- Top of page
- Materials and Methods
The identification of recurrent patterns of specialization in plants and the reduction of the enormous diversity of the natural world into a smaller number of categories have long been major foci of interest in comparative plant ecology (Grime et al., 1997; Westoby, 1998). These have converged into the need to identify a small set of key plant traits. These traits should give maximum information on plant growth and resource-use strategy and, at the same time, should be simple enough to measure, so that they can be recorded for large numbers of species (Díaz & Cabido, 1997; Hodgson et al., 1999; Weiher et al., 1999). One of the most widely accepted of such key traits is specific leaf area (SLA), the light-catching area deployed per unit of previously photosynthesized dry mass allocated to the purpose (Westoby, 1998). SLA has been proven to be strongly linked to relative growth rate and resource use (Garnier, 1992; Lambers & Poorter, 1992; Reich, 1993; Garnier & Laurent, 1994; Grime et al., 1997; Poorter & Van der Werf, 1998; Wilson et al., 1999). Grime et al. (1997) reported that SLA was one of the major contributors to an axis of resource capture, usage and availability.
Variation in SLA depends on changes in leaf tissue density – or leaf water content (LWC), which is closely correlated with tissue density Garnier & Laurent (1994) – and leaf thickness (LT) (Witkowski & Lamont, 1991; Garnier & Laurent, 1994; Shipley, 1995; Cunningham et al., 1999; Pyankov et al., 1999; Wilson et al., 1999). In cool-temperate predominantly herbaceous datasets, the lower SLA of slow-growing species tends to be related more to lower LWC than to higher LT (Dijkstra & Lambers, 1989; Garnier & Laurent, 1994; van Arendonk & Poorter, 1994; Shipley, 1995; Ryser & Aeschlimann, 1999). In datasets dominated by woody perennials, LT has been found to be equally influential (Witkowski & Lamont, 1991; Cunningham et al., 1999; Wright & Cannon, 2001). The leaves of many slow-growing species have thick epidermal walls and cuticle, abundant sclerification, high cell wall/cytoplasm ratio in tissues, and a high ratio of crude fibre to protein (Loveless, 1961, 1962; Fahn, 1982). In these leaves (sclerophyllous leaves) low SLA is accompanied by low LWC. By contrast, the leaves of fast-growing species (tender leaves) have few cell walls per unit leaf area and a high proportion of their volume is occupied by nitrogen-rich, photosynthetically active mesophyll protoplast. These traits are likely to favour carbon assimilation in fast-growing species (Reich, 1993; Garnier & Laurent, 1994). Shifts from tender to sclerophyllous leaves have been reported along regional productivity gradients. For example, along nutrient and water availability gradients in south-east Australia, SLA and LWC decreased and LT increased with decreasing resource availability (Cunningham et al., 1999).
Wilson et al. (1999) have strongly advocated the use of LWC as an indicator of position on an axis of resource use because it is well correlated with SLA, shows less variability between samples, is simpler to measure and does not depend on LT, whose links with plant resource-use strategy are complex. Furthermore, in the British flora, LWC was a better predictor than SLA of position on an independently derived resource-use axis (Hodgson et al., 1999). However, Wilson et al. (1999) based their preference for LWC on the study of a flora with very few succulents, and in which sclerophylly tends to be the most common adaptation to the main source of stress, soil nutrient deficiency. The authors explicitly warn that their conclusions are valid for the flora of Western Europe, but may not necessarily apply to other floras, especially those of arid and semiarid areas in which succulents are common.
Succulents are plants that have thick water-storing tissues in their main photosynthetic organs, so that they can avoid desiccation when the soil is dry. Although some of the most conspicuous families are phylogenetically close (e.g. Cactaceae, Chenopodiaceae, Aizoaceae and Portulacaceae are all in the Order Caryophyllales), succulence appears in several distant clades (Gibson, 1996; The Angiosperm Phylogeny Group, 1998). They are most common in semiarid and arid ecosystems of warm regions (Gibson, 1996; Mabberley, 1997). Most succulent species are perennials, have water contents of 90% or greater in a fully hydrated organ, and have crassulacean acid metabolism as the main photosynthetic pathway (Gibson, 1996). Accordingly, their photosynthetic organs consist of chlorenchyma with large, rounded cells that have a large vacuolar storage space for carboxylic acids and water (Larcher, 1995). The presence of such conspicuous vacuoles and large intercellular air spaces results in comparatively few chloroplasts per surface area (Kluge & Ting, 1978). In many succulents, photosynthetic organs also contain an internal, nonphotosynthetic parenchyma, with high capacity for water storage (Gibson, 1996). Consequently, succulent species are expected to have relatively low SLA (obviously high LT) but high LWC. The use of LWC as a predictor of plant functioning is based on the assumption that LWC is linked to leaf nitrogen content and assimilatory capacity (Reich, 1993; Garnier & Laurent, 1994; Wilson et al., 1999). That may be a wrong assumption in the case of succulents. Interestingly, variations of LWC between slow-growing and fast-growing species have been documented for floras in which succulence is absent or very rare. The analysis of leaf trait variation along a water availability gradient, considering a flora in which succulence is common, thus provides a good test for the universality of LWC as an indicator of plant resource-use strategy.
In this article, we aimed to: analyse the associations among SLA, LWC and LT in the flora of central-western Argentina, in which succulent species are common; to determine which of LWC, SLA or LT has better indicator value for general plant resource-use strategy in that region, and which of these appears to have greater potential for large-scale comparative screening programmes; and to compare the relationships between SLA and LWC in central-western Argentina with those in other regions of the world in which succulents are uncommon (Britain, France, Italy, south-east Canada, south-east Australia and Sri Lanka).
- Top of page
- Materials and Methods
In previous studies, slow-growing species with preferential allocation to storage and defence have shown low SLA associated with low water content. This tendency towards increased sclerophylly is common in nutrient-poor soils (Beadle, 1966; Monk, 1966; Small, 1973; Grime et al., 1997) and also in drought-prone environments where succulents are rare (Cunningham et al., 1999). In regions such as central-western Argentina, where the main proximate cause of low productivity is water deficiency, both sclerophylly and succulence are viable solutions, although short-lived tender-leafed species can also thrive if water deficiency is temporarily ameliorated by seasonal or sporadic precipitation. Our results also confirm the idea (Turner, 1994; Fonseca et al., 2000) that, although the three leaf types represent well-defined constellations of traits, readily recognized by field botanists, there are transitional forms, apparent in Fig. 3 (e.g. succulents with very tough epidermis, such as Agave americana, or tender-leafed plants with rather thick leaves with a high water content, such as Eryngium agavifolium and Carduus thoermeri). Thus, objective definitions of the three types cannot be devised. Indeed, a discriminant analysis of our species, based on LT, LWC and SLA, was unable to reliably separate the three types in every case (results not shown).
Succulence and sclerophylly are different ways of dealing with low water availability, with succulents being more dependent on water pulses (Schwinning & Ehleringer, 2001, and references therein). However, they represent converging strategies in terms of carbon assimilation and nutrient conservation: both succulence and sclerophylly are related to preferential allocation to storage and defence, rather than to new growth (Díaz & Cabido, 1997). In the flora of central-western Argentina, both sclerophyllous and succulent species showed low SLA, but this was accompanied by a low LWC in sclerophyllous species and by a high LWC in succulent species. In the latter, high LWC did not directly reflect high content of photosynthetically active cytoplasm, since most of the water contained in the chlorenchyma is vacuolar water (Kluge & Ting, 1978; Larcher, 1995; Gibson, 1996). Water content was high both in high-SLA tender-leafed species (lower end of resource-use strategy axis) and in low-SLA succulents (higher end of the axis). As a consequence, SLA but not LWC was well correlated with the resource-use strategy axis described by Díaz & Cabido (1997). As pointed out by Wilson et al. (1999) and confirmed by Garnier et al. (2001a), SLA measurements tend to be less reproducible and more difficult to perform than those of LWC. In the British flora, LWC was the best predictor of position on an independently derived resource-use axis (Grime et al., 1997), either alone (for dicots) or combined with SLA (for graminoids; Hodgson et al., 1999). However, our results suggest that SLA, which shows no monotonic association with LWC, appears as a better predictor of a species resource-use strategy than LWC in floras that contain succulent species.
In our data set, LT showed a clear pattern of association with SLA and the resource-use axis. Low-SLA plants, with differential allocation to storage and defence, and more typical of resource-poor habitats (sclerophyllous and succulent plants) tended to have thicker leaves than high-SLA plants with preferential allocation to photosynthesis and growth, more typical of resource-rich habitats (tender-leafed plants). This is in agreement with Cunningham et al. (1999), Meziane & Shipley (1999) and Roderick et al. (2000a). It can be argued, therefore, that in the case of the Argentine dataset, LT could be as useful as SLA as an indicator of plant resource-use strategy. However, Wilson et al. (1999) have suggested a nonmonotonic relationship between LT and plant resource-use strategy in the Northern European flora. The LT can vary for reasons related more to light availability than the use of soil resources (Meziane & Shipley, 1999; Wilson et al., 1999; Roderick et al., 2000a), and both fast-growing plants from fertile habitats and slow-growing plants from shaded habitats can have thin leaves. This was not observed in our case, probably because soil resources represent a much stronger limitation to plant growth than the existence of a dense canopy. This suggests that the usefulness of LT as an indicator of resource use can vary from region to region, and for different ecological reasons. The SLA thus appears as a trait with more direct ecological interpretation in comparative studies. The LT may prove useful at a finer-scale analysis, for example in the distinction between succulence and sclerophylly among low-SLA plants.
Our results did not match the patterns expected on the basis of the generic model linking SLA and LWC developed by Roderick et al. (1999, 2000a,b). This is not surprising, since that model was derived from an empirical dataset containing no succulents. When succulents were excluded from the analysis, there was a good association between SLA and LWC, in accordance with Wilson et al. (1999). This association was similar to those reported for species sets from the UK, Mediterranean region of southern France, Italy, south-east Australia, south-east Canada and Sri Lanka (Fig. 4b), in which succulents are absent or hardly present. The association between SLA and LWC seems to be particularly strong and consistent in floras where the main cause of stress is soil nutrient content, and in which tender-leafed species are common. The strength of the association decreases in floras more dominated by sclerophyllous species (e.g. south-east Australia; Fig. 4a), and disappears in floras with succulents (central-western Argentina; Fig. 4b). The results excluding succulents, both for Argentina and for other floras, matched reasonably well the model proposed by Roderick et al. (1999, 2000a,b).
The Argentine species set presented here encompasses the widest range of values reported to date along the sclerophylly-succulence axis. At the same time, its range of SLA is small and biased toward low SLA values. This situation may not be exceptional, and floras from other arid to semiarid, relatively warm, systems, such as those in Africa, North America, and the Middle East may show similar patterns. This suggests that, although SLA may not always be the best indicator of plant resource-use strategy (e.g. in cool temperate climates), it may be more widely applicable than LWC or LT. Also, the difficulties involved in the measurement of SLA may be compensated by the fact that it is more directly relevant to carbon assimilation and nutrient conservation than LWC or LT (Garnier et al., 2001a). Therefore, among the three leaf traits analysed in this study, SLA appears to be the best candidate for inclusion in large screening programs oriented to regional to global-scale comparisons.