HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (7) Citing Articles via Google Scholar Google Scholar Articles by Read, J. Articles by Stokes, A. Search for Related Content PubMed Articles by Read, J. Articles by Stokes, A. Agricola Articles by Read, J. Articles by Stokes, A.

Plant biomechanics in an ecological context¹

²School of Biological Sciences, Monash University, Victoria 3800, Australia; ³Institut National de la Recherche Agronomique (INRA), Laboratoire Franco-Chinois d'Informatique, d'Automatique et de Mathématiques Appliquées (LIAMA), Chinese Academy of Sciences - Institute of Automation (CASIA), P.O. Box 2728, Haidian District, 100080 Beijing, China

Received for publication March 13, 2006. Accepted for publication August 23, 2006.

ABSTRACT

Fundamental plant traits such as support, anchorage, and protection against environmental stress depend substantially on biomechanical design. The costs, subsequent trade-offs, and effects on plant performance of mechanical traits are not well understood, but it appears that many of these traits have evolved in response to abiotic and biotic mechanical forces and resource deficits. The relationships between environmental stresses and mechanical traits can be specific and direct, as in responses to strong winds, with structural reinforcement related to plant survival. Some traits such as leaf toughness might provide protection from multiple forms of stress. In both cases, the adaptive value of mechanical traits may vary between habitats, so is best considered in the context of the broader growth environment, not just of the proximate stress. Plants can also show considerable phenotypic plasticity in mechanical traits, allowing adjustment to changing environments across a range of spatial and temporal scales. However, it is not always clear whether a mechanical property is adaptive or a consequence of the physiology associated with stress. Mechanical traits do not only affect plant survival; evidence suggests they have downstream effects on ecosystem organization and functioning (e.g., diversity, trophic relationships, and productivity), but these remain poorly explored.

Key Words: allocation patterns • biomechanical plasticity • eco-biomechanics • leaf toughness • plant strategies • sclerophylly • trade-offs • wood density

The mechanical design of a plant is fundamental to its growth and reproductive performance and so to its fitness. Leaves must be positioned to efficiently capture sunlight, commonly requiring support by stems and branches, and roots must be able to push through soil to acquire water and mineral nutrients and to ensure anchorage. The transport system must be able to sustain pressures in conducting fluids, often to tens of meters above ground. Stems and branches need to withstand static loads, such as the mass of the crown, and dynamic loading as a result of wind storms and arboreal animals. Plants must be able to display reproductive organs to biotic pollinators and dispersers (and withstand their mass and movement) or ensure that pollen and diaspores reach effective air currents. In addition, plant organs may require protection from damage by herbivores. However, structural reinforcement can be costly, achieved at the expense of fundamental functions such as carbon gain and reproduction. Furthermore, fitness is more than just survival in a particular abiotic environment; plant design must be considered in the plant's competitive context (Givnish, 1986 ). Hence plant design must be "smart" design or growth and competitiveness will be severely compromised. Consequently, mechanical design from the scale of cellular anatomy to whole-plant architecture is fundamental to survival and likely to be generally under strong selective pressure.

Plant design at any level is likely to be constrained by conflicting influences on different functions, the effects of which will depend on the local environment (Niklas, 1992a ; Press, 1999 ). Furthermore, mechanical properties can co-vary, with trade-offs among some properties (Vogel, 2003 ). Therefore, there may be compromises. In addition, as the number of functions of an organism or organ increases, the number of equally efficient designs may increase, but potentially with a decline in efficiency of performance of any particular function (Niklas, 1992a , 1997 ). Hence, even though there may be mechanical trends along environmental gradients, a variety of mechanical solutions may occur within any environment (Niklas, 1997 ; Press, 1999 ). Some designs may be far from optimal due to constraints imposed by phylogenetic history. Furthermore, if plant parts are exposed to different environments during development there may be a heterogeneous profile of mechanics among and within plants of the same species (Niklas, 1999 ). Plant tissues are more complex than most engineering structures, making their mechanical properties often difficult to measure, interpret, and predict (Atkins and Mai, 1985 ; Vincent, 1990 , 1992 ; Niklas, 1992a , 1999 ; Spatz et al., 1999 ; Vogel, 2003 ; Sanson, 2006 ). All these factors contribute to the complexity of understanding the functional and ecological significance of biomechanical design.

An increasing amount is known about the importance of mechanical properties of individual organs of herbaceous and woody plants, i.e., stems, branches, roots, leaves, and reproductive parts, and how these properties relate to or are affected by the local environment (Niklas, 1992a , 1999 ; Givnish, 1995 ; Telewski, 1995 ; Mattheck and Kubler, 1997 ; Rowe and Speck, 2005 ; Fournier et al., 2006 ; Reith et al., 2006 ). However, our understanding of how these properties interact with each other over time and space is much less evident, especially in terms of plant fitness and distribution. Some plant responses to external mechanical stresses have positive secondary effects on, for example, susceptibility to insect attack (van Emden et al, 1990 ), hardiness to drought and temperature stress (Jaffe and Biro, 1979 ), and the ability to capture sunlight (Berthier and Stokes, 2005 ). Nevertheless, plant investment in withstanding mechanical perturbations can have detrimental effects on plant size (Telewski, 1995 ) and fecundity (Cipollini, 1999 ); thus, fitness may be reduced unless the increased chance of survival compensates for reduced fecundity. Even less is known about downstream implications of plant mechanics on ecosystem structure and function, for example, on habitat quality, species diversity, trophic interactions, and nutrient cycling. In this review, we emphasize the relationships between plant mechanics and environmental conditions and consider their potential ecosystem-level implications.

HOW ENVIRONMENTAL CONDITIONS INFLUENCE PLANT MECHANICAL DESIGN

Fundamental design traits, i.e. structure and architecture, at both the cellular and whole plant level, are directly influenced by the immediate environment in terms of resource supply and biotic/abiotic stress. Plants can respond to stress over evolutionary time (adaptation), but can also respond to temporal and spatial fluctuations in external stresses through adjustment in their shape and structure (phenotypic plasticity), allowing them to counterbalance the effects of the stress over their life span. Self-supporting terrestrial plants can be exposed to a range of changing stresses, such as wind (Sellier and Fourcaud, 2005 ), avalanches (Johnson, 1987 ), and other soil mass movements. Vertical uprooting forces or trampling from grazing animals commonly have to be resisted by herbaceous species (Ennos and Fitter, 1992 ; Striker, et al., 2006 ). In addition, plants may be exposed to changes in resource availability (both temporal and spatial), which may lead to mechanical responses. Significant examples of how plant mechanical properties change depending on the local environment can be found in self-supporting and lianoid forms of the same species, e.g., western poison oak (Toxicodendron diversilobum Torr. & Gray) (Gartner, 1991 ). In such species, seedlings are often self-supporting until certain environmental cues are received, e.g., light availability or the proximity of a host support. The self-supporting structure can then change to a lianoid form, hence altering significantly the ecological function of that individual over a period of time (a subject extensively reviewed by Rowe and Speck [2005 ]).

Abiotic stress
Stem and root mechanical design
In woody plants, wind loading may be the most commonly experienced mechanical abiotic stress. Many species also need to grow upwards to the canopy, so investment in mechanical support, anchorage, and light capture needs to be finely balanced. Trees in particular may reach extreme heights, e.g., Sequoiadendron giganteum Lindl. and Eucalyptus regnans F.Muell, and thus need rigid stems with a high biomass investment to prevent trunk failure by buckling or bending. A tree may achieve such rigidity by, for example, increased stem cross-section or a wood (xylem) structure capable of resisting the imposed static and dynamic mechanical stresses experienced throughout its life. Wood structure in trees is highly diverse both spatially (geographically and between populations and species) and temporally (throughout the life of an individual). This diversity reflects evolutionary history and particular ecological conditions, as well as the various trade-offs among mechanical support, transport, and economy of construction. Wood density and conductivity are basic traits that determine growth and mechanical support and also influence photosynthesis, so that variation in density may be more related to mechanical support of the water column, which is under negative pressure, than to support of the tree (Sperry et al., 2006 ). In addition, wood density in tropical savannah trees (Bucci et al., 2004 ) and Mediterranean broadleaved trees (Barij et al., 2006 ) depends significantly on soil moisture content. Air temperature also affects wood density, by influencing the viscosity of sap as it ascends the stem, which in turn results in changes in cell lumen size (Thomas et al., 2004 ). Therefore, trends in wood density must be strongly related to the geographic location and available water supply of a species, as well as the mechanical constraints imposed on individuals.

Roots are also particularly sensitive to fluctuations in water, nutrients, and mechanical stress. When external dynamic loading on the shoot system occurs due to, for example, wind, the adaptive growth responses in the root system are often more striking than those of the shoots (Stokes et al., 1995 ; Tamasi et al., 2005 ). The vigor of a root system is fundamental to survival of a self-supporting plant, not only with regard to resource uptake, but also to anchorage strength. In many situations, even if the shoot is grazed by animals or damaged through mechanical loading, a well-anchored root system will ensure survival. Root systems of mature trees need to be large and deep enough to withstand overturning due to mechanical loading on the stem. The structural proximal roots tend to anchor the tree to the soil, whereas the finer distal roots serve a more absorptive function. Tree root systems with a higher degree of branching and vertical growth are usually better anchored than more shallow root systems, although soil type has a strong influence on the mode of failure (Moore, 2000 ; Dupuy et al., 2005a ). In young trees and woody shrubs, overturning is less problematic. Saplings and shrubs often display more fibrous root systems, with less secondary thickening and with a relatively large overall surface area. Increased root area and branching not only improve nutrient absorption, but augment soil–root friction and resistance of vertical uprooting (Dupuy et al., 2005b ). Herbs and young woody plants are particularly vulnerable to vertical uprooting by grazing animals, and a well-anchored root system may improve plant survival.

In plants subjected to recurring mechanical stresses such as wind loading or other mechanical perturbation, several responses are observed. Leaf number and area may diminish (Stokes et al., 1995 ; Niklas, 1996 ), which reduces wind-induced drag on the crown (Sellier and Fourcaud, 2005 ), but also reduces photosynthesis. Plants develop a stunted appearance, as stem elongation is reduced but radial growth is increased; hence, trees growing on windy sites tend to have a highly tapered shape (Telewski, 1995 ; Ennos, 1997 ). There is increased allocation of nitrogen (Cordero, 1999 ) and carbon to structural components, usually in the areas where mechanical stresses are highest, i.e., the base of the trunk and branches and along certain lateral roots and tap roots (Telewski, 1995 ; Nicoll and Ray, 1996 ; Stokes et al., 1997 ; Danjon et al., 2005 ; Tamasi et al., 2005 ). Windward lateral roots held in tension during loading have been shown to develop more branches per unit area of soil, thus increasing surface area and root–soil friction (Stokes et al., 1995 ). Plant internal structure also undergoes chemical and morphological changes, with the formation of thick-walled tracheid cells in trees (Telewski, 1989 ). This flexure wood is more resistant than normal wood in compression, but due to the structure of cellulose in the cell walls, is also more flexible, with a lower modulus of elasticity (E), hence allowing the stem and branches to flex more easily during wind sway. Combined, these responses to mechanical loading result in a reduced bending, lodging, or overturning moment, when the stem or crown is subjected to an external force. However, plant responses may differ depending on the abiotic stress applied and the species. Smith and Ennos (2003) pointed out that species plasticity will reflect the evolutionary history and current ecological conditions specific to each species. These authors showed that plants exposed to wind loading and stem flexure react in different and opposite ways. In experiments on Helianthus annuus L., wind-loaded plants were taller with a higher stem hydraulic conductivity whilst flexed plants were shorter with a lower stem hydraulic conductivity. Hence, the plant reaction to wind was the sum of two separate responses: a physiological response to air flow around the plant and the response of the stem to mechanical bending. How plants react to wind loading will therefore depend on organ size and mechanical properties, e.g., leaf stiffness or stem flexibility, and may explain why the responses of different species to wind loading are so variable (Telewski, 1995 ).

Certain trade-offs must therefore exist between different plant functions (Givnish, 1995 ), including survival and reproduction. The mechanical stability of self-supporting plants is clearly important for survival, but can plant mechanical properties shift the emphasis from one factor to another? Examples of shifts from reproduction to stress resistance have been recorded in several species. The fitness of both Capsella bursa-pastoris L. (Niklas, 1998 ) and Brassica napus L. (Cipollini, 1999 ) decreased significantly when subjected to mechanical perturbation. Delays in anthesis and reductions in flower number or size occurred, thus this trade-off is costly to fitness, unless benefits gained through the development of a mechanically hardened phenotype counterbalance costs to flower production. Similarly, Pigliucci (2002) showed that wind loading significantly increased branching in certain populations of Arabidopsis thaliana L. from different geographic locations, which in turn negatively affected fecundity. Plants subjected to 16 h of daily wind loading flowered later and with more basal fruiting branches than those exposed to 0 or 6 h of daily wind. A high degree of genetic differentiation existed among the different populations, suggesting that many genes exist that underlie quantitative traits, thus offering greater phenotypic flexibility.

Although wind loading may decrease fecundity, the response of plants to this ubiquitous physical stress has been shown to improve other traits important to survival. Berthier and Stokes (2005) demonstrated that phototropic movements were significantly augmented in seedlings of Pinus pinaster Ait. subjected to daily intermittent wind loading, thus improving light capture in exposed plants and hence their competitive strategy. Similarly, in inclined saplings of the same species, Berthier and Stokes (2006) found that leaning plants exposed to daily wind loading straightened significantly faster than leaning control plants not exposed to wind. Hence, wind stress may actually confer an ecological advantage to young trees in terms of competitive strategy and survival by stimulating tropic processes. A comparable example can also be found within the forest understory. Tree seedlings and herbs in forests are often damaged by falling debris such as branches and leaves. When a gap appears, e.g., after tree fall, the high levels of light stimulate compensatory growth in damaged understory plants (Bruna and Nogueira Ribeiro, 2005 ). The mechanisms underlying these physiological responses are not yet fully understood, but Berthier and Stokes (2005) suggest that cross-talk exists between different sensory pathways in plant cells and that by stimulating mechanosensors through mechanical loading, several other reactions may be induced, including tropic movements.

Support and protection of leaves
Leaves generally need to support their own mass to align at an optimal angle to the sun, usually more or less horizontally but depending partly on the particular light environment (Niinemets and Fleck, 2002 ). Horizontal alignment maximizes bending forces, so the optimum design for light interception may maximize the risk of bending failure (Niklas, 1997 ). In addition, leaves can be repeatedly exposed to dynamic bending and twisting loads from wind, snow, hail, falling debris, and animals, and abrasion with other foliage, often with considerable force, yet are typically relatively flimsy structures compared with stems and roots. Hence, leaf design must be a compromise between carbon gain (including issues of thermoregulation and water loss) and the costs of sustaining static and dynamic loads (Niklas, 1992b , 1997 ). The leaf structure of many vascular plants is such that forces perpendicular to the surface are spread over a larger area, with small lateral deflections absorbed by the stringers (veins and the cells that connect them to both surfaces) that hold the leaf surfaces together, giving them the property of stress-skin panels (Niklas, 1999 ). Such structures are relatively flexible in bending and torsion and can deform elastically to some extent without damage (Niklas, 1999 ). Hence, even relatively soft leaves can survive a moderate degree of mechanical stress. Even so, there is considerable variation in the structure and mechanics of leaves that is assumed to have adaptive significance in support and/or protection, in conjunction with other functions. However, the mechanical design of leaf laminae is poorly understood compared with that of petioles and stems (Niklas, 1992a , 1999 ).

Some leaves use predominantly hydrostatic support around a skeleton of vascular tissue (Fig. 1a), an effective form of support if leaf water status can be maintained (Niklas, 1992a ). The mechanical properties and arrangements of vascular tissue reflect demand for both supply and support. In large and thin "hydrostatic" leaves, the cost of veins supporting the periphery may be substantial since the vein mass must increase at a higher power than the length, so above a critical leaf size it may be more efficient to produce multiple leaves than a single leaf of similar area (Givnish, 1986 ). However, leaf size and shape (via effects on boundary layers) also influence rates of photosynthesis and water loss (Givnish, 1979 ; Gutschick, 1999 ) and costs of branch support (Givnish, 1979 ), so in some environments it may still be more efficient to invest in large leaves, even if the cost of self-support is relatively high. Plants may have specialized structures that stiffen the leaf, such as a sub-epidermal layer of sclerenchyma, or enhanced I-beams formed by bundle sheath extensions of sclerenchyma meeting stiffened sub-epidermal layers (Fig. 1b), shifting reliance from hydrostatic toward structural support. Because these leaves are less dependent on water for support they may be more common where water is seasonally limiting. The most effective position for stiffening tissues such as collenchyma and sclerenchyma is close to the outer surfaces (distant from the neutral axis of bending). Since this design may compromise carbon gain by both diluting photosynthetic tissue and attenuating light reaching tissues below (Niklas, 1997 ), it may be common only in species adapted to sunny environments. Silica can provide stiffening at potentially low cost, although its density may impose other costs (Raven, 1983 ). Some devices with major support and strengthening functions, such as silica and sclerenchyma bundle sheath extensions, may have other significant roles, such as in water transport to the parenchyma and epidermis for the latter (Heide-Jørgensen, 1990 ) and protection against herbivores.

View larger version (83K): [in this window] [in a new window]   Fig. 1. Contrasting leaf anatomy, mechanics and ecology of two species from woodland in south-eastern Australia. (a) Prostanthera lasianthos Labill. (Lamiaceae); (b) Banksia marginata Cav. (Proteaceae). The light micrographs are cross-sections across a secondary vein (sections provided by P. Peeters). Bar = 100 µm. Mechanical properties were measured following Read et al. (2006) (E. Caldwell, J. Read, G. Sanson, Monash University, unpublished data)

Some capacity of the petiole and/or lamina to bend and twist allows reorientation in response to environmental change (e.g., changed position of light source) or rapid repositioning in response to strong loads such as wind to reduce drag and damage (Vogel, 1989 ; Niklas, 1992b ; Ennos, 1997 ; Vogel, 2003 ), the latter being particularly important for large leaves. When exposed to strong loads leaves must either resist bending and torsion or deform without fracture to escape damage (Niklas, 1999 ). Leaf shape is not necessarily constant; local forces exerted by wind on a leaf depend partly on leaf shape, but shape itself is influenced by wind (Vogel, 2003 ). Thin leaves, leaflets, and leaf clusters readily fold and reconfigure in strong wind, reducing drag and consequent damage (Vogel, 1989 ; Niklas, 1999 ; Vogel, 2003 ). Leaf movement may have the extra advantage of increasing air turbulence close to the leaf surface, thinning the boundary layer and increasing the capacity for heat and gas exchange and photosynthesis (Vogel, 1970 ; Ennos, 1997 ), although leaf flutter under very low wind speeds does not seem to be sufficient to generate this effect (Grace, 1978 ). Strong winds may have adverse effects, reducing carbon gain in a thin leaf by cooling and drying the leaves or curling the leaves and reducing their effective surface area (Ennos, 1997 ). The petiole also contributes substantially to leaf bending and twisting by properties of its material (tissue types and arrangements) and shape (Vogel, 1992 , 2003 ; Niklas, 1992b , 1994 , 1999 ). In particular, its cross-sectional shape can have a marked effect on the ratio of flexural to torsional rigidity. A groove or flattening of the upper petiole surface (a departure from a circular cross-section) can increase the ratio of flexural to torsional rigidity, increasing the capacity to twist into the wind, thereby reducing drag (Vogel, 1992 , 2003 ). Tissues also contrast in their mechanical properties. Parenchyma and collenchyma act hydrostatically; under water stress a petiole can lose stiffness so the leaf wilts, reducing drag as well as exposure to high irradiance and further water loss (Niklas, 1999 ). In contrast, sclerenchyma is a non-hydrostatic tissue with high stiffness in bending and twisting (Niklas, 1999 ). There may be a trend towards flexible petioles with collenchyma in the hypodermis in trees exposed to strong winds but petioles with subepidermal sclerenchyma in understory species (Niklas, 1999 ). Plasticity in lamina and petiole form occurs both between and within plants in response to contrasting exposure to wind and light (Niklas, 1999 ; Niinemets and Fleck, 2002 ). For example, Woodward (1983) showed that some upland species of Phleum have greater sclerification than lowland species, but that sclerification could increase in lowland species in response to wind speed; Niklas (1996) showed that Acer saccharum L. trees on windy sites produce fewer and smaller leaves on shorter but more flexible petioles than on sheltered sites, reducing drag and damage. Wind can also lead to abrasive damage; abrasion from windborne particles or foliar contact during wind experiments resulted in increased epidermal conductance in needles of Picea sitchensis (Bong.) Carr. and Pinus sylvestris L. (van Gardingen et al., 1991 ).

The economics of support and supply is complex, not only at the level of the leaf but at the level of the whole plant, since it is influenced by efficiencies of leaf arrangement in both space and time (Givnish, 1979 , 1986 ). Furthermore, the product of branching plus leaf orientation may be critical in competition, not only because of its primary influence on carbon gain, but by shading competing neighbors below (Givnish, 1979 ). For example, a forest dominated by trees that have vertically oriented leaves (such as many eucalypts: King, 1997 ) can support a richer understory of light-demanding plants than forests of trees with horizontally oriented leaves. There may be other downstream effects on community composition and dynamics. Leaves that use hydrostatic support should have less dilution of photosynthetic tissue by structural tissue, so may be preferred by herbivores for their higher concentration of nutrients that are easier to access. If so, and if some communities in particular environments (e.g., mesic) have a greater proportion of species with hydrostatic leaves, there may be a greater density and diversity of herbivores and their predators. Hence, the ecology of leaf support and protection should influence not just competitive success, but higher trophic levels.

Biotic stress
Leaf mechanics in protection from herbivores
Animals can cause significant damage to plants. Resistance of species to trampling has been associated with higher tensile strength of organs (Shearman and Beard, 1975 ; Kobayashi et al., 1999 ) and stem flexibility (Dale and Weaver, 1974 ; Sun and Liddle, 1993 ). The tensile strength of some organs may increase in plants exposed to trampling (Kobayashi and Hori, 1999 ; Kobayashi et al., 1999 ). However, the most widespread source of biotic damage is herbivory. Here we focus on leaves, but mechanical properties are likely to contribute to defense of other organs. There is strong evidence that leaf mechanical properties, including strength (commonly measured as the force to fracture per unit area over which the force is applied) and toughness (usually regarded as the energy or work to fracture) (Wright and Vincent, 1996 ; Sanson et al., 2001 ; Sanson, 2006 ), deter herbivores. Studies of herbivory have used various mechanical measures, including work to fracture, force to fracture, and strength from both tensile tests and compressive punching and shearing tests. Foliar strength and work to fracture are often, but not always, correlated but are rarely both measured, so it is difficult to know which is of most significance in ecological studies. For these reasons, and since terms are not used consistently among studies (Lucas et al., 2000 ; Sanson et al., 2001 ), we here use "tough" to refer to this range of properties. Variation in these properties among plants results from differences in the amount and composition (e.g., lignification) of cell wall, thickness of the cuticle, leaf thickness and density (influencing leaf mass per area, LMA), and amounts and arrangements of particular tissues, including specialized tissues such as sclerenchyma, each contributing differently to each mechanical property (Grubb, 1986 ; Vincent, 1990 ; Choong et al., 1992 ; Turner, 1994b ; Wright and Illius, 1995 ; Wright and Vincent, 1996 ; Lucas et al., 2000 ; Read et al., 2000 ; Sanson et al., 2001 ). Silica is well known in monocots, in external hooked structures that sharpen blade margins and in cell walls and cytoplasm (Jones and Handreck, 1967 ; Raven, 1983 ). It has suggested roles in support and anti-herbivore defense, both in spinescent structures that must be rigid to be effective deterrants and in leaves (including dicots), and may also protect against microorganisms (Jones and Handreck, 1967 ; Raven, 1983 ; Lucas et al., 2000 ; Sanson, 2006 , in this issue).

There is diverse evidence for the defensive roles of mechanical properties. First, there is often a strong negative association between leaf toughness and herbivore damage across species (Coley, 1983 ), with low levels of herbivory in sclerophyllous vegetation (Morrow, 1983 ). Second, tough leaves may be eaten less than softer leaves of the same species that have developed under different conditions (Lowman, 1985 ; Sagers, 1992 ). Third, mature toughened leaves or stems may be eaten less or more slowly than young expanding tissue of the same species (Feeny, 1970 ; Lowman and Box, 1983 ; Raupp, 1985 ; Hoffman and McEvoy, 1986 ), even when young tissue has higher levels of chemical defense (Coley, 1983 ; Larsson and Ohmart, 1988 ; Choong, 1996 ; Read et al., 2003 ). There are also many examples of specific effects of leaf mechanics on herbivores. For example, tough leaves can have components that wear insect mandibles (Raupp, 1985 ), and substantial energy demands are required by leaf-cutting ants in cutting leaves (Roces and Lighton, 1995 ). Ants may avoid mature leaves because of their toughness, even when better for growth of symbiotic fungi than young leaves (Nichols-Orians and Schultz, 1990 ). Toughness can limit feeding by early instars of insects (Ohmart et al., 1987 ; Hoffman and McEvoy, 1986 ; Larsson and Ohmart, 1988 ) and may increase in leaves produced after herbivore attack (Kudo, 1996 ). Some caterpillars roll leaves around expanding buds to feed on the shaded expanding leaves that are less tough than unshaded leaves (Sagers, 1992 ). Feeding time and rate of development can be longer on tougher leaves (Raupp, 1985 ; Clissold et al., 2006 ). In some cases herbivores avoid tough leaves or grow more slowly than on a diet of soft leaves, but feed successfully when the leaves are ground to a powder (Feeny, 1970 ; Berdegue and Trumble, 1996 ; Bezzobs and Sanson, 1997 ; Clissold et al., 2006 ), implying a mechanical constraint to feeding or assimilation. Mechanical constraints can act in a complex way, such that delayed access to nutrients by herbivores can alter the ratio of assimilated nutrients, thereby constraining growth (Clissold et al., 2006 ). Larger head mass of insects is reported in grass specialists than other foliage feeders of similar size, and feeding on tough grasses can induce development of larger heads in caterpillars (Bernays, 1986 , 1991 ). Mechanical defense might also have indirect effects on herbivore survival, if slowed feeding increases vulnerability to abiotic stress and natural enemies (Hoffman and McEvoy, 1986 ; Bernays, 1991 ), and may have complex effects on herbivore behavior across a range of scales (Sanson, 2006 ).

Protection from herbivory may be conferred directly by mechanical constraints to feeding or indirectly by dilution of nutrients (Timmins et al., 1988 ; Peeters, 2002 ) by the structural features that confer the mechanical properties, and nutrients such as proteins may be less available due to binding with lignin (Swain, 1979 ), i.e., the properties that make the leaf tough are associated with those that reduce digestibility, both of which may affect diet selection (Wright and Illius, 1995 ). In addition, leaf chemistry can vary along gradients in mechanical properties so that soft leaves and even parts of leaves are more nutritious (Langenheim et al., 1986 ; Peeters, 2002 ; Choong, 1996 ; Read et al., 2003 ; Brunt et al., 2006 ), sometimes with lower levels of chemical defenses (Feeny, 1970 ). Hence it can be difficult to discern the exact role of mechanical properties. Mechanical defenses can potentially be overcome by herbivores with appropriate musculature and mouthparts, but there may still be significant energy costs (direct or indirect) incurred during acquisition through to assimilation. Hence mechanical defenses may be best considered not only in absolute terms (e.g., whether a herbivore has the bite force to acquire and process the food), but in terms of the cost of acquisition per nutrition gained.

Relatively little is known about the specific mechanical properties that deter herbivores (discussed by Lucas et al., 2000 ; Sanson, 2006 , in this issue). We might expect that mechanical constraints to feeding relate to the way the herbivore uses its mouthparts, but there is still much to learn about the various ways in which teeth and invertebrate mouthparts work to acquire and mechanically process their plant diet (Peeters, 2002 ; Sanson, 2006 ). Leaf "structural" properties, but not "material" properties (per unit leaf thickness), correlated with feeding by moth larvae across 20 plant species, and shearing properties correlated with feeding more than tensile properties (E. Caldwell and J. Read, unpublished data), the latter consistent with small animals having to cut through individual veins (Vincent, 1990 ). Densities of chewing insects, but not sucking insects, correlated with leaf strength of forest understory species (Peeters et al., in press ). Leaf structure can influence invertebrate herbivores at a fine scale (Peeters, 2002 ) since they are small enough to feed selectively across a leaf, often avoiding the toughest parts (Kimmerer and Potter, 1987 ; Bernays, 1991 ; Choong, 1996 ). A dense wall of hairs can be a significant mechanical barrier to feeding insects (Hoffman and McEvoy, 1986 ; Lucas et al., 2000 ), and lignified bundle sheath extensions may provide physical barriers to movement of leaf miners. Some vertebrates even seem to discriminate among and within leaves in relation to mechanical gradients (Sanson, 2006 , in this issue), although others may use toughness to select among feeding sites rather than among leaves (Hill and Lucas, 1996 ). Hence plants must defend themselves across a range of scales, from insects influenced by cell traits to large mammals that ingest whole leaves or even branches (Sanson, 2006 ).

Plants have a range of anti-herbivore defenses, including chemistry, spinescence, and attraction of natural enemies (Gutschick, 1999 ; Press, 1999 ), but there is relatively little evidence for the manner in which mechanical properties are balanced against these other defenses. Limited data for phenolics vs. toughness across species suggests there is no simple association (J. Read, unpublished data). This is perhaps not surprising since phenolics, although useful for comparisons because of their widespread occurrence, have multiple effects (Swain, 1979 ), including photoprotection (Close and McArthur, 2002 ). However, there is some evidence of "trade-offs" between defenses, for example, in declining levels of chemical defenses as expanding leaves toughen (Brunt et al., 2006 ), and between different forms of physical defense (Björkman and Anderson, 1990 ).

Mechanical properties can provide defense indirectly. For example, slippery wax barriers on stems of Macaranga myrmecophytes limit access by generalist ants that might compete with or prey on the resident "wax-running" ant partners that protect the host from herbivores (Federle and Bruening, 2006 ). Macaranga species are often pioneers, and the deposition of slippery waxes on the stems may incur direct costs, but not the additional indirect costs to photosynthesis associated with leaf toughening, which could be critical to fast-growing pioneers.

If mechanical properties influence plant selection by herbivores, they may significantly influence plant growth, survival, and patterns of species' abundance, but prediction of the latter can be difficult because of the variety of ways in which plants resist herbivory and the complexity of scaling up to community-level properties (Rosenthal and Kotanen, 1994 ; Wright and Illius, 1995 ; Olff and Ritchie, 1998 ; Díaz et al., 2001 ; Pérez-Harguindeguy et al., 2003 ; Cingolani et al., 2005 ). Leaf mechanical and underlying structural traits may also influence local densities and feeding guilds of herbivores (e.g., Raupp, 1985 ; Ezcurra et al., 1987 ; Peeters et al., 2001 , 2002 , in press) and consequently the community abundance and composition of herbivores and potentially their predators. In addition, for herbivores limited by leaf mechanics (or in concert with nutrition and chemical defense) and restricted to feeding on immature leaves, the window of opportunity for feeding may be small, particularly in seasonal climates where leaf expansion is synchronous and restricted in duration (Lowman and Box, 1983 ; Morrow, 1983 ; Aide and Londoo, 1989 ; Brunt et al., 2006 ) or where environmental stress limits production of new leaves (Ohmart et al., 1987 ; Larsson and Ohmart, 1988 ). If temporal variation in the mechanical profile of vegetation influences temporal patterns of herbivore abundance, higher trophic levels may be affected. Hence leaf mechanics should have a significant role in the organization and functioning (e.g., diversity, trophic relationships, and productivity) of aboveground ecosystems, a role that has been little addressed to date.

THE BIOMECHANICS OF REPRODUCTION AND REGENERATION

Reproductive structures
Reproductive structures vary substantially in their mechanics, from often delicate flowers to robust woody fruits, but relatively little is known about their mechanical traits. Animal-pollinated flowers may need to be large if supplying the appropriate food reward for a bird or mammal or clustered to concentrate the reward in space. Larger inflorescences may require co-evolved stem properties for support and to minimize drag to reduce damage, particularly if the perianth is delicate. Flowers on long stems may be more apparent to pollinators or better able to deposit pollen in wind currents, but are more exposed to wind damage. For example, laterally facing daffodil flowers are supported on stems that are elliptical in cross-section, allowing twisting, thereby reducing drag (Etnier and Vogel, 2000 ). Ennos (1993) noted that inflorescences of the lowland sedge Carex acutiformis Ehrh. were borne on tall stems that were triangular in section with low torsional rigidity, allowing twisting in wind, reducing drag and possibly self-pollination; however, mountain species that must resist stronger winds had shorter stems that were more circular in section and better able to resist buckling. Some flowers or their parts (e.g., the labellum of orchids and the style supporting the pollen presenter in Proteaceae species) may need to be stiffened to support the mass and activity of pollinators, but few data are available on this subject.

Fruits have evolved for protection and dispersal of seeds, often with specialized mechanical design that enhances one or both functions. Possibly the greatest degree of protection is afforded against seed predators and fire. Species in regions where wildfires are common are often serotinous, with mature seeds stored in the canopy for an extended period in robust fruits that range from leathery to massive woody structures until the advent of more suitable post-fire conditions for seedling recruitment (Lamont et al., 1991 ). Protection of seeds from the heat of a fire is related to wall thickness and fruit size (Bradstock et al., 1994 ), but woody fruits may also confer protection against insect larvae and seed-eating birds (Groom and Lamont, 1997 ). In Hakea (Proteaceae), protection from seed eaters may provide a better explanation than fire for the particularly dense and thick-walled follicles of some species, and reproductive organs are sometimes also protected by a "nest" of tough, spinescent leaves (Groom and Lamont, 1997 ). In the same family, but from rainforest, Macadamia ternifolia F.Muell. has a follicle with a wall as hard as commercial-grade aluminum, but with twice the strength (stronger than concrete or brick) and half the density (Niklas, 1992a ). In the Australian tropical rainforest a higher content of N has been recorded in seeds (adjusted for mass) that have marked protection, including by woody fruits (Grubb et al., 1998 ). Seeds of some plants have hard seed coats, often associated with seed dormancy and potentially affording protection from both herbivores and pathogens, i.e., enabling a long life span, as well as contributing to dormancy. For animal-dispersed seeds, a hard seed coat or stony endocarp may provide mechanical protection as the seed is handled or passes through the disperser's gut. In a different vein, the changes in fruit texture accompanying ripening of fleshy fruits may be used as a selection cue by primates, in conjunction with visual and olfactory cues (Dominy, 2004 ).

Reproductive parts are often motile to a greater or lesser degree, with some wind-dispersed pollen, fruits, and seeds mechanically designed for passive "flight." Added mass increases wing-loading (mass : area), thereby decreasing dispersal capacity; hence there is a potential trade-off between dispersal distance (by wind) and the nutritional reserves and protective seed coverings (including fruit tissue) that might confer an advantage during dispersal, any dormancy period, and establishment (Augspurger, 1988 ). Wind-dispersed seeds and fruits may have devices such as hairs, fibers, and wings that increase drag, or in larger diaspores lift-to-drag, thereby maximizing distance of dispersal (Matlack, 1987 ; Augspurger, 1988 ; Niklas, 1992a ; Nathan et al., 2002 ; Vogel, 2003 ). Winged fruits of some species glide downwards in a helical manner, in which a lift-generating airfoil, for example, reduces the descent rate (Green and Johnson, 1990 ; Vogel, 2003 ). Dispersal of spores, pollen, and seeds is sometimes achieved by dynamic release, following cellular dehydration in reproductive structures such as fern sporangia, anthers at dehiscence, and fruits (and similarly, the mechanical design of vegetative parts allows sun-tracking by leaves, leaf-folding as in some grasses and legumes, and trap-closing in fly-traps, not discussed in this review), often involving specialized cells and patterns of wall-thickening (Zhang et al., 1991 ; Niklas, 1992a ; Witztum and Schulgasser, 1995 ; Vogel, 2003 ). Dynamic loading by wind can assist in spore and pollen release, and ovulate organs may be designed to capture pollen aerodynamically (Niklas, 1992a ). Some plants have a highly specialized mechanical system of pollen transfer. These include buzz pollination (e.g., in Solanaceae) (Buchmann and Hurley, 1978 ; Harder and Thomson, 1989 ), whereby a stamen is vibrated by rapid contractions of the flight muscles of pollen-collecting insects, releasing small quantities of pollen (avoiding over-exploitation) through poricidal anthers, and piston (Fabaceae) and lever systems (e.g., Salvia), in which the pollinator lands on, or pushes against, a staminal lever that moves to deposit pollen onto the visitor (Reith et al., 2006 ). Hence mechanical design has often been finely tuned in plants to achieve motility of diaspores and pollen.

Vegetative structures in regeneration
The mechanical properties of stem and root systems have in some species evolved to become either highly resistant to breaking or to the contrary easily snapped or pulled, which allows for a greater dispersal of seeds and vegetative parts. For example, the stem base of the fast-growing herbaceous annual Galium aparine L. is highly extensible and thus able to withstand high breaking strains. Uprooting may therefore be avoided as the mericarps hook on to passing mammals (Goodman, 2005 ). In many species, however, the opposite strategy may improve dispersal. Bobich and Nobel (2001) studied the morphology, mechanics, and rooting ability of four species of cholla (Opuntia spp.) cactus in the Mojave and Sonoran deserts. Only one species, O. bigelovii Engelm., showed evidence of vegetative reproduction in the field and was also the only species that had both relatively stiff, but weak terminal joints, as well as joints that demonstrated a high rooting ability. This species also had the highest frequency of spines per area of tubercule, thus increasing the chance that as a vertebrate brushed past the cactus, terminal segments would be broken off at the weak joint and also catch onto the animal's pelt, hence allowing for greater dispersal. In a study of the brittleness (a measure of ease of fracture) of twig bases of willows (Salix spp.), Beismann et al. (2000) found that S. fragilis L. and S. rubens Schrank were the most brittle, thus enabling these riparian species to propagate vegetatively downstream via broken twigs. However, in other species, e.g., S. triandra L. and S. viminalis L., the loss of twigs may serve more as a mechanical safety mechanism, as these riverside species are often submerged during flooding and the loss of twigs will reduce drag forces during inundation, thus preventing stem breakage. In the one subalpine species tested, S. appendiculata Vill., low stresses at fracture and yield were noted. Survival throughout the winter is a priority for this species, which is often buried under heavy snow or subjected to avalanches and so should be able to withstand high strains from either wind or flooding.

Root anchorage quality of saplings and vegetative propagules can also determine the success of reproduction in riparian species. In uprooting experiments (Karrenberg et al., 2003 ), Salix elaeagnos Scop. saplings growing in disturbed sites close to the river channel were found to have a relatively higher resistance to uprooting than Populus nigra L. saplings that occupied more stabilized sites. However, cuttings of P. nigra demonstrated superior growth and anchorage ability under severe flood conditions, thus leading to greater success in vegetative reproduction. Differences in root anchorage were attributed to differences in morphology and tensile strength between saplings and cuttings. Root tensile strength increases with decreasing diameter and is highly related to cellulose content (Genet et al., 2005 ). Herbs with highly fibrous root systems will thus better resist vertical uprooting than species with fewer or thicker roots. Fibrous-rooted grasses are also more deeply rooted than the average herbaceous plant (Schenk and Jackson, 2002 ), thus improving uprooting resistance. Ennos and Fitter (1992) therefore proposed that such species would better resist uprooting by grazing animals. If the root system remains intact in the ground, regrowth of shoots will also be more vigorous after damage (Bruna and Nogueira Ribeiro, 2005 ).

Aquatic macrophytes also have a large number of thin and branched roots. This increases surface area along with tensile strength and ensures a high root–soil friction (Stokes et al., 1995 ; Schutten et al., 2005 ). Aquatic plants are subject not only to pulling forces arising from waves, but also to grazing forces from birds. Although stem breakage will lead to a loss of fitness, uprooting is more damaging. Schutten et al. (2005) demonstrated that aquatic species had either relatively weak stems or a substantial investment in anchorage. In plants that break, e.g. Myriophyllum spicatum L., regrowth of the shoot is possible. Certain aquatic plants may also have evolved a weaker stem section above the sediment surface where breakage occurs, thus sacrificing the shoot, but protecting the perennating root system (Koehl, 1986 ). A large body of literature is available on the mechanics of wave-swept aquatic plants and algae, which is reviewed elsewhere (e.g., Koehl, 1986 ; Denny and Gaylord, 2002 ).

The mode of failure that a tree can undergo during a wind storm can have similar consequences for regeneration. Many broadleaved species are able to resprout when branches are broken or the stem is broken after a storm. By investing in a well-anchored root system, many broadleaved species break in the trunk (when soil conditions are not limiting to root system development) (Putz et al., 1983 ; Putz and Brokaw, 1989 ). Niklas (2000) suggested that broadleaved species invest a large proportion of their biomass in the main stems and establish a class of stems (branches and twigs) prone to mechanical failure during storms, thus providing a margin of safety against catastrophic wind damage. By ensuring good anchorage and a strong stem base, broadleaved trees can resprout after storm damage, thus prolonging their life and maintaining their position in the canopy. However, this hypothesis does not extend to most conifer species, which usually cannot resprout and often die after severe wounding (Stokes et al., 2005 ). Therefore, conifer seedlings need to grow quickly to benefit from gaps caused by storm damage, and in a mixed forest these young conifers will be in competition with resprouting broadleaved species if windstorms cause gaps in the canopy. A long-term dynamic and spatial analysis of regeneration and juvenile tree growth in a forest damaged by a windstorm, along with mechanical properties of each species studied, would be useful in understanding the ecological implications and how to manage such forests.

SPATIAL PATTERNS IN BIOMECHANICS

The mechanical properties of a plant will determine in part where it can survive, and biomechanical ecotypes may even develop within a species. Different environments and niches impose different mechanical stresses on plants so may require different mechanical strategies; for example, vines and lianas benefit from host species for their mechanical support (Rowe and Speck, 2005 ), mangrove forests are subjected to regular tidal cycles (Turner et al., 1995 ; Sun and Suzuki, 2001 ), and trees growing at high altitudes are exposed to strong winds and heavy snowfalls in winter (Körner, 2003 ). Hence, biomechanics can often explain why certain plant forms are found in a given environment, from intra-plant spatial strategies (e.g., in Rosa canina L., fast-growing young branches may use the older axes of the same plant as a trellis for mechanical support [Wissemann et al., 2006 ]), to broad-scale patterns of tree structure and architecture (see section: Spatial patterns in stem and branch mechanics). Nevertheless, niche determination may be governed by water and nutrient supply as priority requirements (Schenk and Jackson, 2002 ), before biomechanical factors. However, form and function are strongly interdependent, which means that biomechanical properties have major consequences, especially for light capture and fecundity. Biomechanical characteristics are also dynamic and plastic within a species and subject to change throughout the life of an individual, which further complicates our understanding of processes driving the spatial patterns in mechanical design.

Biomechanical plasticity appears to be an important trait facilitating occurrence in some habitats, including invasion by alien plant species. Spector and Putz (2006) showed that the pronounced plasticity of Brazilian pepper (Schinus terebinthifolius Raddi) enabled it to invade and dominate both maritime forest and saltmarshes in north Florida, USA. In crowded stands, Brazilian pepper grows like a woody vine, allowing it to emerge and cover neighboring tree crowns. This species can also adopt the form of a sprawling shrub, overtopping and killing saltmarsh plants because its arching stems spread laterally for long distances in the absence of tall neighbors. This success is governed by allometric changes in form that allow the same species to inhabitat diverse ecological niches.

Spatial patterns in stem and branch mechanics
It is difficult to find broad-scale spatial patterns in stem and branch mechanical properties, largely due to the trade-offs that have occurred between wood structure and function throughout the evolution of plants (see review by Sperry et al., 2006 , in this issue). Biomechanical characteristics are highly specific to the ecological context in which any plant species is living and also demonstrate a certain plasticity. Nevertheless, some patterns can be found across the plant kingdom; an example is the application of allometry theory to explain evolutionary changes in size (e.g., Niklas, 1994 ; West et al., 1999 ). Recent models of xylem structure in trees propose that plants have evolved a tapered structure of conduits, with narrower conduits distally, which should minimize the cost of water transport from roots to leaves (West et al., 1999 ; McCulloh et al., 2003 ; Sperry et al., 2006 ). However, these models do not predict the mechanical role of xylem, and further studies on xylem evolution need to be undertaken to better understand the series of trade-offs that have occurred depending on hydraulic and mechanical demands. Furthermore, whole-tree analyses of xylem structure across a range of species and habitats are surprisingly lacking, in particular with regard to the mechanical role of xylem (Anfodillo et al., 2006 ). Nevertheless, several studies of spatial patterns in wood density and specific gravity (SG) for different species have been undertaken, which directly take into account the local mechanical environment. In gymnosperms, Jagels et al. (2003) plotted E and the modulus of rupture (MOR) against SG for green wood of several low-to-moderate-density conifers and classed these conifers into two categories: trees from habitats subjected to low-to-moderate dynamic loading and trees subjected to strong dynamic loading. Trees exposed to the latter had higher specific stiffness (E/SG) than those exposed to weak lateral loading. No relationship was found between loading and MOR. Differences in SG therefore seem to reflect different adaptive strategies in temperate forests. However, the range of SG values is fairly narrow and certainly more limited than that found in tropical forest trees, where the wide range of values has been suggested to be related to niche diversification (Williamson, 1984 ; Woodcock and Shier, 2002 ). In a study of 30 tropical Bolivian tree species, van Gelder et al. (2006) determined that short shade-tolerant trees spend a longer time in the understory, where they are subjected to mechanical damage from falling debris; these species possess strong, dense wood. Tall shade-tolerant species and pioneer species have wood of much lower density and grow rapidly to the canopy, thus escaping long-term mechanical damage. Short pioneers dominate in early stages of succession in which growth can be fast and competition for light is strong, but tall pioneers are often long-lived and require denser wood to persist in the closing canopy. Tall, long-lived climax trees with dense wood will also be able to better withstand strong winds than will less dense woods (King and Loucks, 1978 ). The production of high-density wood found in tropical forests may require higher energy inputs throughout the year than occur in mid- to high latitudes. Therefore, fast-growing tropical pioneers may not be viable where there is a pronounced temperature cycle and shorter growing season. Nevertheless, it must not be forgotten that wood density is also strongly linked to physiological mechanisms discussed earlier, i.e., temperature changes and water supply. Therefore, more work is needed before we are able to understand the function of xylem in both a hydraulic and mechanical context and how this is linked to habitat occupation.

Spatial patterns in leaf mechanics
Patterns in leaf mechanics are evident across a range of spatial scales. Tropical species have been suggested as having tougher leaves than temperate species, in response to greater herbivore pressure (Coley and Aide, 1991 ), but there are few data that allow this scale of comparison free of confounding factors such as vegetation type, soil fertility, etc. We have no evidence to date of this latitudinal trend among evergreen species (Fig. 2), and varying latitudinal trends have been reported within taxa (Siska et al., 2002 ; Andrew and Hughes, 2005 ; Hallam and Read, 2006 ). However, there are strong gradients in leaf mechanics with respect to more specific abiotic stresses. First, in strongly seasonal climates in which a favorable growing season is paired with a harsh, cold, or dry season, the vegetation may have a substantial component of deciduous, annual, or ephemeral species, i.e., the leaves are short-lived (Chabot and Hicks, 1982 ; Kikuzawa, 1995 ; Givnish, 2002 ). Leaf structure is in effect decoupled from the seasonal stresses experienced by perennial evergreen species (Niinemets, 2001 ). Since high rates of photosynthesis may be necessary to accumulate reserves for canopy replacement in deciduous plants or to rapidly complete the life cycle in the case of annual plants, leaves may be relatively thin and soft with a high specific leaf area (SLA, leaf area per mass), facilitating high rates of carbon gain per leaf mass investment (Lambers and Poorter, 1992 ; Reich et al., 1992 ; Aerts, 1995 ; Cornelissen et al., 1999 ; Niinemets, 1999 ; Kohout and Read, 2006 ). These leaves are predicted to be less defended against herbivores than long-lived evergreen leaves (Chabot and Hicks, 1982 ), with any defense weighted more towards mobile chemical defenses that are not wasted at senescence (Coley et al., 1985 ) rather than structural defense, which cannot be reclaimed and may reduce photosynthetic rates by light attenuation or dilution of photosynthetic tissue. Even in evergreen species it may be more efficient in optimal environments to build short-lived low-cost leaves that are discarded as photosynthetic rates decline (Chabot and Hicks, 1982 ; Mooney and Gulmon, 1982 ). Hence, the limited data suggest that in environments that are optimal, or otherwise promote species with short inherent leaf life span, leaves on average have lower strength, work to fracture and stiffness, both as structures and often per unit thickness (Reich et al., 1991 ; Wright and Cannon, 2001 ; Balsamo et al., 2003 ; M. Kohout and J. Read, Monash University, unpublished data) (Fig. 2).

View larger version (21K): [in this window] [in a new window]   Fig. 2. Mean values of leaf toughness (measured as work to shear, with s.e.m) from a range of vegetation types. Unless otherwise indicated, the data refer to leaf strips that include secondary veins, but not midrib or margin, and are unpublished data for evergreen species (J. Read, G. Sanson, Monash University) (number of species given in parentheses). The values of work to shear are expressed per strip width. Specific work to shear is calculated as work to shear per unit thickness (a measure of "material" toughness). Values marked with an asterisk are properties of the lamina between secondary veins, so are lower than if secondary veins were included. Where multiple values are given for a collection, the darker-shaded bars represent the values from the more severe sites. Collections are ordered latitudinally (in degrees from the equator); the dashed line indicates the divide between tropical and temperate species. The Nothofagus species are native to a range of temperate latitudes (grown experimentally in Victoria). NSW, New South Wales. 1Choong et al. (1992) ; 2Turner et al. (1993) ; 3Read et al. (2006) ; 4Read et al. (2005); 5E. Caldwell (Monash University, unpublished data); 6M. Kohout and J. Read (Monash University, unpublished data)