Ontogenetic variation in light requirements of juvenile rainforest evergreens


*Correspondence author. E-mail: clusk@bio.mq.edu.au


  • 1Although shade tolerance is often assumed to be a fixed trait of species, recent work has reported size-related changes in the relative and absolute light requirements of woody taxa. We hypothesized that, in evergreen forests, light requirements of shade-tolerant species that accumulate multiple foliage cohorts will be more stable during juvenile ontogeny than those of intolerant species with short leaf lifetimes.
  • 2We quantified the light environments occupied by three size classes of 13 coexisting evergreens in a temperate rainforest, to determine how size influenced their relative shade tolerance. Minimum light requirements (MLRs) of species were estimated by computing the 10th percentile of the distribution of juveniles in relation to percentage canopy openness, for each size class. Leaf life span in low light (2%–5% canopy openness) was estimated by recording survival of marked leaves over 12 months, or retrospectively on species with clearly discernible foliage cohorts.
  • 3Agreement of ranks of species’ MLR across size classes was significant, although not strong (Kendall's W = 0·159, P = 0·02). MLRs of the most shade-tolerant species changed little between size-classes, whereas those of most of the less-tolerant species rose with increasing size.
  • 4Shift in MLR across size-classes was negatively correlated with leaf life span, possibly because of the effects of leaf life span on biomass distribution and whole-plant carbon balance. Survival of light-demanding species with short leaf lifetimes may thus depend on their encountering increasing light levels as they grow taller, whereas progressive accumulation of an extensive leaf area by late-successional taxa enables them to continue to tolerate low light despite increasing size.
  • 5Results suggest that shade-tolerance differences between evergreens become increasingly apparent with increasing size. In identifying a relationship with leaf life span, this work also provides a basis for predicting changes in species’ light requirements during juvenile ontogeny.


Interspecific variation in light requirements is widely regarded as one of the most important aspects of functional diversity in woody plant communities (Wright 2002; Reich et al. 2003; Gilbert et al. 2006). Succession in humid forests arises mainly because of a well-documented trade-off between survival in low light and growth under open conditions (Kobe et al. 1995; Wright 2002), and as old-growth forests are a shifting mosaic of light environments, differences in shade tolerance contribute to the maintenance of species richness at landscape scale (Montgomery & Chazdon 2002; Poorter & Arets 2003). In some temperate regions, foresters and ecologists have devised shade tolerance scales based on semi-quantitative or subjective assessments (Fowells 1965; Donoso 1989; Ellenberg 1991). More recently, shade tolerance variation has been approached more quantitatively by monitoring survival rates in low light (Kitajima 1994; Kobe et al. 1995; Lusk 2002) or by comparing the range of light environments naturally occupied by juveniles of different species (Davies 1998; Lusk & Reich 2000).

Shade tolerance is often tacitly assumed to be a fixed trait of species, both in classifications devised by foresters and ecologists (Fowells 1965; Ellenberg 1991), and in attempts to explain species coexistence on the basis of light gradient partitioning (Hubbell et al. 1999). However, there is evidence that both absolute and relative light requirements of species can change during development. Givnish (1988) argued that whole-plant light compensation points must increase as woody plants grow bigger, due to a declining ratio of autotrophic to heterotrophic tissues, and mechanical requirements for increasing allocation to stem construction. The allometry of juvenile trees is often inconsistent with the latter of Givnish's arguments (Niklas 1995; Lusk et al. 2006b), and we are not aware of empirical tests of the hypothesized increase in compensation points. However, monitoring of juvenile trees in a West African rainforest showed that individuals of most species tended to occupy brighter environments as they grew taller (Poorter et al. 2005). There are also reports of ontogenetic rank changes in the average light environments occupied by coexisting species (Clark & Clark 1992; Poorter et al. 2005), and in low light survival (Kneeshaw et al. 2006). If such cross-overs were widespread, a multi-dimensional view of species light requirements might be required to understand forest dynamics, and species coexistence in old-growth forests (Grubb 1977; Poorter et al. 2005). This idea is disputed by Gilbert et al. (2006), who found that seedlings and saplings of Neotropical woody plants showed similar growth-survival trade-offs at seedling and sapling stages, despite extensive rank changes.

Interspecific differences in leaf life span might be expected to influence ontogenetic variation in light requirements of broadleaved evergreens (King 1994; Lusk 2004). Young seedlings of pioneer species often develop a large ratio of leaf area to plant biomass (leaf area ratio, LAR), leading to favourable short-term carbon balance and relatively rapid growth even in low light (Kitajima 1994; Walters & Reich 1999; Lusk 2004). However, their short leaf lifetimes lead to a steep ontogenetic decline in low light LAR (Lusk 2004), because in the understorey they are unable to compensate their high leaf loss rates (due to senescence, herbivory and mechanical damage) with high leaf production rates (Bongers & Popma 1990; Lusk 2002). Whole-plant light compensation points (Givnish 1988) of such taxa therefore seem likely to increase rapidly as they grow larger. Species with long-lived leaves, on the other hand, are better able to conserve their ratio of leaf area to biomass as they grow bigger, through accumulation of many overlapping leaf cohorts: this pattern is seen in late-successional shade-tolerant evergreens (Lusk 2004). Light requirements of these species might therefore be expected to be more ontogenetically stable than those of pioneer species with short leaf lifetimes.

Here we relate ontogenetic variation in shade tolerance to leaf life span differences among 13 coexisting evergreens in a temperate rainforest. We compared the minimum light levels naturally occupied by three juvenile size classes of the 13 species, and measured leaf life span in low light. We addressed three questions: (i) Do light requirements vary significantly between size classes? (ii) Do species’ ranks change significantly between size classes? (iii) Is the magnitude of any ontogenetic shift in minimum light requirements (MLRs) correlated with interspecific variation in leaf life span?

Materials and methods

species distributions in relation to canopy openness

Sampling was carried out in the low-altitude forests (350–440 m a.s.l.) of Parque Nacional Puyehue (40°39′ S, 72°11′ W) located in the western foothills of the Andean range in south-central Chile. This area experiences a maritime temperate climate, with an average annual precipitation of around 3500 mm (Almeyda & Saez 1958). The old-growth rainforest at this altitude on the western foothills of the Andes is comprised exclusively of broad-leaved evergreens (Lusk, Chazdon & Hofmann 2006a).

Distributions of juvenile trees were quantified in relation to canopy openness measurements made with a pair of LAI-2000 canopy analysers (Li-Cor, Lincoln, NE). 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. Integration of data from the two instruments enabled estimation of percentage diffuse irradiance at each sampling point within the forest, equivalent to percentage of canopy openness over the quasi-hemispherical (148º) field of view perceived by the LAI-2000 sensors. Measurements were made on overcast days, using the full 148º field of view, over a period of about 4 years from 2000 to 2003. 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 transects run through old-growth stands including tree-fall gaps of varied sizes. Sets of parallel transects were run through accessible stands, spaced at least 20 m apart, the angle and number of transects depending on terrain, access considerations and proximity to forest margins. At 1608 sample points spaced at random intervals (2–10 m apart) along transects, canopy openness measurements were made at 50, 100 and 200 cm height with the LAI-2000. Presence of tree and large shrub species was recorded in three height classes in a circular plot of 1-m diameter, centred on the sample point. Juveniles 10–50 cm tall were recorded as associated with the light level measured at 50 cm of height. Juveniles 50–100 cm tall were referred to the light level measured at 100 cm of height, and individuals 100–200 cm tall were referred to the light level measured at 200 cm of height. Although up to 20 individuals of some species were found in some plots, only presence or absence data are used in the present analysis.

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 light requirements in the present study, however, we counted only juveniles of seedling origin.

quantifying minimum light requirements (MLRs)

The 10th percentile of the distribution of each species in relation to light availability (percentage of canopy openness) was used as an approximation of the lowest light levels tolerated by each species (Fig. 1). This parameter, referred to hereafter as the MLR, was calculated for each size class of each species. Only species represented in each size class on at least 20 sampling plots were considered. We stress that our scores are an inversion of traditional shade tolerance ratings (Donoso 1989) that is, shade-tolerant taxa such as Myrceugenia planipes have low values for this index, and light-demanders such as Aristotelia chilensis score high (Fig. 1).

Figure 1.

MLRs of three juvenile size classes of 13 evergreens in a Chilean rainforest. Solid lines denote species with statistically significant changes (P < 0·05) in MLR between the smallest and largest height classes; dotted lines indicate no significant change across this size range.

statistical analysis

We used a resampling approach to test for significant variation in MLR of each species across size classes (Table 1). The distribution of light environment data was normalized by log 10 transformation before analysis. For each species, we computed the 10th percentile of the distribution of light environments occupied by each of the three size classes, as an indicator of MLR. The absolute difference (corresponding to a two-tail test) in MLR among size classes was then taken as an indicator of changing light requirements. To estimate the significance of observed differences, light levels were randomized among size classes and differences in MLR recalculated. This procedure was repeated 1000 times. The proportion of samples showing greater than observed difference between each pairwise combination of size classes gives the probability that such a difference between size classes could arise by chance (Table 1).

Table 1.  Leaf life spans (mean ± SE) of saplings in low light (2%–5% canopy openness), and results of resampling tests for significant variation in MLRs across three height classes (10–50, 50–100, 100–200 cm) of each of 13 species. MLRs were computed as the 10th percentile of the distribution of light environments (percentage of canopy openness) occupied by each height class
SpeciesLeaf life span (year)
(mean ± SE)
P-values for height class comparisons
1 vs. 31 vs. 22 vs. 3
Aristotelia chilensis (Elaeocarpaceae)0·8 ± 0·10·0060·1780·02
Azara lanceolata (Flacourtiaceae)2·2 ± 0·3< 0·001< 0·0010·097
Dasyphyllum diacanthoides (Asteraceae)2·7 ± 0·4< 0·0010·0170·237
Eucryphia cordifolia (Cunionaceae)2·8 ± 0·20·6970·9070·59
Lomatia ferruginea (Proteaceae)2·8 ± 0·30·0290·9560·055
Caldcluvia paniculata (Cunoniaceae)2·9 ± 0·20·0090·5740·054
Amomyrtus luma (Myrtaceae)3·3 ± 0·3< 0·001< 0·0010·091
Rhaphithamnus spinosus (Verbenaceae)3·6 ± 0·40·2010·0870·679
Luma apiculata (Myrtaceae)3·8 ± 0·20·0030·0230·343
Laureliopsis philippiana (Atherospermataceae)4·7 ± 0·60·8320·9970·834
Myrceugenia planipes (Myrtaceae)5·0 ± 0·20·1320·2390·645
Gevuina avellana (Proteaceae)5·2 ± 0·40·2310·6280·544
Aextoxicon punctatum (Aextoxicaceae)5·3 ± 0·3   0·530·9580·562

Kendall's coefficient of concordance was used to test significance of agreement of species MLR ranks across size classes.

leaf life span

Leaf life spans of 11 of the 13 species were estimated by monitoring leaf survival over 12 months. All leaves were marked on the main stem of five to six 100–200-cm tall saplings of each species, growing at microsites with 2%–5% canopy openness as measured by the LAI-2000. Plants were relocated 12 months later, and leaf mortality during this period used to estimate average leaf life span. As leaf longevities were < 1 year on most individuals of A. chilensis, abscission scars were counted to determine mortality of new leaves initiated after the start of the study period. Leaf longevity (year) was estimated as:


where ni = initial number of leaves, nf = final number surviving from ni, and mn = mortality of new leaves initiated since the first census.

Static demographic methods were used estimate leaf life spans of the other two species, Eucryphia cordifolia and Dasyphyllum diacanthoides. The end of previous seasons’ extension growth in these species is marked by persistent cataphylls (E. cordifolia) or by resting bud scars which remain visible for several years (D. diacanthoides), permitting reconstruction of recent growth history and ready delimitation of foliage cohorts. We estimated mean leaf life spans of these species by inspecting 10 saplings growing at microsites with 2%–5% canopy openness as measured by the LAI-2000, and determining the age of the youngest foliage cohort that had undergone c. 50% mortality. Estimates of mean leaf life span were averaged for the 10 individuals of each species.


Agreement of ranks of species’ MLRs across size classes was significant, although not strong (Kendall's W = 0·159, P = 0·02).

MLRs of seven species increased significantly between height classes 1 and 3 (Fig 1; Table 1). The other six species, which showed no significant ontogenetic variation, included the three most shade-tolerant trees in this forest (M. planipes, Aextoxicon punctatum, Laureliopsis philippiana), which occupied the lowest three rankings in the largest size class. Interspecific variation in MLRs was thus wider in the largest height class (log-variance = 0·051), than in the medium (log-variance = 0·023) and small (log-variance = 0·027) height class.

Leaf life spans ranged from < 1 year in A. chilensis to ≥ 5 years in Gevuina avellana and A. punctatum (Table 1). Leaf life spans were negatively correlated with MLR in all size classes (Fig. 2). However, in the smallest size class the correlation was modest (r = 0·60) and driven mainly by one outlier, A. chilensis (Fig. 2). The correlation between leaf life span and MLR became progressively stronger with increasing size, rising to 0·88 in the largest size class (Fig. 2). As a result, leaf life span was negatively correlated with proportional change in MLR across the three size classes (Fig. 3).

Figure 2.

Correlation of leaf life span with MLRs of rainforest evergreens in three height classes. Black symbols and solid line show relationship for 10–50-cm tall juveniles (r = 0·60, P = 0·032). Grey symbols and long-dashed line = 50–100 cm tall juveniles (r = 0·83, P < 0·0001). White symbols and short-dashed line = 100–200-cm tall juveniles (r = 0·88, P < 0·0001).

Figure 3.

Correlation of leaf life span with proportional change in MLRs across three height classes.


Although most species underwent rank changes, there was significant agreement of species light requirements across height classes (Fig. 1). While some of the minor rank changes could be attributed to random variation and small sample sizes, the more substantial rank changes shown by Azara lanceolata suggest a stronger upward trend in MLR than those of any of its associates. We are aware of few published data directly comparable to ours. A somewhat different problem is addressed by studies documenting variation in the average light environments occupied by suites of tree species during the course of their lives (Clark & Clark 1992; Poorter et al. 2005). Many of the rank changes in those studies appeared to reflect differences in maximum height, with short-statured species eventually being overtopped by taller ones (Clark & Clark 1992; Poorter et al. 2005). However, physiological mechanisms similar to those discussed in the present study could also have some influence on their results. A study of 31 Neotropical evergreens (Gilbert et al. 2006) showed that seedling survival of was strongly correlated overall with survival of the same taxa at the sapling stage, despite numerous (and sometimes substantial) rank changes between size classes. However, Gilbert et al. (2006) did not compare species’ survival in a common light environment, nor were light environments standardized between the two size classes, making it difficult to compare our results directly with theirs.

The effect of size on light requirements appeared to depend on species’ successional status. Species considered or shown to be tolerant of shade, especially M. planipes, A. punctatum and L. philippiana (Donoso 1989; Figueroa & Lusk 2001; Lusk 2002), had low MLRs, which did not change significantly between height classes (Fig. 1, Table 1). In contrast, MLRs of most of the less-tolerant species underwent ontogenetic increases (Fig. 1). Light requirements of the 13 species were thus most clearly differentiated in the largest size class. Again there are few directly comparable data: Kneeshaw et al. (2006) found that shade-tolerance differences between seven boreal tree species became slightly less marked with increasing size, although that study was based on variation in survival rates at a given growth rate, rather than the range of light environments naturally occupied by species.

Ontogenetic changes in light requirements were correlated with interspecific variation in leaf life span (Figs 2 and 3). The accretion of multiple leaf cohorts by shade-tolerant evergreens (Williams, Field & Mooney 1989; King 1994; Lusk 2002), plus a probable ontogenetic decline in allocation to roots (Lusk 2004), enables these species to maintain their LAR as seedlings grow larger (Lusk 2004). This suggests conservation of the relationship of carbon gain to respiratory demands, although progressive loss of older leaf cohorts in juveniles > 1 m tall means that this isometry of leaf area with plant biomass cannot continue indefinitely. In contrast, light-demanding species such as A. chilensis and A. lanceolata have more marked differences between turnover rates of foliage and woody tissues (Bongers & Popma 1990; Lusk 2004). In shaded understoreys, where their relatively fast foliage turnover cannot be compensated by high leaf production rates, increasing age and size of light-demanding evergreens must thus inevitably be associated with an increasing ratio of heterotrophic to autotrophic tissues. In order to survive, they must therefore encounter increasing light levels as they grow taller, enabling higher carbon gain per unit leaf area.

Our approach of relating the distribution of juvenile trees to present canopy openness may slightly underestimate MLRs (Lusk et al. 2006a). Because of the dynamic nature of forest light environments, some individuals were probably sampled in light environments below their whole-plant light compensation points, as carbohydrate reserves could enable juvenile trees to persist for some time in light environments in which their net carbon gain is negative. As there is evidence that shade-tolerant species have larger carbohydrate reserves than their light-demanding associates (Poorter & Kitajima 2007; but see Lusk & Piper 2007), this sort of underestimation could be greater in the former, leading to an over-estimation of species differences in MLR. However, this sort of error would not affect our conclusions about ontogenetic trends.

It is unlikely that spatial autocorrelation of forest light environments has any bearing on our conclusions. Nicotra, Chazdon & Iriarte (1999) found that light environments in an old-growth tropical rainforest showed spatial dependence at scales of up to 25 m. This scale of pattern suggests some degree of dependence between successive observations spaced 2–10 m apart on our transects. If some species (e.g. those with large seeds) are more clumped than others (Dalling, Hubbell & Silvera 1998), spatial autocorrelation of light environments will not affect all species equally. However, it is not evident how this sort of pattern could influence our measurements of ontogenetic variation in species’ light requirements.

Other traits besides leaf life span could also influence ontogenetic shifts in light requirements. Leaf mass per area increases during juvenile ontogeny (Sack, Maranon & Grubb 2002; Lusk 2004), at least partly as a result of increasing leaf thickness (Kenzo et al. 2006). Unless offset by shifts in other traits (e.g. declining allocation to roots), the higher respiration rates associated with thicker leaves, as well as the rising cost of producing photosynthetic surface area, are likely to inflate whole-plant light compensation points as plants grow taller. Self-shading is likely to reduce the efficiency of light interception as plants grow taller (Lusk et al. 2006b), although some broadleaved species are surprisingly adept at minimizing self-shading by repositioning old leaves through bending of petioles, with substantial consequences for carbon gain (Gálvez & Pearcy 2003). Different architectural models (Hallé, Oldeman & Tomlinson 1978) could also have differing consequences for the allometry of light interception, and hence for ontogenetic variation in light requirements. Finally, ontogenetic trends in herbivory could also influence both the absolute and relative light requirements of species. Herbivore pressure has been shown to significantly influence light requirements (DeWalt, Denslow & Ickes 2004); if increasing size makes plants more attractive to vertebrate herbivores (Boege & Marquis 2005), their impact could contribute to ontogenetic increases in the MLRs of species with short-lived, relatively palatable foliage.

If differences in leaf life span do drive ontogenetic divergence in the behaviour of light-demanding and shade-tolerant species, there is no reason to expect such a pattern in deciduous forests. The gulf between turnover rates of foliage and woody tissues in all deciduous species should result in an ontogenetic decline in LAR, irrespective of shade tolerance level. Light requirements of all deciduous trees should therefore increase as they grow taller. Kneeshaw et al. (2006) reported that saplings of four out of five boreal deciduous trees had lower understorey survival than conspecific seedlings, although their expression of survival in function of growth (rather than light availability) makes comparisons problematic.

Evergreen conifers are likely to show a different pattern again. Small leaves and lack of petiole development make it difficult for conifers to accumulate multiple leaf cohorts without incurring considerable self-shading (Lusk et al. 2006b). Thus, although some conifers accumulate as many as 20 foliage cohorts (Lusk 2001), and are able to conserve their LAR for many years during juvenile ontogeny (Lusk et al. 2006b), this does not necessarily translate into isometry of light interception and carbon gain with plant mass. Light requirements of most evergreen conifers therefore seem likely to rise with increasing size, although this ontogenetic trend may be steeper in light-demanders than in shade-tolerant taxa.

Understanding coexistence of tree species, and development of silvicultural programs in mixed forests, will require knowledge of species light requirements throughout their developmental trajectories (Poorter et al. 2005). This paper shows that work on saplings can detect shade tolerance differences not evident from comparisons of seedlings, but which may nevertheless have considerable bearing on forest regeneration patterns. In identifying a relationship with leaf life span, this work also provides a basis for predicting changes in species’ light requirements during juvenile ontogeny.


We thank FONDECYT for support through grant no. 1980084, CONAF for facilitating access to Parque Nacional Puyehue, and an anonymous reviewer for constructive criticism.