Generality Of Leaf Trait Relationships: A Test Across Six Biomes

Ecology, Sept, 1999 by Peter B. Reich, David S. Ellsworth, Michael B. Walters, James M. Vose, Charles Gresham, John C. Violin, William D. Bowman

INTRODUCTION

The idea that disparate plant species from different biomes may share similar leaf form and function has a long history in ecology, but has rarely been tested in a quantitative manner (Beadle 1966, Chabot and Hicks 1982). It is clear that leaf life span, specific leaf area (SLA), nitrogen (N) concentration, leaf diffusive conductance ([G.sub.s]), and photosynthetic rate are fundamental plant traits that vary greatly among species, often by orders of magnitude (Field and Mooney 1986, Reich et al. 1992). This variation may represent adaptation to environmental heterogeneity that exists both locally and globally. It is well established that these leaf traits are generally correlated, based on examination of variation among and within species (e.g., Lugg and Sinclair 1981, Field and Mooney 1986, K6rner 1989, Reich et al. 1992), but a comprehensive, quantitative evaluation of these relationships among biomes is lacking. If combinations of leaf traits recur in distantly related taxa across a wide biogeographic range it would suggest that selection constrains the collective leaf traits possessed by every species (Reich et al. 1992, Reich 1993) and provide evidence for convergent evolution.

Studies from a variety of species and ecosystems, among plant groups taxonomically either broad (Field and Mooney 1986, Evans 1989, Reich et al. 1991, 1992, Reich 1993, Abrams et al. 1994, Mulkey et al. 1995) or narrow (Chazdon and Field 1987, Williams et al. 1989) have generally reported similar patterns of relationships among leaf traits. Mass-based photosynthetic capacity ([A.sub.mass]) and leaf nitrogen ([N.sub.mass] are usually positively correlated, both are positively correlated with SLA, and all three traits decline with increasing [TABULAR DATA FOR TABLE 1 OMITTED] leaf life span. We have theorized that these relationships are universal at the broadest, continental to global scales (Reich et al. 1992, Reich 1993) because of the functional interdependency among traits, plus the constraints placed by biophysics and natural selection that lead to ecological trade-offs (Coley et al. 1985, Coley 1988, Field and Mooney 1986, Reich et al. 1992, Mulkey et al. 1995).

Thus, we hypothesize that interspecific proportional scaling relationships (e.g., log Y = a b log X) among leaf traits will be quantitatively similar among diverse biomes, supporting the idea of convergent evolution. Alternatively, leaf trait correlations may occur but vary among vegetation types or along climatic or edaphic gradients. To address these issues we quantified leaf gas exchange rates, SLA, leaf N, and leaf life span, and their relationships for 10-43 species within each of six sites representing different biomes in the Americas. Our sites represent a range of biomes (Table 1) that vary in growing season length, mean air temperature, elevation, water availability, and soil fertility. A brief summary paper (submitted after the original submission of this manuscript) based in part on these same data highlights the idea that the slopes of interspecific trait relationships were similar among biomes and among data sets (Reich et al. 1997). In this paper we take the opportunity to (1) present the biome-specific data, (2) explicitly contrast slopes among sites, (3) compare the elevations of these scaling equations, (4) contrast relationships among functional groups, and (5) make multivariate analyses of species leaf traits; collectively these represent the main objectives of this paper and were beyond the scope of the previous summary publication.

METHODS

Sites were selected to provide a wide range of environmental conditions and terrestrial ecosystem types (Table 1). Two sites (in Colorado and Wisconsin) were located at the ecotone between biomes and include species common to both biomes. At several sites, measurements were made in more than one ecosystem type or study area. The study site in Colorado was located at Niwot Ridge in the Front Range of the Rocky Mountains. Study plots were located in both wet and dry meadow tundra communities at 3510 m elevation, and in open subalpine forest-alpine meadow transition at 3200 m. The species studied include common conifers, hardwood shrubs, and herbs. Soils at the site were largely coarse-textured Inceptisols. The main study area in Wisconsin was the University of Wisconsin Arboretum, in Madison, Wisconsin (275-m elevation). Study plots were located in a mosaic of natural and restored ecosystems, including forest, hedgerow, savanna and tall grass prairie. Soils were largely mediumtextured silt loams (Alfisols). We also examined species in a cedar-tamarack swamp and adjacent bog at Cedarburg Bog, at the University of Wisconsin-Milwaukee Field Station in Saukville, Wisconsin, [approximately]130 km northeast of Madison. The climate in southern Wisconsin is humid continental, with cold winters and warm summers. Species studied included a number of common prairie and forest understory forbs, deciduous hardwood and coniferous forest tree species and broadleafed evergreen bog shrubs.