal responses, cuticular conductance was determined in the laboratory on shoots collected in the field. Branches (c. 300 mm long) were cut from trees, immediately re-cut under water, and transported to the laboratory where they were re-saturated over-night at 4°C. Measurements were made on whole leaves as described by Burghardt & Riederer (2003).

Steady state gas exchange

Steady state gas exchange was determined on CO2 and light response curves of net photosynthesis (A) using a Li-Cor 6400 and 20 × 30 mm broadleaf chamber with integrated light source (LI-6400–02B). For each leaf a CO2 response curve (A-ci curve) was generated at a saturating Q of 1500 µmol m−2 s−1, leaf temperature was controlled at 15 ± 1°C by setting the block temperature, and leaf-to-air vapour pressure deficit was less than 0.8 kPa. Leaves were exposed to 360 µmol mol−1 CO2 in air and a Q of 1500 µmol m−2 s−1 until rates of photosynthesis and transpiration were steady. After this, the CO2 concentration in the chamber ca was increased to 1500 µmol mol−1 and an A-ci curve was generated by stepwise (Fig. 2) decreasing ca to 50 µmol mol−1. At each ca photosynthesis was allowed to stabilise for at least 3 min, and three successive measurements were made at 1-min intervals to ensure stability.

Figure 2.

Steady state photosynthesis in upper canopy (open circles) and coppice leaves (closed circles) of Nothofagus cunninghamii growing in a cool temperate rainforest. Medians and median deviations of n = 4 trees. (a) The response of net photosynthesis (A) to incident photosynthetic photon flux density (Q) The light-response curve was fitted to a nonrectangular hyperbola.(b) The response of net photosynthesis (A) to internal CO2 concentration (ci).

Light response curves were measured under the same chamber conditions as A-ci curves, except that chamber CO2 concentration was held constant at 360 µmol mol−1 and Q was varied. Leaves were enclosed in the dark chamber, and the rate of dark respiration was recorded once rates of gas exchange were steady. Thereafter, Q was increased stepwise (Fig. 2) to a maximum of 1500 µmol m−2 s−1, which was sufficient to light-saturate photosynthesis. Measurements were logged once rates of gas exchange were completely stable, which took 5–15 min at each point depending on the preceding conditions.

The light response of photosynthesis was fitted to a nonrectangular hyperbola:

image( Eqn 1)

where inline image is the maximum quantum efficiency of carbon fixation, Θ is the curvature factor, Asat is the maximum light saturated photosynthesis rate, and Rd is the respiration rate.

Simulated sunfleck treatments

Simulated sunflecks were imposed with two LiCor 6400 s – one fitted with a 20 × 30 mm broadleaf chamber with integrated light source (6400–02B; measured surface 6 cm2) and the other with an integrated fluorescence chamber head (LI-6400–40; measured surface 2 cm2). Both units had been recently calibrated. While there were no significant differences or consistent trends in the measured rates between the two chamber types, leaves enclosed in the fluorometer chamber did not always equilibrate well. When we could not equilibrate enclosed leaves, measurements were repeated on different leaves.

Leaves were enclosed in the leaf chamber at a Q of 20 µmol m−2 s−1, and steady-state rate of photosynthesis and stomatal conductance were recorded once the reading was stable (after a minimum of 5 min). Thereafter, Q was increased to 1500 µmol m−2 s−1 (in one step) and gas exchange readings were logged every 10 s. Because the block temperature, not leaf temperature, was directly controlled, leaf temperatures increased by a maximum of 1°C when Q was increased and remained higher for the duration of the simulated sunfleck. This led to a maximum decrease in the calculated stomatal conductance by 6–9%. We did not correct for this effect as it was similar for all treatments and may also be expected in natural sunflecks.

Fluorescence parameters were logged less frequently (see below). Initial experiments using simulated sunflecks of 20 min duration showed that > 95% of the maximal changes were reached after 10 min. Therefore, all subsequent evaluations were based on simulated sunflecks of 10-min duration.

Chlorophyll fluorescence

Chlorophyll fluorescence measurements were conducted with the modulated fluorometer chamber of the LiCor 6400. Nomenclature follows Maxwell & Johnson (2000). F0 and Fm were determined after 10 min dark adaptation. Because all leaves had been pre-exposed to low irradiance during the overcast measurement days, this dark adaptation time proved long enough to fully restore maximal quantum efficiency of PS II (Fv : Fm) to 0.81. Overnight dark adaptation did not yield higher Fv : Fm. After determination of F0 and Fm, leaves were allowed to stabilise at a Q of 20 µmol photons m−2 s−1 and fluorescence parameters were recorded. After the simulated sunfleck was started, fluorescence parameters were determined by subsequent saturation pulses at 1–3 min intervals. PS II photochemical quantum yield (ΦPSII) was calculated as inline image, nonphotochemical quenching (NPQ) as Fm : inline image, and the potential quantum yield of PS II as inline image : inline image in irradiated leaves and Fv : Fm in the dark, respectively.

Saturation pulses affected gas exchange readings for up to 30 s after the saturation pulse. These data points were identified visually (in an induction curve graph similar to Fig. 3) and excluded from gas exchange evaluation.

Figure 3.

The response of net photosynthesis (A), stomatal conductance (gs) and intercellular CO2 concentration (ci) to a simulated sunfleck in Nothofagus cunninghamii. Typical data for a coppice leaf are shown. Leaves were exposed to 20 µmol photons m−2 s−1 (in the 400–700 nm range) until rates of gas exchange were steady, then a simulated sunfleck was imposed by increasing irradiance to 1500 µmol m−2 s−1 for 10 min. The simulated sunfleck started at time 0. For clarity, only every fifth to tenth data point is drawn. Dotted lines indicate fitted functions (exponential for A and logistic for gs).

Pigment determinations

Single leaves were placed in prelabelled vials and immediately (within seconds) immersed in liquid nitrogen. Samples were kept under liquid nitrogen or at −80°C until they were lyophilised. They were then protected from humidity and stored at −20°C prior to analysis.

Pigments were extracted in ice-cold acetone and extracts analysed on a gradient-HPLC using the method described in Tausz et al. (2003).

To determine the response of pigments to the simulated sunfleck treatment, leaves were sampled at ambient (low irradiance) conditions, immediately after a 10 min simulated sunfleck treatment, after 10 and 30 min of recovery in darkness, and after overnight recovery. PS II quantum efficiency was measured simultaneously on adjacent leaves.

Fitting of response curves

The response of net photosynthesis (A) to a sudden increase in Q from 20 to 1500 µmol m−2 s−1 (Fig. 3) was described by the equation:

image(Eqn 2)

where Ashade is the initial photosynthetic rate in the shade (20 µmol m−2 s−1), Amax is the maximal photosynthesis rate achieved in response to the simulated sunfleck, t is the time from the beginning of the simulated sunfleck, and t1 is a characteristic time constant (time until 63% of change occurs). Several parameters were then calculated using the resulting equation: Amax (maximal photosynthesis rate calculated from the induction curve), A60 (photosynthesis rate 60 s into the simulated sunfleck), t90% (t until 90% of Amax is reached), IS60s (induction state 60 s into the sunfleck) calculated as A60 : Amax (Allen & Pearcy, 2000b).

Stomatal responses were best described by a logistic curve (compare Fig. 3):

image(Eqn 3)

where gs shade is the initial stomatal conductance in low irradiation (20 µmol m−2 s−1), gs max the maximum stomatal conductance reached during the induction treatment, t1 a characteristic time constant, and p a parameter determining the curve shape. The time until 90% of change was reached (t90%) was calculated from the resulting function (Naumberg & Ellsworth, 2000).

The increase in NPQ and decrease in inline image were fitted using an exponential curve analogous to the one used for photosynthesis rates (eqn 2, Fig. 4). ΦPSII was not fitted by a mathematical function.

Figure 4.

The response of chlorophyll fluorescence parameters to a simulated sunfleck in Nothofagus cunninghamii. Data are shown for a typical coppice leaf. Leaves were exposed to 20 µmol photons m−2 s−1 (in the 400–700 nm range) until rates of gas exchange were steady, then a simulated sunfleck was imposed by increasing irradiance to 1500 µmol m−2 s−1 for 10 min. The simulated sunfleck (1500 µmol m−2 s−1) started at time 0. Dotted lines are exponential functions fitted to the data points. (a) open circle, PS II effective quantum yield (ΦPSII); closed circle, potential quantum yield (inline image); NPQ, nonphotochemical quenching.


Our experimental units were leaves from different branches of four trees selected at each canopy level. In some cases we made several measurements on different branches of one tree. Differences between upper canopy leaves and coppice leaves were evaluated using t-tests in cases where sample sizes were greater than five (simulated sunflecks evaluated) and normality assumptions could be met. Alternatively, Mann–Whitney U-tests were applied. For multiple comparisons among different sampling times (e.g. Fig. 5), significance of differences was assessed by the Kruskal–Wallis test followed by Conover's cross comparisons (Bortz et al., 1990). When there were more than five replicates, means and standard deviations are reported; where there were fewer than five replicates, medians and median deviations are presented (following the suggestions in Sachs, 1992).

Figure 5.

The responses of carotenoids and fluorescence parameters to a simulated sunfleck and subsequent recovery in coppice leaves (a, left-hand panels) and upper canopy leaves (b, right-hand panels) of Nothofagus cunninghamii. Carotenoid contents (open circle, neoxanthin; closed circle, V + A + Z; open square, lutein; closed square, β-carotene), de-epoxidation state (DEPS) of the xanthophyll cycle (calculated as (Z + 0.5 * A)/(V + A + Z) * 100), and potential quantum efficiency of PS II (Fv : Fm, inline image) are shown as medians and median deviations of n = 4 trees. Horizontal bars indicate irradiance. Shaded section, Q = 20 µmol photons m−2 s−1; white section, simulated sunfleck (Q = 1500 µmol m−2 s−1); black section, darkened. A, antheraxanthin; V, violaxanthin; Z, zeaxanthin. Time from start of the simulated sunfleck. Asterisks indicate significant differences between upper canopy leaves and coppice leaves (Mann–Whitney U-test). Letters indicate significant differences between sampling times; points that have no letter in common are significantly different from each other (Kruskal–Wallis test with Conover's posthoc comparisons).

Modelled effects of sunflecks and photoinhibition on leaf carbon gain

Although a quantification of carbon gain in the investigated leaves was beyond the scope of the present paper, we investigated limited scenarios that served to highlight the relative importance of the observed phenomena. Complete models of carbon uptake under dynamic light regimes have been suggested (Pepin & Livingston, 1997; Stegemann et al., 1999, or Naumberg & Ellsworth, 2000) and generally require detailed studies of dynamic induction gain and loss, temperature dependencies, and detailed environmental data. A theoretical model evaluating potential effects of photoinhibition on carbon gain has been recently published (Zhu et al., 2004).

  • 1To estimate the relative contribution of sunflecks to carbon gain in situ, potential carbon gain of coppice and upper canopy leaves during 1 d was roughly calculated from direct (sunfleck) and diffuse Q, using the nonrectangular hyperbola (eqn 1), thereby disregarding effects of photosynthetic induction and photoinhibition.
  • 2The potential effect of a persistent decrease in Fv : Fm (photoinhibition) on carbon gain was evaluated by comparing estimates of carbon gain in the absence of photoinhibition (i.e. Fv : Fm at maximum), with estimates for photoinhibited leaves. We assumed a proportional relationship between decreases in Fv : Fm (maximum quantum efficiency of PS II) and decreases in inline image (maximum quantum efficiency of carbon fixation) as outlined in Zhu et al. (2004). The coupled decrease in Θ and inline image was described by the following equation (Zhu et al., 2004):
    image(Eqn 4)
    This relationship was then used to recalculate steady state photosynthesis at a given Q using the nonrectangular hyperbola (eqn 1).
  • 3To evaluate the potential overestimation of carbon gain from sunflecks (which would occur if effects of photosynthetic induction were neglected), we used the following methods to calculate carbon gain during a 10 min sunfleck of Q = 1500 µmol m−2 s−1:
    • (a) Steady state rates of photosynthesis from the nonrectangular hyperbola (eqn 1, i.e. neglecting the induction phase).
    • (b) The measured induction kinetics (eqn 5, Table 2).
      image(Eqn 5)
    • (c) The measured induction kinetics (eqn 5, Table 2) assuming a photoinhibited leaf at Fv : Fm = 0.7. In this case, photosynthesis rates (Amax, Ashade) used in eqn 5 were recalculated from the nonrectangular hyperbola as outlined in (2).
Table 2.  Responses of net photosynthesis and stomatal conductance of coppice and upper canopy leaves of Nothofagus cunninghamii to simulated sunflecks
 Coppice leavesUpper canopy leavesP
  1. Leaves were exposed to 20 µmol photons m−2 s−1 (in the 400–700 nm range) until rates of gas exchange were steady, then a simulated sunfleck was imposed by increasing irradiance to 1500 µmol m−2 s−1 for 10 min. Data are means and standard deviations of n = 17 (coppice leaves) and n = 12 (upper canopy leaves) sunflecks (between three and five replicate measurements on different shoots of four trees at each canopy level). Ashade is the steady state rate of net photosynthesis at 20 µmol photons m−2 s−1 (in the 400–700 nm range), Amax is the maximum rate of net photosythesis at 1500 µmol photons m−2 s−1 calculated from the induction curve, A60s is the rate of net photosynthesis 60 s into the sunfleck, IS60s is the photosynthetic induction state calculated 60 s into the sunfleck, t1 is the time constant in fitting function (time to 63% change), t90% is time to 90% of Amax calculated from the induction curve, gs shade is the steady state stomatal conductance at 20 µmol m−2 s−1, gs max is the maximum stomatal conductance reached during the 10 min simulated sunfleck. Cuticular conductance (gcut) was measured only in upper canopy leaves. The significance of differences between upper canopy and coppice leaves (P-values) were examined with t-tests or Mann–Whitney U-test (as appropriate).

Ashade[µmol CO2 m−2 s−1] 0.9 ± 0.5 0.8 ± 0.6  0.688
Amax[µmol CO2 m−2 s−1] 4.5 ± 2.2 8.0 ± 1.5  0.004
A60s[µmol CO2 m−2 s−1] 2.6 ± 1.3 6.3 ± 1.0< 0.001
t1[s]113 ± 63 42 ± 20  0.027
t90%[s]230 ± 127 92 ± 44  0.032
IS60s[%] 49 ± 21 77 ± 16  0.013
gcut[mmol H2O m−2 s−1]   5 ± 1 
gs shade[mmol H2O m−2 s−1] 76 ± 52123 ± 30  0.115
gs max[mmol H2O m−2 s−1]104 ± 88167 ± 24  0.190
t90%gs[s]157 ± 133 65 ± 19  0.140


Light environment at the positions of the measured leaves

Hemispherical photographs confirmed that coppice leaves received less irradiance than upper canopy leaves. Assuming clear days, total daily Q was 5.1 ± 1.6 mol m−2 d−1 at coppice level vs 8.3 ± 3.0 mol m−2 d−1 in the upper canopy (in the 400–700 nm range; median ± median deviations of four positions at each canopy level). The proportion of incident irradiance received directly (relative to total) was similar at both canopy levels (81 ± 6 vs 86 ± 3%).

The main difference between upper canopy and coppice level was the duration of uninterrupted periods of high irradiance (Q > 1000 µmol m−2 s−1) received on clear days (Fig. 1). The total duration of high irradiance periods (sunflecks) was longer in upper canopy leaves (c. 1 h vs 46 min). Single sunflecks were shorter and more frequent (3 ± 2 min) for coppice leaves, compared with upper canopy leaves (18 ± 14 min; Fig. 1). Maximum diffuse irradiance received by coppice leaves was c. 25 µmol m−2 s−1, and at upper canopy leaves c. 85 µmol m−2 s−1. This was confirmed by spot measurements with a LI-190 quantum sensor on overcast days, which indicated that Q = 20 µmol m−2 s−1 at coppice level and 50 µmol m−2 s−1 at upper canopy level.

Figure 1.

Diurnal photosynthetic photon flux density (Q) at upper canopy (a) and coppice (b) leaves of Nothofagus cunninghamii at the investigated rainforest site on a clear day (20 October). Typical positions at each canopy level are shown.

Nonetheless, clear days are rare in this ecosystem in early spring. The closest long-term meteorological record (Healesville, c. 20 km from the field site) reports, on average, only 3.2 clear days in September and 2.8 in October (Australian Government, Bureau of Meteorology, http://www.bom.gov.au). Hence, the general case is that sunflecks occur during occasional sunny periods that are in turn interspersed among much longer overcast periods.

Leaf morphology and pigments in upper canopy and coppice leaves

Upper canopy leaves had a significantly lower specific leaf area than coppice leaves (Table 1). On an area basis, chlorophyll a and b concentrations, and chlorophyll a : b ratios, were greater in upper canopy leaves than in coppice leaves. On a dry mass basis, differences in pigment composition were smaller. On a total chlorophyll basis, the upper canopy leaves had higher concentrations of the xanthophyll cycle carotenoids (V + A + Z), whereas neoxanthin, lutein, and β-carotene did not differ significantly between upper canopy and coppice leaves (Fig. 5).

Table 1.  Chlorophyll concentrations and specific leaf area (S) of coppice and upper canopy leaves of Nothofagus cunninghamii
 Coppice leavesUpper canopy leavesP
  1. Data are means and SD of n = 16 (S) to 20 (chlorophyll) leaves from different shoots of eight trees (four and five leaves per tree, respectively, four sample trees at each canopy level). d. wt leaf dry weight. The significance of differences between upper and lower canopy leaves (P-values) were examined with t-tests (chlorophyll) or Mann–Whitney U-tests (S).

Chlorophyll a [µmol m−2]  450 ± 68 594 ± 42< 0.001
Chlorophyll b [µmol m−2]  139 ± 24 173 ± 12< 0.001
Chlorophyll a : b 3.26 ± 0.193.44 ± 0.13  0.002
Chlorophyll a [µmol g−1 d. wt] 4.98 ± 0.754.60 ± 0.32  0.065
Chlorophyll b [µmol g−1 d. wt] 1.54 ± 0.261.34 ± 0.09  0.005
S [m2 kg−1 d. wt]11.05 ± 1.197.75 ± 0.80< 0.001

Steady state characteristics of photosynthesis

There were large differences between upper canopy and coppice leaves in their steady-state light- and CO2-responses (Fig. 2). Upper canopy leaves had higher A than coppice leaves when Q was above 200 µmol m−2 s−1, but there was little difference at lower Q. The initial slope of the A-ci response and the maximum A at high ci were both greater in upper canopy leaves than in coppice.

Responses of gas exchange and chlorophyll fluorescence to simulated sunflecks

When Q was increased from 20 to 1500 µmol m−2 s−1, photosynthesis increased quickly and after 10 min was near (at least 80%) or at steady state rates (Fig. 3, Table 2). The time taken to 90% of maximum net photosynthesis varied considerably, but was significantly faster in the upper canopy than in the low-level coppice (Table 2).

Stomatal conductance at a Q of 20 µmol m−2 s−1 was highly variable among leaves and stomatal responses to simulated sunflecks were similarly variable. In some leaves there was no response, while others showed increases in stomatal conductance to varying degrees (Fig. 3 Table 2). There was no consistent pattern. The stomatal response (or lack thereof) was unrelated to initial stomatal conductance (at Q = 20 µmol m−2 s−1), and leaf-to-air vapour pressure deficit (data not shown).

We plotted A as a function of internal CO2 concentrations (ci) to determine if there was a stomatal response during the simulated sunfleck (Pepin & Livingston, 1997; Allen & Pearcy, 2000a; Fig. 6). The slope dA/dci was negative during the sunfleck and the steady-state A-ci curve (compare Fig. 3) was reached within 10 min. A turn towards positive dA/dci, which would indicate stomatal opening (Pepin & Livingston, 1997; Allen & Pearcy, 2000a), was not observed.

Figure 6.

Steady-state and simulated sunfleck response of the relationship between net photosynthesis (A) and intercellular CO2 concentration (ci). The time-course of the simulated sunfleck response (following an increase in irradiance from 20 to 1500 µmol photons m−2 s−1) is shown by open symbols. The steady state A-ci curve for a leaf on the same branch is shown by closed symbols connected with a line. Data are typical results of a coppice leaf of Nothofagus cunninghamii. Dotted line shows sunfleck response that would occur if stomata opened during the light fleck (compare e.g. Fig. 2 in Allen & Pearcy, 2000b; Fig. 6 in Pepin & Livingston, 1997).

In leaves acclimated to 20 µmol m−2 s−1, there was no difference in inline image, ΦPSII, or NPQ between upper canopy and coppice leaves (Table 3). In response to simulated sunflecks there was a rapid increase in NPQ and a concomitant fast reduction in inline image. ΦPSII decreased rapidly to nearly zero and increased slowly thereafter (Fig. 4). The rate of increase in NPQ was significantly slower in the upper canopy leaves (Table 3).

Table 3.  Responses of chlorophyll fluorescence parameters of coppice and upper canopy leaves of Nothofagus cunninghamii to simulated sunflecks
 Coppice leavesUpper canopy leavesP
  1. Leaves were exposed to 20 µmol photons m−2 s−1 (in the 400–700 nm range) until rates of gas exchange and fluorescence were steady, then a simulated sunfleck was imposed by increasing irradiance to 1500 µmol m−2 s−1 for 10 min. t1 is the time constant in the fitting equation (time to 63% change). Data are means and SDs of n = 10 (lower canopy) and n = 6 (upper canopy) sunflecks (on between one and three different branches from four trees at each canopy level). The significance of differences between upper canopy and coppice leaves (P-values) were examined with t-tests or Mann–Whitney U-test (as appropriate).

NPQshade0.55 ± 0.300.48 ± 0.140.622
NPQmax3.88 ± 0.623.92 ± 0.700.934
t1[s] 111 ± 28 178 ± 420.034
0.73 ± 0.050.74 ± 0.030.549
0.40 ± 0.070.44 ± 0.030.130
t1[s]  58 ± 29  94 ± 610.189
ΦPSII shade0.71 ± 0.030.72 ± 0.020.402
ΦPSII(10 min)0.03 ± 0.020.06 ± 0.040.130

The allocation of absorbed Q was calculated according to Logan et al. (1997), assuming an optimum PSII quantum efficiency of 0.81 (N. cunninghamii leaves after overnight darkening), translating to a constitutive light dissipation of 19%. At Q = 20 µmol m−2 s−1, coppice leaves allocated 71% of the absorbed Q to electron transport and 8% to variable energy dissipation (photoprotection or, possibly, photodamage). This changed to 3% allocation to electron transport and 41% variable energy dissipation at the end of the 10 min 1500 µmol m−2 s−1 simulated sunfleck. Upper canopy leaves allocated 72% of absorbed Q to electron transport and 7% to variable energy dissipation in the shade, and 6% vs 37% at the end of a 10-min simulated sunfleck.

Relaxation of reduced photochemical efficiency and xanthophyll de-epoxidation

Xanthophyll cycle pigments in upper canopy and coppice leaves were significantly more de-epoxidised immediately after a 10-min simulated sunfleck (Fig. 5). The increase in the de-epoxidation state to 40–50% was mainly due to conversion of violaxanthin to zeaxanthin – antheraxanthin concentrations were not affected (data not shown). When leaves were subsequently darkened, the de-epoxidation state decreased, with the rate of this decrease being faster in coppice leaves than in upper canopy leaves (Fig. 5). In both upper canopy and coppice leaves there was no significant change in the de-epoxidation state after 10 min recovery in the dark. After 30 min dark recovery, the de-epoxidation state returned to shade values in coppice leaves but remained unchanged in upper canopy leaves. In both upper canopy and coppice leaves, high violaxanthin levels and a corresponding low de-epoxidation state were reached overnight.

The potential quantum efficiency of PS II (Fv : Fm and inline image in darkened and illuminated leaves, respectively) decreased when Q increased from 20 to 1500 µmol m−2 s−1 (Fig. 4). The recovery of Fv : Fm to presunfleck values upon darkening was faster in coppice leaves than in upper canopy leaves (Fig. 5). In coppice leaves, Fv : Fm returned to presunfleck values after 10 min of darkening, whereas in upper canopy leaves Fv : Fm remained lower for at least 30 min (Fig. 5).

Estimated effects of sunflecks and photoinhibition on leaf carbon gain

According to our simple calculation of daily carbon gain using the steady-state characteristics of photosynthesis, upper canopy leaves gain about twice as much C as coppice leaves (Table 4) on a clear spring day. Direct irradiance (sunflecks) accounts for 37% (in coppice leaves) and 41% (in upper canopy leaves) of the total C gain on a clear spring day. The remaining 63 and 59% of daily C gain is due to diffuse irradiance (Table 4). Hence, C gain from diffuse irradiance is larger than that from sunflecks.

Table 4.  Modelled C gain of coppice and upper canopy leaves of Nothofagus cunninghamii using different scenarios.
 Coppice leavesUpper canopy leaves
  1. Estimated daily C gain was calculated assuming a clear spring day using irradiance curves from 20 October (as in Fig. 1) from four positions at both canopy levels and steady-state photosynthesis characteristics (eqn 1). Medians ± median deviations.

  2. Estimated effect of photoinhibition on steady-state values was calculated assuming a maximum Fv : Fm = 0.81 and a photoinhibited Fv : Fm = 0.70. Percent values are relative to steady-state (Fig. 2).

  3. C gain during a 10-min sunfleck was estimated using steady-state rates from eqn 1, using the observed induction kinetics (eqn 5), and using eqn 5 corrected for the effect of photoinhibition at Fv : Fm = 0.70. Percent values are relative to observed induction kinetics.

Calculated daily C gain
[mmol C m−2d−1]
Total53.9 ± 6.8108.4 ± 21.1
From diffuse Q37.0 ± 3.0 66.0 ± 19.2
From direct Q19.9 ± 9.0 44.3 ± 3.7
Calculated effect of photoinhibition on photosynthesis rates (A)
[µmol m−2 s−1]
At Q = 50 µmol m−2 s−1
A (steady state, Fv : Fm = 0.81)2.44100%1.28100%
A (photoinhibited, Fv : Fm = 0.70)2.1990%0.9070%
At Q = 1500 µmol m−2 s−1    
A 1500 (steady state, Fv : Fm = 0.81)5.95100%8.73100%
A 1500 (photoinhibited, Fv : Fm = 0.70)5.9099%8.6799%
Estimated C gain during a 10-min sunfleck
[mmol C m−2]
Assuming immediate induction3.57153%5.24117%
Using observed induction2.34100%4.47100%
Assuming Fv : Fm = 0.72.3098%4.46100%

Estimated total carbon gain during a 10 min sunfleck of 1500 µmol photons m−2 s−1 was higher in upper canopy leaves compared with coppice leaves (Table 4). Neglecting the induction phase (assuming an immediate increase of photosynthesis rate with Q) would lead to an overestimation of the carbon gain during a 10-min sunfleck by 53% in coppice leaves and by 17% in upper canopy leaves.

The effect of photoinhibition on photosynthesis was estimated based on an Fv : Fm of 0.7 (cf. presunfleck Fv : Fm = 0.81), which was measured in upper canopy leaves after a 10-min sunfleck and persisted for at least 30 min (Fv : Fm = 0.7, Fig. 5). At high irradiance this reduction in Fv : Fm would decrease photosynthesis only marginally (e.g. at Q = 1500 µmol m−2 s−1 by a maximum of 1%), whereas at low irradiance (Q = 50 µmol m−2 s−1) the same reduction in Fv : Fm led to a 30% reduction in photosynthesis. The same reduction in Fv : Fm would also decrease photosynthesis in coppice leaves under low irradiance, but the effect is somewhat smaller (10% reduction at Q = 50 µmol m−2 s−1). Photoinhibition at Fv : Fm = 0.7, would decrease C gain during a sunfleck only marginally (Table 4).


Response of gas exchange to simulated sunflecks

The photosynthetic response of N. cunninghamii to simulated sunflecks was rapid – around 90% complete within 3 min (Fig. 3, Table 2). This is amongst the fastest reported responses (photosynthetic induction time constants, t90%A) for tropical and temperate shade-tolerant woody angiosperms (between 3 and 37 min, Naumberg & Ellsworth, 2000).

The fast response of N. cunninghamii is all the more remarkable given ambient Q prior to simulated sunflecks was less than 20 µmol m−2 s−1 (coppice leaves) or 50 µmol m−2 s−1 (upper canopy leaves). These are at the lower end of ‘shaded’ irradiances reported in other studies (Naumberg & Ellsworth, 2000) and are a consequence of the southerly aspect and deeply incised topography of the study site. Hemispherical photographs show that the understorey light environment is similar to many other closed forests (Watling et al., 1997), insofar as leaves are highly shaded much of the time but this deep shade is irregularly punctuated by sunflecks of high irradiance (Fig. 1).

The fast response of photosynthesis to sunflecks can be of major significance to carbon balance (Pearcy, 1987) and thus may be important for the regeneration of N. cunninghamii underneath shaded canopies. In accordance with results in other ecosystems (Pearcy, 1987; Watling et al., 1997), we estimate that direct irradiance accounts for c. 40% of the daily C gain of leaves at both canopy levels. Since most high intensity sunflecks in the understorey are of short duration (in the range of minutes, Fig. 1) at the investigated site, induction processes with response times of > 10 min would contribute little to efficient light use in situ.

The high induction state of N. cunninghamii leaves, even under overcast conditions (Table 2), means that errors incurred by assuming immediate induction are smaller than in other species with lower induction states. In N. cunninghamii, carbon gain during a 10-min sunfleck would be overestimated by c. 50% in coppice leaves and 17% in upper canopy leaves if immediate induction were assumed, much less than the 86.5% overestimation reported for tropical understorey species by Stegemann et al. (1999). With greater background irradiance prior to sunflecks, the induction state would increase (Pearcy, 1987; Stegemann et al., 1999), and any potential overestimation of carbon gain would decrease still further.

In N. cunninghamii there was little evidence of stomatal control of the photosynthetic response to simulated sunflecks (Fig. 6) – by contrast to results reported for many other species (Pepin & Livingston, 1997; Allen & Pearcy, 2000a; Naumberg & Ellsworth, 2000). Large variation in stomatal response times have, however, been reported (Allen & Pearcy, 2000a; Naumberg & Ellsworth, 2000) but in N. cunninghamii stomatal responses were small. The absence of a large stomatal response is almost certainly due to consistently high stomatal conductance, even under low light conditions. The high gs is due to partially open stomata rather than cuticular water loss because cuticular conductance was approximately an order of magnitude less than gs shade (Table 2). Our results are not unprecedented: in the wet season in seasonally dry tropical rainforests, gs of shade leaves was similarly high and photosynthetic induction was also independent of the very small stomatal responses (Allen & Pearcy, 2000a,b). The high and variable gs shade (see SDs in Table 2), suggests poor stomatal control in N. cunninghamii– an assertion supported by a previous study showing that gs was highly variable and not related to variations in ambient vapour pressure deficit of water (compare Fig. 1 in Hovenden & Brodribb, 2000).

Photosynthetic induction was faster in upper canopy leaves than coppice leaves (Table 2). Stomatal responsiveness is an unlikely cause because there was little evidence that stomatal responses governed photosynthetic induction in general. We did not investigate relative contributions of biochemical and stomatal limitations in detail (cf. Allen & Pearcy, 2000a), but the steady-state A-ci curves (Fig. 2) suggest that the faster response of upper canopy leaves was due to biochemical factors (higher photosynthetic capacity). There are few within-species comparisons of photosynthetic responses to sunflecks in sun vs shade foliage and results from species differing in shade tolerance are inconsistent. It has been suggested that in shade-tolerant species, stomata open more rapidly than in shade-intolerant species, but this view was not supported by experimental data on conifers (Pepin & Livingston, 1997) or conifers and angiosperms (Naumberg & Ellsworth, 2000). Alternatively, plants adapted to conditions of higher evaporative demand may benefit from fast stomatal responses that afford improvements in water use efficiency under fluctuating light (Allen & Pearcy, 2000b).

Leaf pigmentation and morphology

Leaf morphology and the composition of chloroplast pigments exhibited by N. cunninghamii leaves varied between upper canopy and coppice leaves, but variation was smaller than in other species. Upper canopy leaves of N. cunninghamii have lower specific leaf area, more chlorophyll per unit area, and more carotenoids (per mol total chlorophyll) than coppice leaves. This result is consistent with results from many other tree species (Larcher, 2003) including other Nothofagus species (Niinemets et al., 2004). The almost 2-fold variation of the xanthophyll cycle pigments from c. 50 mmol mol−1 chlorophyll in coppice leaves to c. 100 mmol mol−1 in upper canopy leaves is comparable with some northern hemisphere broadleaved species (e.g. five species from laurel forests, a cloud and fog adapted ecosystem; González-Rodríguez et al., 2004), but considerably less than the 4-fold increase from less than 30 mmol mol−1 chlorophyll to close to 150 mmol mol−1 reported for many other forest species (Niinemets et al., 2003). Differences in leaf morphology and pigmentation are possibly ‘driven’ (Niinemets et al., 2003) by the gradient in Q from the top to bottom of the canopy. One obvious explanation for small differences in morphology and pigments is a small gradient in Q. The upper canopy leaves of N. cunninghamii studied here were themselves frequently shaded by the sparse, yet much taller, overstorey of Eucalyptus regnans. Hence, total daily Q decreased from upper canopy leaves to coppice leaves by c. 50%, a far smaller reduction than the up to 30-fold decrease measured across some forest canopies (Niinemets et al., 2003). Q received by an individual leaf is also determined by shoot architecture and this can ‘compensate’ for gradients in Q. We observed that leaf angle and the shoot orientation in the upper canopy of N. cunninghamii are close to vertical, while coppice leaves are nearly horizontal. While this observation has not been quantified in the present study, recent measurements on N. cunninghamii (Kern et al., 2004) and other Nothofagus species in New Zealand (Niinemets et al., 2004) showed the effect of such differences in canopy morphology on interception of Q by individual leaves.

Photoinhibition during simulated sunflecks

Nonphotochemical quenching (NPQ), that is the dissipation of absorbed light energy as heat (Baroli & Niyogi, 2000; Maxwell & Johnson, 2000), responded quickly to simulated sunflecks, as has been demonstrated for tropical rainforest species (Logan et al., 1997; Watling et al., 1997). Effective quantum yield of PS II electron transport (ΦPSII) dropped initially to near zero, which was likely due to simultaneous closure of available reaction centres. Thereafter, effective PS II quantum yield slowly increased as photosynthetic induction proceeded (Fig. 4, compare Watling et al., 1997). The majority of the change in NPQ is probably due to pH-dependent quenching, which engages and reverts quickly (Maxwell & Johnson, 2000). Our data indicate higher maximal photosynthetic capacity (Fig. 2) and, correspondingly, a slightly slower engagement of nonphotochemical quenching in upper canopy leaves. After 10 min of high light exposure c. 50% of the absorbed quanta were unaccounted for by either constitutive energy dissipation, electron transport, or variable energy dissipation – which would include photoprotective mechanisms and photodamage (calculations after Logan et al., 1997). This fraction may result in an accumulation of excitation energy in the pigment bed and lead to the formation of triplet chlorophyll or singlet oxygen. If not detoxified by carotenoids or tocopherols, these compounds can lead to the inactivation of reaction centres (Logan et al., 1997).

Chlorophyll fluorescence studies on tropical understorey plants suggest sunflecks cause photoinhibition (defined as a persistent decline in maximum PSII quantum efficiency, Watling et al., 1997), and we observed such photoinhibition in N. cunninghamii. However, the extent and rate of recovery varied between upper canopy and coppice leaves. In the upper canopy the reduction in PS II efficiency at the end of a 10-min sunfleck was sustained for > 30 min in the dark. In coppice leaves, reductions in PS II efficiency were effectively reversed within this same period. The reduction in PS II efficiency during the high light period was accompanied by rapid zeaxanthin formation leading to a de-epoxidation state of the xanthophyll cycle of c. 40% in both upper and lower canopy – similar to that observed in tropical understorey species during high intensity sunflecks (Watling et al., 1997). As with the changes in PSII quantum efficiency, the recovery of xanthophyll de-epoxidation state to presunfleck values was slower in upper canopy leaves than in coppice leaves. We did not observe persistent decreases in Fv : Fm once the de-epoxidation state of the xanthophyll cycle had reverted to presunfleck values, which suggests xanthophyll-cycle-related quenching was solely responsible for the more persistent photoinhibition in the upper canopy. This contrasts with observations by Watling et al. (1997), who reported persistent reductions in Fv : Fm that were unrelated to de-epoxidation state. The authors suggested that the photoinactivation of reaction centres (‘photodamage’) might have been responsible for this effect. By comparison, N. cunninghamii successfully avoided such photodamage during simulated sunflecks and the observed modulated efficiency was related to de-epoxidation state. We speculate that differences in enzyme activities (lower epoxidase or high de-epoxidase activities) or in the required regulation factors (ascorbate, thylakoid pH-gradient; Demmig-Adams, 2003) are responsible for the observed differences in the xanthophyll cycle between coppice and upper canopy leaves. Prolonged maintenance of low photochemical efficiency following a simulated sunfleck seems a sound rationale for upper canopy leaves that are more likely to experience longer uninterrupted periods (up to hours) of full sun than the coppice leaves, which rarely experience Q > 1000 µmol m−2 s−1 for > 10 min (compare Fig. 1).

Effect of photoinhibition on carbon gain

A recent theoretical study suggested that reduced photochemical efficiency (photoinhibition) may have a significant effect on leaf carbon balance (Zhu et al., 2004) and this is likely also the case in N. cunninghamii. The reduction in Fv : Fm observed in the present study would have little effect on carbon gain of leaves under high irradiances. On the other hand, carbon gain would be markedly reduced under lower irradiances (deep shade). At both canopy levels, we estimated that carbon gain from diffuse irradiance accounted for more than 50% of the total C gain on a clear spring day. Diffuse Q reached up to 85 µmol m−2 s−1 at upper canopy leaves, whereas it rarely exceeded 20 µmol m−2 s−1 at coppice level. At such Q, photosynthesis would be reduced by 14% in coppice leaves and by c. 18% in upper canopy leaves. Photoprotection may be more important than optimal light use efficiency in upper canopy leaves, although it comes at a considerable cost (Zhu et al., 2004).


The photosynthetic response of N. cunninghamii to sunflecks is amongst some of the faster photosynthetic responses reported and is consistent with our hypothesis that N. cunninghamii responds quickly to sunflecks. There was no evidence that stomata were controlling this photosynthetic response. Instead, stomatal conductance remained high even in deep shade. There was thus a small stomatal response to simulated sunflecks. High stomatal conductance in the shade generally reduces water-use efficiency, but increases the efficiency of sunfleck use. Hence, poor drought tolerance and the ability to survive and regenerate under a shaded canopy may well be functionally related to high stomatal conductance in the shade in this species.

We expected that coppice leaves would have faster sunfleck responses than upper canopy leaves, because of putative differences in the relative importance of sunflecks for daily carbon gain. However the reverse was the case – photosynthesis of upper canopy leaves responded more rapidly to simulated sunflecks than that of coppice leaves. This was related to the higher photosynthetic capacity of upper canopy leaves, and might relate to the lack of difference between upper canopy and coppice leaves in the relative importance of sunflecks for daily carbon gain (Table 4).

Photoprotective mechanisms (xanthophyll cycle) led to a more persistent decline in photochemical efficiency after the simulated sunfleck in upper canopy leaves than in coppice leaves. No persistent photoinhibition (decrease in photochemical efficiency) was observed in coppice leaves after simulated sunflecks, suggesting that fast recovering pH- and xanthophyll-related photoprotection mechanisms prevented photodamage to the photosynthetic apparatus and ensured high efficiency for subsequent shade periods. In upper canopy leaves there was a more persistent decline of maximum quantum efficiency and roughly modelled estimates suggest this could decrease photosynthesis by more than 30% during a shade period after a sunfleck.


We thank Parks Victoria for the permission to conduct this research and Frank Jones for technical assistance in the field. Financial support from the Australian Research Council is gratefully acknowledged.