Seedling size influences relationships of shade tolerance with carbohydrate-storage patterns in a temperate rainforest


†Author to whom correspondence should be addressed. E-mail:


  • 1Carbohydrate storage has been attributed an important role in the ability to tolerate shade, yet empirical support for this idea has been patchy. We asked if carbohydrate-storage patterns of seedling evergreens in low light are correlated with variation in shade tolerance, and how these patterns change with seedling size.
  • 2We measured biomass distribution and total non-structural carbohydrate (NSC) concentrations of leaves, stems and roots of two seedling size classes of six evergreens growing in a temperate rainforest understorey. Light requirements of the six species were quantified by calculating the 10th percentile of the distribution of established seedlings in relation to canopy openness.
  • 3NSC averaged 14% of the dry mass of small seedlings (40–60 mm tall), and 22% of that of large seedlings (400–600 mm tall). This difference was entirely due to variation in starch reserves, which on average accounted for 60% of NSC in small seedlings and 84% in large seedlings.
  • 4NSC concentrations of leaves and roots (but not stems) of large seedlings were negatively correlated with species’ shade tolerance, but no such pattern was found in small seedlings. Leaf NSC on an area basis was not related to species’ shade tolerance in either size class.
  • 5Partitioning of the NSC pool between leaves, stems and roots of small seedlings was closely related to variation in shade tolerance. Small seedlings of shade-tolerant species had a relatively low proportion of their NSC pool in leaves and a high proportion in roots. This is likely to ensure the retention of the greater part of the NSC pool even in the event of extensive defoliation, and the availability of reserves to replace lost leaves. In contrast, the large leaf-mass fraction of large seedlings of shade-tolerant species (46–47% of biomass) meant that these plants had a large proportion of their NSC pool in foliage.
  • 6Results suggest that, in Chilean rainforest evergreens, any adaptive relationship of carbohydrate storage with shade tolerance may be confined to young seedlings, involving interspecific variation in the partitioning of reserves between leaves and other organs, rather than especially high NSC concentrations in shade-tolerant species.


Interspecific variation in the ability to establish under shade is a fundamental determinant of temporal and spatial patterns in the distribution of woody plants (Smith & Huston 1989; Niinemets & Valladares 2006). Physiological ecologists have explored the possible roles of a range of seedling and juvenile traits in determining this variation (Givnish 1988; Walters & Reich 1999). Early work focused on leaf-level gas exchange (Grime 1965; Loach 1967), and this emphasis has found renewed support in recent reports that shade-tolerant species tend to have lower respiration rates and light-compensation points than light-demanding associates, suggesting more favourable leaf-level carbon balance in low light (Lusk 2002; Craine & Reich 2005). In addition, advocacy of a whole-plant perspective (Givnish 1988) has led to increased interest in biomass distribution, foliage display and tissue turnover as determinants of light interception and carbon balance (King 1994; Walters & Reich 1999; Lusk 2004; Pearcy et al. 2004).

The balance of carbon allocation between storage and growth could also play an important role in shade tolerance (Kitajima 1994; Kobe 1997). It has been argued that the difficulty in recovering from defoliation and other damage in low-energy understorey environments, and the low opportunity cost of storage in low light, will select for ample storage of carbohydrates in roots and stems of seedlings and juvenile trees (Kobe 1997). The selective importance of defoliation in the understorey is given credence by evidence that shade leaves suffer higher herbivory rates than sun leaves (Niesenbaum 1992; Dudt & Shure 1994). Carbohydrate storage would serve as a buffer against defoliation in low light, and might at least partly explain why pioneer and late-successional species differ in survival at a given growth rate (Kobe et al. 1995). Allocation to storage might also help explain biomass-distribution patterns of small seedlings. Large root-mass fraction has been widely reported in young seedlings of shade-tolerant trees (Kitajima 1994; Paz 2003; Lusk 2004), which might indicate early allocation to below-ground storage in these taxa (Kitajima 1994).

Despite the plausibility of the argument linking shade tolerance to carbohydrate storage, empirical support to date has been patchy. Non-structural carbohydrates (NSC) of seedlings and juvenile trees have been measured in few comparative studies involving more than two species (Kobe 1997; DeLucia et al. 1998; Canham et al. 1999), and these have produced quite diverse results. These discrepancies could be partly attributable to age and size differences between plants studied by different authors, as well as variation in leaf habit. Large seedlings and saplings of shade-tolerant evergreens are probably exposed to a low risk of catastrophic defoliation, at least by insect herbivores. These taxa often accumulate five or more foliage cohorts in shaded environments (King 1994; Williams, Field & Mooney 1989; Lusk 2002); as mature, hardened foliage is generally much less attractive than young, expanding leaves to chewing and sucking insects (Coley 1983), established juveniles of evergreens with slow foliage turnover expose only a small fraction of their photosynthetic surface area to a high risk of herbivory at a given time. In contrast, deciduous species are, by implication, highly vulnerable during leaf-out, which might constitute a stronger selection pressure for maintenance of extensive reserves in low light. First-year seedlings of evergreens are in a position analogous to deciduous species, in that they depend on a single cohort of leaves that are vulnerable to herbivory during expansion. If herbivory is a major selective pressure on carbohydrate storage in rainforest understoreys, we might thus expect NSC concentrations of evergreens to show close relationships with shade-tolerance variation in very young seedlings, but not necessarily in older established juveniles.

The distribution of carbohydrate-storage pools could also change during juvenile ontogeny. Root-mass fraction of shade-tolerant evergreens is initially high, but in some cases declines later to the point where shaded juveniles approaching 1 m tall have only ≈20% of their biomass below ground (Lusk 2004; Machado & Reich 2006). Roots therefore seem unlikely to be the main storage organs of established juveniles of these taxa. Because of uncertainties associated with the daily NSC dynamics of leaves, leaf tissues are often omitted from studies of carbohydrate storage. In rainforest understoreys, however, the large leaf-mass fraction of established juveniles of shade-tolerant species is likely to contain a large fraction of these plants’ total NSC reserves.

Here we examine the NSC reserves of shade-grown seedlings of six temperate rainforest evergreens shown to differ widely in shade tolerance. We asked if concentrations of starch and soluble sugars, and the distribution of NSC reserves between different organs, vary systematically in relation to shade-tolerance level, and how these patterns change during seedling ontogeny. These questions were addressed by quantifying the biomass distribution of two different seedling size classes, and by measuring starch and sugar concentrations of roots, stems and leaves.

Materials and methods

study area and species

The study was carried out in the lowland forest of Parque Nacional Puyehue (40°39′ S, 72°11′ W) located in the western foothills of the Andes in south-central Chile. The lowland forests are dominated by broadleaved evergreen trees including Laureliopsis philippiana, Aextoxicon punctatum, Eucryphia cordifolia and Nothofagus dombeyi.

Plant material was obtained from 40-year-old even-aged stands located on an alluvial terrace at 350 m a.s.l., in the Anticura sector of the Park. Although the overstorey of these stands is dominated almost exclusively by N. dombeyi, seedlings and saplings of a wide variety of other species are common in the understorey (Lusk 2002).

We chose six common evergreens that differ widely in shade tolerance (Table 1). Three of these have abundant seedlings and saplings in shaded understoreys (Figueroa & Lusk 2001; Lusk, Chazdon & Hofmann 2006), and their low mortality rates indicate a high level of shade tolerance (Lusk 2002). The other three regenerate mainly in well lit environments associated with gaps or forest margins (Figueroa & Lusk 2001; Lusk et al. 2006), and two of them (Aristotelia chilensis, E. cordifolia) have been shown to suffer high mortality rates in low light (Lusk 2002). Eucryphia cordifolia is regarded as mid-tolerant by Chilean foresters, and the third light-demanding species (N. dombeyi) as intolerant (Donoso 1989). As all species belong to different genera, they are referred to henceforth by their generic names.

Table 1.  Species chosen for measurement of total non-structural carbohydrates and biomass distribution in low light (2–5% canopy openness), in a temperate rainforest in south-central Chile
SpeciesFamilyMax. height (m)Index of seedling light requirementsMean dry mass (g) of seedlings
Small (40–60 mm tall)Large (400–600 mm tall)
  1. Index of light requirements estimated as 10th percentile of seedling distributions in relation to canopy openness in old-growth stands (n = number of quadrats containing each species). Last two columns show mean and SE of seedling dry mass.

Aristotelia chilensisElaeocarpaceae  103·4 (n = 56)0·015 ± 0·002 3·7 ± 0·5
Nothofagus dombeyiFagaceae> 453·2 (n = 17)0·041 ± 0·004 4·0 ± 0·6
Eucryphia cordifoliaCunoniaceae> 402·1 (n = 52)0·031 ± 0·004 4·8 ± 1·1
Laureliopsis philippianaAtherospermataceae  351·2 (n = 38)0·029 ± 0·00410·7 ± 0·9
Aextoxicon punctatumAextoxicaceae  351·2 (n = 129)0·092 ± 0·09 9·7 ± 1·1
Myrceugenia planipesMyrtaceae  200·8 (n = 228)0·082 ± 0·10 9·1 ± 0·9

During the past two decades, comparative ecologists and ecophysiologists have become more aware of the importance of phylogenetic relatedness for comparative studies (Harvey & Pagel 1991). If most of the ecological or physiological variation of interest in the data set is associated with one or a few old divergences, a simple cross-species analysis treating more closely related taxa (representing more recent divergences) as independent data points can yield misleading conclusions about adaptive patterns (Ackerly & Reich 1999). This problem is often avoided by restricting comparisons to congeners (Pearcy et al. 2004), or by using phylogenetically independent contrasts (Kelly 1995). Neither of these options was feasible for our study, as congeneric woody plants rarely coexist in Chilean temperate rainforests, and because it was not practical to sample a large enough group of species to carry out independent contrasts. By sampling the three most shade-tolerant woody tree species of the region, and three of the most light-demanding species (each from six different angiosperm families), we ensured a strong ecological contrast among our study species. However, it is possible that more extensive studies taking phylogeny into account could come to different conclusions about general patterns of adaptation to shade.

quantification of species light requirements

The natural distributions of seedling juveniles in relation to light availability in neighbouring old-growth stands (Lusk et al. 2006) were used to provide an index of species shade tolerance (Lusk & Reich 2000). Distributions of large seedlings (100–1000 mm tall) were quantified in relation to canopy openness measurements made with a pair of LAI-2000 canopy analysers (Li-Cor, Lincoln, NE, USA). One instrument was used to take measurements at each sampling point, while the other, placed at the centre of a 2-ha clearing, was programmed to take readings at 30-s intervals. By comparing measurements made within the forest with simultaneous readings taken in the clearing, we calculated percentage diffuse irradiance at each sampling point within the forest, equivalent to percentage canopy openness over the quasi-hemispherical (148°) field of view perceived by the LAI-2000 sensors. Measurements with the LAI-2000 are a good surrogate of spatial variation in mean daily photosynthetic photon flux density within a stand (Machado & Reich 1999). Sampling was carried out on a series of parallel transects run through old-growth stands, including a wide range of light environments. At sample points spaced at random intervals (2–10 m apart) along transects, readings were taken at 500 and 1000 mm height with the LAI-2000. Juveniles 100–1000 mm tall were recorded in a circular quadrat of 1 m diameter, centred on the sample point. Juveniles <500 mm tall were referred to the light environment measured at 500 mm height; those >500 mm tall were referred to the light environment measured at 1000 mm height. One of the study species (Eucryphia) frequently reproduces by basal shoots and root suckers as well as seedlings (Donoso, Escobar & Urrutia 1985). For the purposes of evaluating shade tolerance in the present study, however, we counted only juveniles of seedling origin.

The 10th percentile of the distribution of each species along the light gradient (percentage canopy openness) was used as an approximation of the lowest light levels tolerated by each species (Table 1). Species’ scores on this index generally corresponded closely to their ratings on the five-point scale devised by Donoso (1989). However, we stress that our scores are an inversion of traditional shade-tolerance ranking systems: shade-tolerant taxa such as Myrceugenia have low values for this index, and light-demanders such as Nothofagus score high (Table 1).

sampling design for nsc and biomass

In early autumn (March) 2005, seedlings of two size classes in the forest understorey were selected and tagged. The small size class (40–60 mm tall) corresponds to first-year seedlings of relatively large-seeded Aristotelia, Aextoxicon and Myrceugenia, and to second-year seedlings of small-seeded Nothofagus, Eucryphia and Laureliopsis. The large size class (400–600 mm tall) contains a wider range of ages. Sampling points were set out at random intervals along a transect run through the stand. At each sample point, the nearest seedling of each size class of each species was located and tagged. Individuals with extensive browsing or other damage were excluded.

The LAI-2000 canopy analysers were used to quantify canopy openness within a quasi-hemispherical field of view, immediately above the apex of each tagged seedling. In order to minimize the influence of variation in irradiance on the study parameters, only plants growing within the range of 2–5% canopy openness were used (Lusk 2004). Sampling was continued until 10 such individuals of both size classes were obtained for each species.

nsc and biomass determinations

All plant material was collected between 13 : 00 and 18 : 00 h. Each seedling was excavated by removing a sod of sufficient width and depth to include the root system. After removal of soil from roots, each seedling was separated into leaves, stem and roots. Leaves were removed and photographed with a digital camera, and their total area measured later using the software SigmaScan (Systat Software Inc., Richmond, CA, USA).

Complete roots, stems and leaves were used for NSC analyses. Within 2 h after sampling, tissues were cooked at 600 W for 90 s in a microwave oven at a field station to denature enzymes (Popp et al. 1996), then dried to a constant mass at 80 °C. Dry mass of each organ was recorded, and dried samples were then ground to a fine powder.

Samples were analysed for total soluble sugars and starch using ethanol and perchloric acid (Hansen & Moller 1975). Total soluble sugars were extracted from tissue in 86% v/v ethanol at 80 °C for 1 h. The supernatant was collected after centrifugation and the concentration of total soluble sugars was determined spectrophotometrically by the Resorcinol method (Roe 1934) at a wavelength of 520 nm, using sucrose as standard. Starch was extracted from the ethanol-insoluble fraction by agitating for 1 h with 35% v/v perchloric acid (Sutton, Ting & Sutton 1981). This method of extraction can yield starch values higher than those estimated by more accurate enzyme methods (Rose et al. 1991), probably as a result of hydrolysis of some cell wall components. However this is not a major problem for our study, as we were interested primarily in the relative concentrations of different species, size classes and organs. The protocol for starch determination in the extract was similar to that used for sugars, but using glucose as standard. Starch and soluble sugars in each plant component were added together to determine total NSC in mg g−1 dry mass.

Diurnal variation in leaf NSC levels (Upmeyer & Koller 1973) will result in differences between individual plants sampled at different times. However, as we excavated all plants at each sample point before moving on to the next point, sampling time should not contribute to interspecific variation.

statistical analyses

Two-way anova showed no difference (P = 0·74) between light environments of small and large seedlings selected for NSC and biomass measurements (mean 3·3 and 3·2% canopy openness, respectively). Nor did mean light environments differ significantly between species (P = 0·10), and there was no significant interaction of size with species. Light environment was not correlated with any dependent variable of interest, and so was not used as a predictor variable in any analysis.

Two-way anovas were used to test for interspecific variation in NSC traits and biomass, and for differences between leaves, stems and roots. We then examined the correlations of NSC and biomass distribution traits with species light requirements. All analyses were carried out using jmp Statistical Software (SAS Institute, Cary, NC, USA).


biomass distribution

There was significant interspecific variation in dry mass of both small and large seedlings (Table 2). Shade-tolerant species tended to be heavier, on average, than those of light-demanders (Table 1), reflecting the more robust construction of the former.

Table 2.  Summary of anova testing for interspecific variation in seedling biomass and carbohydrate storage traits, plus differences between organs
EffectSmall seedlingsLarge seedlings
  1. NSC, non-structural carbohydrate. Degrees of freedom in all analyses: 5 for species; 2 for organ; 10 for the interaction.

Species 6·430·00020·970·000
Organ × species 2·440·00912·140·000
log Biomass
Organ 0·350·70721·60·000
Organ × species 1·470·152 5·40·000
log NSC pool
Organ 5·810·00422·00·000
Organ × species 2·480·008 3·40·001

The distribution of biomass between leaves, stems and roots of small seedlings did not differ systematically between species, as indicated by the non-significant interaction of species and organ (Table 2). Despite this lack of statistical significance, leaf-mass fraction tended to be higher in light-demanding species than in shade-tolerators (Fig. 1a).

Figure 1.

Biomass distribution (a,b) and total non-structural carbohydrate (NSC) partitioning (c,d) of seedlings of six temperate rainforest trees growing in low light (2–5% canopy openness). Small seedlings (a,c) were 40–60 mm tall; large seedlings (b,d) were 400–600 mm tall.

Biomass distribution of large seedlings differed widely between species, as shown by the highly significant interaction of species and organ (Table 2; Fig. 1b). Biomass-distribution parameters of large seedlings were strongly correlated with interspecific variation in shade tolerance (Table 3). Shade-tolerant species had a higher leaf-mass fraction, and lower stem mass fraction, than light-demanders (Table 3; Fig. 1b).

Table 3.  Correlations of carbohydrate storage and biomass distribution traits with seedling light requirements of six rainforest evergreens (= 10th percentile of seedling distributions in relation to canopy openness)
TraitSmall seedlingsLarge seedlings
  1. Marginally significant correlations in bold type; stronger correlations bold and italic.

  2. NSC, non-structural carbohydrate. r = Pearson's correlation coefficient.

Leaf 0·620·19 0·810·05
Stem−0·150·77 0·540·27
Root−0·260·62 0·920·01
Fraction of NSC pool
Leaf 0·810·05−0·600·21
Stem−0·470·35 0·600·21
Root0·900·015 0·070·90
Biomass fraction
Leaf 0·630·180·990·0002
Stem−0·380·43 0·880·02
Root−0·550·26 0·580·23

nsc concentrations

NSC contributed on average ≈14% (8–28%) of the dry mass of small seedlings, and 22% (range 14–34%) of that of large seedlings (Fig. 2). This difference was entirely due to variations in starch concentration, as average sugar levels were actually lower in large seedlings. Starch was invariably the dominant fraction of NSC in both size classes of all species, on average accounting for ≈60% of NSC in small seedlings, and ≈84% in large seedlings (Fig. 2).

Figure 2.

Whole-plant concentrations of starch and total sugars (mg g−1 DW) of (a) small and (b) large seedlings of six temperate rainforest trees growing in low light (2–5% canopy openness).

NSC concentrations of both size classes differed significantly between species (Table 2). In both size classes there was evidence that average concentrations differed systematically between organs, with leaves having the highest average levels in both cases (Table 2). The significant interaction terms (Table 2) reflected wide interspecific variation in NSC levels of leaves in large seedlings, and in both leaves and stems of small seedlings (Fig. 3).

Figure 3.

Non-structural carbohydrate (NSC) concentrations of leaves, stems and roots of (a) small and (b) large seedlings of six temperate rainforest trees growing in low light (2–5% canopy openness).

Whereas NSC concentrations of small seedlings were not related to light requirements, there was evidence that large seedlings of light-demanders stored more starch than shade-tolerant species (Fig. 3; Table 3). In large seedlings, light requirements were marginally correlated with NSC levels of leaves, and strongly correlated with those of roots (Table 3). However, the relationship of leaf NSC with light requirements disappeared when NSC concentration was expressed on an area basis (Fig. 4). This reflects wide interspecific variation in specific leaf area, which was higher in light-demanding species than in their shade-tolerant associates.

Figure 4.

Non-structural carbohydrate (NSC) concentrations per unit leaf area of (a) small and (b) large seedlings of six temperate rainforest trees growing in low light (2–5% canopy openness). Scale on y-axis differs between (a) and (b).

distribution of nsc pools

The size of NSC pools differed significantly between species in both size classes (Table 2), reflecting appreciable interspecific variation in seedling dry mass. The significant interaction of species and organ means that partitioning of NSC between leaves, stems and roots differed significantly among species (Table 2; Fig. 1c,d).

In small seedlings, partitioning of the NSC pool was related to variations in shade tolerance (Table 3; Fig. 1c). Light requirements were marginally positively correlated with the fraction of the NSC pool in leaves, and strongly negatively correlated with NSC partitioning to roots (Table 3). Thus small seedlings of the three most shade-tolerant species had 33–36% of their NSC pool in roots, compared with 21–27% in the more light-demanding species. Leaves contained 29–38% of the NSC pool of small seedlings of the shade-tolerant species, well below the corresponding figures of 47–52% in the light-demanders (Fig. 1c).

In large seedlings there was no statistically significant relationship of NSC partitioning with species light requirements (Table 3). However, the fraction of NSC in leaves tended to be largest in species of intermediate-to-high shade tolerance (Fig. 1d), reflecting the large foliage biomass of these taxa (Fig. 1b). One of the shade-tolerant species, L. philippiana, had more than half (54%) its NSC in leaves (Fig. 1d).


In both size classes, seedlings of all species had substantial NSC reserves. Machado & Reich (2006), working with larger juveniles of two evergreen and one deciduous species, reported NSC levels that were highly variable, but generally lower than those reported here. A study of 2-year-old seedlings growing in low light in a deciduous forest understorey (Canham et al. 1999) reported root and stem NSC concentrations somewhat higher than those found in small seedlings (40–60 mm tall) of our evergreen species, but only slightly higher on average than those seen in our large (400–600 mm tall) seedlings. Even in low light, juveniles of both evergreen and deciduous temperate trees therefore appear to accumulate substantial amounts of NSC. Non-structural carbohydrates have other functions in addition to energy storage, such as the involvement of sugars in cold resistance of plants exposed to freezing temperatures (Sakai & Larcher 1987). However, this function cannot explain the large amounts of insoluble starch accumulated by the plants we studied (Fig. 2), as starch is not known to have a role in cold-resistance mechanisms. We sampled towards the end of the growing season, when conditions were still favourable for carbon gain but when the strength of growth sinks was probably waning. This partial asynchrony of supply and demand (Chapin, Schulze & Mooney 1990) is probably an important contributor to the high NSC levels we report here, although overestimation of starch by perchloric acid extraction could also be involved (Rose et al. 1991).

NSC concentrations of large seedlings were higher than those of small seedlings, especially in the light-demanding taxa (Figs 2 and 3). This pattern was entirely due to differences in starch concentration (Fig. 2), and indicates a progressive accumulation of carbohydrate reserves during seedling development. As far as we are aware, no previous studies have compared carbohydrate reserves of different juvenile size classes for a comparable number of species. However, average NSC levels of seedlings in a tropical rainforest at Panama (Wurth, Winter & Korner 1998) were much lower than the average NSC content of adult trees (Wurth et al. 2005). As light environment appears to have a relatively weak long-term effect on tissue NSC levels (Kobe 1997), at least part of this difference reflects an ontogenetic increase in the relative magnitude of NSC reserves in trees.

We found no evidence that shade-tolerant species stored more carbohydrate than similarly sized light-demanders growing in the same light environment. Average NSC levels of small seedlings showed no trend in relation to species light requirements, and those of large seedlings were actually higher in light-demanding species (Figs 2 and 3). This contrasts with results obtained from saplings of four northern temperate species: in separate comparisons of deciduous and evergreen pairs, Kobe (1997) reported that shade-tolerant Acer saccharum and Tsuga canadensis had higher low-light NSC concentrations than intolerant Fraxinus americana and Pinus strobus, respectively, although the effect size was minimal in the evergreen comparison. A study of four northern temperate deciduous species found that low-light survival of young seedlings was correlated with NSC variation across species and across experimental treatments (Canham et al. 1999). However, neither survival, nor NSC concentrations, nor NSC pool size was related to species’ reputed shade tolerance. In contrast, a study of germinants of seven neotropical rainforest species found that shade tolerance was positively correlated with NSC pool size in low light, but not with tissue concentrations (J. A. Myers & K. Kitajima, unpublished data). Machado & Reich (2006) found that understorey saplings of shade-tolerant Abies balsamea and relatively intolerant P. strobus had similar whole-plant NSC concentrations: although leaf NSC was higher in A. balsamea, P. strobus had the higher concentration in roots. The limited evidence available to date therefore does not indicate any consistent relationship between light requirements and low-light NSC concentrations.

Our leaf NSC data are likely closely to reflect net assimilation rates. Against initial expectations, NSC per unit leaf mass was actually lower in shade-tolerant species than in light-demanding taxa (Fig. 3; Table 3). A study of saplings of four deciduous species (Niinemets 1997) reported that the more light-demanding taxa tended to accumulate more leaf NSC in high light, but that all four species had similar leaf NSC levels in low light. This presumably reflects an intimate link with rates of carbon gain, inasmuch as the greater photosynthetic capacity of the light-demanding species is fully expressed only in high light, species of differing shade tolerance having similar net assimilation rates in low light (Poorter 1999). A consideration of species differences in specific leaf area suggests that an essentially similar mechanism may explain our leaf NSC data. In contrast to the four deciduous species studied by Niinemets (1997), our six evergreens differed substantially in specific leaf area: as a result, light-demanders and shade-tolerant taxa showed a similar range of NSC concentrations per unit leaf area (Fig. 4). This may indicate that the light-demanding species had higher assimilation rates per unit mass of leaf tissue, but similar rates per unit area. Marenco, Gonçalves & Vieira (2001) report a similar pattern in their comparison of two evergreens of differing light requirements: in both high and low light, leaf starch and sugar concentrations of shade-tolerant Dipteryx odorata and mid-tolerant Swietenia macrophylla were different on a mass basis, but almost identical on an area basis.

Although tissue NSC concentrations did not conform to the adaptive pattern hypothesized by Kobe (1997), partitioning of the NSC pool between leaves, stems and roots of small seedlings did show an interesting relationship with species’ shade tolerance. Small seedlings of intolerant and mid-tolerant species had, on average, about half their NSC pool in leaves (Fig. 1c), largely because of their large leaf-mass fraction (Fig. 1a). They therefore risk a major depletion of reserves if defoliated. In contrast, small seedlings of shade-tolerant species had, on average, only about one-third of their NSC pool in leaves: this ensures the retention of the greater part of the NSC pool even in the event of extensive defoliation, and availability of reserves to subsequently replace lost leaves. This pattern was partially reversed in large seedlings: the accretion of many leaf cohorts by shade-tolerant Myrceugenia, Aextoxicon and Laureliopsis (Lusk 2002), plus a probable ontogenetic decline in allocation to roots (Lusk 2004), gives large seedlings of these taxa a large leaf-mass fraction in low light (Fig. 1). As a result, large seedlings of the shade-tolerant species had 36–54% of their NSC in foliage (Fig. 1), despite relatively low NSC concentrations in leaf tissues (Fig. 3). There was thus some support for our initial prediction that any relationship of carbohydrate-storage traits with shade tolerance of evergreens will be more evident in small seedlings than in larger juveniles. The latter, having several cohorts of hardened leaves, are probably less vulnerable to defoliation than first-year seedling evergreens, or deciduous species of any size. Relationships of NSC storage with shade tolerance might therefore be a more persistent life-history trait in deciduous species than in evergreens.

Differences in total NSC pool size may also influence early seedling survival. J. A. Myers & K. Kitajima (unpublished data) showed that survival of germinants of seven tropical tree species was correlated with interspecific variation in NSC pool size (although not with NSC concentrations), reflecting the ability of large-seeded, shade-tolerant species to supply young seedlings with abundant carbohydrate reserves. On the other hand, we did not find a significant correlation of light requirements with NSC pools of young seedlings of our six species (P = 0·13; data not shown), but as we standardized seedling height across species, our study was not likely to be very sensitive to interspecific variation in this trait.

We conclude that the importance and precise nature of any relationship of carbohydrate storage with variations in shade tolerance may depend on species’ leaf habit and on ontogenetic stage. Although our study of evergreens did not show the postulated relationship of NSC concentrations with shade tolerance, the partitioning of NSC reserves between the organs of small seedlings was broadly consistent with an adaptive explanation invoking defoliation as a major selective pressure on early ontogenetic stages (Kitajima 1994). No such pattern was evident in large seedlings, which had a biomass distribution and NSC partitioning traits very different from small seedlings. This study, the first to examine low-light carbohydrate-storage patterns in more than one size class, adds to recent evidence that consideration of multiple life-history stages may help further understanding of the determinants of variations in shade tolerance in humid forests (Lusk 2004; Niinemets 2006).


We thank Carolina Jara for help with field work and sample processing, Luis Corcuera for use of his laboratory, and Lourens Poorter and an anonymous referee for helpful feedback. This paper was partially funded by FONDECYT grant 1980084, and by the Millennium Center for Advanced Studies in Ecology and Biodiversity, Grant No. P02-051-F ICM.