stoichiometry of foliar carbon consistuents varies along light

Summary Foliar morphology and chemical compositionwere examined along a light gradient in the canopies of fivedeciduous temperate woody species, ranked according toshade-tolerance as Populus tremula L. < Fraxinus excelsior L.< Tilia cordata Mill. = Corylus avellana L. < Fagus sylvatica L.Foliar carbon was divided between structural (cell-wall poly-saccharides, lignin) and nonstructural (proteins, ethanol-sol-uble carbohydrates, starch) fractions. Foliar morphology of allspecies was strongly affected by irradiance. Both leaf dry massper area (MA), a product of leaf density and thickness, and leafdry to fresh mass ratio (Dw), characterizing the apoplastic leaffraction, increased with increasing relative irradiance (Isum,calculated as the weighted mean of fractional penetration ofdiffuse and direct irradiance). Though the relationships werequalitatively identical among the taxa, more shade-tolerantspecies generally had lower values of MA than shade-intolerantspecies, and their morphological relationships with irradiancewere curvilinear; however, there were no signs of saturationeven at the highest irradiances in shade-intolerant species. Inall species, lignin concentrations increased and cell-wall poly-saccharide concentrations decreased with increasing irradi-ance. Consequently, biomass investment in structural leafcomponents appeared to be relatively constant along lightgradients. The relationship between irradiance and structuralcompounds tended to be asymptotic in the more shade-tolerantspecies, whereas MA was linearly correlated with concentra-tions of structural leaf components, suggesting that similarfactors were responsible for the curvature in the morphologicaland chemical relationships with irradiance. Because ligninincreases tissue elastic modulus thereby rendering leaves moreresistant to low leaf water potentials, we conclude that changesin stoichiometry of cell wall components were related to foliageacclimation to the gradients of water deficit that develop in thecanopy and inherently accompany light gradients. We alsoconclude that increased lignification decreased leaf expansiongrowth, and that species differences in lignification were partlyresponsible for the observed interspecific variability in mor-phological plasticity.

Analysis of structural leaf compounds provided no indica-tion of how shade-intolerant species with low investments inlignin acclimated to gradients of water availability in the can-

opy. Because shade-intolerant species generally had highercapacities for photosynthesis than shade-tolerant species, wepostulated that they should also have a greater ability forosmotic adjustment of leaf water potential with photosyn-thates. The concentrations of soluble carbohydrates increasedwith increasing irradiance in all species; however, the osmoticadjustment achieved in this way was similar in all species,except for shade-intolerant F. excelsior, which had a lowerpotential for osmotic adjustment with carbohydrates than theother taxa. Although we did not determine whether the gradi-ents of stem water potential and leaf water deficits were similarin canopies of different species, we demonstrated that waterrelations play a central role in determining foliar structure andcomposition along light gradients in the canopy.

Keywords: foliar morphology, irradiance, lignin, nitrogen,nonstructural carbohydrates, osmotic adjustment, shade-tol-erance, structural carbohydrates, water requirement.

Introduction

Morphological modifications in foliar structure dominate theacclimation of foliar photosynthetic capacity per area (Amax

a ) tolong-term light conditions in woody canopies (Ellsworth andReich 1993, Niinemets 1996b, Niinemets and Tenhunen 1997,Niinemets et al. 1998). Numerous studies have shown thatAmax

a (Wallace and Dunn 1980, Chazdon and Field 1987, Wal-ters and Field 1987, Ellsworth and Reich 1992, Niinemets andTenhunen 1997) and the content of assimilative compoundsper unit area, e.g., nitrogen (Walters and Field 1987, Ellsworthand Reich 1993, Niinemets 1995, 1997b, Niinemets and Ten-hunen 1997), scale positively with irradiance. Because photo-synthetic capacity per mass (Chazdon and Field 1987, Waltersand Field 1987, Ellsworth and Reich 1992, 1993) and nitrogenconcentration (Chazdon and Field 1987, Walters and Field1987, Ellsworth and Reich 1993, Niinemets 1995, 1997b) tendto be relatively constant along light gradients, correlationsbetween Amax

a and growth irradiance are mainly attributable toan accumulation of photosynthesizing mass per area, reflectedin the positive relationship between leaf dry mass per area, MA,and irradiance (Wallace and Dunn 1980, Walters and Field1987, Ellsworth and Reich 1992, Kull and Niinemets 1993,

Stoichiometry of foliar carbon constituents varies along light gradientsin temperate woody canopies: implications for foliage morphologicalplasticity

ÜLO NIINEMETS and OLEVI KULL

Estonian Institute of Ecology, Riia 181, Tartu EE 2400, Estonia

Received March 17, 1997

Tree Physiology 18, 467--479© 1998 Heron Publishing----Victoria, Canada

Downloaded from https://academic.oup.com/treephys/article-abstract/18/7/467/1687125by gueston 11 February 2018

Wayne and Bazzaz 1993, Niinemets 1997b), rather than tophysiological adjustments. This phenotypic plasticity in MA

plays a principal role in whole-canopy carbon acquisition.Modeling studies demonstrate that canopy photosynthesis isoptimized if MA increases with increasing height in the canopy(Gutschick and Wiegel 1988, Baldocchi and Harley 1995)----abalanced response between the benefits of low MA at lowirradiance, resulting in a greater surface area for light capture,and the advantages of high MA at high irradiance, resulting inmore efficient quantum utilization in carbon assimilation.

Linear functions have been used to fit the relationship be-tween MA and irradiance in deciduous woody species (e.g.,Kull and Niinemets 1993, Niinemets 1997b). However, thereis evidence that morphological relationships with light may becurvilinear. For example, the dependence of MA on irradiancesaturates at high light in Acer saccharum Marsh. (Tjoelker etal. 1995) and Fagus sylvatica L. (Ducrey 1981), and leafthickness--irradiance relationships have an upper asymptote inAcer pseudoplatanus L., Tilia cordata Mill. and Fraxinus ex-celsior L. (Starzecki 1975). Witkowski and Lamont (1991)concluded that the relationship between MA and irradiance iscomplex, because MA is the product of leaf density and thick-ness, both of which may vary independently. Changes in leafthickness are the major response of leaf structure to lightgradients (Pieters 1974, Goryshina et al. 1979, Nobel andHartsock 1981, Oberbauer et al. 1987, Abrams and Mostoller1995, Niinemets and Kull 1995), but leaf density tends toincrease with increasing irradiance as well (recalculated fromthickness and MA: Goryshina et al. 1979, Oberbauer et al.1987, Abrams and Mostoller 1995). In contrast, studies incontrolled conditions show that leaf density is independent ofirradiance (Chabot et al. 1979), and the parameters responsiblefor differences in leaf density----average cell size (Pieters 1974,Dengler 1980, Körner and Pelaez Menendez-Riedl 1990) andintercellular air space volume (Chabot et al. 1979, Dengler1980)----are relatively constant at different irradiances. Otherenvironmental factors that vary inherently along natural lightgradients may explain the discrepancy between these field andlaboratory studies. For example, with increasing irradiance inthe canopy, air temperature increases and humidity decreases,resulting in a greater water vapor pressure deficit, (e.g., Eliáš1979, Chiariello 1984, Shuttleworth et al. 1985). Wind speedsalso increase with increasing height (Chiariello 1984). Conse-quently, the mean boundary layer conductance for water trans-fer increases with increasing irradiance, further augmentingevaporative demand and potential water stress. Covariationbetween environmental factors probably explains why bothdrought stress and increasing relative irradiance modify leafstructure in a similar manner. Jones (1985) hypothesized thatthe same physiological mechanism underlies both drought-and light-induced modifications in leaf structure and involveslowered leaf water potential. Expansion growth is the physi-ological process most readily affected by water stress (Hsiaoet al. 1985), and an increase in leaf density with increasingirradiance may be a direct consequence of enhanced waterstress during leaf development.

A plethora of adaptive alterations to increasing water defi-cits along light gradients in the canopy has been observed,

including greater investment of foliar biomass in supportingtissues (Myers et al. 1987), greater ability for osmotic adjust-ment (Myers et al. 1987, Oberbauer et al. 1987), and lowersymplasmic water fraction (Oberbauer et al. 1987); however,the cost of these alterations in terms of changing foliar chem-istry and photosynthetic potential has not been determined. InFagus sylvatica, leaves in the upper canopy have greater ligninconcentrations than shade leaves (Kausch and Haas 1966).Because increased lignification may increase resistance forCO2 transfer from leaf intercellular air spaces to the sites ofcarboxylation (Syvertsen et al. 1995), increased lignificationprobably curbs foliar photosynthesis per unit biomass invest-ment in leaves. Enhanced diffusive resistances to carboxyla-tion sites with increasing leaf thickness and MA may constrainphotosynthesis more than the increase in accumulation ofphotosynthetic machinery per unit area enhances it. Thus, wepostulate that foliar morphology--light relationships saturate athigh irradiances because the light-induced effects on structuraland chemical limitations for gaseous diffusion affect photo-synthesis in an opposite way to the light-induced effects onbiomass accumulation per unit area. From another perspective,the drawback of an enhanced investment of foliar biomass insupporting compounds is that it results in decreases in assimi-lative compounds and photosynthetic capacity per mass.

Little is known about how different species cope with waterstress gradients in the canopy or the significance of thesespecies-specific responses in forest development and succes-sion. There is evidence of lower values of MA in more shade-tolerant species than in less shade-tolerant species (Abramsand Kubiske 1990, Niinemets and Kull 1994). Because leavesof shade-intolerant species generally have higher nitrogen con-centrations and photosynthetic capacities per mass than leavesof shade-tolerant species (Küppers 1994, Reich et al. 1994,Niinemets 1997b, Niinemets and Tenhunen 1997), shade-in-tolerant species may have a greater potential for osmotic ad-justment of leaf water potential by means of photosynthates,and lower requirements for expensive investments in foliarcompounds, which increase resilience to low water potentials.An enhanced capacity for osmotic adjustment may also enablegrowth in leaf thickness to continue longer in shade-intolerantspecies than in shade-tolerant species. Thus, a fundamentaldifference between shade-tolerant and shade-intolerant speciesin how morphology--light relationships are affected by waterlimitations may underlie the low morphological plasticityoften encountered in shade-tolerant species (Fetcher et al.1983, Walters and Field 1987, Ducrey 1992).

We studied five temperate deciduous species differing inshade-tolerance to test the following hypotheses: (1) foliarchemical composition is affected along the natural light (andwater) gradients in the temperate woody canopy; (2) changingleaf chemistry is associated with foliar morphological plastic-ity; (3) relationships between foliar morphology, chemicalcomposition and irradiance depend on the shade tolerance ofspecies; and (4) there is a link between the tolerance of light-related water stress gradients across the canopy and the succes-sional position of a species.

468 NIINEMETS AND KULL

TREE PHYSIOLOGY VOLUME 18, 1998Downloaded from https://academic.oup.com/treephys/article-abstract/18/7/467/1687125by gueston 11 February 2018

Material and methods

Studied species

Populus tremula L. and Fraxinus excelsior are early succes-sional pioneer species, whereas Tilia cordata and Fagus sylva-tica are common members of late-successional temperateforests (Ellenberg 1988). The shade-tolerance of the studiedtrees (see Canham et al. 1994, Niinemets 1996b) increases inthe order: P. tremula < F. excelsior < T. cordata < F. sylvatica(Ellenberg 1988, Otto 1994). Corylus avellana L., a shrubspecies, is a dominant component of the understory of Esto-nian natural deciduous stands (Laasimer 1965), but it may alsocolonize large gaps. In Central Europe, C. avellana growsmostly at intermediate irradiances (Ellenberg 1991); however,in Estonian forests, it also grows in deep shade (Kull andNiinemets 1993) and tolerates low irradiances at least as wellas T. cordata (O. Kull and Ü. Niinemets, unpublished observa-tions).

There is evidence that T. cordata outcompetes F. sylvatica inmore xeric habitats, and C. avellana is frequently found inmore xeric habitats than T. cordata even though it has a lessdeep root system than T. cordata (Pigott and Pigott 1993).According to Ellenberg (1988), the common occurrence of thespecies from xeric to mesic sites is T. cordata > P. tremula >F. excelsior = F. sylvatica. Populus tremula, F. excelsior andT. cordata have similar drought tolerances when growing inmesic habitats (Pigott 1991, Pigott and Pigott 1993), andF. excelsior and C. avellana often co-dominate in Estonianalvar forests on thin rendzinas, which may suffer severedroughts during the growing season (Laasimer 1975).

Experimental sites and foliar sampling

In mid-September 1991, F. sylvatica was sampled in a forestdominated by Picea abies (L.) Karst. at Oberwarmensteinach(49°59′ N, 11°47′ E; elevation about 760 m a.s.l.), Fichtelge-birge, Germany. The stand was located on a plateau-like crestof a hill with acid (pHCaCl2

in the topsoil 3.2--3.8, base satura-tion 4--8%, and C/N ratio 18.7) podsolic and brownpseudopodsolic soils formed on phyllite (Hantschel 1987,Türk 1992). Mean total height ( ± SE) of the sampled trees(n = 7) was 7.8 ± 1.0 m, and age, estimated from incrementcores at 1.3 m, was 28.3 ± 2.3 years. The highest mean relativesampling height (height in the tree per total height) was 0.88 ±0.02. Further details of foliage sampling in this species aregiven in Niinemets (1995).

Fraxinus excelsior, P. tremula and C. avellana were studiedin August 1994 in a naturally established mixed deciduousstand near Ülenurme (58°18′ N, 26°42′ E, elevation about60 m a.s.l.), Estonia. The forest was 15--18 m high, and inaddition to P. tremula, Betula pendula Roth. dominated theoverstory. F. excelsior and Sorbus aucuparia L. were sub-canopy species, and C. avellana, F. excelsior and Padus aviumMiller dominated the understory. Where the overstory was lessdense, as a result of thinning about 10 years ago, the largestindividuals of C. avellana (about 8 m) and F. excelsior (about12 m) had also reached the upper canopy. The soil, formed ona reddish-brown sandy clay moraine, was a sandy loam

pseudopodsol with pHKCl of 3.9--4.6, base saturation of 42--60% and C/N molar ratio of 9.4--9.7 in the A horizon (0--26 cm). There was no litter layer. Sampled trees of F. excelsior(n = 10) were 7.2 ± 1.1 m tall and 22.0 ± 4.8 years old, sampledtrees of P. tremula (n = 4) were 16.3 ± 0.6 m tall and 26.5 ± 3.4years old. For C. avellana, which has a polycormic growthhabit, the largest stems of the sampled shrubs (n = 5) had amean height of 6.0 ± 0.7 m and were 18.8 ± 2.3 years old. Amobile lift was used for foliage sampling. The highest sam-pling height per total tree height was 0.90 ± 0.07 in F. excel-sior, 0.95 ± 0.03 in P. tremula, and 1.0 in C. avellana.

Foliage of T. cordata was collected in August 1994 nearTartu (58°15′ N, 26°45′ E, elevation about 60 m a.s.l.), Estoniain a planted mixed stand. The overstory consisted of T. cordataand Picea abies, which grew widely apart, such that more lightpenetrated the canopy than in the other stands. There was nowoody understory; however, the herb cover, dominated byDactylis glomerata L., Phleum bertolonii D.C. and Poapratensis L. was dense (height 0.7--1.0 m, coverage 100%).Compared with the other stands, the soil, which was a brownpseudopodsol formed on a sandy clay moraine, had a thicker(about 45 cm) and a less acidic humus horizon with pHKCl of6.0--6.5, base saturation of 91--98% and C/N ratio of 18.3.Studied trees (n = 4) were 15 ± 2 m tall and 55 ± 5 years old.The highest samples were taken from the tops of the trees.

Foliage was sampled between 1500 to 1600 h on cloudydays from the south side of the canopy in F. sylvatica, andbetween 0900 to 1100 h with no special regard to weatherconditions and compass direction from the canopies of the fourother species. In all cases, five to 12 leaves were taken persampling location, and three to seven canopy locations per treewere analyzed.

Relative irradiance of the sampling locations

Long-term light conditions were determined by a hemispheri-cal photographic technique (Anderson 1964, Rich et al. 1993)as modified by Nilson and Ross (1979). Several hemisphericphotographs were taken from sample locations immediatelyafter foliage collection. Diffuse site factor (Idif, proportion ofpenetrating diffuse solar radiation of open sky) was calculatedfrom the measurements of relative area of canopy gaps foruniformly overcast sky conditions. Direct site factor (Idir, pro-portion of potential penetrating direct solar radiation of opensky) was calculated from canopy gap fractions along solartracks during 2-month intervals from the summer solstice.Both Idif and Idir calculations took account of cosine of inci-dence effects. The fraction of penetrating irradiance in thephotosynthetically active spectral region, Isum, was found as:

Isum = pdifIdif + (1 − pdif)Idir, (1)

where pdif is the fraction of diffuse irradiance of total irradi-ance in the photosynthetically active spectral region (400--700 nm) above the canopy. The value of pdif depends onlong-term cloudiness as well as on the differences in spectralquality of light of diffuse and direct irradiance, and was takento be 0.58 for the Ülenurme and Tartu stands (Tôravere Mete-

LEAF MORPHOLOGY AND CARBON INVESTMENT IN DECIDUOUS TREES 469

TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.comDownloaded from https://academic.oup.com/treephys/article-abstract/18/7/467/1687125by gueston 11 February 2018

orological Station, Estonia, 58°16′ N, 26°28′ E). In F. sylva-tica, where the sample points were always facing south anddata for pdif were missing, Isum was used as an estimate oflong-term irradiance (relative irradiance). Because Idif and Idir

were strongly correlated (r2 = 0.78, P < 0.001 for the wholematerial), the conclusions were not qualitatively altered byusing different light estimates.

Foliar chemical composition

Foliar carbon (Cm) and nitrogen (Nm) concentrations wereestimated with an elemental analyzer (F. sylvatica: Model1500, Carlo Erba, Italy; other species: CHN-O-Rapid, FossHeraeus GmbH, Hanau, Germany). Concentration of non-structural carbohydrates (ethanol-soluble carbohydrates plusstarch) was determined by anthrone reaction as described pre-viously (Niinemets 1995, 1997a). Ash content was estimatedafter combustion of the sample in a muffle furnace at 500 °Cfor 3 h, and mineral content (M) was calculated by assumingthat the fraction of minerals in ash equaled 0.67 (Vertregt andPenning de Vries 1987). The contribution of ethanol-solublecarbohydrates to leaf osmotic potential (Ψ) was calculatedaccording to van’t Hoff’s equation:

ΨESC = −cESC DwσRT, (2)

where cESC is soluble carbohydrate concentration (kmol kg−1),Dw is leaf dry to fresh mass ratio (kg kg−1, apoplastic waterfraction is assumed to be nil), σ is water density (kg m−3), R isthe universal gas constant and T is absolute temperature (avalue of 293 K was used).

The concentrations of structural carbohydrates and acid-in-soluble lignin were determined by a gravimetric method. Thesamples were extracted with 100% acetone, 96% ethanol, andethanol:benzene solution (1:2, v/v) at 40 °C to remove acid-in-soluble compounds. The extraction solution was mixed fre-quently, and the solvent replaced every hour. The extractionprocess was repeated four times with each solvent, or until thesolution remained colorless. The residue (1) was dried at 70 °Cfor 24 h and weighed. Acid hydrolysis was carried out accord-ing to Effland (1977), the acid-resistant residue (2) wasweighed after drying at 70 °C for 24 h, and lignin concentrationwas calculated after correcting for mineral content. Structuralcarbohydrate concentration was found as the mass differencebetween the residues (1) and (2) minus starch and insolubleprotein concentrations (Dijkstra and Lambers 1989). The lattervalue was the difference between total and soluble proteins.Total protein concentration was calculated as 6.25Nm, assum-ing that all foliar nitrogen was in proteins. Soluble proteinswere determined in fresh leaf material according to Bradford(Bradford 1976, Robinson 1979) as described in Niinemets etal. (1998). Only lignin concentration was determined in F. syl-vatica.

Foliar morphology and parameters for NSC-free leaf drymass

In F. sylvatica, projected leaf area was measured with a videoarea meter (DIAS, Delta-T Devices, Cambridge, U.K.). In the

other four species, leaf circumference was traced with a com-puter digitizer (QD-1212, QTronix, Taiwan) and projected areacalculated. The two methods gave similar results. Foliage drymass was determined after drying to a constant mass at 70 °C.

Parameters for NSC-free dry mass were computed as de-scribed by Niinemets (1995, 1997a).

Data analysis

Correlation and regression analyses were used to test the influ-ence of irradiance on foliar morphology and chemistry. Forsimplicity, linear techniques were used where possible. Be-cause area per leaf showed a strongly skewed distribution,natural logarithmic transformation was applied to normalizethe distributions before statistical analysis. In several species,irradiance affected foliar structure and chemical compositionin a nonlinear manner, and so these relationships were fitted bysecond-order polynomial regressions.

Interspecific differences in foliar structure and chemicalcomposition were analyzed by one-way analysis of variance.Significant species influences were separated by Bonferronitest. When the parameter was significantly altered by irradi-ance according to the regression analysis, irradiance was in-cluded in the statistical model as a covariate. Because thesampled light ranges differed among the species (e.g., P. tre-mula possessed no leaves below an Isum of 0.19, and C. avel-lana did not reach an Isum greater than 0.79), only the data foran Isum range of 0.17--0.82 were included in the covariationanalysis. This routine also eliminated the difficulties associ-ated with the nonlinear effects of light on foliar structure,because the parameters were effectively linear over the trun-cated light range (cf. Tsutakawa and Hewett 1978). When theslopes did not differ between species (detected as the interac-tion term), a common-slope ANCOVA model was used.Throughout the paper, the term ‘at common irradiance’ de-notes intercept differences. Covariation analysis was also usedto compare parameters between species at a common MA; allvalues were included in these statistical models. However,these comparisons should be interpreted with caution, becausethe ranges for MA differed among species (see Figure 1). Allstatistical tests were considered significant at P < 0.05.

Results

Foliar morphology and dry matter content versus irradiance

In all species, MA was positively correlated with irradiance(Figure 1). The relationship tended to saturate at high irradi-ances, and second-order polynomial regressions gave gener-ally higher r2 values than linear fits. The dependence of MA onirradiance was most strongly curved in T. cordata, where MA

saturated at an Isum of about 0.5. Conversely, there were nosigns of MA saturation even at the highest irradiances in P. tre-mula (Figure 1). Populus tremula had a lower slope (P < 0.001)but larger MA than the other species over the whole range ofirradiances. Thus, species ranking according to MA at commonirradiance (the other species were compared by a common-slope ANCOVA followed by Bonferroni test) was: P. tremula> F. excelsior > T. cordata = C. avellana > F. sylvatica (Ta-



ble 1) was inversely related to their shade-tolerance ranking.Generally, the proportion of explained variance was highest inthe relationships with Isum (Equation 1). However, considera-tion of diffuse irradiance alone gave a slightly lower proportionof explained variance (r2 = 0.86) than Isum in F. excelsior,whereas direct site factor gave a better fit with MA (r2 = 0.95)than either the diffuse or combined site factor in C. avellana(cf. Figure 1).

In all species, Dw scaled positively with irradiance, parallel-ing the MA--irradiance relationships (Figure 1). Again, thisrelationship was strongly curved in T. cordata (Figure 1). Atcommon irradiance (intercept differences), species ranking onthe basis of Dw was: P. tremula > C. avellana > F. sylvatica >T. cordata = F. excelsior (Table 1). The relationship betweenMA and Dw was linear in all species (Table 2) and the r2 valueswere always higher than for the Dw--irradiance relationships(cf. Figure 1 and Table 2). When Dw values were compared atcommon MA, the ranking was: F. sylvatica > C. avellana >

P. tremula = T. cordata > F. excelsior.In T. cordata, area per leaf (S) decreased with increasing Isum

(r2 = 0.26, P < 0.01), but it was independent of light conditionsin the other species. Leaves of C. avellana and P. tremula wereof similar size (30.2 ± 1.0 cm2) and were significantly (P <0.001) larger than leaves of the other taxa (19.9 ± 1.0 cm2; inF. excelsior, leaflet area was used in the comparisons).

Relationships between irradiance and nonstructural foliageconstituents

Concentrations of ethanol-soluble carbohydrates (ESC,mostly mono- and oligosaccharides) increased with increasingirradiance in all species (Figure 2, upper five panels). Based onESC at common irradiance, the ranking of the species was:T. cordata > P. tremula = C. avellana > F. sylvatica > F. excel-sior (Table 1). The contribution of soluble carbohydrates toleaf osmotic potential (Equation 2) was negatively related toirradiance in all species (r2 values ranged from 0.55 to 0.73,

Figure 1. Dependence of leaf drymass per area (MA, filled sym-bols) and leaf dry to fresh mass ra-tio (Dw, open symbols) on relativeirradiance in five temperate de-ciduous woody species. In F. syl-vatica, the fraction of penetratingdiffuse solar irradiance was usedas an estimate of relative irradi-ance; in the other four species,relative irradiance was calculatedas the weighted mean of fractionalpenetration of diffuse and direct ir-radiance (Equation 1).

Table 1. Interspecific differences in foliar structural and morphological variables: results of one-way ANCOVA (species as the main effect,irradiance as the covariate). Parameters with the same letter are not significantly different (Bonferroni test, P > 0.05).

Leaf parameter1 Mean ± SE

P. tremula F. excelsior C. avellana T. cordata F. sylvatica

Dry mass per area (MA)2 0.102 ± 0.004a 0.0788 ± 0.006b 0.0534 ± 0.005c 0.0759 ± 0.005c 0.0302 ± 0.0023dDry to fresh mass ratio (Dw) 0.422 ± 0.008a 0.336 ± 0.011d 0.392 ± 0.014b 0.356 ± 0.008d 0.365 ± 0.009cTotal nitrogen (Nm) 0.0215 ± 0.0003b 0.0182 ± 0.0004c 0.0205 ± 0.0007b 0.0238 ± 0.0004a 0.0242 ± 0.0005aEthanol-soluble carbohydrates (ESC) 0.125 ± 0.007b 0.0711 ± 0.0022d 0.104 ± 0.006b 0.153 ± 0.008a 0.101 ± 0.005cPart of osmotic potential from ESC (ΨESC) --0.72 ± 0.05ab --0.321 ± 0.017c --0.53 ± 0.05b --0.79 ± 0.05a --0.49 ± 0.04abStructural carbohydrates (SP) 0.361 ± 0.006b 0.476 ± 0.009a 0.324 ± 0.010c 0.303 ± 0.012c nd3

Lignin (Lm) 0.185 ± 0.005d 0.122 ± 0.006e 0.238 ± 0.007b 0.192 ± 0.008c 0.266 ± 0.003aStructural carbon compounds (SC, SP + Lm) 0.546 ± 0.005b 0.560 ± 0.005a 0.562 ± 0.005b 0.494 ± 0.006c nd

1 All parameters are in kg kg−1, except for MA (kg m−2) and ΨESC (MPa, Equation 2). The means are for the whole set of data, but only the valuesfor an Isum range of 0.17--0.82 were included in the covariation analyses to account for different light ranges for different species, as well as toeliminate the influence of nonlinearity (Figures 1 and 3) on statistical calculations.

2 Interaction term was significant, indicating that P. tremula had a lower slope than the other species (Figure 1); however, it had greater MA overthe whole relative light range than the other species (see text).

3 nd = Not determined.



P < 0.001 for all). The ΨESC plateaued at high irradiances inT. cordata, but the relationship was linear for the other species.Fraxinus excelsior had lower ΨESC than the other four species(Table 1). The similarity in ΨESC among the four species wasthe result of low symplasmic leaf fractions (characterized by1/Dw) in species with low ESC (cf. Figures 1 and 2, upper fivepanels).

The relationships between ST and irradiance differed quali-tatively among the species. Starch concentration increasedwith increasing irradiance in P. tremula and C. avellana, butdecreased in F. excelsior and T. cordata, and was independentof irradiance in F. sylvatica, where ST was lower than in theother species over most of the light range. However, similarly

low ST was reached in F. excelsior and T. cordata at highirradiances (Figure 2, upper five panels). Despite decliningstarch concentrations with increasing irradiance in T. cordata,total nonstructural carbohydrate concentration (NSC, ESC +ST) was positively related to Isum (r

2 = 0.19, P < 0.05), becausethe increase in ESC exceeded the decrease in ST. In F. excel-sior, NSC decreased with increasing irradiance (r2 = 0.20,P < 0.05). Species ranking according to NSC at common irra-diance was similar to the ranking based on ESC (F. excelsior,which exhibited the opposite slope to the other species, was notincluded in the ANCOVA).

With increasing irradiance, Nm increased in P. tremula anddecreased in F. sylvatica (Figure 2, lower three panels), but was

Table 2. Linear dependencies of foliar structural and nonstructural constituents on leaf dry mass per area (MA, kg m−2) in four temperate deciduouswoody species. Asterisks: * = P < 0.05, ** = P < 0.01, *** = P < 0.001, and ns = not significant.

Variable1 Populus tremula Fraxinus excelsior Corylus avellana Tilia cordata

Intercept Slope r2 Intercept Slope r2 Intercept Slope r2 Intercept Slope r2

Dw 0.239*** 1.79*** 0.77 0.213*** 1.56*** 0.85 0.228*** 2.70*** 0.90 0.222*** 1.97*** 0.90Nm 0.0188*** 0.0261ns 0.10 0.0165*** 0.0221ns 0.15 0.0196*** 0.0172ns 0.02 0.0239*** 0.00246ns 0.00ESC --0.0277ns 1.50*** 0.67 0.0577*** 0.171* 0.24 0.0535*** 0.941*** 0.56 0.0424* 1.50*** 0.63ST --0.0250* 0.850*** 0.72 0.0873*** --0.305* 0.38 0.0373*** 0.572*** 0.55 0.101*** --0.509*** 0.37NSC --0.0526ns 2.26*** 0.78 0.145*** --0.147ns 0.08 0.0907*** 1.51*** 0.69 0.143*** 0.990** 0.35SP 0.481*** --1.17*** 0.56 0.555*** --0.999*** 0.51 0.415*** --1.71*** 0.65 0.463*** --2.15*** 0.72Lm 0.109*** 0.747** 0.34 0.0594*** 0.789*** 0.64 0.175*** 1.18*** 0.73 0.0737*** 1.59*** 0.65SC 0.590*** --0.424ns 0.14 0.720*** 0.178ns 0.09 0.590*** --0.524* 0.23 0.537*** --0.562* 0.22

1 All units and symbols as in Table 1, except for ST (starch) and NSC (total nonstructural carbohydrates, ESC + ST; all in kg kg−1).2 The MA relationships for F. sylvatica are reported in Niinemets (1995) except for Lm ( y = 0.241 + 0.818x; r2 = 0.33, P < 0.01).

Figure 2. Changes in foliar mo-bile carbon components alongthe light gradient in the canopy.Upper five panels: ethanol-sol-uble carbohydrate (ESC, filledsymbols) and starch concentra-tions (ST, open symbols).Lower three panels: nitrogenconcentration (Nm). Speciessymbols as shown in the upperfive panels. All of the nitrogenwas assumed to be present inproteins, and crude protein con-tent was calculated as 6.25Nm.



independent of irradiance in the other species. When nitrogenconcentration was calculated per unit NSC-free dry mass, itwas positively related to irradiance in C. avellana (r2 = 0.21,P < 0.05) and P. tremula (r2 = 0.66, P < 0.001). Although therelationships between nitrogen and irradiance were poor, therewas considerable variability in Nm among and within species(Figure 2, lower three panels). Assuming that all nitrogen waspresent in proteins and that proteins contain 53.5% carbon(Vertregt and Penning de Vries 1987), proteins made up a largefraction of leaf nonstructural carbon. For the whole material,proteins contained on average 15.5 ± 0.2% (range 11.1--21.1%) of total leaf carbon. At common irradiance and MA,species ranking on the basis of Nm was: F. sylvatica = T. cor-data > P. tremula = C. avellana > F. excelsior (Table 1).

With the exception that Nm was independent of MA in P. tre-mula, and NSC was independent of MA in F. excelsior, allrelationships between mobile leaf compounds and MA weresimilar (Table 2), indicating that MA is a good descriptor oflight climate (Figure 1).

Dependence of structural leaf components on irradiance andleaf morphology

In all species, there was a negative correlation between Lm andstructural polysaccharide (SP, cellulose and hemicellulose)(r2 = 0.76 for the whole material, range 0.48--0.79, P < 0.001for all), because the stoichiometry of foliar structural compo-nents was affected by irradiance. Lignin concentration in-creased and SP decreased with increasing relative irradiance inall species (Figure 3). Consequently, the sum of SP and Lm

(SC) was independent of irradiance in T. cordata and F. excel-sior, and there was a slight negative correlation between SCand irradiance in P. tremula and C. avellana (Figure 3). How-ever, when the carbon investment in structural leaf compoundswas calculated taking the carbon concentration of hardwoodlignin as 63.3% (calculated from Nimz 1974) and that ofstructural carbohydrates as 44.4% (based on cellulose as themodel compound), structural carbon investment was inde-pendent of irradiance in all species (P > 0.1). The dependencies

of SP and Lm on irradiance were generally linear, curving onlyin T. cordata (Figure 3). All relationships of structural com-pounds with MA were linear (Table 2), suggesting that struc-tural chemicals scaled with foliar morphology rather thandirectly with irradiance.

At common MA and irradiance, shade-tolerant species hadhigher values of Lm than shade-intolerant species; the speciesranking in this respect was: F. sylvatica > C. avellana >T. cordata > P. tremula > F. excelsior (Figure 3, Tables 1and 2), whereas the ranking was reversed for structural poly-saccharides (Tables 1 and 2). Interspecific differences in foliarbiomass allocation to structural compounds were the result ofvariable investment patterns in both lignin and SP. At commonIsum and MA, the biomass investment in structural compounds(SC, Lm + SP) was larger in shade-intolerant species, rankingas F. excelsior > P. tremula = C. avellana > T. cordata (Ta-ble 1). This was similar for the structural carbon investment,except that it was significantly higher in C. avellana than inP. tremula (P < 0.001).

In all species, dry matter content of the leaves was positivelyrelated to Lm (r2 values ranged from 0.29 to 0.70, P < 0.02),except for F. sylvatica (r2 = 0.17, P > 0.08), and negativelyrelated to structural polysaccharides (r2 values from 0.39 to0.62, P < 0.005). As a result of the opposing trends in Lm andSP with Dw, dry matter content was independent of totalbiomass investment in foliar structure in all species, except forT. cordata (r = --0.46, P < 0.02).

Carbon investment in structural versus nonstructural leafcompounds

Species-specific investment patterns in structural compounds(Figure 3, Tables 1 and 2) had major influences on concentra-tions of assimilative chemicals such as proteins. There was atrade-off in investment between the structural and assimilativefoliar chemicals. The species with the highest proportions ofleaf resources in structural polysaccharides had the lowestconcentrations of foliar nitrogen (r2 = 0.52, P < 0.001 for thewhole material). Although the correlation between Nm and Lm

Figure 3. Effects of irradianceon cell-wall constituents. Con-centration of leaf structural carb-on components (SC, shadedsymbols) was calculated as thesum of cell-wall polysaccharides(SP, open symbols) and lignin(Lm, filled symbols). The orderof regression equations followsthat of the curves.



was slightly positive (r2 = 0.22, P < 0.001), Nm and structuralleaf components (Lm + SP) were negatively related (r2 = 0.68,Figure 4).

Discussion

Foliar morphology versus irradiance

In temperate deciduous species, foliar morphology versus irra-diance relationships measured along a canopy light gradientsaturate at high irradiances (Figure 1; cf. Starzecki 1975,Ducrey 1981, Tjoelker et al. 1995). Ducrey (1981) observedthat MA saturates at lower irradiance in more shade-tolerantF. sylvatica than in shade-intolerant Quercus sessiliflora L. exLiebl., and we found that shade-tolerant species had low plas-ticity at high irradiances (Figure 1). In contrast, the nonlinear-ity of light versus leaf morphology dependencies is lesspronounced in both shade-tolerant and shade-intolerant spe-cies when the changes in MA are followed across forest habitatsdiffering in exposure (Niinemets 1996a, 1997b) or in labora-tory and field experiments with artificial shading and frequentirrigation (Chabot et al. 1979, Gulmon and Chu 1981, Nobeland Hartsock 1981, Wayne and Bazzaz 1993). This discrep-ancy suggests that growth in leaf thickness in the upper canopyat high irradiance is limited by environmental stress factorsassociated with the light gradients. Several environmental pa-rameters are inherently correlated with irradiance in the can-opy. For example, a positive relationship between evaporativedemand and irradiance in the canopy is consistently observed.However, other environmental factors may not necessarily becorrelated with irradiance across different forest habitats (seeAbrams and Mostoller 1995). For example, soil drought chro-nologies may differ between understory and open habitats(Bazzaz and Wayne 1994), but the gradient in water availabil-ity along the canopy occurs at a constant soil water potential.Such differences may explain why foliar morphology respondsdifferently to light gradients in the canopy under differentenvironmental conditions. Although the main effect of reducedwater availability is a reduction in leaf area growth, growth inleaf thickness is also inhibited by low soil water potentials(Nobel 1977). Thus, reduced water availability at high irradi-

ance may be partly responsible for the curvature in the irradi-ance versus leaf morphology relationships.

Relationships between irradiance and nonstructural foliageconstituents

Because carbon translocation from leaves is relatively constantduring the day (Huber and Huber 1996) and irradiance andphotosynthesis are positively related, ESC depends on lightconditions within the canopy (Servaites et al. 1989, Takahashiet al. 1993, Niinemets 1995, 1997a). Correlations betweenNSC and irradiance also alter MA and the concentrations ofother plant metabolites (Chatterton et al. 1972, Niinemets1995, 1997a). We found that r2 values for most relationshipswere improved by accounting for the dilution effects of NSCon leaf chemicals. For example, in C. avellana, Nm was inde-pendent of Isum when expressed per unit total dry mass, but wassignificantly related to irradiance if expressed on an NSC-freedry mass basis.

Nitrogen concentrations were relatively constant in thecanopies of all species (Figure 2, lower three panels) (cf.Chazdon and Field 1987, Walters and Field 1987, Ellsworthand Reich 1993, Niinemets 1997b), indicating that variabilityin MA with irradiance in the canopy (Figure 1) is a moreimportant factor for photosynthetic acclimation to growth irra-diance than changes in leaf chemistry, because the relativeconstancy of Nm assures that nitrogen per area scales linearlywith MA. Nevertheless, there was a tendency for shade-tolerantspecies to possess higher values of Nm at low irradiances(F. sylvatica, Figure 2, lower three panels; Niinemets 1997b),and for shade-intolerant species to possess higher values of Nm

at high irradiances (P. tremula, Figure 2, lower three panels;Chazdon and Field 1987, Niinemets 1997b). This variabilityamong species----reflecting the balance between foliage pro-duction and the cost of foliage in terms of nitrogen, and thedifferences in nitrogen investment in light versus dark reac-tions of photosynthesis----was particularly pronounced in acomparative study of habitats differing in exposure (Niinemets1997b).

Relationship between NSC and photosynthetic capacity

Interspecific variability in the carbohydrate pool size (Fig-ure 2, upper five panels) may result from differences in photo-synthetic capacity among species. Shade-intolerant speciesgenerally have higher photosynthetic rates per dry mass(Chazdon and Field 1987, Reich et al. 1994), and greater poolsof nonstructural carbohydrates (Niinemets 1997b) than shade-tolerant species. However, T. cordata, one of the most shade-tolerant species studied, was an exception to this rule. Amongthe species examined, T. cordata had the highest NSC and ESCas well as a larger Nm than the three less shade-tolerant species.However, Niinemets et al. (1998) found that, at constant sto-matal limitation, T. cordata and C. avellana have lower photo-synthetic capacity per mass than P. tremula. Similarly, atmoderate soil water deficits, F. excelsior has higher photosyn-thetic capacity both at common growth irradiance and com-mon leaf thickness than the more shade-tolerant speciesT. cordata (recalculated from Starzecki 1975). Based on these

Figure 4. Relationship between the concentrations of leaf nitrogen andstructural carbon components (sum of cell-wall polysaccharides andlignin). Data for four species pooled.



estimates of photosynthetic capacity and the ranking of speciesaccording to ESC as T. cordata > P. tremula = C. avellana >F. sylvatica > F. excelsior (Table 1), we conclude that interspe-cific variability in photosynthetic capacity was not responsiblefor the observed species differences in soluble carbohydrates.Moreover, the decline in starch concentrations with increasingirradiance in T. cordata and F. excelsior (Figure 2, upper fivepanels) does not support the hypothesis that NSC concentra-tions are primarily determined by photosynthetic production.

Osmotic adjustment and interspecific differences in solublecarbohydrates

An enhanced requirement for osmotica for osmotic adjustmentof leaf water potential at high irradiances may provide analternative explanation for the observed variability in ESC,because nonstructural carbohydrates form a major part of leafosmotica (Morgan 1984) and their concentrations increase inresponse to water stress, particularly at high irradiances(Munns and Weir 1981, Blom-Zandstra and Lampe 1985). InF. sylvatica, starch concentrations were low over the wholeirradiance range and similar to those in T. cordata and F. excel-sior at high irradiance (Figure 2), indicating that low starchconcentrations may be related to osmotic regulation of waterpotential in these species. Starch degradation in response towater stress may account for the accumulation of sucrose thattypically occurs in drought-stressed leaves (Fox and Geiger1986). However, despite the differences in NSC, ΨESC (Equa-tion 2) varied little among the species, suggesting that thecarbohydrate contribution to osmotic adjustment was similaramong the studied species, except for F. excelsior which hadlower values. The contribution of other osmotica may havediffered among the species; although given that neutral osmo-tica such as sucrose do not alter the ionic medium for theenzymatic reactions and are therefore more advantageous thanions (Lawlor and Leach 1985, Lawlor 1995), this seems un-likely.

Evidence of greater water stress in leaves of shade-tolerantspecies compared with shade-intolerant species comes fromdata on stomatal conductance, which is a sensitive index ofwater availability (Schulze and Hall 1982, Schulze et al. 1987).The ratio of intercellular to ambient CO2 concentrations(Ci/Ca) calculated from the photosynthetic measurements at aCa of 325 µmol mol−1 and saturating light, was highest inC. avellana, lowest in T. cordata and intermediate in P. tre-mula, suggesting that photosynthesis was most limited bystomata in T. cordata (Niinemets et al. 1998). High Ci/Ca inC. avellana resulted from high stomatal conductances and lowphotosynthetic rates at low irradiances. Stomatal limitation ofphotosynthesis was similar in C. avellana and T. cordata athigh irradiances. The light (and water) gradients across thecanopy also altered Ci/Ca to a greater extent in C. avellana andT. cordata, where it decreased with increasing Isum, than inP. tremula, where it was independent of Isum (Niinemets et al.1998).

Foliar biomass investment in structural leaf constituentsversus irradiance

Kausch and Haas (1965, 1966) found that, in F. sylvatica, theproportion of lignin in cell wall substances is greater in sunleaves (47.1%) than in shade (42.3%) leaves, and that shadeleaves have higher concentrations of cell-wall polysaccharidesin total dry mass. Increasing irradiance also resulted in in-creased lignin concentrations in Salix (Waring et al. 1985) andBoquila trifoliata (Steubing et al. 1979).

In many species, there is a negative relationship betweenleaf thickness and the fraction of epidermal leaf tissue (F. syl-vatica: Lichtenthaler 1985, Lonicera maackii (Rupr.) Maxim.:Luken et al. 1995, 26 mesic woody species: Philpott 1956,Lolium varieties: Charles-Edwards et al. 1974, Helianthusannuus L.: Dengler 1980 and Fragaria virginiana Duchesne:Chabot et al. 1979, Jurik et al. 1982). An increased fraction ofmetabolically less competent epidermal structures may partlyaccount for the increase in cell-wall polysaccharide concentra-tion at low irradiance (Figure 3).

In contrast, the proportion of leaf tissues in vascular ele-ments increases with increasing irradiance (Charles-Edwardset al. 1974, Jurik et al. 1982). Of the studied species, this hasbeen documented for C. avellana (recalculated from Gries etal. 1987). Changes in fractional biomass allocation to vasculartissues are probably caused by disproportionally increasingrequirements for water conduction in leaves at high irradi-ances. In contrast with our results, Vance and Zaerr (1991)reported that shading had no effect on foliage lignin concen-tration and resulted in a lower cellulose concentration in Pinusponderosa, suggesting that other factors co-varying with irra-diance may have altered the stoichiometry of foliar structuralcompounds in our study. There is evidence that mechanicalstress caused by increased wind speeds also modifies foliagecomposition by increasing the investment in structural tissues(Grace and Russell 1977, Woodward 1983, Retuerto andWoodward 1992).

Influences of changing stoichiometry of structural foliarcompounds on leaf water relations

We assessed the significance of positive relationships betweenfoliar lignin concentration and irradiance across the canopy(Figure 1) in terms of tissue mechanical properties. Tissuemechanical properties vary along the light gradient in thecanopy. Tissue elastic modulus (ε) near full hydration (definedas turgor change per fractional change of the mass of symplas-mic water; Koide et al. 1989) increases with height in thecanopy (Oberbauer et al. 1987), and it also tends to scalepositively with MA (Abrams et al. 1994). Because high ε resultsin large changes in tissue water potential with small changesin water content in symplasms, water potential gradients andwater flow can be maintained with proportionally less variablecell volume and at lower leaf water deficits. Lignificationincreases the physical strength of the cell walls (Wallace andFry 1994), but also renders cell walls less permeable to water,thereby reducing the extent to which cellulose fibrils canhydrate (Preston 1974). This decreases the effects of fluctuat-ing apoplastic water contents on cell-wall elastic modulus



(Niklas 1991). Thus, enhanced lignification should result in aless variable ε, and improve tissue resistance to low waterpotentials.

Leaf water content is generally greater at low irradiancesthan at high irradiances (Figure 1, Bourdeau and Laverick1958, Bongers et al. 1988, Ingestad and McDonald 1989,Stewart et al. 1990). This correlation may result from a lowerapoplastic water fraction caused by increased lignification athigh irradiance, although increased desiccation stress may alsodiminish water content. We detected a positive relationshipbetween leaf water content and cell-wall polysaccharides, anda negative relationship with lignin concentration. Increased Dw

with increasing Lm is compatible with the role of lignin indecreasing cell wall hydration; however, the observed correla-tions may also result from covariation of Dw and investmentpatterns in structural compounds with irradiance.

Increased lignification in leaves should be more importantin species that suffer from severe water stress at a commonevaporative demand. Because shade-intolerant species havehigher photosynthetic capacities than shade-tolerant species,their potential for osmotic adjustment is also larger and mayallow for a smaller investment in compounds responsible formechanical strength of tissues.

Although F. excelsior had the lowest lignin concentrationsof the species studied, it had the highest total investment incell-wall components. High investment in cell-wall formingchemicals is compatible with reduced volume of symplasmand may increase ε (Salleo 1983, Lawlor and Leach 1985).Quantitative adjustment costs more in terms of carbon, andalso results in a lowered fraction of assimilative structures inleaves compared with qualitative adjustment (Figure 4).

Although enhanced lignification resulted in mechanicallymore resistant leaves, it probably increased diffusive resistanceto CO2. High internal resistance from the intercellular space tothe site of ribulose-1,5-bisphosphate (RUBP) carboxylation(rias) are correlated with leaf structure, and increase with in-creasing leaf density and thickness (Epron et al. 1995, Syvert-sen et al. 1995). Niinemets et al. (1998) have shown that, inT. cordata, rias is likely to increase with MA and Isum. Thenegative effect of increased lignin concentration on CO2 trans-fer conductance may mean that osmotic adjustment of waterpotential is more beneficial in terms of carbon capture thanelastic acclimation.

Lignification versus MA and light relationships

Expansion growth depends on cell-wall extensibility and ef-fective turgor pressure (turgor minus yield threshold term, Y;Dale 1988). The finding that Y is larger in shade-tolerantspecies than in shade-intolerant species (Taylor and Davies1986) is in keeping with their higher lignin concentrations(Figure 3). Although it is possible to compensate for increasedturgor requirements for cell expansion through osmotic adjust-ment of water potential (Morgan 1984, Dale 1988), thismechanism is inherently constrained in shade-tolerant speciesbecause of their low photosynthetic capacities. Interspecificdifferences in Y and the ability to adjust osmotically may notbe important early in the growing season, when stored carbo-

hydrates and nutrients are being transported to the leaves, thesoil is close to field capacity, and the developing water stressalong the light gradient in the canopy is small. Because epider-mal cell division ceases earlier than mesophyll cell division(Dengler 1980, Dale 1988, Körner and Pelaez Menendez-Riedl 1990), leaf size may be less affected by water availabilitygradients in the canopy than leaf thickness. During the periodof major changes in leaf thickness later in the season, thegradients in water availability in the canopy become moreaccentuated, and growth in leaf thickness is more constrainedas a result of increased cell-wall lignification in response towater stress. A positive relationship between Lm and Y providesan explanation for the curvilinear relationships of MA versusirradiance, and for the interspecific variability in these depend-encies.

Interplay between drought tolerance and shade tolerance

Because assimilation rates are generally less affected by de-creasing water availability in early successional pioneer spe-cies, which tend to be shade intolerant, than in shade-tolerantspecies (Bazzaz 1979, Bahari et al. 1985, Abrams and Mostol-ler 1995), an inverse relationship between shade tolerance anddrought tolerance has been hypothesized (Bazzaz 1979,Abrams 1994, Niinemets and Tenhunen 1997). In response toincreasing irradiance, our shade-tolerant species exhibited sev-eral modifications in foliar chemistry (high lignin and osmoti-cally active carbohydrate concentrations) and physiology(lower Ci/Ca), suggesting increased water limitation of photo-synthesis at high irradiances. However, a species ranking basedon their occurrence along a water availability gradient differedfrom that based on shade tolerance.

This apparent discrepancy may be associated with differingtolerances to low soil and leaf water potentials. Low soil wateravailability may favor species with low stomatal conductancesand deep root systems. Fluctuations in leaf water status duringthe day may not reflect soil water status, because water stresswill develop more quickly at high irradiances (Abrams et al.1992) when evaporative demand is higher, especially in shade-tolerant species with a low potential for osmotic adjustment.However, shade-tolerant species may become drought tolerantwhen grown at low irradiance, because water stress is likely todevelop more gradually, thereby allowing acclimation to soilwater depletion. Prunus serotina J.F. Ehrh. seedlings closedstomata later and developed lower osmotic potentials in shade,where water availability declined slowly, than in sun condi-tions (Abrams et al. 1992). Hardening in response to waterstress results in an increased capacity for osmoregulation andincreased tolerance of low water potentials (Morgan 1984).

Acknowledgments

This study was financed by the Estonian Science Foundation (Grants1199, 1597, 2048) and by the German Federal Minister of Researchand Technology (BMBFT, Grant BEO 51-0339476A). We thank AnneJôeveer (Tôravere Meteorological Station, Estonia) for the solar radia-tion data for the study area. Skilled technical assistance of Anne Aan,Sirje Kattel, Raja Kährik, Eve Niinemets, Asko Noormets and IgnaRooma is appreciated.



References

Abrams, M.D. 1994. Genotypic and phenotypic variation as stressadaptations in temperate tree species: a review of several casestudies. Tree Physiol. 14:833--842.

Abrams, M.D. and M.E. Kubiske. 1990. Leaf structural characteristicsof 31 hardwood and conifer tree species in Central Wisconsin:influence of light regime and shade-tolerance rank. For. Ecol.Manag. 31:245--253.

Abrams, M.D. and S.A. Mostoller. 1995. Gas exchange, leaf struc-ture and nitrogen in contrasting successional tree species growingin open and understory sites during a drought. Tree Physiol.15:361--370.

Abrams, M.D., B.D. Kloeppel and M.E. Kubiske. 1992. Ecophysi-ological and morphological responses to shade and drought in twocontrasting ecotypes of Prunus serotina. Tree Physiol. 10:343--355.

Abrams, M.D., M.E. Kubiske and S.A. Mostoller. 1994. Relating wetand dry year ecophysiology to leaf structure in contrasting temper-ate tree species. Ecology 75:123--133.

Anderson, M.C. 1964. Studies of the woodland light climate. I. Thephotographic computation of light conditions. J. Ecol. 52:27--41.

Bahari, Z.A., S.G. Pallardy and W.C. Parker. 1985. Photosynthesis,water relations, and drought adaptation in six woody species ofoak--hickory forests in central Missouri. For. Sci. 31:557--569.

Baldocchi, D.D. and P.C. Harley. 1995. Scaling carbon dioxide andwater vapour exchange from leaf to canopy in a deciduous forest. II.Model testing and application. Plant Cell Environ. 18:1157--1173.

Bazzaz, F.A. 1979. The physiological ecology of plant succession.Annu. Rev. Ecol. Syst. 10:351--371.

Bazzaz, F.A. and P.M. Wayne. 1994. Coping with environmentalheterogeneity: the physiological ecology of tree seedling regenera-tion across the gap--understory continuum. In Exploitation of Envi-ronmental Heterogeneity by Plants. Ecophysiological ProcessesAbove- and Belowground. Eds. M.M. Caldwell and R.W. Pearcy.Physiological Ecology, Academic Press, San Diego, pp 349--390.

Blom-Zandstra, M. and J.E.M. Lampe. 1985. The role of nitrate in theosmoregulation of lettuce (Lactuca sativa L.) grown at differentlight intensities. J. Exp. Bot. 36:1043--1052.

Bongers, F., J. Popma and S. Iriarte-Vivar. 1988. Response of Cordiamegalantha Blake seedlings to gap environments in tropical rainforest. Funct. Ecol. 2:379--390.

Bourdeau, P.F. and M.L. Laverick. 1958. Tolerance and photosyntheticadaptability to light intensity in white pine, red pine, hemlock, andailanthus seedlings. For. Sci. 4:196--207.

Bradford, M.M. 1976. A rapid and sensitive method for the quantita-tion of microgram quantities of protein utilizing the principle ofprotein-dye binding. Anal. Biochem. 72:248--254.

Canham, C.D., A.C. Finzi, S.W. Pacala and D.H. Burbank. 1994.Causes and consequences of resource heterogeneity in forests:interspecific variation in light transmission by canopy trees. Can. J.For. Res. 24:337--349.

Chabot, B.F., T.W. Jurik and J.F. Chabot. 1979. Influence of instanta-neous and integrated light-flux density on leaf anatomy and photo-synthesis. Am. J. Bot. 66:940--945.

Charles-Edwards, D.A., J. Charles-Edwards and F.I. Sant. 1974. Leafphotosynthetic activity in six temperate grass varieties grown incontrasting light and temperature environments. J. Exp. Bot.25:715--724.

Chatterton, N.J., D.R. Lee and W.E. Hungerford. 1972. Diurnalchange in specific leaf weight of Medicago sativa L. and Zeamays L. Crop Sci. 12:576--578.

Chazdon, R.L. and C.B. Field. 1987. Determinants of photosyntheticcapacity in six rainforest Piper species. Oecologia 73:222--230.

Chiariello, N. 1984. Leaf energy balance in the wet lowland tropics. InPhysiological Ecology of Plants of the Wet Tropics. Eds. E. Medina,H.A. Mooney and C. Vázquez-Yánes. Proc. Int. Symp. Tasks forVegetation Science 12. Dr. W. Junk Publishers, The Hague, pp 85--98.

Dale, J.E. 1988. The control of leaf expansion. Annu. Rev. PlantPhysiol. Plant Mol. Biol. 39:267--295.

Dengler, N.G. 1980. Comparative histological basis of sun and shadeleaf dimorphism in Helianthus annus. Can. J. Bot. 58:717--730.

Dijkstra, P. and H. Lambers. 1989. Analysis of specific leaf area andphotosynthesis of two inbred lines of Plantago major differing inrelative growth rate. New Phytol. 113:283--290.

Ducrey, M. 1981. Étude bioclimatique d’une futaie feuillue (Fagussilvatica L. et Quercus sessiliflora Salisb.) de l’Est de la France. III.Potentialités photosynthétiques des feuilles à différentes hauteursdans le peuplement. Ann. Sci. For. 38:71--86.

Ducrey, M. 1992. Variation in leaf morphology and branching patternof some tropical rain forest species from Guadeloupe (French WestIndies) under semi-controlled light conditions. Ann. Sci. For.49:553--570.

Effland, M.J. 1977. Modified procedure to determine acid-insolublelignin in wood and pulp. TAPPI 60:143--144.

Eliáš, P. 1979. Stomatal activity within the crowns of tall deciduoustrees under forest conditions. Biol. Plant. 21:266--274.

Ellenberg, H. 1988. Vegetation ecology of central Europe, 4th Edn.Cambridge Univ. Press, Cambridge, 989 p.

Ellenberg, H. 1991. Zeigerwerte der Gefäßpflanzen (ohne Rubus). InZeigerwerte von Pflanzen in Mitteleuropa. Eds. H. Ellenberg,R. Düll, V. Wirth, W. Werner and D. Paulißen. Scripta Geobotanica,18. Erich Goltze KG, Göttingen, pp 9--166.

Ellsworth, D.S. and P.B. Reich. 1992. Leaf mass per area, nitrogencontent and photosynthetic carbon gain in Acer saccharum seed-lings in contrasting forest light environments. Funct. Ecol. 6:423--435.

Ellsworth, D.S. and P.B. Reich. 1993. Canopy structure and verticalpatterns of photosynthesis and related leaf traits in a deciduousforest. Oecologia 96:169--178.

Epron, D., D. Godard, G. Cornic and B. Genty. 1995. Limitation of netCO2 assimilation rate by internal resistances to CO2 transfer in theleaves of two tree species (Fagus sylvatica L. and Castanea sativaMill.). Plant Cell Environ. 18:43--51.

Fetcher, N., B.R. Strain and S.F. Oberbauer. 1983. Effects of lightregime on the growth, leaf morphology, and water relations ofseedlings of two species of tropical trees. Oecologia 58:314--319.

Fox, T.C. and D.R. Geiger. 1986. Osmotic response of sugar beetsource leaves at CO2 compensation point. Plant Physiol. 80:239--241.

Goryshina, T.K., L.N. Zabotina and E.G. Pruzhina. 1979. Osoben-nosti assimilyatsionnyh tkanei i plastidnogo apparata lista vraznyh chastyah krony u nekotoryh drevsnyh porod v lesostepnoidubrave. (Characteristics of assimilatory tissues and plastid appa-ratus in leaves of various parts of crown in some tree species ofthe oakwood). Biol. Ser. Vestnik LGU 0:67--76. In Russian.

Grace, J. and G. Russell. 1977. The effect of wind on grasses. III.Influence of continuous drought or wind on anatomy and waterrelations in Festuca arundinacea Schreb. J. Exp. Bot. 28:268--278.

Gries, C., R. Lösch and L. Kappen. 1987. Untersuchungen zum pflan-zlichen Wasserhaushalt von Corylus avellana in einer Hollsteinis-chen Wallhecke. In Verhandlungen der Gesellschaft für Ökologie,Vol. 16. Ed. L. Steubing. Gesellschaft für Ökologie, Göttingen,pp 227--230.



Gulmon, S.L. and C.C. Chu. 1981. The effects of light and nitrogen onphotosynthesis, leaf characteristics, and dry matter allocation inthe chaparral shrub, Diplacus aurantiacus. Oecologia 65:214--222.

Gutschick, V.P. and F.W. Wiegel. 1988. Optimizing the canopy photo-synthetic rate by patterns of investment in specific leaf mass. Am.Nat. 132:67--86.

Hantschel, R. 1987. Wasser- und Elementbilanz von geschädigten,gedüngten Fichtenökosystemen im Fichtelgebirge unter Berück-sichtigung von physikalischer und chemischer Bodenhetero-genität. Bayreuther Bodenkundliche Berichte, Bd. 3. Lehrstuhl fürBodenkunde und Bodengeographie der Univ. Bayreuth, Bayreuth,219 p.

Hsiao, T.C., W.K. Silk and J. Jing. 1985. Leaf growth and waterdeficits: biophysical effects. In Control of Leaf Growth. Eds. N.R.Baker, W.J. Davies and C.K. Ong. Soc. Exp. Biol. Semin. Ser. 27.Cambridge Univ. Press, Cambridge, pp 239--266.

Huber, S.C. and J.L. Huber. 1996. Role and regulation of sucrose-phosphate synthase in higher plants. Annu. Rev. Plant Physiol. PlantMol. Biol. 47:431--444.

Ingestad, T. and A.J.S. McDonald. 1989. Interaction between nitrogenand photon flux density in birch seedlings at steady-state nutrition.Physiol. Plant. 77:1--11.

Jones, H.G. 1985. Adaptive significance of leaf development andstructural responses to environment. In Control of Leaf Growth.Eds. N.R. Baker, W.J. Davies and C.K. Ong. Soc. Exp. Biol. Semin.Ser. 27. Cambridge Univ. Press, Cambridge, pp 155--173.

Jurik, T.W., J.F. Chabot and B.F. Chabot. 1982. Effects of light andnutrients on leaf size, CO2 exchange, and anatomy in wild straw-berry (Fragaria virginiana). Plant Physiol. 70:1044--1048.

Kausch, W. and W. Haas. 1965. Chemische Unterschiede zwischenSonnen- und Schattenblättern der Blutbuche (Fagus sylvatica L. cv.Atropunicea). Naturwissenschaften 52:214--215.

Kausch, W. and W. Haas. 1966. Ligningehalte der Zellwände beiSonnen- und Schattenblättern der Blutbuche (Fagus sylvatica cv.Atropunicea). Naturwissenschaften 53:89.

Körner, C. and S. Pelaez Menendez-Riedl. 1990. The significance ofdevelopmental aspects in plant growth analysis. In Causes andConsequences of Variation in Growth Rate and Productivity ofHigher Plants. Eds. H. Lambers, M.L. Cambridge, H. Koningsand T.L. Pons. SPB Academic Publishing, The Hague, pp 141--157.

Koide, R.T., R.H. Robichaux, S.R. Morse and C.M. Smith. 1989. Plantwater status, hydraulic resistance and capacitance. In Plant Physi-ological Ecology. Field Methods and Instrumentation. Eds. R.W.Pearcy, J.R. Ehleringer, H.A. Mooney and P.W. Rundel. Chapmanand Hall, London, pp 161--183.

Kull, O. and Ü. Niinemets. 1993. Variation in leaf morphometry andnitrogen concentration in Betula pendula, Corylus avellana andLonicera xylosteum. Tree Physiol. 12:311--318.

Küppers, M. 1994. Canopy gaps: competitive light interception andeconomic space filling----a matter of whole-plant allocation. InExploitation of Environmental Heterogeneity by Plants. Ecophysi-ological Processes Above- and Belowground. Eds. M.M. Caldwelland R.W. Pearcy. Physiological Ecology, Academic Press, SanDiego, pp 111--144.

Laasimer, L. 1965. Eesti NSV taimkate. (The plant cover of EstonianSSR). Valgus, Tallinn, 400 p. In Estonian.

Laasimer, L. 1975. Eesti lood ja loometsad, nende kaitse. (Die Alvarenund die Alvarwälder in Estland und ihr Schutz). In Eestiloodusharulduste kaitseks. Eds. E. Kumari, V. Hang, A. Mäemetsand O. Renno. Valgus, Tallinn, pp 90--103. In Estonian.

Lawlor, D.W. 1995. The effects of water deficit on photosynthesis. InEnvironment and Plant Metabolism: Flexibility and Acclimation.Ed. N. Smirnoff. Environ. Plant Biol. Ser. Bios Scientific PublishersLtd., Oxford, pp 129--160.

Lawlor, D.W. and J.E. Leach. 1985. Leaf growth and water deficits:biochemistry in relation to biophysics. In Control of Leaf Growth.Eds. N.R. Baker, W.J. Davies and C.K. Ong. Soc. for Exp. Biol.Semin. Ser. 27. Cambridge University Press, Cambridge, pp267--294.

Lichtenthaler, H.K. 1985. Differences in morphology and chemicalcomposition of leaves grown at different light intensities and quali-ties. In Control of Leaf Growth. Eds. N.R. Baker, W.J. Davies andC.K. Ong. Soc. Exp. Biol. Semin. Ser. 27. Cambridge Univ. Press,Cambridge, pp 201--221.

Luken, J.O., T.C. Tholemeier, L.M. Kuddes and B.A. Kunkel. 1995.Performance, plasticity, and acclimation of the nonindigenousshrub Lonicera maackii (Caprifoliaceae) in contrasting light envi-ronments. Can. J. Bot. 73:1953--1961.

Morgan, J.M. 1984. Osmoregulation and water stress in higher plants.Annu. Rev. Plant Physiol. 35:299--319.

Munns, R. and R. Weir. 1981. Contribution of sugars to osmoticadjustment in elongating and expanded zones of wheat leaves dur-ing moderate water deficits at two light levels. Aust. J. PlantPhysiol. 8:93--105.

Myers, B.J., R.H. Robichaux, G.L. Unwin and I.E. Craig. 1987. Leafwater relations and anatomy of a tropical rainforest tree species varywith crown position. Oecologia 74:81--85.

Niinemets, Ü. 1995. Distribution of foliar carbon and nitrogen acrossthe canopy of Fagus sylvatica: adaptation to a vertical light gradi-ent. Acta Oecol. 16:525--541.

Niinemets, Ü. 1996a. Changes in foliage distribution with tree sizeand relative irradiance: differences between the saplings of Acerplatanoides and Quercus robur. Ecol. Res. 11:269--281.

Niinemets, Ü. 1996b. Importance of structural features of leaves andcanopy in determining species shade-tolerance in temperate decidu-ous woody taxa. Dissertationes Biologicae Universitatis Tartuensis.Tartu University Press, Tartu, Estonia, 149 p.

Niinemets, Ü. 1997a. Distribution patterns of foliar carbon and nitro-gen as affected by tree dimensions and relative light conditions inthe canopy of Picea abies. Trees 11:144--154.

Niinemets, Ü. 1997b. Role of foliar nitrogen in light harvesting andshade tolerance of four temperate deciduous woody species. Funct.Ecol. 11:518--531.

Niinemets, Ü. and K. Kull. 1994. Leaf weight per area and leaf size of85 Estonian woody species in relation to shade tolerance and lightavailability. For. Ecol. Manag. 70:1--10.

Niinemets, Ü. and O. Kull. 1995. Effects of light availability and treesize on the architecture of assimilative surface in the canopy ofPicea abies: variation in needle morphology. Tree Physiol. 15:307--315.

Niinemets, Ü. and J.D. Tenhunen. 1997. A model separating leafstructural and physiological effects on carbon gain along lightgradients for the shade-tolerant species Acer saccharum. Plant CellEnviron. 20:845--866.

Niinemets, Ü., O. Kull and J.D. Tenhunen. 1998. Growth irradianceinfluences the nitrogen investment in light harvesting more than inRUBISCO and photosynthetic electron transport in temperate de-ciduous woody species of contrasting shade tolerance. Tree Physiol.Vol. 18. In Press.

Niklas, K.J. 1991. Effects of tissue volume and location on the me-chanical consequences of dehydration of petioles. Am. J. Bot.78:361--369.



Nilson, T. and V. Ross. 1979. Characterization of the transparency ofa forest canopy by fish-eye photographs. In Spruce Forest Ecosys-tem Structure and Ecology, Vol. 2. Basic Data on the EstonianVooremaa Project. Ed. T. Frey. Estonian IBP Report. 12. Academyof Sciences of the Estonian S.S.R. Estonian Republican Committeefor IBP, Tartu, pp 114--130.

Nimz, H. 1974. Das Lignin der Buche----Entwurf eines Konstitutions-schemas. Angew. Chem. 86:336--344.

Nobel, P.S. 1977. Internal leaf area and cellular CO2 resistance: pho-tosynthetic implications of variations with growth conditions andplant species. Physiol. Plant. 40:137--144.

Nobel, P.S. and T.L. Hartsock. 1981. Development of leaf thicknessfor Plectranthus parviflorus----influence of photosynthetically ac-tive radiation. Physiol. Plant. 51:163--166.

Oberbauer, S.F., B.R. Strain and G.H. Riechers. 1987. Field waterrelations of a wet-tropical forest tree species, Pentaclethramacroloba (Mimosaceae). Oecologia 71:369--374.

Otto, H.-J. 1994. Waldökologie. Verlag Eugen Ulmer, Stuttgart, 391 p.Philpott, J. 1956. Blade tissue organization of foliage leaves of some

Carolina shrub-bog species as compared with their AppalachianMountain affinities. Bot. Gaz. 118:88--105.

Pieters, G.A. 1974. The growth of sun and shade leaves of Populuseuramericana ‘Robusta’ in relation to age, light intensity and tem-perature. Meded. Landbouwhogesch. Wageningen, 74:1--107.

Pigott, C.D. 1991. Biological flora of the British Isles. J. Ecol.79:1147--1207.

Pigott, C.D. and S. Pigott. 1993. Water as a determinant of the distri-bution of trees at the boundary of the Mediterranean zone. J. Ecol.81:557--566.

Preston, R.D. 1974. The physical biology of plant cell walls. Chapmanand Hall, London, 650 p.

Reich, P.B., M.B. Walters, D.S. Ellsworth and C. Uhl. 1994. Photosyn-thesis-nitrogen relations in Amazonian tree species. I. Patternsamong species and communities. Oecologia 97:62--72.

Retuerto, R. and F.I. Woodward. 1992. Effects of windspeed on thegrowth and biomass allocation of white mustard Sinapis alba L.Oecologia 92:113--123.

Rich, P.M., D.B. Clark, D.A. Clark and S.F. Oberbauer. 1993. Long-term study of solar radiation regimes in a tropical wet forest usingquantum sensors and hemispherical photography. Agric. For. Me-teorol. 65:107--127.

Robinson, T. 1979. The determination of protein in plant extracts thatcontain polyphenols. Plant Sci. Lett. 15:211--216.

Salleo, S. 1983. Water relations parameters of two Sicilian species ofSenecio (groundsel) measured by the pressure bomb technique.New Phytol. 95:179--188.

Schulze, E.-D. and A.E. Hall. 1982. Stomatal responses, water lossand CO2 assimilation rates of plants in contrasting environments. InPhysiological Plant Ecology, Vol. 12B. Eds. O.L. Lange, P.S. Nobel,C.B. Osmond and H. Ziegler. Encyclopedia of Plant Physiology.Springer-Verlag, Berlin, pp 181--230.

Schulze, E.-D., R.H. Robichaux, J. Grace, P.W. Rundel and J.R.Ehleringer. 1987. Plant water balance. Bioscience 37:30--37.

Servaites, J.C., D.R. Geiger, M.A. Tucci and B.R. Fondy. 1989. Leafcarbon metabolism and metabolite levels during a period of sinusoi-dal light. Plant Physiol. 89:403--408.

Shuttleworth, W.J., J.H.C. Gash, C.R. Lloyd et al. 1985. Daily vari-ations of temperature and humidity within and above Amazonianforest. Weather 40:102--108.

Starzecki, W. 1975. Ecophysiological investigation on photosyntheticproductivity of leaves of chosen tree species. Pol. Ecol. Stud.1:51--63.

Steubing, L., C. Ramirez and M. Alberdi. 1979. Artenzusammen-setzung, Lichtgenuß und Energiegehalt der Krautschicht des Valdi-vianischen Regenwaldes bei St. Martin. Vegetatio 39:25--33.

Stewart, G.R., C.A. Gracia, E.E. Hegarty and R.L. Specht. 1990.Nitrate reductase activity and chlorophyll content in sun leaves ofsubtropical Australian closed-forest (rainforest) and open-forestcommunities. Oecologia 82:544--551.

Syvertsen, J.P., J. Lloyd, C. McConchie, P.E. Kriedemann and G.D.Farquhar. 1995. On the relationship between leaf anatomy and CO2diffusion through the mesophyll of hypostomatous leaves. PlantCell Environ. 18:149--157.

Takahashi, T., N. Tsuchihashi and K. Nakaseko. 1993. Estimation ofassimilation activity by daily increase in water soluble sugar con-tent in spring wheat canopy. Jpn. J. Crop Sci. 62:319--323.

Taylor, G. and W.J. Davies. 1986. Yield turgor of growing leaves ofBetula and Acer. New Phytol. 104:347--353.

Tjoelker, M.G., J.C. Volin, J. Oleksyn and P.B. Reich. 1995. Interac-tion of ozone pollution and light effects on photosynthesis in aforest canopy experiment. Plant Cell Environ. 18:895--905.

Tsutakawa, R.K. and J.E. Hewett. 1978. Comparison of two regessionlines over a finite interval. Biometrics 34:391--398.

Türk, T. 1992. Die Wasser- und Stoffdynamik in zwei unterschiedlichgeschädigten Fichtenstandorten im Fichtelgebirge. BayreutherBodenkundliche Berichte. Bd. 22. Lehrstuhl für Bodenkunde undBodengeographie der Universität Bayreuth, Bayreuth, 231 p.

Vance, N.C. and J.B. Zaerr. 1991. Influence of drought stress and lowirradiance on plant water relations and structural constituents inneedles of Pinus ponderosa seedlings. Tree Physiol. 8:175--184.

Vertregt, N. and F.W.T. Penning de Vries. 1987. A rapid method fordetermining the efficiency of biosynthesis of plant biomass.J. Theor. Biol. 128:109--119.

Wallace, G. and S.C. Fry. 1994. Phenolic components of the plant cellwall. A Survey of Cell Biol. Int. Rev. Cytol. 151:229--267.

Wallace, L.L. and E.L. Dunn. 1980. Comparative photosynthesis ofthree gap phase successional tree species. Oecologia 45:331--340.

Walters, M.B. and C.B. Field. 1987. Photosynthetic light acclimationin two rainforest Piper species with different ecological amplitudes.Oecologia 72:449--456.

Waring, R.H., A.J.S. McDonald, S. Larsson, T. Ericsson, A. Wiren,E. Arwidsson, A. Ericsson and T. Lohammar. 1985. Differences inchemical composition of plants grown at constant relative growthrates with stable mineral nutrition. Oecologia 66:157--160.

Wayne, P.M. and F.A. Bazzaz. 1993. Birch seedling responses to dailytime courses of light in experimental forest gaps and shadehouses.Ecology 74:1500--1515.

Witkowski, E.T.F. and B.B. Lamont. 1991. Leaf specific mass con-founds leaf density and thickness. Oecologia 88:486--493.

Woodward, F.I. 1983. The significance of interspecific differences inspecific leaf area to the growth of selected herbaceous species fromdifferent altitudes. New Phytol. 95:313--323.



Downloaded from https://academic.oup.com/treephys/article-abstract/18/7/467/1687125by gueston 11 February 2018

stoichiometry of foliar carbon consistuents varies along light

Documents