Sapling strength and safety: the importance of wood density in tropical forests

Tropical forests typically achieve a leaf area index of 4–5 m2 leaf per m2 ground. Such numbers make these forests highly productive, but also indicate the intensity of shading and competition for light among individual plants. Light levels at the forest floor can be as low as 1–2% of open sky. In a large gap, however, light levels may reach 30% or more (Denslow, 1987). A persistent goal in tropical ecology has been to understand how this gradient of recruitment opportunity is exploited by different species (Turner, 2001). Wood density (WD) is one trait commonly assumed to indicate a species’ position along this gradient (Ter Steege & Hammond, 2001), yet empirical evidence supporting this assumption is sparse. Augspurger (1984) showed that saplings with a high WD were better able to resist infection by pathogens, suggesting the main structural benefits of dense wood were in making hard impenetrable stems. More recent studies have confirmed density to be correlated with growth rate and mortality (Enquist et al., 1999; Wright et al., 2003; King et al., 2006), but the mechanisms leading to increased longevity for high WD species were left open. New research by van Gelder et al., in this issue of New Phytologist (pp. 367–378), shows how WD can influence a wide range of structural characteristics likely to benefit saplings in shaded understoreys. ‘A prerequisite, however, is to quantify how differences in wood density, strength and safety-factors relate to overall mortality rates and to those from specific hazards’ The key to success for juveniles growing in shaded understoreys is not only efficient light capture and photosynthesis, but also a robust stem capable of withstanding a range of hazards. van Gelder and colleagues quantified three properties of wood influencing stem structural integrity on saplings of 30 tropical rainforest species at La Chonta, Bolivia: (1) modulus of rupture (MOR) – the amount of force required to snap a stem of given thickness; (2) modulus of elasticity (MOE) – the deflection per unit force before breakage (stiffness); and (3) compressive strength (CS) – the maximum load that can be borne along the main axis before permanent deformation. Each was tightly correlated with wood density (r2 = 0.56–0.91). Further dry density (i.e. measured as dry mass per green volume) explained 20–30% more variance in strength and stiffness than density measured using green or saturated mass, consistent with the notion that it is total investment in fibres and lignin that gives improved mechanical support. Despite general acceptance of WD as an indicator of strength, these results provide the first comparison across a moderate number of species. The properties of wood combine with the geometry of the stem to give a measure of overall strength (Niklas, 1992). One way of comparing strength among plants differing in a range of structural parameters is to estimate a safety factor, namely the ratio of an observed parameter to that which would induce mechanical failure. Typically, such safety factors are expressed as a function of height or diameter. For example, the Euler-Greenhill formula for ideal columns can be used to estimate the maximum height possible (Hcrit) at a given diameter at breast height (DBH) before the stem will collapse under its own weight (McMahon, 1973): The safety factor is then the ratio of Hcrit : Hobs (Hobs, the observed height). Similarly, a bending safety factor can be calculated by comparing the observed diameter to that which would cause a branch or a stem to snap under a dynamic load. WD plays a potential role first by influencing the mass of the stem segment and thus the size of the load, and second via correlation with measures of strength: MOE (buckling) or MOR (bending). A unique feature of the study by van Gelder et al. was that they quantified all the components necessary to calculate safety factors on similarly sized saplings for a set of co-existing species. Buckling safety factors ranged from 1.55 to 3.31, with Hobs accounting for more variation than WD or MOE. For bending, however, safety factors were significantly correlated with WD via influences on MOR. Importantly, bending safety factors for stems (1.75–5) were, on average, 160% those of branches (1.39–2.43). This makes good sense – the main stem is integral for long-term survival but greater risks can be taken with individual branches. Branches may also be shed later during growth, so any additional investments would be wasted. The second important contribution by van Gelder and colleagues was to establish a solid link between structural properties of species and their tendency to occur in high- or low light. The term ‘shade-tolerance’, although commonly used in ecology, is problematic because there are no standards or procedures outlining how it should be measured, or what exactly is being measured. Often the knowledge of experienced botanists is sought, although this restricts the ability of researchers to move into new locations and opens the door to differences in interpretation among studies. An approach gaining popularity is to quantify crown exposure (CE) experienced by plants using a simple ranking from 1 to 5 (Clark & Clark, 1992). CE is less exact for each individual than direct measurements of canopy openness (using hemispherical photos for example), but the ability to survey a large number of individuals more than compensates. van Gelder et al. use measurements of CE from an average of 660 individuals per species to estimate CE for a juvenile at a standard height of 2 m (CEjuv). Measured in this way, CEjuv is not a trait but an ecological outcome; it quantifies the combined effects of differential recruitment and survival in alternative light environments (Fig. 1). Schematic overview of relationships among traits influencing light capture strategy and resulting trade-offs. Filled arrows indicate physically enforced trade-offs that should extend to a range of vegetation types and localities. These trade-offs can result in partitioning of light environment among species, dependent on the importance of that trade-off in a given environment (open arrows). The work by van Gelder et al. is one of several recent publications (Poorter & Bongers, 2006; Poorter et al., 2006; Sterck et al., 2006) to illustrate the importance of CE as a link between functional traits and ecological success at La Chonta. WD, MOR, CS, buckling and bending safety factors were all correlated with CEjuv, the tightest relationship being with WD (r2 = 0.53). At first this may seem surprising because we might expect correlation with WD via only one of the fitness outcomes, such as strength. However, as there is no single risk that determines survival in the understorey, WD is under selection for its combined effect on survival via a range of structural measures. WD also has direct effects on growth rate, the presumed trade-off of increased strength. Other results show that WD is part of a broader suite of traits related to CEjuv. In the study of van Gelder et al., the height:diameter ratios of saplings were an important determinant of buckling safety-factor and seemed to be co-ordinated with shade-tolerance. With data for an additional 24 species, Poorter & Bongers (2006) found strong co-ordination between CEjuv and leaf economics traits, with species in low light tending towards long-lived leaves with slow turnover rates. Similarly, low light species had deeper crowns (Poorter et al., 2006). Height–growth rates and mortality were also quantified for all 54 species and were correlated with CEjuv (r2 = 0.34, 0.39), demonstrating that variation in functional traits is associated with distinct differences in growth and survival. Taken together, these relationships can be thought of as a trade-off between growth potential and survival (Walters & Reich, 1999). At one end are species that grow rapidly, can secure ownership of a gap and then produce seeds promptly, cashing in on the favourable conditions. At the other end are species that grow slowly, but persistently, in low light conditions over years, possibly decades, en route to reproductive maturity. Here, chronic resource-shortage makes carbon economy crucial. Small losses owing to herbivory, falling debris or passing browsers could easily push carbon budgets into the negative. In a simulation model, Sterck et al. (2006) show how incorporation of leaf traits into a carbon-based growth model can account for some of the variation in this growth–survival trade-off (r2 = 0.2), but it seems likely that inclusion of WD would significantly improve predictions. A prerequisite, however, is to quantify how differences in WD, strength and safety factors relate to overall mortality rates and to those from specific hazards. Maximum adult height (Hmax) is also thought to be an indicator of ecological strategy (Fig. 1), yet Hmax was not correlated with structural traits such as WD. Dividing species into those that sort out along a successional axis and those that sort out along a vertical light gradient may help to clarify relationships (Falster & Westoby, 2005), as van Gelder et al. demonstrate (see Fig. 4, pp. 367–378). However, the main message from a plot of Hmax vs. CEjuv (Fig. 2) is that adult height is not co-ordinated with variation in juvenile growth strategy, at least among the species studied. van Gelder et al. report a weak positive correlation of CEjuv with Hmax, but this arises mainly as a result of the absence of short pioneers, rather than to a consistent decrease in height with decreasing CEjuv. Overall, a wide range of strategy combinations is observed; species of intermediate height (20–35 m) are found at all points along the successional spectrum. Plot of adult height vs juvenile crown exposure for 54 species of rainforest tree from La Chonta, Bolivia (data from Poorter et al., 2006). In this issue of New Phytologist, van Gelder et al. (pp. 367–378) present data for 30 of these species (closed symbols). Shade-tolerance classes are: pioneers (circles), long-lived pioneers (squares), partially shade-tolerant (triangles), and shade-tolerant (diamonds). van Gelder et al. have provided good evidence linking WD to stem strength, stiffness and safety. Such physically enforced trade-offs probably apply broadly in a range of environments (Fig. 1). Their results also suggest an important role for stem mechanics in determining juvenile light-use strategy in tropical forests via a growth–survival trade-off. Some outstanding challenges are: (1) to link strength and safety to observed mortality rates; (2) to understand whether a growth–mortality trade-off leads to partitioning of juvenile light environment in other localities and vegetation types; and (3) to understand how the different aspects of light-use strategy combine to put an envelope round the overall diversity of trait values at a site.