| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ecology |
2School of Biological Sciences, Monash University, Victoria 3800, Australia; 3Institut Agronomique néo Calédonien, CIRAD UPR 22 Gestion intégrée de la faune, Station de Port-Laguerre, BP 73, Païta 98890, New Caledonia; and 4IRD–Laboratoire de Botanique et d'Écologie Végétale Appliquée, Institut de recherche pour le développement, Centre de Nouméa, BP A5 Nouméa 98848, New Caledonia
Received for publication April 6, 2006. Accepted for publication August 23, 2006.
ABSTRACT
The defining characteristics of sclerophylly are mechanical (e.g., hardness, toughness, stiffness), but little is known about how they vary in contrasting environments and contribute to the adaptiveness of sclerophylly. Here we investigate how the degree and nature of sclerophylly in terms of leaf mechanics differ between vegetation of two contrasting stressful environments, maquis on nutrient-deficient, moist sites and dry forest on moderate-nutrient, drier sites. We measured toughness, strength, and stiffness at the level of the whole leaf ("structural") and per unit thickness ("material"). Leaves of maquis plants were on average structurally stiffer, stronger, and tougher than those in dry forest. There was little difference in material properties between habitats, and leaf thickness was the main contributor to differences in structural mechanical properties between habitats. Flexural stiffness varied most among species and habitats, correlating strongly with leaf mass per area and thickness. We suggest that having thicker leaves allows efficient packaging of biomass to reduce branching costs in sunny but stressful environments, with subsequent impacts on structural mechanical properties. Sclerophylly is probably a complex phenomenon, however, with its mechanical constitution arising from both evolved mechanical properties that confer protection or resistance to stress and nonadaptive mechanical consequences of adaptation to stressful environments.
Key Words: dry forests • ecobiomechanics • leaf toughness • plant biomechanics • sclerophylly • ultramafic vegetation
Sclerophylly (having hard, tough, stiff, leathery leaves; Schimper, 1903
) occurs in plants native to a wide range of environments, including tropical to temperate and lowland to alpine, where there is some identifiable stress (Grubb, 1986
; Turner, 1994b
; Salleo and Nardini, 2000
). Its adaptive significance is still being debated (Lamont et al., 2002
). It has been suggested to be an adaptation to specific stresses such as seasonal water deficits (Oertli et al., 1990
) and excessive solar radiation (Jordan et al., 2005
) and as an adaptation to, or consequence of, low-nutrient soils (Whittaker, 1954
; Loveless, 1961
; Beadle, 1966
, 1968
; Jaffré, 1980
; Kruckeberg, 1984
). Propagation of ice may be delayed in tissues with densely lignified or cutinized barriers, so sclerophylly may also provide resistance to frost damage by allowing some supercooling (Larcher, 2005
). Sclerophylly may be more likely in environments where multiple stresses are operating (Mooney and Dunn, 1970
). Sclerophylly may alternatively, however, be a nonspecific response to a range of stresses (Salleo and Nardini, 2000
). This "nonspecific" response may be directly adaptive, for example by protection from damage, thereby enhancing leaf longevity and long-term photosynthetic efficiency across a range of stressful environments (Chabot and Hicks, 1982
; Turner, 1994b
). It might be nonadaptive, however, occurring as the by-product of structures or physiology that increase resistance to stress.
The defining characteristics of a sclerophyll are textural (Schimper, 1903
), potentially the sum of a suite of constituent mechanical properties such as strength, toughness, stiffness, and hardness (Choong et al., 1992
; Edwards et al., 2000
; Read and Sanson, 2003
). The difficulty of interpreting the functional significance of sclerophylly is threefold. First, relatively little is still known about the mechanical properties of leaf blades (Niklas, 1999
). Second, mechanical properties will derive from anatomical design across multiple scales, including from the cell to whole leaf. Third, the mechanical design of a leaf will involve compromises between the need to withstand static and dynamic loads (Niklas, 1999
) while maximizing physiological performance (Givnish, 1979
; Parkhurst, 1986
; Gutschick, 1999
; Press, 1999
), not just at the level of the leaf, but of the whole plant (Givnish, 1979
). For example, in optimal environments leaves may be thinner to minimize internal self-shading and resistance to CO2 diffusion, and larger to enhance photosynthetic rates (if the thicker boundary layer elevates leaf temperature to optimal levels) and reduce relative branching costs (Givnish, 1979
), thereby increasing growth rates. In suboptimal environments, however, abiotic stresses may lead to a range of evolved variations in leaf design that may impact mechanical properties. For example, leaves that are smaller, more lobed, or serrated with a thinner boundary layer may maximize photosynthetic gains relative to transpiratory costs in infertile, dry, or cold environments (Givnish, 1979
). The increased cost of branching associated with smaller leaves may be amortized by having a lesser number of thicker leaves, at least in sunny, open environments where the negative effects of internal self-shading and resistance to CO2 diffusion (Givnish, 1979
; Parkhurst, 1986
, 1994
) may be reduced. Leaf photosynthetic function may even be enhanced by thickening leaves in sunny environments (Roderick et al., 1999
), with reduced rates of water loss (Parkhurst, 1994
). Increasing leaf thickness will have consequences for mechanical properties, increasing the leaf strength, toughness and stiffness. Hence, sclerophylly may be at least partly a consequence of efficient biomass partitioning in stressful environments (Read et al., 2005
).
In addition, in stressful environments the rate of photosynthesis is lower, and longer leaf lifespans may be necessary to maximize returns per investment (Givnish, 1979
; Chabot and Hicks, 1982
; Mooney and Gulmon, 1982
; Turner, 1994b
). The plant may invest in defensive traits, including against herbivores, to increase the likelihood of leaves reaching the optimal lifespan, including chemical and biotic defense, spinescence, and pubescence (Gutschick, 1999
; Press, 1999
), and mechanical defense (strength and toughness), i.e., sclerophylly.
These are general trends that might be predicted in many stressful environments. In contrasting stressful environments, however, different selection pressures may operate on leaf design, and thus the constitution of sclerophylly may differ in terms of its mechanical components. For example, Wright and Westoby (2002)
found that sclerophylly as judged by leaf toughness (work to shear) differed in its relationship with leaf mass per unit area (LMA) between wet and dry sites (dry site species having a lower work to shear for a given LMA), consistent with different anatomical and physiological strategies being employed in contrasting conditions of water availability.
In this study we investigate how the degree of sclerophylly varies in two different types of stressful environment and whether sclerophylly is the same phenomenon in terms of its mechanical properties in these environments. There has been little detailed study of the mechanical components of sclerophylly in contrasting environments, notably in studies by Choong et al. (1992)
, Turner et al. (1993)
, Edwards et al. (2000), Wright and Westoby (2002)
, and Read et al. (2005)
, and even less study of a wide range of mechanical traits (Edwards et al., 2000
; Read et al., 2005
). Here we compare the leaf mechanics of plants growing in maquis, a shrub-dominated vegetation on ultramafic soils characterized by low levels of some plant nutrients and high levels of metals such as nickel and iron (Jaffré, 1980
, 1993
; McCoy et al., 1999
) in a tropical wet climate (Table 1), with those of plants in "dry forests" on soils of moderate nutrient status in a drier climate (Jaffré et al., 1993
) (Table 1). There is little overlap in species between the two vegetation types, with species apparently specialised to these contrasting habitats. If sclerophylly is a specific adaptation to, or consequence of, low nutrients, then the maquis plants should have more scleromorphic leaves than those of plants in dry forest. Conversely, if sclerophylly is an adaptation to, or consequence of, seasonal water deficits then the dry-forest plants should have more scleromorphic leaves than those of maquis plants. The main questions are: (1) Is sclerophylly (specifically, strength, toughness, and flexural stiffness) more strongly developed in species growing on low-nutrient soils (as found in a study of temperate species; Read et al., 2005
), with a negative correlation between leaf mechanics and foliar concentrations of N and P? Or alternatively, is sclerophylly more strongly developed in species subjected to more severe seasonal water deficits? (2) Does the mechanical constitution of sclerophylly vary between these contrasting stressful environments?
|
We investigated the mechanical components of sclerophylly using fracture tests to measure toughness and strength, and bending tests to measure stiffness, both at the level of the whole leaf ("structural") and at the level of "material" properties (expressed per unit leaf thickness) (Appendix 1). We also investigated traits that may be associated with the mechanical components of sclerophylly and are therefore relevant to their interpretation, including foliar nutrients, cell wall concentration, and some chemical defenses that may potentially impact alternative investment in leaf mechanical properties.
MATERIALS AND METHODS
Study region and sites
Leaves were collected in November 2003 from up to five sites in each of dry forest and maquis. Maquis is extensive in southern New Caledonia, typically on massifs because of its association with ultramafic soils, but occurs across a wide range of elevations including near sea level (Jaffré 1980
, 1992
, 1993
). Dry forest (also called "sclerophyll forest") is restricted to fragments along the drier west coast below 300 m a.s.l., mostly on sedimentary rock (Jaffré et al., 1993
, 1998a
; Bouchet et al., 1995
). These forests were once widespread but have been reduced to c. 2% of their former distribution and are significantly threatened (Bouchet et al., 1995; Jaffré et al., 1998a
).
The soil of all maquis sites was ultramafic (on peridotite or serpentinite), ranging from brown eutrophic hypermagnesian soils (syn. vertic cambisols) to ferrallitic soils of the ferritic type (i.e., deficient in aluminum clays, syn. acric ferralsols), including intermediate forms. The dry forests occurred on brown eutrophic soils (eutric cambisols) derived largely from basic sedimentary rock, with varying levels of Ca due to juxtaposition with calcareous and acid siliceous rock; Ca levels were particularly elevated at one site where the soil was intermediate between a rendzina and a brown eutrophic soil. Soil nutrient levels were low in the maquis, particularly for P, K and Ca, and levels of C, N, P, K, Ca and Ca : Mg were significantly higher (3–58-fold) in the dry forest (Table 1). Soil pH was slightly higher and Fe and Ni 9- and 17-fold higher, respectively, in the maquis (Table 1). There was no difference between the two vegetation types in soil Mg, with high variability among the maquis sites.
All sites were in the same region (Fig. 1), but since dry forest is typically closer to the coast and the maquis is restricted to ultramafic soils associated with massifs, the sites differed in climate, particularly in rainfall (Table 1) because of a strong inland rainfall gradient. The annual rainfall across the collection sites ranges from c. 850–1060 mm in the dry forests to c. 1400–2700 mm in the maquis (from Météo France de Nouvelle-Calédonie: rainfall stations, or derived from isohyets) (Table 1). The driest months in this region are September to November, with the duration and severity more marked on the drier west coast than on the massifs (Jaffre et al., 1993
).
|
Fully expanded leaves less than c. 12 months old and adult in form (some species are heteroblastic), were collected from sunlit branches <2 m above ground. Leaves were collected in the early morning, wrapped in wet paper towel, and sealed in a plastic bag to ensure full hydration, with mechanical and morphological measurements undertaken over the next 12–24 h. For species with compound leaves, the unit of measurement was a leaflet. Leaves were haphazardly chosen from this sample for measurement of mechanics, morphology, and chemistry, and for fixation for later anatomical study.
Indices of sclerophylly, leaf morphology, and chemistry
Leaf mass per area (LMA) was used as an indirect index of sclerophylly (Witkowski and Lamont, 1991
; Groom and Lamont, 1999
; Edwards et al., 2000
). LMA is not an ideal index because it reflects mass allocation rather than necessarily mechanical properties, but is often positively correlated with these properties (Choong et al., 1992
; Read and Sanson, 2003
; Read et al., 2005
), and correlates with an index derived from botanists' rankings (Edwards et al., 2000
; Read and Sanson, 2003
). Succulent species may also have high LMA (Turner, 1994b
; Cowling and Campbell, 1983
), but none were recorded in this study. LMA was measured on one leaf per replicate plant. Leaf area was measured by image analysis (Bioscan Image Analyser, School of Biological Sciences, Monash University), and leaves were then dried to constant mass at 80°C. Water content was calculated per leaf dry mass. Leaves collected for chemical analysis were dried at 50°C then later freeze-dried and ground to a powder using a Retsch Mixer Mill MM 301 (Haan, Germany). The Loveless sclerophylly index (SI: 100 x crude fiber dry mass/crude protein dry mass) was devised as an estimate of the ratio of cell wall to cell content (Loveless, 1961
, 1962
). We determined cell wall as neutral detergent fiber (NDF) following Van Soest et al. (1991)
. Foliar N was measured using a Leco CHN-2000 analyzer (St Joseph, Michigan) and presented per unit dry mass, and per unit water content as an estimate of protoplasmic concentration. SI was calculated as the ratio of NDF dry mass to protein dry mass (N x 6.25). Foliar P was measured to allow correlation with sclerophylly indices, using the molybdenum-blue colorimetric method (Grimshaw et al., 1989
) following digestion by the sulphuric-peroxide procedure (Grimshaw, 1987
) and expressed per unit dry mass and per unit mass of water.
"Total phenolics" were extracted in 50% acetone (Cork and Krockenberger, 1991
) and assayed by the Prussian-blue method (Price and Butler, 1977
) as modified by Graham (1992)
. Concentration was expressed as gallic acid equivalents (GAE) per leaf dry mass. Because tannins vary in their capacity to bind proteins, we report binding of protein per unit leaf dry mass rather than concentration of tannins, using the blue BSA (bovine serum albumen) method (Asquith and Butler, 1985
) with bovine gamma globulin as the standard. The presence of cyanogenic glycosides was investigated in fresh foliage using picrate paper (Harborne, 1998
), with apple seed controls.
Leaf mechanical properties
Fracture tests (punching, tearing, and shearing) and bending tests were undertaken following Read and Sanson (2003)
using a purpose-built, portable force-tester. The mechanical properties were derived from force-displacement curves using the software LeafME (M. Logan, Monash University) (Appendix 1). We used work to punch, work to shear and work to tear (the latter two measured on a 5-mm strip cut from between the leaf margin and midrib) as estimates of leaf toughness (Read and Sanson, 2003
). Strength was measured in a punch test as the maximum force divided by the area over which the force was applied, and force to tear and tensile strength were measured on notched leaf strips. We present both "structural" properties (normalized to width of the test piece, but not thickness) and "material" properties (i.e., normalized to both width and thickness and termed "specific") (Appendix 1). The flexural Young's modulus (E) and flexural stiffness (EI, where I is the second moment of area) (Appendix 1) were measured in a three-point bending test (Jackson, 1992
) on a 5-mm-wide leaf strip. A pilot study was undertaken on Acacia spirorbis to estimate the optimal span to thickness ratio (S/T) to minimize shear deformation (Jackson, 1992
). A ratio of 40–80 was deemed suitable, but for some other species a ratio of 40–50 seemed unreliable. We generally confined our tests to strips with S/T of 70. Because we used a standard strip for this test, we are not considering differences among species in whole leaf stiffness, which would be influenced by other characteristics, including midrib, margin and petiole characteristics, and leaf shape and size, but we consider that these have less to do with the textural issues of sclerophylly.
Statistical analyses
Nested ANOVA was used to test for differences between vegetation types (fixed factor) and among species (random factor) in each mechanical property, treating species as though they were fully independent, i.e., using means of species rather than of sites. ANCOVA was used to test whether the relationship between mechanical properties and LMA varied between vegetation types (log LMA used as the covariate). Principal components analysis (PCA) was used to reduce the set of mechanical variables to major components. Mechanical properties were highly intercorrelated, so hierarchical partitioning (Chevan and Sutherland, 1991
) was used to determine the independent contribution (IHP) of each mechanical property to LMA (as an indirect index of sclerophylly). Z-scores were calculated from randomization of IHP values, with statistical significance based on the normal distribution (Mac Nally and Horrocks, 2002
). For all analyses, assumptions were checked and transformations used where necessary. Hierarchical partitioning was undertaken with hier.part v. 0.5–1 and rand.hp (Mac Nally and Walsh, 2004
) in R (R Development Core Team 2004
), and all other analyses with SYSTAT v. 10 and v. 11 (Point Richmond, California, USA), with a critical value of 
= 0.05 used for hypothesis testing.
Analyses that treat species as independent data points do not take into account that species may share traits through common descent and may lead to correlations that are artifacts of relatedness (Harvey and Pagel, 1991
; Purvis and Rambaut, 1995
; Mazer, 1998
). Therefore, in addition to analyses treating each species as an independent data point (TIPs), some trends in leaf traits were assessed using phylogenetically independent contrasts (PICs) computed by CAIC v. 2.6.8 (Purvis and Rambaut, 1995
), an approach developed from Felsenstein (1985)
. A phylogenetic tree was derived using APG (2003)
, with finer resolution for Apocynaceae and Euphorbiaceae based on current subfamily classification (a polytomy was used for subfamilies of Euphorbiaceae), and used to compute independent standardized contrasts for all variables. Comparisons of leaf traits between vegetation types were undertaken using t tests of standardized contrasts generated by CAIC using the "Brunch" option (11 contrasts). To test associations among leaf properties, standardized contrasts for each continuous variable were generated using the "Crunch" option (34 contrasts), followed by regression through the origin (Purvis and Rambaut, 1995
).
RESULTS
Leaf chemistry and morphology
There was significant variation in all foliar chemical and morphological traits among species, but not all traits varied between vegetation types (the same conclusions were reached from analysis of PICs) (Table 2). Foliar N varied 5-fold among species (0.6–3.0%), and was 2-fold higher on average in dry-forest species than maquis species (Table 2). Foliar P was also generally low, but varied 7-fold among species (0.02–0.14%) and was 3-fold higher in dry-forest species than maquis species. There was 8-fold variation among species in N per unit water (2.7–22.0 mg · g–1) and 12-fold variation in P per unit water (0.1–1.2 mg · g–1), with the same contrast between vegetation types as when expressed per unit dry mass (2- and 3-fold higher in dry forest, respectively; Table 2). The high foliar N : P ratio (>20) in maquis species (21.6–42.2) (Table 2), and in Acacia spirorbis and Cloezia artensis var. artensis on maquis soils (Table 3), suggests P limitation (Güsewell, 2004
) in maquis vegetation, more so than in the dry forest where only 25% of species had N : P as high as 20–24.
|
|
|
Acacia spirorbis and Cloezia artensis var. artensis occurred on both soils, although the latter was found only at one dry-forest site. For both species, higher LMA and N : P were recorded in maquis plants, and P concentration per water was higher in dry-forest plants (Table 3). Acacia spirorbis also had a higher protein to protein precipitation ratio in dry-forest plants. Cloezia artensis var. artensis had a higher P and N per dry mass and N per water in plants from dry forest and a higher SI and thickness in maquis plants (Table 3).
Leaf mechanics
There was significant variation among species in all mechanical properties, but not all properties varied between vegetation types (Table 2). "Structural" properties varied 7–26-fold among species, other than EIW, which varied 136-fold, and were always higher on average in the maquis species (from 1.4-fold higher to 6.4-fold higher for EIW). Punch strength and work to tear, however, did not differ between vegetation types using phylogenetic contrasts (Table 2). The "materials" properties (expressed per unit leaf thickness) varied 7–63-fold among species, but the only differences between vegetation types were higher specific work to shear in maquis species (but not significantly different for phylogenetic contrasts), and higher specific punch strength in dry-forest species. Ratios of the major structural mechanical properties (work : strength, EIW : work and EIW : strength) also differed between vegetation types, always higher on average in maquis species (up to 5.4-fold higher in EIW : strength) (Table 2).
Both A. spirorbis and C. artensis var. artensis showed higher work to shear, EIW, work : strength and EIW: strength in maquis plants (Table 3). Acacia spirorbis also had a higher specific work to shear and EIW : work in maquis plants, and C. artensis var. artensis had a higher work to punch in maquis plants, but a higher specific punch strength in dry-forest plants (Table 3). Variation ranged to 3.8-fold in EIW for C. artensis var. artensis between habitats, but the differences within these species (when present) was generally smaller than differences between the averages of species of the two vegetation types.
Relationships among mechanical properties and with LMA
Many of the mechanical properties were strongly intercorrelated. The main patterns were as follows: (1) for all test types, strong leaves were also strong per unit leaf thickness (R = 0.58, P < 0.001), and tough (high work to fracture) leaves were tough per unit leaf thickness (R = 0.73–0.86, P < 0.001), but stiff leaves were not stiff per unit leaf thickness (R = 0.26, P = 0.264); (2) strong leaves were also tough (R = 0.65–0.88, P < 0.001), with better correlations between punching and shearing tests, than of either with tearing tests; and (3) stiff leaves (high EIW) were strong and tough (R = 0.45–0.80, P = 0.003 to < 0.001) (more so with results from punching and shearing tests than those from tearing tests).
A range of mechanical properties contributed to the first component axis of the PCA (explaining 60% of the total variation among species), with the strongest contributions by specific work to shear, punch strength, force to tear, specific work to punch and work to shear (Fig. 3). EIW contributed most strongly to component 2, followed by specific punch strength (Fig. 3). There was a significant difference between maquis and dry-forest species in component 1 scores (F = 6.0, P = 0.019), but much more so along component 2 (F = 46.0, P < 0.001).
|
|
|
|
|
The adaptive significance of sclerophylly
The leaves of maquis species were on average stronger (but not after phylogenetic correction), tougher, and particularly stiffer (higher EIW) than leaves of dry-forest species. Thus, sclerophylly (specifically flexural stiffness, toughness, and strength) was more strongly developed in species growing on low-nutrient soils than in those on sites of moderate nutrient status but more severe seasonal water deficit. Flexural stiffness made the strongest contribution to sclerophylly as judged by LMA, consistent with other studies (Read and Sanson, 2003
; Read et al., 2005
). High stiffness can also occur in other leaf forms, however, such as succulent foliage, so that not all stiff leaves are sclerophylls, and flexural stiffness should not be used on its own as a measure of sclerophylly. The range of values overlaps values recorded previously in other species, but none of the maquis leaves were as stiff, tough, or strong as those of some woody species occurring on low-nutrient soils in Western Australia (Fig. 6).
|
, 1962
; Beadle, 1968
) or by promotion of nutrient-use efficiency by extending leaf lifespan (Chabot and Hicks, 1982
; Turner 1994b
). Similarly, structural mechanical properties correlated negatively with foliar nutrients, whether expressed per unit biomass, or per unit water content as a better estimate of protoplasmic concentration. These patterns appear to conflict with the stronger trend of LMA (as a sclerophylly index) with mean annual rainfall than foliar N and P found by Lamont et al. (2002)
in southwestern Australia and the Cape of South Africa. Negative correlations of mechanical properties or LMA with foliar P and/or N concentration have been demonstrated elsewhere, however, across a variety of scales of investigation (Loveless, 1961
, 1962
; Sobrado and Medina, 1980
; Mooney and Gulmon, 1982
; Specht and Rundel, 1990
; Reich et al., 1991
; Choong et al., 1992
; Turner, 1994a
, b; Wright and Cannon, 2001
; Wright et al., 2004
; Read et al., 2005
). Furthermore, the absence of a correlation between LMA and foliar P or N does not necessarily signify the absence of a causal relationship between sclerophylly and soil-nutrient status if plants can maintain high protoplasmic levels of nutrients through adaptation to low fertility (Read et al., 2005
).
Even though we show that sclerophylly is more pronounced in the infertile, rather than the drier habitat, we do not have sufficient evidence to confirm that sclerophylly is a specific adaptation to nutrient deficiency. First, the ultramafic soils are also characterized by nutrient imbalances (high Mg : Ca) and high levels of elements such as Ni that are potentially toxic (Latham et al., 1978
; Latham, 1980
; Jaffré, 1992
), and the roles of these factors in influencing leaf form are uncertain. The high N : P ratios in maquis plants, however, does suggest that the plants are P-limited. Second, the greater development of sclerophylly in maquis species may simply reflect a greater degree of stress (resource deficit and/or toxicity) than experienced by species of the dry forest, i.e., where sclerophylly is a nonspecific response to environmental stress (Salleo and Nardini, 2000
). Third, the leaves of maquis species might be exposed to strong evaporative loads during the drier season because the canopy is more open than in dry forests, i.e., the maquis plants may experience stronger water deficits than apparent from the rainfall data. In addition, although maquis is currently widespread in wet climates in New Caledonia, it is predominantly a secondary formation in these areas, with fire associated with human settlement as the major determinant of its distribution (Jaffré, 1980
; Morat et al., 1986
; Jaffré et al., 1998b
; McCoy et al., 1999
). Exceptions are small areas of primary maquis on ultramafic soils on drier sites (Virot, 1956
; Jaffré, 1980
; Morat et al., 1981
); hence, some maquis species may have evolved under drier conditions than suggested by their current range. Therefore, although our data suggest that sclerophylly is better developed in species native to low-nutrient soils, they may also reflect the greater degree of stress to which these plants are exposed, either nutritional, or of a more complex and multidimensional nature.
If sclerophylly is a nonspecific response to environmental stress, it may still be directly adaptive. For example, in suboptimal environments the reduced rate of photosynthesis may increase the time needed to maximize carbon gain per unit investment (Chabot and Hicks, 1982
; Mooney and Gulmon, 1982
). Therefore, the adaptive significance of sclerophylly may result from leaf protection, particularly from herbivores, thereby enhancing leaf longevity and long-term carbon gain efficiency (Turner, 1994b
). In this scenario, sclerophylly would potentially be advantageous only in environments where light was not limiting photosynthetic rates. Shaded conditions would reduce photosynthetic rates and therefore internal carbon availability for structural development, thereby imposing a direct constraint on sclerophylly. Furthermore, the anatomical components of sclerophylly may reduce light transmission through the leaf (affecting carbon balance), a particular problem in shaded conditions. High levels of total phenolics and tannin activity in maquis species also suggest higher investment in protection from herbivores (Harborne, 1998
) or from UV radiation and its consequences (Hagerman et al., 1998
; Mazza et al., 2000
), consistent with the view that the maquis species experience greater stress (although we have measured a limited range of defenses, and investment in these C-based defenses may reflect greater soil nutrient limitation: Coley et al., 1985
). Only three species were spinescent (all in the dry forest), consistent with the absence of native mammalian folivores in New Caledonia (de Garine-Wichatitsky et al., 2003
).
Sclerophylly might alternatively be nonadaptive, i.e., occurring as a nonspecific response to environmental stress as the by-product of adaptation to a stressful environment. Flexural stiffness has been found to be a better predictor of sclerophylly (as judged by LMA) than strength or toughness, here and in previous studies (Read and Sanson, 2003
; Read et al., 2005
). While strength and toughness are potentially associated with protection, particularly from herbivores (Choong et al., 1992
; Lucas et al., 2000
; Sanson et al., 2001
), it is more difficult to provide an adaptive explanation for the stronger contribution of stiffness, because values exceed what is necessary for support. We suggest that the explanation lies in the strong effect of leaf thickness on leaf mechanical properties, particularly stiffness, and on biomass-allocation efficiency. Thickening of leaves to maintain photosynthetic biomass while minimizing branching costs (and water loss) and potentially to enhance leaf performance (Roderick et al., 1999
; Niinemets, 1999
) should increase leaf strength, toughness, and flexural stiffness, and indeed there is a positive correlation of strength and toughness with leaf thickness (R = 0.35–0.67, P = 0.024 to <0.001). However, the strongest correlation is with flexural stiffness (EIW) (R = 0.88, P < 0.001); that is, stiffness is influenced more consistently by thickening than by altered material properties (EIW did not correlate with E across all species, whereas strength and toughness correlated with their respective material properties). The particularly strong effect of leaf thickness on flexural stiffness is due to its contribution to the shape variable I (the second moment of area), such that flexural stiffness is proportional to the cube of thickness (I = WT3/12: Appendix 1), whereas structural strength and toughness are linearly related to thickness. In addition, increased thickness has the effect of moving the stiffer outer leaf layers (epidermis and cuticle) away from the neutral axis of bending, further enhancing EIW. Much of the phenomenon of sclerophylly may therefore have evolved as a means of optimally allocating biomass to maximize carbon gain and growth in stressful environments, with the mechanical properties only a consequence rather than of primary adaptive significance.
The range of specialized leaf anatomies among sclerophyllous species, particularly in regard to sclerification (Esau, 1977
; Sobrado and Medina, 1980
; Grubb, 1986
; Choong et al., 1992
; Groom et al., 1997
; Cunningham et al., 1999
; Read et al., 2000
; Peeters, 2002
; Jordan et al., 2005
), however, suggests that sclerophylly is more than just selection for efficient biomass allocation. It is probably a sum of (1) the consequences of efficient biomass allocation, with secondary selection by herbivory given that thicker leaves are also tougher and stronger structures, (2) primary selection of mechanical properties that reduce herbivory (Lucas et al., 2000
; Sanson et al., 2001
) (if protection from damage is a major component of sclerophylly, then the strong contribution of stiffness to sclerophylly could also be a consequence of thickening the leaf to increase leaf strength and toughness); (3) consequences of structures or physiology that increase resistance to stress: examples include a thick cuticle and epicuticular wax layer (Gutschick, 1999
), which can reduce water loss, leaching of nutrients, microbial attack, and overheating, and increase water repellence, which may increase frost resistance, and the increased tissue density due to small cell size (Loveless, 1961
; Beadle, 1966
) that may accompany the physiology associated with phosphorus deficiency. The role of sclerophylly in resistance to water deficits is still being debated (Salleo and Nardini, 2000
; Niinemets, 2001
).
The two species sampled from both habitats showed plasticity in some leaf traits, including significant plasticity of some mechanical properties. When plasticity occurred, however, the degree of variation was generally smaller (or inconsistent between species) than the overall variation between vegetation types. Therefore, the differences we have recorded between vegetation types are likely to reflect the traits of differentially adapted species, rather than only plastic responses to these contrasting environments. In addition, analyses of phylogenetically independent contrasts indicate that these trends generally remain strong after any relatedness effect (potential "phylogenetic inertia") is removed.
The constitution of sclerophylly in contrasting environments
Because the mechanical design of a leaf may involve compromises among requirements for efficient biomass allocation, protection from damage, and design that enhances physiological performance, the mechanical constitution of sclerophylly may vary between contrasting stressful environments. The summary provided by the PCA showed that species varied considerably in their combinations of properties, irrespective of the habitat in which they occurred. Even though species varied in their constitution of sclerophylly (also shown by Edwards et al., 2000
; Read and Sanson, 2003
; Read et al., 2005
), our analysis of mechanical properties using LMA as a covariate suggested that the mechanical constitution of sclerophylly did not vary overall between the maquis and forest plants. Leaves of maquis plants, however, had lower tensile strength and E adjusted for EIW. There are several possible explanations. Dry-forest species may differ in anatomy, for example, if thickening of leaves of maquis species occurs primarily by thickening palisade and spongy mesophyll (tissue with low stiffness compared to epidermal tissues), E may be lower relative to EIW. In particular, because tissue density can have marked effects on internal CO2 diffusion (Niinemets, 1999
), the lower E adjusted for EIW in maquis plants may be the result of reduction in tissue density necessary to offset any negative impacts of thickening on internal CO2 diffusion. Tissue density did not differ between species from dry forest and maquis. However, density correlated negatively with leaf thickness in maquis species, but not in dry-forest species (Fig. 2). Hence, there is some difference in the constitution of leaf mechanics between these two environments, but not in the structural mechanical properties that best correlate with sclerophylly. There is little evidence to date of differences in mechanical trends associated with sclerophylly among the vegetation of different climatic regions (e.g., Fig. 5), but few data sets are yet available to test trends in the mechanical constitution of sclerophylly at a global scale.
The stronger correlation of structural properties with "sclerophylly" is to be expected when using LMA as its index, because an increase in leaf mass per area should (on average, but depending on any associated change in anatomical design) increase strength, toughness, and stiffness irrespective of whether thickness or density (the components of LMA) is increasing. LMA is more likely to correlate strongly with material properties if density is contributing significantly to the increase in LMA. Density is generally unlikely to increase with increasing LMA beyond a certain point, however, because of its negative effect (depending on its anatomical basis) on gaseous diffusion (Niinemets, 1999
), which becomes even more limiting in a thicker leaf (as suggested by our data). It is sometimes assumed that "material" toughness (toughness per unit thickness, or specific work to fracture) is the mechanical essence of sclerophylly. Indeed, it is difficult to argue that a thick but otherwise mesophytic leaf is anything other than a thick mesophytic leaf if there is no change in anatomy. In contrast, a leaf that differs in material properties must differ in cell wall volume fraction or chemistry, or in anatomy, e.g., in sclerification (but not necessarily in density), so perhaps a more useful sclerophylly index would be one that correlates strongly with material mechanical properties. Our studies to date, however, have shown that leaf structural mechanical properties (the consequence of thickness plus material properties) correlate best with botanists' rankings of sclerophylly, differ between contrasting environments more than do material properties, and that there may be various ways to achieve sclerophylly anatomically even though some general trends in mechanical traits appear to be relatively consistent among the contrasting stressful environments investigated to date. Despite these consistent trends across vegetation types, sclerophylly may be best considered as a complex and variable phenomenon at the species level, in which the leaf form is an evolved compromise between efficient carbon gain (at the level of the whole plant), resistance to stress of various types, and protection from damage.
APPENDIX 1.
Measured and derived mechanical properties used in this study (modified after Sanson et al., 2001
; Jackson, 1992
; Read and Sanson, 2003
). A, area under the force-displacement curve (N · m = J); C, cross-sectional area of punch (m2); D, displacement of moving head of test machine (m); Fmax, maximum force recorded (N); G, gradient of force-displacement curve (N · m–1); S, span length between supports in a bending test (m); T, thickness of leaf at position of test (m); W, width of leaf in plane of shear, tear, or bend (m)
APPENDIX 2.
Species from (A) maquis and (B) dry forest used in this study, and their habit. Collection details are given (the number of replicate plants sampled, followed by the number of sites from which they were collected). Nomenclature follows Jaffré et al. (2001)
. L, liane; S, shrub, or shrublike (including small but single-trunked forms); T, tree. The first letter indicates the habit in which the species was commonly observed on the study sites
FOOTNOTES
1 The authors are grateful to IAC/CIRAD and IRD (New Caledonia) for their support, the invaluable assistance of J. Spaggiari in the field, and F. Clissold, S. Kerr, and D. Griepsma for analyses of foliar and soil chemistry. This project was funded by a Monash Small Grant (J.R. and G.S.) and the Programme de Conservation des Forêts Sèches (M.G.). 
5 Author for correspondence: (jenny.read{at}sci.monash.edu.au
)
LITERATURE CITED
Angiosperm Phylogeny Group (APG).. 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Linnean Society 141: 399-436.[CrossRef]
Asquith T. N. Butler L. G.. 1985. Use of dye-labeled protein as spectrophotometric assay for protein precipitants such as tannin. Journal of Chemical Ecology 11: 1535-1544.[CrossRef][ISI]
Beadle N. C. W.. 1966. Soil phosphate and its role in molding segments of the Australian flora and vegetation, with special reference to xeromorphy and sclerophylly. Ecology 47: 992-1007.[CrossRef][ISI]
Beadle N. C. W.. 1968. Some aspects of the ecology and physiology of Australian xeromorphic plants. Australian Journal of Science 30: 348-355.
Bouchet P. Jaffré T. Veillon J.-M.. 1995. Plant extinction in New Caledonia: protection of sclerophyll forests urgently needed. Biodiversity and Conservation 4: 415-428.[CrossRef][ISI]
Chabot B. F. Hicks D. J.. 1982. The ecology of leaf life spans. Annual Review of Ecology and Systematics 13: 229-259.
Chevan A. Sutherland M.. 1991. Hierarchical partitioning. American Statistician 45: 90-96.
Choong M. F. Lucas P. W. Ong J. S. Y. Pereira B. Tan H. T. W. Turner I. M.. 1992. Leaf fracture toughness and sclerophylly: their correlations and ecological implications. New Phytologist 121: 597-610.[CrossRef][ISI]
Coley P. D. Bryant J. P. Chapin III F. S.. 1985. Resource availability and plant antiherbivore defense. Science 230: 895-899.
Cork S. J. Krockenberger A. K.. 1991. Methods and pitfalls of extracting condensed tannins and other phenolics from plants: insights from investigations on Eucalyptus leaves. Journal of Chemical Ecology 17: 123-134.[CrossRef][ISI]
Cowling R. M. Campbell B. M.. 1983. The definition of leaf consistence categories in the fynbos biome and their distribution along an altitudinal gradient in the south eastern Cape. Journal of South African Botany 49: 87-101.
Cunningham S. A. Summerhayes B. Westoby M.. 1999. Evolutionary divergences in leaf structure and chemistry, comparing rainfall and soil nutrient gradients. Ecological Monographs 69: 569-588.[CrossRef]
De Garine-Wichatitsky M. Duncan P. Labbe A. Suprin B. Chardonnet P. Maillard D.. 2003. A review of the diet of rusa deer Cervus timorensis russa in New Caledonia: are the endemic plants defenceless against this introduced, eruptive ruminant?. Pacific Conservation Biology 9: 136-143.
Edwards C. Read J. Sanson G.. 2000. Characterising sclerophylly: some mechanical properties of leaves from heath and forest. Oecologia 123: 158-167.[CrossRef][ISI]
Esau K.. 1977. Anatomy of seed plants, 2nd ed Wiley, New York, New York, USA.
Felsenstein J.. 1985. Phylogenies and the comparative method. American Naturalist 125: 1-15.
Givnish T.. 1979. On the adaptive significance of leaf form. In O. T. Solbrig, P. H. Raven, S. Jain, and G. B. Johnson [eds.] Topics in plant population biology 375-407 Columbia University Press, New York, New York, USA.
Givnish T. J.. 1986. Biomechanical constraints on crown geometry in forest herbs. In T. J. Givnish [ed.] On the economy of plant form and function 525-583 Cambridge University Press, Cambridge, UK.
Graham H. D.. 1992. Stabilization of the Prussian blue color in the determination of polyphenols. Journal of Agricultural and Food Chemistry 40: 801-805.[CrossRef]
Grimshaw H. M.. 1987. The determination of total phosphorus in soils by acid digestion. In A. P. Rowland [ed.] Chemical analysis in environmental research, 92-95 Institute of Terrestrial Ecology, Abbots Ripton, UK.
Grimshaw H. M. Allen S. E. Parkinson J. A.. 1989. Nutrient elements. In S. E. Allen [ed.] Chemical analysis of ecological materials, 2nd ed 81-159 Blackwell Scientific Publications, Oxford, UK.
Groom P. K. Lamont B. B.. 1999. Which common indices of sclerophylly best reflect differences in leaf structure?. Ecoscience 6: 471-474.[ISI]
Groom P. K. Lamont B. B. Markey A. S.. 1997. Influence of leaf type and plant age on leaf structure and sclerophylly in Hakea (Proteaceae). Australian Journal of Botany 45: 827-838.[CrossRef]
Grubb P. J.. 1986. Sclerophylls, pachyphylls and pycnophylls: the nature and significance of hard leaf surfaces. In B. Juniper and R. Southwood [eds.] Insects and the plant surface 137-150 Edward Arnold, London, UK.
Güsewell S.. 2004. N : P ratios in terrestrial plants: variation and functional significance. New Phytologist 164: 243-266.[CrossRef][ISI]
Gutschick V. P.. 1999. Biotic and abiotic consequences of differences in leaf structure. New Phytologist 143: 3-18.[CrossRef][ISI]
Hagerman A. E. Riedl K. M. Jones G. A. Sovik K. N. Ritchard N. T. Hartzfeld P. W. Riechel T. L.. 1998. High molecular weight plant polyphenolics (tannins) as biological antioxidants. Journal of Agricultural and Food Chemistry 46: 1887-1892.[CrossRef]
Harborne J. B.. 1998. Phytochemical methods: a guide to modern techniques of plant analysis, 3rd ed Chapman & Hall, London, UK.
Harvey P. H. Pagel M. D.. 1991. The comparative method in evolutionary biology Oxford University Press, Oxford, UK.
Jackson A. P.. 1992. Bone, nacre, and other ceramics. In J. F. V. Vincent [ed.] Biomechanics—materials: a practical approach 33-56 Oxford University Press, Oxford, UK.
Jaffré T.. 1980. Étude écologique du peuplement végétal des sols dérivés de roches ultrabasiques en Nouvelle Calédonie. Collection Travaux et Documents de l'ORSTOM no. 124 Paris, France.
Jaffré T.. 1992. Floristic and ecological diversity of the vegetation on ultramafic rocks in New Caledonia. In A. J. M. Baker, J. Proctor, and R. D. Reeves [eds.] The vegetation of ultramafic (serpentine) soils 101-107 Intercept Ltd., Andover, UK.
Jaffré T.. 1993. The relationship between ecological diversity and floristic diversity in New Caledonia. Biodiversity Letters 1: 82-87.[CrossRef]
Jaffré T. Bouchet P. Veillon J.-M.. 1998a. Threatened plants of New Caledonia: is the system of protected areas adequate?. Biodiversity and Conservation 7: 109-135.[CrossRef][ISI]
Jaffré T. Morat P. Veillon J.-M.. 1993. Etude floristique et phytogéographique de la forêt sclérophylle de Nouvelle-Calédonie. Bulletin du Museum National d'Histoire Naturelle, Paris, 4e séries 15, section B, Adansonia 1–4: 107-146.
Jaffré T. Morat P. Veillon J.-M. Rigault F. Dagostini G.. 2001. Composition et caractéristiques de la flore indigène de la Nouvelle-Calédonie. Documents Scientifiques et Techniques II 4 Institut de recherche pour le développement (IRD), Nouméa, New Caledonia.
Jaffré T. Rigault F. Dagostini G.. 1998b. Impact des feux de brousse sur les maquis ligno-herbacés des roches ultramafiques de Nouvelle-Calédonie. Adansonia 20: 173-189.
Jordan G. J. Dillon R. A. Weston P. H.. 2005. Solar radiation as a factor in the evolution of scleromorphic leaf anatomy in Proteaceae. American Journal of Botany 92: 789-796.
Kruckeberg A. R.. 1984. California serpentines: flora, vegetation, geology, soils, and management problems University of California Press, Berkeley, California, USA.
Lamont B. B. Groom P. K. Cowling R. M.. 2002. High leaf mass per area of related species assemblages may reflect low rainfall and carbon isotope discrimination rather than low phosphorus and nitrogen concentrations. Functional Ecology 16: 403-412.[CrossRef][ISI]
Larcher W.. 2005. Climatic constraints drive the evolution of low temperature resistance in woody plants. Journal of Agricultural Meteorology 61: 189-202.[CrossRef]
Latham M.. 1980. Les oxydisols dans quelques milieux insulaires du Pacifique sud: caractérisation, conditions de formation, fertilité, classification. Cahiers Office de la Recherche Scientifique et Technique Outre-Mer (ORSTOM), série Pédologie 18: 305-316.
