Functional Ecology 2001 15, 113 –123 Light-dependent changes in biomass allocation and their importance for growth of rain forest tree species Blackwell Science, Ltd L. POORTER Department of Plant Ecology, Utrecht University, P.O. Box 80084, 3508 TB Utrecht, The Netherlands, and Programa Manejo de Bosques de la Amazonía Boliviana (PROMAB), Casilla 107, Riberalta, Bolivia Summary 1. Sapling growth of six rain forest tree species was compared to evaluate whether species respond in a similar way to a natural light gradient. Saplings were measured non-destructively; production and loss of leaves, stem and branches were analysed in detail. 2. Sapling height growth was positively related to light environment and leaf area. No single descriptor of light environment explained sapling growth best. Direct or diffuse light could explain plant growth, depending on species. 3. Seventeen percent of the saplings had negative relative biomass growth rates, although they occurred in fairly bright conditions. Negative growth rates were caused by leaf shedding and stem breakage. Sapling relative growth rate increased with irradiance, mainly because of an increase in net assimilation rate. 4. On a shoot basis, shaded plants had a smaller leaf mass fraction (LMF) and a larger specific leaf area, resulting in similar leaf area ratios (LAR) to those of sun plants. This contrasts with the results of seedling studies under controlled conditions, where LMF and LAR increased with shade. 5. Biomass partitioning to leaf growth decreased with irradiance and relative growth rate of the sapling. This leaf partitioning ratio was better correlated with RGR than with irradiance. 6. Species differed in the effect of light-dependent changes in specific leaf area (SLA) on growth. This underscores the importance of SLA in explaining differences in species performance in a forest environment. Nevertheless, the effect of SLA was not related to the shade tolerance of the species. Key-words: Biomass partitioning, Bolivia, growth analysis, shade tolerance, tropical rain forest Functional Ecology (2001) 15, 113–123 Introduction In tropical rain forests, light is probably the most important environmental factor affecting plant establishment, growth and survival. Plant performance is enhanced through morphological and physiological acclimation to the light environment. Growth analyses of tree seedlings under controlled conditions indicate how such plants adjust to the light environment (e.g. Popma & Bongers 1988; Osunkoya et al. 1994; Veenendaal et al. 1996; Poorter 1999). At low irradiance, shade plants enhance interception of light by a large biomass fraction in leaves (leaf mass fraction, LMF). In combination with a large leaf area per unit leaf biomass (specific leaf area, SLA), this leads to a © 2001 British Ecological Society Present address: Silviculture and Forest Ecology, Department of Environmental Sciences, P.O. Box 342, 6700 AH Wageningen, The Netherlands. large interceptive leaf area per unit plant mass (leaf area ratio, LAR). Shade leaves have slow respiration and light-saturated photosynthetic rates (Langenheim et al. 1984; Oberbauer & Strain 1985). Little physiological activity means smaller maintenance costs, and in this way, carbon losses in the understorey are reduced and potential relative growth rates (RGR) enhanced (Lehto & Grace 1994; Sims, Gebauer & Pearcy 1994). At high irradiance, nutrient and water availability may limit plant growth. Accordingly, sun plants invest relatively more biomass in roots (i.e. they have a large root mass fraction, RMF). Carbon fixation is increased by the formation of thick leaves with fast light-saturated photosynthetic rates, increasing biomass growth per unit leaf area (net assimilation rate, NAR) and potential RGR. Plant growth in the understorey is limited by the amount of light intercepted for photosynthesis. Growth in the understorey may depend on the amount 113 114 L. Poorter of incident radiation and the leaf area (Oberbauer et al. 1988). Incident radiation is composed of direct light and diffuse light. Because 50–80% of the total daily radiation in the forest understorey may be as sunflecks (Pearcy 1987; Chazdon 1988), direct light is probably a more important determinant of plant growth than diffuse light. However, this is not always supported by field observations (Clark, Clark & Rich 1993). A plant’s carbon balance is affected by the partitioning of fixed carbon to photosynthesizing tissue (leaf partitioning ratio, LPR) and its longevity (i.e. the leaf lifespan). At low irradiance, a greater biomass partitioning to leaves increases the leaf area. As carbon gain in the understorey is slow, leaf longevity should be increased to return the construction costs of the leaves (Chabot & Hicks 1982; Williams, Field & Mooney 1989). Pioneer species have short-lived leaves (Bongers & Popma 1990; Reich, Ellsworth & Uhl 1995). They must therefore replace lost leaf area quickly to sustain growth (King 1994), a prerequisite that is difficult to meet in the forest understorey. Plant growth analysis is useful to evaluate how plants differ in growth rate, either inherently or dependent on the environment (Lambers 1998). Few studies have analysed growth of saplings in this way (but see King 1991, 1994). To evaluate whether (i) species differ in their response to the light environment, and (ii) similar patterns are found for large saplings in the field compared to small seedlings grown under controlled conditions, a field study was carried out on growth and biomass partitioning of six rain forest tree species differing in shade tolerance. Naturally established saplings occurring along a light gradient were selected, and the growth and turnover of shoots and their components were analysed. The following questions were addressed: 1. How is height growth related to light environment, and what factors cause this relationship? It is hypothesized that sapling growth depends on the amount of incident radiation and leaf area, and that direct light is a more important determinant for plant growth than diffuse light. 2. How does RGR and its components (NAR, LAR, LMF and SLA) vary with irradiance? Do species differ in light-dependent changes in biomass allocation, and what is their effect on growth? It is expected that NAR increases with irradiance, due to faster photosynthetic rates, and that LMF, SLA and LAR increase with shade to enhance light interception. 3. How do LPR and leaf lifespan vary with light environment? LPR and leaf longevity are postulated to increase with shading. Because leaf turnover of pioneer species exceed those of shade-tolerant species, it is also expected that they will have a larger leaf partitioning ratio. Materials and methods   The field study was carried out in forest reserve ‘El Tigre’ (10°59′ S, 65°43′ W). El Tigre is a research site of the Programa Manejo de Bosques de la Amazonía Boliviana (PROMAB), located 170 m above sea level in the Amazon region of Bolivia. Mean annual temperature is 26 °C, and annual rainfall is about 1780 mm with a distinct drier period (< 100 mm mo–1) from May to September. Vegetation in the region can be classified as a tropical lowland moist forest, with a canopy 25–35 m high. Some of the canopy trees are deciduous during the dry season. Six tree species were selected based on their light requirements for establishment and survival (Table 1). Classification of the species – as pioneer, non-pioneer light-demanding, and shade-tolerant – was based on observations of the abundance of saplings in different microhabitats. Cecropia ficifolia and Bellucia pentamera are typical pioneer species, which establish soon after disturbance. Tachigali vasquezii is a non-pioneer lightdemanding species with intermediate light requirements. Table 1. The studied species and their characteristics, including strategy, adult height, seedling light compensation point (LCPNAR, see Poorter 1999), mode of lateral crown expansion, sapling height and direct site factor. Mean values followed by a different letter were significantly different at P = 0·05 (Student Newman–Keuls test) Species Family Cecropiaceae Cecropia ficifolia Warb. Bellucia pentamera Naud. Melastomataceae Tachigali vasquezii J.J. Pipeloy Fabaceae Cariniana micrantha Ducke Lecythidaceae Capirona decorticans Spruce Rubiaceae Theobroma speciosum Willd. Sterculiaceae ex Sprengel © 2001 British P = Pioneer. Ecological Society, NPLD = Non-pioneer light demander. Functional Ecology, S =113–123 Shade-tolerant. 15, Sapling height (m) Direct site factor Strategy Adult height (m) LCPNAR (%) Expansion mode Mean Range Mean Range n P P NPLD S S S 12 20 –25 25 –30 35 25 –30 15 2·0 1·8 1·2 0·3 – 0·1 petiole branch rachis branch branch branch 1·6a 1·6a 1·7a 1·8a 1·6a 1·6a (0·9–2·6) (0·9–2·8) (1·0–2·9) (0·7–3·3) (0·6–2·5) (0·8–3·0) 25·4a 24·2a 7·0b 5·5b 9·2b 7·3b (2·8–68·2) (4·6–64·7) (1·3–18·7) (1·1–18·4) (2·5–38·7) (2·9–26·2) 16 14 14 16 15 15 115 Light, biomass allocation and growth Cariniana micrantha, Capirona decorticans and Theobroma speciosum are shade-tolerant species. The species differ in adult stature, sapling architecture and whole-plant light compensation point, i.e. the irradiance at which net growth is zero (Table 1; Poorter & Werger 1999). Henceforth, the species will be referred to by their generic name.    © 2001 British Ecological Society, Functional Ecology, 15, 113–123 Approximately 15 saplings per species were selected. For each species, care was taken to select saplings along the whole light gradient, including individuals in large gaps, small gaps and understorey. All saplings occurred on the same soil type (Ferrasols) and presumably experienced the same nutrient availability. Saplings were 0·6–3·3 m tall (Table 1) and had no signs of stem breakage or resprouting. Saplings were measured non-destructively to determine the production and losses of stems, branches and leaves over the growth interval. Saplings were first measured between October and December 1996 and remeasured 11 months later (mean interval 335 days, range 293–389 days). Diameter was determined at three to five permanently marked positions along the stem, including one at 10 cm from its base. The diameter of each branch was measured 5 cm from its base. Stem and branch diameters were measured in two perpendicular directions, using a caliper. Lengths of stem, branches and spacer organs like petioles (Cecropia) and rachae (Tachigali) were also determined. At the final measurement, stem height was measured from the highest marked ring onwards, to accurately estimate height growth. Leaves at the terminal growing points were marked to monitor leaf production and loss. For all species but Cariniana (which has very small leaves), leaf lengths and widths, plus the height of the insertion points of petioles on the stem, were recorded. Leaf damage by herbivory was estimated visually in 10% area interval classes. For Cariniana, leaves were simply counted. Crown width in two perpendicular directions and length of each internode, from the lowest leaf onward, were measured. After the second non-destructive measurement, three leaves per sapling were selected along the stem and their areas determined with a portable leaf area meter (CI-202, CID Inc., Vancouver, WA, USA). These leaves were oven-dried for 48 h at 70 °C. To estimate stem biomass and leaf area from the non-destructive measurements, 10 additional, similarsized saplings per species were selected in the same area. Five saplings were sampled in bright light and five saplings in shade. Three leaves were sampled from each sapling and a 10 cm section was taken from the upper and lower part of the stem. Leaf area and length and width of each leaf blade were measured. An allometric relation was established between ln(leaf area), ln(leaf length) and ln(leaf width) for each species. In this way, total leaf area per plant could be estimated from individual non-destructive leaf measurements. The mean coefficient of determination for the regression equations was 0·96 (range 0·91–0·99). Volume of the stem sections was determined and stem parts were ovendried for 48 h at 70 °C. Specific stem mass was calculated as the stem mass per unit stem volume (SSM, in g cm–3). For species-specific values of SSM, see Table 3 in Poorter & Werger (1999). Stem and branch volume of the monitored saplings were calculated based on size measurements and volume equations (see Poorter & Werger 1999). All areas were measured in cm2, volumes in cm3 and masses in g. Stem mass was obtained by multiplying stem volume by SSM, branch mass by multiplying branch volume by SSM, and leaf mass by dividing the leaf area by the specific leaf area (SLA, the leaf area per unit leaf mass) of the sapling. The leaf mass fraction was calculated as leaf mass per unit above-ground plant mass (LMF, g g–1), and the net above-ground relative growth rate (RGRn, mg g–1 d–1) as (lnM1 – lnM0)/t, in which M0 and M1 are the total shoot masses at the beginning and end of the experiment, respectively, and t is the time interval (days). The RGRn can be factored into a morphological component, the leaf area ratio (LAR, leaf area per unit above-ground mass, m2 kg–1), and a physiological component, the net assimilation rate (NAR, the net above-ground biomass growth per unit leaf area on average present during that time interval, g m–2 d–1) (Hunt 1978). Due to practical limitations, the growth analysis in this study is restricted to the above-ground plant parts. Biomass invested in roots is not taken into account and the reported NAR will therefore be less than the whole-plant NAR. The LMF is a static descriptor of biomass allocation to leaves. It describes the fraction of shoot biomass present as leaves, and is the end result of the production and loss of leaf, stem and branch biomass. In contrast, the leaf partitioning ratio is a dynamic descriptor of biomass allocation to leaves. It indicates the fraction of newly produced, above-ground biomass allocated to the production of new leaves (LPR, g g–1). The LPR is an approximation of the real proportion of carbon gain allocated to the leaves, as it is based on net biomass growth. Note that carbon costs for growth (i.e. to convert fixed CO2 into dry matter) and maintenance respiration are not included in the LPR. Average leaf lifespan (days) was calculated as the ratio of the mean leaf area present at the beginning and the end of the study period (m2), to the mean of the leaf area production and loss over the study period (m2 d–1) (King 1994).     Hemispherical photographs were taken above the saplings in December 1996. Photographs were taken mainly around sunrise or sunset, and sometimes when the sky was overcast. Photos were taken with a Canon AE-1 camera with a 7·5 mm fish-eye lens using an 116 L. Poorter Ilford 125 Asa black and white film. The camera was mounted on a pole and kept horizontal using a levelling device. Negatives were scanned with a Sony XC77CE black and white CCD camera attached to a VIDAS image analysis system (Kontron/Zeiss, Eching, Germany) and analysed with Winphot 5 (Tropenbos Foundation, Wageningen, The Netherlands; ter Steege 1997). With Winphot, the direct site factor (DSF) and the indirect site factor (ISF) were calculated. They indicate what proportion of radiation above the forest canopy reaches the sapling in the form of direct radiation (i.e. sunflecks) and diffuse radiation, respectively (Mitchell & Whitmore 1993).  © 2001 British Ecological Society, Functional Ecology, 15, 113–123 Individuals should be of similar sizes and have experienced similar light conditions in order to make direct comparisons between species’ morphology and growth responses to light. Mean sapling height ranged from 1·6 to 1·8 m, and was similar for all species (Table 1, one-way , F5,84 = 0·29, P > 0·05). The DSF differed considerably between species (Table 1, one-way , F5,84 = 6·1, P < 0·001); the mean DSF of the two pioneer species (24 –25%) was significantly greater than the DSF of the other species (5·5 –9·2%). DSF was similar for all species when individuals from large gaps (DSF > 30%) were excluded from the analysis (one-way , F5,73 = 1·4, P > 0·05). Therefore, two types of analysis were performed. For intraspecific comparisons, all saplings were included. For interspecific comparisons, only the restricted data set was used. To analyse which plant features and abiotic factors were associated with intraspecific differences in height growth, a backward multiple regression was carried out for each species, using plant growth as dependent variable and DSF, ISF and initial plant height and leaf area as independent variables. Initial plant height was included as an independent variable because growth rate may be size-dependent. Height growth is a function of internode length and production rate, both of which are probably influenced by sapling size and the light environment. Therefore, for each species, three regression analyses were carried out: two multiple regressions with the internode production rate or mean internode length as dependent variables and height and DSF as independent variables; and one multiple regression of height growth against internode length and internode production rate. A growth response coefficient (GRC) was calculated to evaluate how responses in biomass partitioning to light affect the RGR of a plant. The GRC indicates how intraspecific variation in RGR is associated with variation in one if its components. If RGR is analysed as the product of NAR and LAR, then the sum of the GRCs of NAR and LAR should equal unity. Alternatively, RGR can be defined as the product of NAR, LMF and SLA, and their three corresponding GRCs should also sum to unity. A GRC value of 1 indicates that a change in the growth parameter being considered leads to a proportional change in RGR. GRCs can be > 1 if the increase in growth parameter is stronger than the increase in RGR, 0 if an increase in the growth parameter does not lead to a change in RGR, and negative if an increase in the growth parameter is associated with a decrease in RGR (Poorter & Nagel 2000). The GRCs can be obtained by calculating the regression slope of ln(NAR) or ln(LAR) against ln(RGR) for each species. Saplings with a negative RGR were excluded from this regression analysis. For more information on GRCs see Poorter & Van der Werf (1998) and Poorter & Nagel (2000). To evaluate whether species differ in their GRC, an  was carried out in which the ln-transformed growth parameter was used as dependent variable, species as a factor and ln(RGR) as a covariate. A significant species– covariate interaction indicates that species differ in their GRC. To test whether species differ in the effect of lightdependent changes in allocation on growth, the speciesspecific GRC was correlated with the whole-plant light compensation point (LCPNAR), estimated from an independent growth experiment (Poorter 1999; Table 1). The LCPNAR of Capirona is not known, but as this is a highly shade-tolerant species, it was set to 0·1%. All statistical analyses were carried out using SPSS 6·0 (SPSS Inc., Chicago, IL, USA; Norusis 1993). Results Seven of the 91 saplings died during the study period. Cecropia had the poorest survival (69%), whereas 92–100% of the saplings of the other species survived. Initial plant height or DSF had no effect on survival probability when all species were pooled (logistic regression, P > 0·05). Height growth was positively related to light environment, leaf area, or a combination of the two (Table 2), whereas initial plant height had no effect. The regression models explained 76% of the variation, on average (range 27– 98%). Leaf area affected the growth of all species but Theobroma. The light environment affected growth of all species but Cariniana. For the other four species, the combined effects of light and leaf area explained more than either alone. There was no single measure of the light environment that best explained height growth for all species; for some species, height growth was more closely related to DSF, whereas for other species it was more closely related to ISF. An increased height growth at higher irradiance was due to the combined effect of a greater internode production rate and longer internodes (Table 3). For some species, the enhanced height growth was largely due to an increase in internode production rate (e.g. Cariniana). For other species (e.g. Theobroma), it was mainly due to increased internode length. Fourteen of the 84 surviving individuals (17%) had a net negative RGR, as a result of leaf shedding, branch loss or stem breakage. Saplings which showed negative 117 Light, biomass allocation and growth Table 2. Effect of direct site factor (DSF) indirect site factor (ISF), initial height and initial leaf area on height growth of six rain forest tree species. A backward multiple regression was carried out, and standardized partial regression coefficients are shown for those variables which contributed significantly to the model Species DSF (%) ISF (%) Height (cm) Leaf area (m2) r2 Cecropia Bellucia Tachigali Cariniana Capirona Theobroma 0·46* – – – 0·87*** 0·52* – 0·81*** 0·77*** – – – – – – – – – 0·56* 0·24** 0·33* 0·78*** 0·36*** – 0·84 0·98 0·85 0·60 0·92 0·27 – P > 0·05; * P < 0·05; ** P < 0·01; *** P < 0·001. Table 3. Effect of initial plant height and light environment (DSF) on internode production rate (IPR) and internode length (IL) of six rain forest tree species. Regression coefficients (b) indicate the effect of 1 cm change in height or 1% change in direct light on the dependent variables. Light has an effect on height growth (expressed as cm change in height per year per percent change in DSF) via the IPR and via the IL Internode production rate (year−1) Internode length (cm) Height (cm) DSF (%) Height (cm) DSF (%) Species b P b P r2 b P b P r2 Via IPR Via IL (cm year−1 %−1) Cecropia Bellucia Tachigali Cariniana Capirona Theobroma 0·06 0·00 0·04 –0·03 0·01 0·01 * ns * ns ns ns 0·37 0·15 0·40 2·04 0·19 –0·03 *** *** * ** *** ns 0·91 0·90 0·61 0·84 0·72 0·21 0·01 –0·02 0·00 0·00 0·02 0·01 ns ns ns ns ns ns 0·06 0·46 0·65 0·18 0·29 0·87 * *** *** ns * ns 0·52 0·95 0·67 0·46 0·36 0·29 2·1 1·4 3·5 5·3 1·7 –0·1 Effect DSF on height growth 0·9 3·7 3·3 1·0 0·9 0·9 – P > 0·05; * P < 0·05; ** P < 0·01; *** P < 0·001. Table 4. Relation between sapling light environment (direct site factor, %) and net RGR (RGRn ), NAR, LAR, LMF, SLA and leaf lifespan. LAR and LMF are given for the initial measurement, SLA for the final measurement. Pearson’s correlation coefficients are shown and significance levels are indicated by * Species RGRn (mg g–1 d–1) NAR (g m–2 d–1) LAR (m2 kg–1) LMF (g g–1) SLA (m2 kg–1) Lifespan (yr) Cecropia Bellucia Tachigali Cariniana Capirona Theobroma 0·70* 0·95*** 0·85*** – 0·89*** – 0·90*** 0·95*** 0·89*** – 0·90*** 0·51* – – – – – – 0·64** 0·73** – – 0·74** – –0·79** –0·81** –0·58* – –0·70** –0·66** –0·74** – –0·70** – – – – P > 0·05; * P < 0·05; ** P < 0·01; *** P < 0·001. © 2001 British Ecological Society, Functional Ecology, 15, 113–123 growth rates were not confined to severely shaded microsites; theywere also growing in relatively brighter conditions (mean DSF = 5·0%, range 2·9 – 9·5). In general, net RGR (RGRn ) and NAR were positively correlated with DSF (Table 4, Fig. 1). Responses of RGRn and NAR did not show a saturating response to light over the light range evaluated here. There was a striking lack of significant correlations between LAR and light environment as LMF and SLA showed opposite responses to irradiance. For three species, the LMF increased with irradiance whereas for five species, the SLA decreased. An increase in RGR with irradiance was closely paralleled by an increase in NAR (Fig. 1); the GRCNAR of the species ranged from 0·88 to 1·23 (Fig. 2a), and was significantly different from zero in all cases. For four species, an increase in LMF with irradiance led to an increase in RGR (and positive GRCLMF; Fig. 2c). This was counterbalanced by a decrease in SLA with irradiance and a negative GRCSLA. As a consequence, adjustments in LAR had little effect on RGR, and GRCLAR values were close to zero (Fig. 2b). Species tended to differ in their GRCs, but this was only significant for GRCSLA (GRCNAR, F = 2·0, P = 0·089; GRCLAR, F = 2·1, P = 0·081; GRCLMF, F = 2·2, P = 0·071; GRCSLA, F = 3·1, P = 0·016). The two pioneer species, Cecropia and Bellucia, had the largest GRCLMF and GRCSLA (Fig. 2), but overall, the GRCs were 118 L. Poorter Fig. 1. Relationship between above-ground RGRn, NAR, LAR, LMF, SLA and sapling light environment (DSF) for Cecropia (d) and Bellucia (s) (a); Tachigali (d) and Cariniana (s) (b); and Capirona (d) and Theobroma (s) (c). LAR and LMF are given for the initial measurement, SLA for the final measurement. Continuous regression lines refer to the filled symbols, and broken regression lines to the open symbols. Regression lines are only shown if they are significant; + P < 0·054; * P < 0·05; ** P < 0·01; *** P < 0·001. © 2001 British Ecological Society, Functional Ecology, 15, 113–123 not correlated with the whole-plant light compensation point of the species (Spearman’s rank correlation between GRCs and LCPNAR of the six species: GRCNAR, r = – 0·17, P > 0·10; GRCLAR, r = 0·17, P > 0·10; GRCLMF, r = 0·78, P = 0·066; GRCSLA, r = – 0·32, P > 0·10). Species differed considerably in their LPR (one-way , F5,64 = 4·74, P < 0·001). On average, Cecropia had the greatest LPR (92%), and Theobroma had the smallest (17%) (Fig. 3a). Species also had different leaf lifespans (one-way , F5,63 = 5·6, P < 0·001), from 0·4 years for Cecropia to 3·2 years for Tachigali (Fig. 3b). LPR was negatively correlated with irradiance and RGRn, although correlations were stronger with RGRn (three significant relationships) than with DSF (two significant relationships) (Fig. 4, Table 5). The leaf lifespan of Cecropia and Tachigali was related negatively to light environment (Table 4); plants in shade had leaf lifespans which were, on 119 Light, biomass allocation and growth Fig. 3. Interspecific variation in leaf partitioning ratio (LPR) and leaf lifespan for six rain forest tree species. Back-transformed logarithmic means and corresponding standard errors are shown. Bars with a different letter were significantly different at P = 0·05 (Student Newman–Keuls test). Species are ordered along the X-axis according to their shade tolerance. Only saplings with a direct site factor < 30% were included. average, half (Cecropia) to three times (Tachigali) longer than plants in a high-light environment (data not shown). Discussion © 2001 British Ecological Society, Functional Ecology, 15, 113–123 Fig. 2. Growth response coefficient (GRC) (mean ± SE) of net assimilation rate (NAR) (a); leaf area ratio (LAR) (b); leaf mass fraction (LMF) (c); and specific leaf area (SLA) (d) for six rain forest tree species. LAR and LMF are taken as the average over the growth interval. Species are ordered along the X-axis according to their shade tolerance, with the most shade-tolerant species on the right. * indicates whether the GRC is significantly different from 0 (P < 0·05). Note that the scaling of the Y-axis differs between graphs. Despite careful selection of saplings on the basis of their light environment, the average irradiance experienced by the pioneer species still exceeded those of the non-pioneer species (Table 1). Unintentionally, this reflects differences in microhabitat distribution between pioneer and shade-tolerant species (cf. Hubbell & Foster 1986; Clark et al. 1993). When working with naturally established populations in the field, it is difficult to establish whether observed differences are caused by inherent adaptations to the light environment or the acclimation of plants to the contrasting environments in which they occur. Interspecific comparisons should 120 L. Poorter Table 5. Correlation between leaf partitioning ratio (LPR), the light environment (DSF) and carbon balance (RGRn) of saplings of six rain forest tree species. Pearson’s correlation coefficient and significance levels are shown LPR (g g–1) Species DSF (%) RGRn (mg g–1 d–1) Cecropia Bellucia Tachigali Cariniana Capirona Theobroma –0·94*** 0·18 ns –0·44 ns –0·44 ns –0·55* –0·05 ns –0·70* 0·19 ns –0·49 ns –0·63* –0·70** 0·20 ns ns P > 0·05; * P < 0·05; ** P < 0·01; *** P < 0·001. Fig. 4. Leaf partitioning ratio (LPR) vs sapling RGRn for Cecropia (d) and Bellucia (s) (a); Tachigali (d) and Cariniana (s) (b); and Capirona (d) and Theobroma (s) (c). Continuous regression lines belong to the filled symbols, and broken regression lines belong to the open symbols. Regression lines are only shown if they are significant; * P < 0·05; ** P < 0·01; *** P < 0·001. Note that the scaling of the X-axis differs between graphs. therefore be made only in the shared light range in which both species groups occur. the findings of other studies (Oberbauer et al. 1988, 1989; Sterck et al. 1999) and supports the hypothesis that sapling growth in a forest environment is limited by the amount of light intercepted for photosynthesis. Leaf loss due to herbivory or leaf shedding has, therefore, important repercussions for plant growth. Using regression equations (data not shown), it can be calculated that for an average plant (leaf area = 0·5 m2, DSF = 7·8), a 10% reduction in leaf area also implies a 10% reduction in height growth. In the forest understorey, 50–80% of the PPFD is in the form of sunflecks (Pearcy 1987; Chazdon 1988). Therefore, it was expected that sapling growth would be most closely related to DSF. This hypothesis was not confirmed by the results, however; for two species, even ISF was a better predictor of sapling height growth than DSF (Table 2). Few studies make an explicit distinction between the effect of direct and diffuse radiation on plant growth (e.g. Canham 1988; King 1991; Oberbauer et al. 1993; Clark et al. 1993), and from these, no clear patterns emerge; if growth is affected by light environment, then growth can be correlated with direct light, diffuse light, or both. In the understorey, a close relation is found between direct light and diffuse light. This relation becomes weaker at stronger irradiances (Clark et al. 1993), and perhaps there is sufficient uncoupling between the two light descriptors to distinguish between the effects of direct and diffuse light on plant growth. Gap studies (Sipe & Bazzaz 1995; Poorter 1998) have shown that diffuse light is a better predictor of plant growth than direct light, probably because peak irradiances above the photosynthetic light saturation point are not used for carbon fixation (e.g. Wayne & Bazzaz 1993), and can inhibit plant growth.             © 2001 British Ecological Society, Functional Ecology, 15, 113–123 For all species but Theobroma, height growth was positively related to sapling leaf area, light environment, or a combination of the two (Table 2). This agrees with Most species reduced their internode length in the shade (Table 3, cf. King 1994; Sterck 1999). In this way, smaller masses of woody support tissue were invested for the new production of a unit leaf tissue. This is at the risk of an increased self-shading of the 121 Light, biomass allocation and growth crown (Horn 1971). The advantage of smaller support mass is that more biomass was available for new leaf growth. This shift in biomass allocation from stem to leaf mass is also reflected in the high LPR of understorey saplings (Table 5). Morphologically, height growth is a function of the length of the internodes and their rates of production. Increased height growth in bright light was attained mainly through an increase in internode production rates, which were up to seven times higher in plants in gaps compared to those in the understorey (Table 3). For herbaceous stoloniferous plants, light quantity mainly affects the production rate, and light quality mainly affects the length of the modules (Stuefer & Huber 1998). However, herbaceous plants increase their module length with shade, whereas tree species decrease their internode length with shade.    -    Eight percent of the saplings died. In addition, 17% of the surviving saplings had a net negative RGR, mostly because of leaf shedding and occasionally because of branch loss or stem breakage. Interestingly, these saplings occurred not only in severely shaded microsites of the forest understorey, but also in brighter sites. Although seedling experiments in controlled conditions have shown that whole-plant light compensation points may occur at 0·5–2% of incident radiation (Boot 1993; Poorter 1999), saplings growing in a forest understorey respond differently. LCPs of forest-grown saplings of several rain forest tree species are between 1% and 5% incident irradiance (King 1994). Differences in LCP between seedling studies in controlled conditions and saplings growing under natural conditions may be partly because of size differences among the plants. The whole-plant light compensation point increases with plant size, due to a smaller LAR and smaller carbon gain per unit plant mass (Givnish 1988; Veneklaas & Poorter 1998). Another cause for the high LCP of forest-grown plants is a stochastic one: under natural conditions, plants lose biomass through herbivory or falling debris. Leaf losses due to herbivory can be considerable: in this study, herbivory accounted for 4% (Cecropia) to 23% (Bellucia) of the standing leaf biomass. Falling debris from the canopy may also damage saplings. In a Costa Rican rain forest, 25% of understorey saplings are exposed annually to such events (Clark & Clark 1989).   R G R n      © 2001 British Ecological Society, Functional Ecology, 15, 113–123 Net RGR and NAR of saplings increased with irradiance (Fig. 1, Table 4), agreeing with other studies describing above-ground sapling growth (King 1991, 1994; Moad 1992 cited in Ackerly 1996). Increases in RGRn with light were largely due to increases in NAR (cf. Poorter & Nagel 2000; Fig. 2a), because bright light increases photosynthetic capacity (Rijkers et al. 2000) and rate. In general, pioneer species show a stronger increase in photosynthetic capacity with irradiance than shade-tolerant species (Strauss-Debenedetti & Bazzaz 1996). Therefore, it was expected that they would also have a larger GRCNAR, yet this was not the case. Sapling LMF increased and SLA decreased with irradiance (Fig. 1, Table 4). LAR remained constant over the light range regarded, as a result of opposing influences of LMF and SLA. Similar results were obtained by Sterck (1997) when studying whole-plant allocation of understorey and clearing saplings of three rain forest tree species. The LAR of Chlorocardium saplings even increases with irradiance, mainly due to an increase in LMF (ter Steege et al. 1994). The greater standing leaf biomass of the sun plants can be explained by their considerably faster leaf production rates, and the similar or shorter leaf lifespan of sun plants compared to shaded plants (Tables 3 and 4, King 1994; Rijkers 2000). The observed patterns are in sharp contrast with the hypothesis. They do not corroborate the results from seedling experiments stating that LMF and LAR of shaded plants exceed those of sun plants (Popma & Bongers 1988; Osunkoya et al. 1994; Poorter 1999). Note that LMF is affected not only by the biomass production, but also by the loss of the different shoot components. In seedling studies, leaf shedding does not occur to a significant extent because the experiments are often too short, and leaf cohorts of freshly germinated seedlings are too young. Species tended to differ in the effect of light-dependent changes in their allocation on growth (Fig. 2), but this was only significant for SLA. SLA is considered to be a key factor in explaining growth differences between species (Poorter & Van der Werf 1998), especially at low irradiances (Poorter 1999). The two pioneer species were most plastic in the response of LMF and SLA to irradiance, and this had important repercussions for their growth (cf. Veneklaas & Poorter 1998). Nevertheless, GRCLMF and GRCSLA were not correlated consistently with species’ shade tolerance, probably because of too few data. Large plasticity in SLA has the potential to increase sapling RGR in the shade. Pioneer species are very plastic in their SLA, but shade-adapted species are not likely to show such a plastic response; large SLA increases the chances of herbivory, and lost leaf material is difficult to replace in a light-limited environment (Kitajima 1996; Walters & Reich 1999).     Observed patterns in standing biomass fractions (LMF, LAR) do not support the idea that shaded plants respond to the light environment by investing more biomass in leaf area for light interception. However, standing biomass fractions reflect the end result of allocation, rather than the process of biomass 122 L. Poorter allocation itself. If the LPR is considered, then it appears that a larger proportion of the biomass growth of shaded plants was allocated to leaves, although only significantly for two out of six species (Table 5). LPR correlated more strongly with RGRn than with DSF. Similar trends in LPR were observed for other rain forest saplings (King 1991, 1994). This suggests that leaf partitioning is influenced by the carbon balance of the plant rather than by the light environment, or that RGRn is a better reflection of the actual light environment than DSF. The standing biomass fraction in leaves is a poor indicator of the biomass partitioning to leaves because the LPR was up to three times larger than the LMF (compare Figs 1 and 3a). The actual LPR may be considerably larger if maintenance respiration is taken into account (Veneklaas & Poorter 1998). Leaf longevity increased with shading for two out of six species (Table 4). Similar results were obtained for seedlings growing under controlled conditions (Bongers & Popma 1990; Ackerly & Bazzaz 1995) and for saplings growing in the field (Sterck 1999). At low irradiance, carbon gain is limited. In theory, therefore, leaf construction costs can be recovered only if leaves live longer (cf. Williams et al. 1989; Rijkers 2000), which is consistent with these observations.     Mean LPR ranged from 20% to 90% (Fig. 3a). It was hypothesized that leaf partitioning rates of pioneer species would exceed those of shade-tolerant species, because their greater leaf turnover rates require more carbon partitioning to leaves in order to replace those that are shed. This hypothesis does not hold because there was neither a systematic difference in leaf lifespan between pioneer and shade-tolerant species (Fig. 3b), nor a significant correlation between species’ LPR and leaf lifespan (Pearson’s r = – 0·60, P > 0·05, n = 6). In a study on several moist forest species in Panama (King 1994), such a negative relation between LPR and leaf lifespan was found (Pearson’s r = – 0·65, P < 0·05, n = 10), probably because more species were included. Interspecific variation in LPR also reflects interspecific variation in sapling architecture. Species with the smallest LPR (Bellucia and Theobroma) tend to form relatively heavy branches; species with the largest LPR have lateral extension units that are cheap to construct (Cecropia) or virtually absent (Capirona) (Poorter & Werger 1999). Conclusions © 2001 British Ecological Society, Functional Ecology, 15, 113–123 Intraspecific variation in sapling growth is determined by irradiance and leaf area per plant. Gap saplings may have standing leaf areas similar to those of understorey saplings, because of a fast rate of leaf production. For understorey saplings, the formation of short internodes, a large biomass partitioning to leaves, a large SLA and long leaf lifespan may enhance light interception and increase the chance of persistence in shade. Species tend to differ in the effect of lightdependent changes in biomass allocation on sapling growth, but this is not related to the shade tolerance of the species. Acknowledgements Staff of Programa Manejo de Bosques de la Amazonía Boliviana are acknowledged for their logistic support, especially Don Nico Divico, Miguel Cuadiay, Rene Aramayo and Luis Apaza. Maarten Terlou of the Department of Image Analysis and Design facilitated the scanning of the fish-eye photographs; John Grace, Marielos Peña, Frank Sterck, Erik Veneklaas, Marinus Werger, Pieter Zuidema and two anonymous reviewers gave helpful comments on the manuscript. This research was partly funded by grant BO 009701 from the Netherlands Development Assistance. References Ackerly, D.D. (1996) Canopy structure and dynamics: Integration of growth processes in tropical pioneer trees. Tropical Forest Plant Ecophysiology (eds S. S. Mulkey, R. L. Chazdon & A. P. Smith), pp. 619–658. Chapman & Hall, New York. Ackerly, D.D. & Bazzaz, F.A. (1995) Leaf dynamics, selfshading, and carbon gain in seedlings of a tropical pioneer tree. Oecologia 101, 289–296. Bongers, F. & Popma, J. (1990) Leaf dynamics of rainforest seedlings in relation to canopy gaps. Oecologia 82, 122 –127. Boot, R.G.A. (1993) Growth and survival of tropical rain forest tree seedlings in forest understorey and canopy openings. Implications for forest management (Tropenbos Documents 6). Tropenbos Foundation, Wageningen. Canham, C.D. (1988) Growth and canopy architecture of shade-tolerant trees: response to canopy gaps. Ecology 69, 786 –795. Chabot, B. & Hicks, D.J. (1982) The ecology of leaf life spans. Annual Review of Ecology and Systematics 13, 229–259. Chazdon, R.L. (1988) Sunflecks and their importance for forest understorey plants. Advances in Ecological Research 18, 1–63. Clark, D.B. & Clark, D.A. (1991) The impact of physical damage on canopy tree regeneration in tropical rain forest. Journal of Ecology 79, 447–457. Clark, D.B., Clark, D.A. & Rich, P.M. (1993) Comparative analysis of microhabitat utilization by saplings of nine tree species in Neotropical rain forest. Biotropica 25, 397–407. Givnish, T.J. (1988) Adaptation to sun and shade, a wholeplant perspective. Australian Journal of Plant Physiology 15, 63–92. Horn, H.S. (1971) The Adaptive Geometry of Trees. Princeton University Press, London. Hubbell, S.P. & Foster, R.D. (1986) Canopy gaps and the dynamics of a Neotropical forest. Plant Ecology (ed. M. J. Crawley), pp. 77–96. Blackwell Scientific Publications, Oxford. Hunt, R. (1978) Plant Growth Analysis. Edward Arnold, London. King, D.A. (1991) Correlations between biomass allocation, relative growth rate and light environment in tropical forest saplings. Functional Ecology 5, 485–492. King, D.A. (1994) Influence of light level on the growth and morphlogy of saplings in a Panamanian forest. American Journal of Botany 81, 948–957. 123 Light, biomass allocation and growth © 2001 British Ecological Society, Functional Ecology, 15, 113–123 Kitajima, K. (1996) Ecophysiology of tropical tree seedlings. Tropical Forest Plant Ecophysiology (eds S. S. Mulkey, R. L. Chazdon & A. P. Smith), pp. 559–597. Chapman & Hall, New York. Lambers, H. (1998) Epilogue: research on the control of plant growth – where do we go next? Inherent Variation in Plant Growth. Physiological Mechanisms and Ecological Consequences (eds H. Lambers, H. Poorter & M. M. I. van Vuuren), pp. 567–581. Backhuys Publishers, Leiden. Langenheim, J.H., Osmond, C.B., Brooks, A. & Ferrar, P.J. (1984) Photosynthetic responses to light in seedlings of selected Amazonian and Australian rainforest tree species. Oecologia 63, 215–224. Lehto, T. & Grace, J. (1994) Carbon balance of tropical tree seedlings: a comparison of two species. New Phytologist 127, 455–463. Mitchell, P.L. & Whitmore, T.C. (1993) Use of hemispherical photographs in forest ecology. O.F.I. Occasional Papers 44. Oxford Forestry Institute, Oxford. Moad, A. (1992) Dipterocarp juvenile growth and understorey light in Malaysian tropical forest. PhD Thesis, Harvard University. Norusis, J.M. (1993) SPSS for Windows: Advanced Statistics, Release 6.0. SPSS Inc., Chicago. Oberbauer, S.F., Clark, D.B., Clark, D.A. & Quesada, M. (1988) Crown light environments of saplings of two species of rain forest emergent trees. Oecologia 75, 207–212. Oberbauer, S.F., Clark, D.A., Clark, D.B. & Quesada, M. (1989) Comparative analysis of photosynthetic light environments within the crowns of juvenile rain forest trees. Tree Physiology 5, 13–23. Oberbauer, S.F., Clark, D.B., Clark, D.A., Rich, P.M. & Vega, G. (1993) Light environment, gas exchange, and annual growth of saplings of three species of rain forest trees in Costa Rica. Journal of Tropical Ecology 9, 511–524. Oberbauer, S.F. & Strain, B.R. (1985) Effects of light regime on growth and physiology of Pentaclethra macroloba (Mimosaceae) in Costa Rica. Journal of Tropical Ecology 1, 303–320. Osunkoya, O.O., Ash, J.E., Hopkins, M.S. & Graham, A.W. (1994) Influence of seed size and seedling ecological attributes on shade-tolerance of rain-forest tree species in Northern Queensland. Journal of Ecology 82, 149–163. Pearcy, R.W. (1987) Photosynthetic gas exchange responses of Australian tropical forest trees in canopy, gap and understorey micro-environments. Functional Ecology 1, 169–178. Poorter, L. (1998) Seedling growth of Bolivian rain forest tree species in relation to light and water availability. PhD Thesis, Utrecht University, Utrecht, The Netherlands. Poorter, L. (1999) Growth responses of 15 rain forest tree species to a light gradient: the relative importance of morphological and physiological traits. Functional Ecology 13, 396– 410. Poorter, H. & Nagel, O. (2000) The role of biomass allocation in the growth response of plants to different levels of light, CO2, nutrients and water: a quantitative review. Australian Journal of Plant Physiology 27, 595– 607. Poorter, H. & Van der Werf, A. (1998) Is inherent variation in RGR determined by LAR at low light and by NAR at high light? Inherent Variation in Plant Growth. Physiological Mechanisms and Ecological Consequences (eds H. Lambers, H. Poorter & M. M. I. van Vuuren), pp. 309 –336. Backhuys Publishers, Leiden. Poorter, L. & Werger, M.J.A. (1999) Light environment, sapling architecture and leaf display in six rain forest tree species. American Journal of Botany 86, 1464–1473. Popma, J. & Bongers, F. (1988) The effect of canopy gaps on growth and morphology of seedlings of rain forest species. Oecologia 75, 625– 632. Reich, P.B., Ellsworth, D.S. & Uhl, C. (1995) Leaf carbon and nutrient assimilation and conservation in species of differing successional status in an oligotrophic Amazonian forest. Functional Ecology 9, 65–76. Rijkers, T. (2000) Leaf function in tropical rain forest canopy trees. The effect of light on leaf morphology and physiology in different-sized trees. PhD Thesis, Wageningen Agricultural University, Wageningen, The Netherlands. Rijkers, T., Pons, T.L. & Bongers, F. (2000) The effect of tree height and light availability on photosynthetic leaf traits of four neotropical tree species differing in shade tolerance. Functional Ecology 14, 77–86. Sims, D.A., Gebauer, R.L.E. & Pearcy, R.W. (1994) Scaling sun and shade photosynthetic acclimation of Alocasia macrorrhiza to whole-plant performance – II. Simulation of carbon balance and growth at different photon flux densities. Plant, Cell and Environment 17, 889–900. Sipe, T.W. & Bazzaz, F.A. (1995) Gap partitioning among maples (Acer.) in central New England: survival and growth. Ecology 76, 1587–1602. ter Steege, H. (1997) Winphot 5.0. A programme to analyze vegetation indices, light and light quality from hemispherical photographs. Tropenbos-Guyana reports 97–3. Tropenbos Foundation, Wageningen, the Netherlands. ter Steege, H., Bokdam, C., Boland, M., Dobbelsteen, J. & Verburg, I. (1994) The effect of man made gaps on germination, early survival and morphology of Chlorocardium rodiei seedlings in Guyana. Journal of Tropical Ecology 10, 245–260. Sterck, F.J. (1997) Trees and light. Tree development and morphology in relation to light availability in a tropical rain forest in French Guiana. PhD Thesis, Wageningen Agricultural University, Wageningen, The Netherlands. Sterck, F.J. (1999) Crown development in tropical rain forest trees in gaps and understorey. Plant Ecology 143, 89– 98. Sterck, F.J., Clark, D.B., Clark, D.A. & Bongers, F. (1999) Light fluctuations, crown traits, and response delays for tree saplings in a Costa Rican lowland rain forest. Journal of Tropical Ecology 15, 83–95. Strauss-Debenedetti, S. & Bazzaz, F. (1996) Photosynthetic characteristics of tropical trees along successional gradients. Tropical Forest Plant Ecophysiology (eds S. S. Mulkey, R. L. Chazdon & A. P. Smith), pp. 162–186. Chapman & Hall, New York. Stuefer, J.F. & Huber, H. (1998) Differential effects of light quantity and spectral light quality on growth, morphology and development of two stoloniferous Potentilla species. Oecologia 117, 1–8. Veenendaal, E.M., Swaine, M.D., Lecha, R.T., Falsch, M.F., Abebrese, I.K. & Owusu-Afriyie, K. (1996) Responses of West African forest tree seedlings to irradiance and soil fertility. Functional Ecology 10, 501–511. Veneklaas, E.J. & Poorter, L. (1998) Carbon partitioning strategies of tropical tree seedlings in contrasting light environments. Inherent Variation in Plant Growth. Physiological Mechanisms and Ecological Consequences (eds H. Lambers, H. Poorter & M. M. I. van Vuuren), pp. 337 –361. Backhuys Publishers, Leiden. Walters, M.B. & Reich, P.B. (1999) Low-light carbon balance and shade tolerance in the seedlings of woody plants: do winter deciduous and broad-leaved evergreen species differ? New Phytologist 143, 143–154. Wayne, P.M. & Bazzaz, F.A. (1993) Birch seedling responses to daily time courses of light in experimental forest gaps and shadehouses. Ecology 74, 1500–1515. Williams, K., Field, C.B. & Mooney, H.A. (1989) Relationships among leaf construction cost, leaf longevity, and light environment in rain-forest plants of the genus Piper. American Naturalist 133, 198–211. Received 15 May 2000; accepted 18 September 2000