Stomatal and non-stomatal limitations on leaf carbon assimilation in beech (Fagus sylvatica L.) seedlings under natural conditions

Limitations to diffusion and biochemical factors affecting leaf carbon uptake were analyzed in young beech seedlings (Fagus sylvtica L.) growing in natural gaps of a beech-wood at the southern limit of the species. Half of the seedlings received periodic watering in addition to natural rainfall to reduce the severity of the summer drought. Plant water status was evaluated by measuring predawn water potential. Basic biochemical parameters were inferred from chlorophyll fluorescence and photosynthesis-CO2 curves (A-Cc) under saturating light. The curves were established on three dates during the summer months. The main variables studied included: stomatal and mesophyll conductance to CO2 (gs and gm respectively), maximum velocity of carboxylation (Vcmax) and maximum electron transport capacity (Jmax). The gm was estimated by two methodologies: the curve-fitting and J constant methosds. Seedlings withstood moderate water stress, as the leaf predawn water potential (Ψpd) measured during the study was within the range –0.2 to –0.5 MPa. Mild drought caused gs and gm to decrease only slightly in response to Ψpd. However both diffusional parameters explained most of the limitations to CO2 uptake. In addition, it should be highlighted that biochemical limitations, prompted by Vcmax and Jmax, were related mainly to ontogenic factors, without any clear relationship with drought under the moderate water stress experienced by beech seedlings through the study. The results may help to further understanding of the functional mechanisms influencing the carbon fixation capacity of beech seedlings under natural conditions.

or maximum rate of electronic transport, J max ) might differ between well-watered and water-stressed plants (Flexas et al., 2006;Grassi et al., 2009).For an accurate estimation of V cmax and J max , we need to account for the mesophyll conductance of CO 2 .Indeed, any change in the estimation of V cmax and J max could modify the way that models, such as that from Farquhar et al. (1980), are applied, and their outcomes in process-based modelling from leaves to ecosystems (Bernacchi et al., 2002;Ethier and Livingston, 2004;Keenan et al., 2010).
Under conditions of low water availability in the soil or atmosphere, plants first trigger mechanisms aimed to minimize water loss.Of these, stomatal closure is one of the most extensively studied and widely recognized (Chaves et al., 2002;Brodribb and Jordan, 2008).However, the stomatal control of water loss incurs a penalty, since CO 2 diffusion into the leaf is concomitantly limited, leading to reduced carbon uptake potential (Wilson et al., 2000a;Aranda et al., 2000;Medrano et al., 2002,).In addition, mesophyll conductance of CO 2 can become non-negligible and impair carbon fixation during drought periods (Flexas and Medrano, 2002;Niinemets et al., 2004;Warren, 2006).Though it has been postulated that g m and g s respond to the same environmental variables and in a similar manner (Flexas et al., 2008), the mechanistic linkage between both types of diffusive conductance is unclear, as is their impact on other functional processes such as the water use efficiency (Hanba et al., 2003).While a decrease of g s and g m under water stress has been reported under controlled conditions (Galmés et al., 2007;Galle et al., 2009), these responses have been less studied in seedlings of forest tree species in natural environments.

Introduction
Water scarcity is recognized as one of the main environmental factors limiting leaf CO 2 fixation, and in turn growth and yield in plants (Chaves, 1991).The principal limitations to carbon uptake operate at the leaf level, which represents the main control point in the process of carbon fixation by plants.Although great advances have been made since pioneering studies there are some uncertainties that remain in our understanding of how the factors limiting CO 2 fixation are modulated (Grassi and Magnani, 2005;Diaz-Espejo et al., 2007).For instance, the importance of CO 2 diffusion from the leaf inter-cellular spaces into the chloroplast and its effect on photosynthesis (see Flexas et al., 2008 for a comprehensive review) has only recently been recognized.Technical advances in the measurement of gas exchange and fluorescence, and isotopic techniques, have provided more-accurate means to assess successive resistances across the entire CO 2 diffusion pathway through the leaf, prompting their importance for carbon uptake to be reconsidered (Flexas et al., 2002;Ennahli and Earl, 2005;Warren, 2006).In this context, it is important to elucidate the changes in biochemical factors and diffusion resistances during photosynthesis when plants are submitted to naturally stressful conditions, such as drought (Niinemets et al., 2005;Galmés et al., 2007;Flexas et al., 2009).This dual limitation influences the potential of seedlings to maintain a positive leaf carbon balance, and should be accounted for when assessing the ultimate consequences of water stress on ecological succession and niche partitioning under sub-Mediterranean environments (Kunstler et al., 2005;Robson et al., 2009).Moreover, the basic physiological parameters that drive the process of carbon uptake (i.e.maximum velocity of carboxylation, V cmax , Limitations to carbon fixation in leaves of beech seedings physiology since the 1990's (Madsen, 1994;Tognetti et al., 1994;Fotelli et al., 2001;Leuschner et al., 2001).However, only recently has the importance of an increased risk of drought across large areas of the species range started to be considered (Leuzinger et al., 2005;Geßler et al., 2007;Granier et al., 2007), as extreme weather events have become more common at sites which were historically unperturbed by drought.Beech is known for its high sensitivity to water stress (Bréda et al., 2006 and references therein), but previous studies were focused mainly on stomatal closure as the main limitation to carbon uptake capacity at different scales; from the leaf (Backes andLeuscher, 2000, Aranda et al., 2002) to the ecosystem (Granier et al., 2000).However, the contribution of other non-stomatal factors on carbon balance in beech leaves is poorly understood (Epron et al., 1995;Warren et al., 2007;Montpied et al., 2009).
The main aim of this paper is to quantify the biochemical and diffusional limitations on leaf carbon assimilation by beech seedlings growing in natural gaps and exposed to two contrasting soil moisture regimes.We tested three hypotheses: that i) g m co-limits carbon uptake to a similar degree as g s under non-waterstressed conditions; ii) g s and g m decrease in response to moderate water-stress but at a different pace; iii) moderate drought involves a higher penalty on leaf carbon uptake incurred via an increase in CO 2 diffusion limitations through g s and g m rather than via biochemical limitations (e.g.decrease of V cmax and J max ).

Site Characteristics
The study was carried out in the beech-oak forest of Montejo de la Sierra (41°1'N 3°5'W 1,400 masl), composed of a mixture of temperate and sub-Mediterranean broadleaved tree species.The forest is at the southwestern limit of European beech (Fagus sylvatica L.) distribution in Europe, and it is subjected to moderate drought.The site has previously been described in detail (Aranda et al., 2000(Aranda et al., , 2002(Aranda et al., , 2005;;Rodríguez-Calcerrada et al. 2008a,b, 2010;Robson et al., 2009).

Experimental design
Two-year old beech seedlings were randomly selected in the spring of 2009 from a plantation of beech nuts carried out in the winter of 2007, in three plots in natural gaps created by fallen canopy trees.Each plot was split into two 1.3 × 2 m sub-plots, and each subplot was randomly assigned to either natural rainfall (D) or natural rainfall plus periodic watering (WW).Watered plants (WW) were separated from their unwatered counterparts (D) by a 0.4 m un-watered buffer zone.Irrigation started on June 27 th .It consisted on adding 40 L water per m 2 of ground area every 7-10 days, and it finished two days before the last sampling date at the middle of August.Additional rainfall events were recorded during the summer months until the beginning of July, afterwards rainfall was almost absent (see Robson et al., 2009, andFigure 1a in Rodríguez-Calcerrada et al., 2010 for more details on the design and watering regime).Hemispherical photographs were taken during late summer, when the overstorey trees were in full leaf, to characterise the light environment for seedlings at two points in each subplot.A Global Site Factor (GSF%) for each plot was calculated, using an atmospheric transmitivity to solar radiation of 0.8 and 0.1 diffuse:direct radiation (canopy analysis software Hemiview 2.1, Delta-T devices Ltd, USA).GSF is an indicator of light availability that ranges between 0 (full canopy closure) and 1 (full sun light).GSF was 0.43 ± 0.06 (10.3 ± 1.5 mol m -2 day -1 ).There was no difference in radiation received between dry and watered sub-plots (F 1,6 = 0.02, P = 0.890).

Gas exchange measurements
On three dates during the summer in June (18-20), July (16-18), and August (20-22), gas exchange and chlorophyll fluorescence were measured in four to six seedlings per treatment.One attached, first-flush and fully expanded leaf per plant receiving direct sunlight, was selected for measurements.In June, a failure in the chlorophyll fluorescence system precluded chlorophyll fluorescence measurement.
Light-saturated CO 2 assimilation rate (A) was measured using portable photosynthesis system equipped with a blue-red light source (LI-6400; Li-Cor Inc., NE; USA) under different CO 2 concentrations.Measurements were carried out at constant light of 1,200 mmol m -2 s -1 .This level of irradiance has been shown to be enough to saturate photosynthesis in leaves of beech seedlings in the field without eliciting photoinhibition (Aranda et al. ment time to a temporal window between 9:00 a.m. and 13:00 p.m. Afterwards it was not possible to reasonably maintain the target temperature in the chamber and, in addition, midday stomatal closure was observed in some of non-watered plants even though water stress was not very intense.Measurements were carried out during three consecutive days.After, allowing 15 minutes at 400 ppm CO 2 concentration (C a ) for gas exchange rates to stabilize, gas exchange were recorded over a range of intercellular CO 2 (C i ) resulting from changing the CO 2 supply in twelve steps from 50 to 1,800 ppm.The supply of CO 2 was reduced step-wise to the minimum value; then returned to 400 ppm again, and increased step-wise from that concentration to complete the A-C i curve at the high C a end.Five records were taken at each target CO 2 concentration when photosynthesis and transpiration showed a CV lower than 5%.This was usually reached after three to four minutes.
Chlorophyll fluorescence was measured simultaneously with gas exchange at each target CO 2 concentration for A-C c curves.Steady-state fluorescence (F s ) and maximum fluorescence (F m ') were measured, in the case of F m ' after applying a saturating pulse of actinic light.The photochemical efficiency of PSII (F PSII ) was then calculated according to Genty et al., (1989) and Kramer et al., (2004) as: The rate of electron transport through PSII (ETR) was calculated following Rosenqvist and van Kooten (2003) as: (13) A value of 0.85 for total leaf absorptance was assumed (Evans and Loreto, 2000), and a factor of 0.5 for the partitioning of light between the two photosystems (Laisk and Loreto, 1996).
A non-linear least squares fitting procedure was applied to measured A -C c curves, to estimate the maximum rate of carboxylation (V cmax ) and the light-saturated maximum rate of RUBP-regeneration-limited electron transport rate (J max ).Regression models were constructed according to equations from Farquhar et al., (1980), including mesophyll conductance and some other modifications (see von Caemmerer, 2000) in which A n was modelled as the minimum value of Rubisco-limited (A c ) and RuBP-limited (A j ) rate of photosynthesis according to (1), ( 2), (3), and without 2002).Leaf temperature was maintained close to 25°C (actual leaf temperature: 25.8 ± 0.1°C) by controlling the temperature of the chamber.This constrained measure-Limitations to carbon fixation in leaves of beech seedings considering limitation by triose phosphate regeneration (TPU) which takes place only under very high C i .
R d is the mitochondrial respiration in light.The concentration of oxygen (O) was considered 20 kPa.Temperature-dependent parameters K c (Michaelis-Menten coefficient of Rubisco for CO 2 ) and K o (Michaelis-Menten coefficient of Rubisco for O 2 ) and the CO 2 compensation point in the absence of mitochondrial respiration in light (Γ*) were calculated following the equations derived by Bernacchi et al., (2002).All the parameters estimated were recalculated to a standard temperature of 25°C (Sharkey et al., 2007).
We used an application for Microsoft Excel developed by Sharkey et al., (2007) for calculation of photosynthetic parameters.This application implements the curvefitting method to iteratively calculate mesophyll conductance (g m ; see Warren 2006 andFlexas et al., 2008 for a comprehensive review on the methodologies to estimate g m ).The reliability of the method was checked by comparing the values of g m in July and August according to the curvature method, with those estimated from chlorophyll fluorescence measurements in parallel to gas exchange, which allowed g m to be estimated by the J constant method (Harley et al., 1992;Warren, 2006;Flexas et al., 2008).Because the estimation of g m is sensitive to errors in both R d and G* (Harley et al., 1992), we used the R d at the leaf temperature given from empirical relationships between R d and temperature obtained in a parallel experiment on the same plants (Rodríguez-Calcerrada et al., 2010).
By taking into account the mean values of those variables involved in the Farquhar et al., (1980) leaf photosynthetic model, which after modification include mesophyll conductance (Harley et al., 1992), Grassi and Magnani (2005) developed a method to evaluate the limitations to photosynthesis during a plant's vegetative period by the amount, activity and kinetics of Rubisco (eq 1).They partitioned the decline of optimum photosynthesis by three main limitations: these are stomatal limitation (S L ), mesophyll-conductance limitation (MC L ) and biochemical limitation (B L ).In turn, these parameters can be subdivided into the contribution of each relative limitation to the recorded difference from the reference value.The relative limitations are identified as: stomatal limitations (l s ), mesophyll limitations (l mc ) and biochemical limitations (l b ).The complete mathematical formulae and full theoretical development of the model are given by Grassi and Magnani, (2005).We compared drought-treatment plants (D) with watered ones (WW) on each measurement date to circumvent any seasonal effect on the different parameters, and to better assess the role played by our water treatments irrespective of ontogenic influences.

Water potential and soil moisture
A pressure chamber (PMS Instrument Co. 1000, Corvallis, USA) was used to take measurements of leaf water potential.These were carried out on the same leaves previously used for A-C i curves.The same leaves were kept hydrated for twelve hours, and used to estimate specific leaf mass per area (LMA), and nitrogen content on a per mass basis (N m ) by the Kjeldhal method after oven drying.Nitrogen content on a leaf area basis (N a ) was estimated from LMA and N m .
Volumetric soil moisture was measured at 10 and 30 cm depths several times during the summer months using a Time Domain Reflectometer (TDR, Trase System I, Soil Moisture Equipment Corp., Santa Barbara, USA).Soil moisture was recorded in two well-separated points in each sub-plot within the three main plots (n = 6).

Statistical Analysis
The effect of drought and time during the season on each physiological parameter was tested using a twoway analysis of variance.All computations were performed in Statistica 6.0.The pair-wise comparison between drought treatments on each date was tested by a post-hoc F test (LSD test).Linear regression models and Pearson correlation were used to analyse the relationships between variables.

Climatic conditions and water status of seedlings
Temperature and relative humidity were moderate during the course of the experimental period.Tem-perature seldom reached above 25°C and the maximum VPD was never higher than 1.5 kPa.These represent moderate climatic conditions during summer months in central Spain, since much higher temperatures and evaporative demands have been encountered in previous years at the same site (e.g.Aranda et al., 2002, Aranda et al., 2004).
Soil moisture measured at 20 cm depth followed a very different pattern between treatments.In the subplots receiving additional water the soil moisture was within the range 15-20%, whereas in the subplots receiving just natural rainfall, soil moisture decreased to around 7.5% by the middle of August (Figure 1).However, the Ψ pd remained similar between seedlings in the two treatments, and the water stress endured should only be considered moderate for the three dates (Ψ pd over -0.5 MPa on average).Seedlings receiving additional water only attained a significantly higher Ψ pd than in the un-watered plots on the last sampling date (T).The slightly higher Ψ pd in WW treatment in August than on previous dates could be explained by the lowest VPD during the night immediately prior to that predawn water potential measurement.Otherwise, the absence of larger drops in the Ψ pd at the end of summer for D seedlings, even though soil moisture reached the minimum value at this time, may be explained by deeper rooting of seedlings to below the depths where soil moisture was recorded.Overall, the dry period of the summer was not sufficiently intense to elicit decreases in predawn water potential as large as those reported in previous studies at the same stand in other years (Aranda et al., 2001, Aranda et al., 2002, Robson et al., 2009).

Gas exchange
There were no clear differences between treatments in A n and g s (measured at ambient 400 ppm) during June and July.In July as much A n or g s were higher than in June, despite water status at dawn in both treatments was slightly worst.Only on the last date (August), there was a tendency for both parameters to decrease in those seedlings enduring the natural rainfall regime (D), compared with those receiving additional water (WW: Table 2).However, differences were not statistically significant when considering all dates and treatments.
Both g s and g m were related to the draw-down of CO 2 from the air (C a ) to the interior of leaf (C a -C i ) or chloroplast (C i -C c ) (Table 2).The largest drop from ambient CO 2 (C a = 400 ppm) to that in the intercellular spaces of leaves (C i ) was caused by g s (range 154 ± 14 to 123 ± 11).The resistance to diffusion from the intercellular spaces into the chloroplasts (g m ) also promoted a decrease in CO 2 concentrations, albeit lower (C i -C c : range 51 ± 7 to 105 ± 17).The g m estimated by the fitting-curve method always gave higher values than the g m * estimated by the method of the J constant.In both cases, the trend was to maintain higher values of g m than of g s of CO 2 (Table 2).
The decrease in g s from July to August had a greater effect on IWUE than the concomitant changes in g m .This suggests a large effect of stomatal regulation on water use efficiency, reaching into the range of water stress endured by plants.This expectation was consistent with the negative relationship between g s and IWUE, whereas there was not a clear relationship between IWUE and g m (Figure 2) nor with the biochemical variables influencing the carbon uptake potential, V cmax and J max (data not shown).However, this relationship should be viewed cautiously as estimated IWUE and g s are not independent.
There were small seasonal changes in those parameters driving the uptake of carbon within choloroplasts (V cmax and J max ), with the lowest values tending to occur on the last sampling date in August, even for plants receiving additional water and with good water status (Ψ pd ~ -0.2MPa).However, non statistical differences A weak negative relationship was observed between g s and g m , and Ψ pd (g s = 0.10 + 0.06 Ψ pd , R 2 = 0.23, P = 0.07; g m = 0.18 + 0.13 Ψ pd , R 2 = 0.27, P = 0.08).However, neither V cmax nor J max showed any relationship with Ψ pd , suggesting that leaf-age could have had a greater effect in driving changes in biochemical regarding diffusional parameters (Table 2).In this respect, as much V cmax as J max were positively correlated with the nitrogen content on a leaf area basis (N a ) (Figure 3), and not with Ψ pd .
Following the approach of Grassi and Magnani (2005), by comparing the relative effect of dry conditions (D plants) as a proportion of normal WW plant traits, we observed that stomatal limitations increased significantly in July and even more so in August.Values of S L accounted for 51 and 59% and MC L for 6 and 11% of photosynthetic down-regulation, in July and August respectively; while the all rest was due to biochemical limitations B L of 43 and 30%.So in August, the reduction of 27% in net photosynthetic rate was mainly due to diffusional limitations (S L + MC L contributed 70% of this).The relative contribution of each single limitation to net photosynthesis was 33% by the stomatal limitation (l s ), for all treatments with the exception of water stress treatment on the last date; 22% from mesophyll related limitations (l mc ); and 45% from biochemical related limitation (l b ).

Diffusional limitations to photosynthesis under moderate water stress
The range of water stress beech seedlings had to cope with in 2009 was lower than during previous studies carried out in the recent past at the same site (minimum Ψ pd around -0.5 MPa), but nevertheless enough to prompt changes in the stomatal conductance to water vapour at the end of summer (Aranda et al., 2002).In this respect, the present study furthers our understanding of the role played by different diffusional limitations and biochemical variables during CO 2 assimilation by leaves of beech seedlings growing under natural environments.We provide new information on how these processes operate under natural forest conditions, which compliments and builds upon previous reports on the same species under semi-controlled conditions and environmental manipulations (e.g.Epron et al., 1995;Dreyer et al., 2001;Warren et al., 2007).The main factor constraining photosynthesis under moderate water stress was diffusional limitation through stomata, as previously reported (Aranda et al., 2002, 2004, Gallé and Feller, 2007;Robson et al., 2009).In the range of water stress endured by beech seedlings, stomatal closure comprised one of the main limiting factors to carbon uptake (Chaves et al., 2002;  Medrano et al., 2002).Water stress of Ψ pd = -0.5 MPa was enough to prompt significant partial stomatal clo-Table 2. Gas exchange variables estimated from A-Cc curves in leaves of seedlings growing in three canopy gaps.Half of the plants received natural rainfall during the summer months (D) and the other half had supplementary watering several times throughout the summer (WW).Mean values ± SE. are displayed (n = 3-6).Mesophyll conductance to CO 2 was estimated by the fitting curve method (g m* ) after Ethier et al. (2004) and the J constant method (g m ) following to Harley et al. (1992).Values were normalized to 25°C following the equations of Sharkey et al. (2007).Failure of the fluorescence system in June meant that g m could not be calculated.Significant differences are indicated by different letters (LSD-test after ANOVA )

June
July August
On other hand, it's difficult to explain the lower values of gas exchange recorded in June compared with July.Predawn water potential was slightly higher in June, though a slight effect of water stress at the late spring could not be ruled out as a possible cause of the low gas exchange maintained in WW and D plants.Other possible explanation could be leaves of seedlings in both treatments had not achieved the full physiological competence in the first measurement date.In fact, a similar result has been observed previously in beech where a seasonal lag in maximum gas exchange rates was observed despite leaf unfolding have been completed by the middle-end of June (Aranda et al., 2000).
The stomatal limitation to carbon uptake, even under moderate water stress, sums to a low mesophyll diffusion conductance to CO 2 .The low g m , common to woody plant species (Wilson et al., 2000a;Grasssi and Magnani, 2005), may be responsible for the low photosynthetic capacity of beech seedlings (Valladares et al., 2002;Aranda et al., 2004;Balandier et al., 2007).This finding agrees with the typically low photosynthetic capacity of shade-tolerant tree species.Accordingly, relative mesophyll limitation (l mc ) accounted for 22% of the relative photosynthesis limitation, a little lower than the 30% value proposed by Epron et al., (1995) using a different approach.On the other hand, stomatal resistance was a bit higher (33-40% in our case vs. 30% from Epron et al., 1995).In our case, g m was slightly higher than the g s , with values close to those previously reported by Epron et al., (1995), and following the same pattern as that observed by Warren et al., (2007) when comparing sun and shade leaves in mature trees.These results were consistent for changes in g m , and qualitatively similar whether the J constant or the curvature method was used to estimate g m .In conclusion, both components of diffusional limitation comprised a high percent of the overall limitation to carbon uptake.
It has been noted that as stress intensifies there is a reduction in the mesophyll conductance of CO 2 (Medrano et al., 2002;Chaves et al., 2003;Flexas et al., 2008).This increases the overall diffusional limitations imposed at the first step by stomata (Medrano et al., 2002).In the present study, where moderate water stress was suffered by beech seedlings, only a sea-sonal reduction in g s in response to water stress was evident.There was no consistent pattern in g m in response to the water stress imposed, except on the last date when the differences in Ψ pd between treatments were greatest.Taking into account the absolute limitations to photosynthesis, it is clear that the main reduction was due to diffusional resistances, mainly through the stomata, accounting for over 50% of the photosynthetic decrease.Under moderate water stress, stomatal limitation of photosynthesis is thought to be the main restriction on carbon uptake (Lawlor andCornic, 2002, Medrano et al., 2002;Grassi and Magnani, 2005;Díaz-Espejo et al., 2007;Grassi et al., 2009).

Limitations to carbon uptake imposed by biochemical factors and time of year
Beech seedlings had a low biochemical capacity for photosynthesis, as previously reported in a comparative study with other co-occurring species (Dreyer et al., 2001).In addition to the increase in diffusional limitations to carbon uptake imposed by water stress, there was also a seasonal reduction in seedlings' photosynthetic capacity (Wilson et al., 2000a;Balandier et al., 2007).V cmax and J max decreased slightly, though only significantly for V cmax , between July and August irrespective of watering and caused a reduction in the capacity to fix carbon.Seasonally-induced decreases in photosynthetic capacity by the end of summer in beech have been reported before (Balandier et al., 2007), and they add to the impairment of carbon uptake caused by increased stomatal limitation under natural conditions of moderate water stress.The trend in the degree of down regulation of V cmax and J max was similar between July and August in plants enduring the natural rainfall pattern and those receiving supplementary water, reinforcing the idea that this was an ontogenic effect.Accordingly, both variables showed a stronger relationship with the leaf nitrogen content (Balandier et al. 2007), than with the water stress experienced.The direct down-regulation of V cmax as consequence of the moderate drought has been reported elsewhere (Wilson et al., 2000b, Xu and Baldocchi, 2003, Damour et al., 2009), and although not statistically significant, there was also a tendency towards consistently lower values in D than in WW on the last two measurement dates in the present study.However, the ontogenic effect should be recognized as the most plausible reason for the decrease of both biochemical parameters in the present study.

Conclusions
The combination of shade tolerance at juvenile stages and a positive reaction to higher light levels makes beech very resilient in both wet and mesic sites, conditioning the ecology and silviculture of the species (see Wagner et al., 2010 for a comprehensive review).However, even moderate drought may change the competitiveness of the species when water is not limiting (Cornic, 1994).Thus, carbon uptake at the leaf level was compromised in seedlings by moderate soil moisture causing stomatal closure which prevailed as the main limitation to net photosynthesis under moderate water stress, what is a well-known fact.Nevertheless, internal conductance of CO 2 was also an important limitation to carbon uptake comprising a 22% of the total limitation to carbon assimilation.This diffusional limitation could continue to increase, like those related with biochemical parameters, in extremely dry years.
The high sensitivity of beech to just moderate water stress is clearly apparent from this and previous studies (Madsen, 1994;Aranda et al., 2004;Robson et al., 2009), but also underscores the importance that the relatively low g m has on the carbon potential uptake of beech leaves whichever water stress endured by seedlings.Ultimately this sensitivity could jeopardize the future of the species in currently marginal beech stands, where future climatic conditions are expected to worsen, and where an increase in the temperature together with a decrease in the seasonal rainfall could compromise much more the low carbon uptake capacity of young beech seedlings.

Figure 1 .
Figure 1.Seasonal changes in air temperature (°C, top panel), volumetric soil moisture content measured as average value between 10 and 30 cm depth (%, middle panel); white points (gap W) depict well-watered seedlings, black points (gap) refer to unwatered seedlings receiving only natural precipitation, and air vapour pressure deficit (KPa, bottom panel) during the summer months in a gap at the Montejo de la Sierra beechwood.

Figure 3 .Figure 2 .
Figure3.Positive relationship between basic leaf biochemical parameters: maximun rate of carboxilation, V cmax (µmolCO 2 m -2 s -1 ) and light saturating maximum rate of RUBP regeneration limited electron transport, J max (µmolCO 2 m -2 s -1 ), and nitrogen content on a leaf area basis (N a -g m -2 ).It's depicted a unique relationship for well watered (black points) and un-watered seedlings (white points).No clear trends were observed when considering watering treatments, thus a unique linear relationship was fitted to the pooled data (P < 0.05).

Table 1 .
Mean values ± SE for water status (Ψ pd , leaf predawn water potential) and leaf morphological traits (LMA, leaf mass per area; N m , nitrogen content on a leaf dry mass basis) measured on beech seedlings growing in three canopy gaps.Half of the plants received natural rainfall during the summer months (D) and the other half had supplementary watering several times throughout the summer (WW).Significant differences are indicated by different letters (LSD-test after ANOVA) were observed for treatments, being only significant for date when grouping data of both treatments within each date (after ANOVA).