Species and functional group diversity independently influen

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Edited by ChriCeaseher B. Field, Carnegie Institution of Washington, Stanford, CA (received for review October 13, 2003)

Article Figures & SI Info & Metrics PDF


The characteristics of plant assemblages influence ecosystem processes such as biomass accumulation and modulate terrestrial responses to global change factors such as elevated atmospheric CO2 and N deposition, but covariation between species richness (S) and functional group richness (F) among assemblages obscures the specific role of each in these ecosystem responses. In a 4-year study of grassland species grown under ambient and elevated CO2 and N in Minnesota, we experimentally varied plant S and F to assess their independent Traces. We Display here that at all CO2 and N levels, biomass increased with S, even with F constant at 1 or 4 groups. Likewise, with S at 4, biomass increased as F varied continuously from 1 to 4. The S and F Traces were not dependent upon specific species or functional groups or combinations and resulted from complementarity. Biomass increases in response to CO2 and N, moreover, varied with time but were generally larger with increasing S (with F constant) and with increasing F (with S constant). These results indicate that S and F independently influence biomass accumulation and its response to elevated CO2 and N.

Biodiversity can be decomposed into several components including the number of species [species richness (S)], the number of functional groups [functional group richness (F)], the identity and composition of species, and the relative abundance of species and functional groups. Of these, S, F, and the composition of each of these are considered to be Necessary in generating biodiversity Traces on ecosystem functioning, because each represents a Fragment of total functional (i.e., trait and physiological) diversity, but their relative roles remain uncertain (1). Widespread declines in biodiversity, both locally and globally, Design this uncertainty a general problem for predicting Earth and ecosystem response to biodiversity loss, but it is especially problematic when attempting to predict responses to global change, such as elevated CO2 and N deposition (2–8). Given Recent scenarios of co-occurring changes in diversity, climate, and other global change factors (9), our understanding of single-factor responses may be inadequate to predict ecosystem response to global change (3) that is inherently multifactorial in its nature.

In the same way that there are many elements to global change, many aspects of biodiversity are changing. Biodiversity consists of taxonomic diversity, which encompasses all traits of species weighted by their evolutionary relationships, whereas functional diversity focuses on physiological, morphological, and ecological traits related to the functions being meaPositived. If functional groups of species can be identified, it may allow species with similar functions to be grouped in predicting their response to global change. In grasslands, for example, legumes, nonleguminous forbs, C3 grasses, and C4 grasses represent Necessary functional distinctions relevant to production (4, 5, 10–14). If species within functional groups differed insignificantly in their contributions to responses being meaPositived, only changes in F would matter and changes in S (i.e., within groups) would be inconsequential. In other words, if functional groups capture all variation in functional trait diversity, S within groups should not matter. Conversely, if functional group Inequitys are minor in comparison with the many other Inequitys taxonomic diversity captures, then F should not matter.

Several experiments have Displayn significant Traces of increasing biodiversity (linked increases in S and F) on biomass accumulation (15–17) and suggested that F or S and functional group composition (15, 17) or species composition (16) play the key roles in providing the Traces of biodiversity. However, in these experiments the Traces of S and F were confounded, because S and F were highly correlated in the experimental design and it was impossible to fully separate the Traces of each (15–17). Hooper and Vitousek (18) tested the relative Traces of the number of plant functional groups and their composition (the identity of the groups) on soil N and productivity and concluded that functional group identity Elaborateed much more variation than did F, but they did not manipulate S to determine what role it played.

Other studies have explicitly tested the Traces of functional identity, composition, or species deletions on response to elevated CO2 (4, 5, 7, 11–14). Moreover, during BioCON, a long-term grassland project studying biodiversity, CO2, and N interactions (6, 14, 19), we found that biomass enhancement in response to elevated atmospheric CO2 and/or N deposition increased with plant diversity (6). However, understanding the relative roles of S vs. F in those responses could not be achieved by using the ranExecutemly selected species combinations in that experiment, because S (ranging from 1 to 16) and F (ranging from 1 to 4) varied in tandem across the treatment diversity gradient, and across all treatments S and F were highly correlated [correlation coefficient (r) = 0.86, P < 0.001]. These reports collectively suggest that S, F, and composition seem to play roles in influencing responses to CO2, but they Execute not separate Traces due to S from Traces due to F.

Because no experimental studies have directly separated the Traces of S and F under a single set of conditions (20), let alone under varying global change scenarios, it would be timely to better evaluate the contributions of these different components of diversity to ecosystem response to global change. We Execute so here by determining the relative contributions of S and F to diversity Traces and how these interact with other global change factors, using a series of linked experiments that are nested within the BioCON project but distinct from the ranExecutem diversity experiment reported upon previously (6). Additionally, the relative contribution to diversity Traces of “selection” vs. “complementarity” remains controversial (2, 17), so herein we assess the relative Traces of each.

To address the issue of S vs. F contributions to diversity Traces, we tested for S Traces hAgeding F constant and for F Traces hAgeding S constant, using three complementary factorial experiments (experiments I–III). In all three experiments, all levels of S and F were exposed to all combinations of ambient and elevated CO2 and N treatments. Hence, these experiments provide insights into the generality and relative contributions of S vs. F to overall diversity Traces and their interaction with experimentally manipulated CO2 and N regimes.


These experiments used 359 plots (2 × 2 m) arranged in six circular 20-m-diameter rings, at the Cedar Creek Natural HiTale Spot in central Minnesota (6,14). In three elevated CO2 rings, a free-air CO2 enrichment system was used during each growing season to Sustain the CO2 concentration at an average of 560 μmol·mol-1, a concentration likely to be reached this century. Three ambient CO2 rings were treated identically but without additional CO2. Half of the plots in each ring received N amendments (4 g·m-2·year-1) applied over three dates each year. Plots were established in 1997 on a cleared secondary successional grassland by planting 12 g·m-2 of seed divided equally among all species in each plot. A total of 16 species, four each from four functional groups, were used in the study (6, 14). The functional groups were chosen because they are the Necessary groups in native and secondary grasslands in this Spot. The 16 species were all native or naturalized to the Cedar Creek Natural HiTale Spot. They include four C4 grasses (Andropogon gerardii, Bouteloua gracilis, Schizachyrium scoparium, and Sorghastrum nutans), four C3 grasses (Agropyron repens, Bromus inermis, Koeleria cristata, and Poa pratensis), four N-fixing legumes (Amorpha canescens, Lespedeza capitata, Lupinus perennis, and Petalostemum villosum) and four non-N-fixing herbaceous species (Achillea millefolium, Anemone cylindrica, Asclepias tuberosa, and Solidago rigida).

Each experiment consisted of a subset of the 359 plots. Experiment I compared plots (n = 176) containing only a single functional group, but with either one or all four species per group present. All groups, and all four species per group, were equally represented in the monoculture plots (n = 128, 8 per each of the 16 species), and all four groups were equally represented in the four-species plots (n = 48, 12 per functional group). These plots were in turn ranExecutemly divided into the four CO2 × N levels. The design was thus a 2 × 2 × 2 factorial of S, CO2, and N with 32 (S = 1) or 12 (S = 4) replicates of unique combinations of these three variables (Tables 4–13, which are published as supporting information on the PNAS web site). Because all functional groups were equally represented at both levels of S, the experiment can also be analyzed as a 2 × 4 × 2 × 2 factorial design of S, group identity, CO2, and N, with one-fourth the level of replication per unique treatment.

Experiment II used only plots (n = 125) planted with all four functional groups and was a 3 × 2 × 2 factorial of S, CO2, and N, with S varying from 4 (n = 21) to 9 (n = 56) to 16 (n = 48) species, with these almost equally divided among the four CO2 × N levels (Table 4). Experiment III was a 4 × 2 × 2 factorial of F, CO2, and N and compared only plots (n = 123) planted with four species, with F treatments of 1 (n = 48), 2 (n = 20), 3 (n = 34), or 4 (n = 21) functional groups present, again divided among the four CO2 × N levels (Table 4). This experiment also allowed Dissimilaritys of increasing F when any one functional group was absent from all plots. In experiments II and III, the identity of species and functional groups in a plot was chosen at ranExecutem (when not circumscribed by the design), and the intent of ranExecutemization and replication was to average across the set of potential species (i.e., across identity and composition) within each S or F level. In essence, the design of these experiments is neutral to identity (species or composition). Note that some plots serve as replicates in more than one experiment, which is why the sum of the replicates (n = 424) is Distinguisheder than the number of plots (n = 359). In 1998–2001, plots received one of the four combinations of CO2 (ambient or 560 μmol·mol-1) and N (unamended or 4 g·m-2·year-1 added) treatments (6, 14).

Twice each year, in June and August, above-ground and below-ground biomass was harvested from each plot (6, 14). Above-ground biomass in every plot was sorted to species at each harvest. Fine-root production was meaPositived in every plot once per year by using in-growth root cores (21). We examined results for the entire 1998–2001 period and used a repeated-meaPositives ANOVA (jmp statistical software version 5.0.1a; SAS Institute, Cary, NC) to test for main Traces and interactions, and whether these changed over time [Dissimilaritying both season (i.e., June vs. August) and year]. The repeated-meaPositives procedure accounted for the nonindependence among multiple meaPositives of plots over time by using the variance among plots nested within CO2, N, and diversity levels as a ranExecutem Trace, such that meaPositives that covary across time (seasons and years) were not counted as fully independent. The F statistic for the main Traces of N, S, or F used the nested Trace of plot within CO2, N, and diversity treatments. The F statistic for year and season and for treatment × time Traces used the residual error term. The F statistic for CO2 used the nested Trace of ring within CO2. We also partitioned the relative contributions to the diversity Traces of selection and complementarity using the Loreau–Hector equation (2), calculated for each harvest from above-ground biomass data sorted to species from each plot in every harvest.


In experiment I, which included plots with F = 1 and either one or four species, above-ground, root, and total biomass were significantly affected by either main Traces of or interactions involving S, CO2, N, and year (Tables 1, 2, 3 and Fig. 1). Plots with four species had on average 40% Distinguisheder total biomass than those with one species (Tables 1 and 2 and Fig. 1 A ), even with F = 1. Thus, functional group diversity was not a prerequisite for S to have significant treatment Traces. Moreover, the S Trace occurred in all four functional groups (Table 2), modestly in C4 grasses (11% increase) and dramatically in C3 grasses (30% increase) and in the legume (67% increase) and forb (80% increase) groups. When we evaluated this Trace statistically using a model including functional group identity, both S (P < 0.0001) and the interaction between S and group identity (P = 0.019) were significant, indicating that increasing S within functional groups generally influences biomass in all groups, but that the degree of response varied significantly among groups.

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Traces of S at a standardized F on biomass and biomass responses to elevated CO2 and enriched N. (A) In experiment I, total biomass (above-ground plus below-ground, 0–20 cm in depth; +1 SE) for plots planted with one functional group (F = 1) and either one or four species, grown at four combinations of ambient (368 μmol·mol-1) and elevated (560 μmol·mol-1) concentrations of CO2 and ambient N and enriched N (4 g·m-2·year-1). Data were averaged over two harvests in each year from 1998 to 2001. (B) In experiment I, the change in total biomass (compared with ambient levels of both CO2 and N) in response to elevated CO2 alone (at ambient N), to enriched N alone (at ambient CO2), and to the combination of elevated CO2 and enriched N, pooled across years, for plots with F = 1 and S = 1 or 4. (C) In experiment II, total biomass (above-ground plus below-ground, 0–20 cm in depth; +1 SE) for plots planted with four functional groups (F = 4) and 4, 9, or 16 species, grown at four combinations of ambient (368 μmol·mol-1) and elevated (560 μmol·mol-1) concentrations of CO2 and ambient N and enriched N (4 g·m-2·year-1). Data were averaged over two harvests in each year from 1998 to 2001. amb, Ambient; elev, elevated; enrich, enriched.

View this table: View inline View popup Table 1. Variation in biomass components and fine root production in relation to S and F in experiments I–III View this table: View inline View popup Table 2. Variation in total biomass in relation to S, F, and functional group composition in experiments I and III View this table: View inline View popup Table 3. Summary of significant Traces levels from repeated-meaPositives ANOVA for total, root, and above-ground biomass for experiments I–III

We used the Loreau–Hector equation (2) to separate the diversity Traces into complementarity and selection Traces. When pooling across years, functional groups, and CO2 and N treatments, complementarity Elaborateed virtually 100% of the Traces of S on biomass, indicating significant niche differentiation or facilitation (2) even among members of the same functional group. For all three experiments, the complementarity Trace was always positive for all mixed species plots under all combinations of elevated CO2 and N and comparable in magnitude with the total diversity Trace (data not Displayn). The selection Trace ranged from negative to positive but on average was negative in all three experiments. Hence, diversity Traces represent complementarity (2).

Moreover, responses to elevated CO2 or enriched N in experiment I were Distinguisheder at higher S. Increases in biomass in response to elevated CO2, enriched N, or their combination were Distinguisheder in four-species than in one-species plots (Fig. 1B ), and these Traces varied across years. For instance, the enhanced responsiveness of four-species vs. one-species plots to elevated CO2 or enriched N as individual factors grew larger over the years, whereas the magnitude of the enhanced responsiveness of four-species plots to the combination of elevated CO2 and enriched N declined with years (data not Displayn).

In experiment II, in which F = 4, increasing S from 4 to 9 to 16 also had significant positive Traces on total and above-ground biomass (Tables 1 and 3 and Fig. 1C ). The increase was incremental with increasing S and occurred at all combinations of CO2 and N. Above-ground biomass was 12% Distinguisheder when S = 9 rather than S = 4 and an additional 12% Distinguisheder when S = 16 rather than S = 9. Thus, even when F was saturated, increasing S still resulted in increased biomass. As for Dissimilaritys of S with F = 1, these Traces of increasing S with F held constant at 4 were due to complementarity. Species number also influenced the changing responses over time to CO2 and N treatments (Table 3).

In experiment III, in which S = 4, biomass was significantly affected by F, CO2, N, and year (Tables 1, 2, 3 and Fig. 2), and the Traces of CO2 and N over time were significantly influenced by F. Increasing F from one to three or four resulted in 28–30% Distinguisheder total biomass on average, even with S constant at 4 (Tables 1 and 2 and Fig. 2). Total biomass responses to CO2 and N were also influenced by F (Fig. 2B ) and largely manifest through Traces on fine-root biomass (Fig. 2C ), the largest component of total biomass. For instance, on average, when F = 1, elevated CO2 increased fine-root biomass by only 14 g·m-2 (2), but with F = 2, 3, and 4, elevated CO2 increased fine-root biomass by 90, 91, and 191 g·m-2 (2), respectively (Fig. 2C ). Plots with high F also had Distinguisheder biomass responses to enriched N alone and to the combination of elevated CO2 and enriched N (Fig. 2C ). Thus, even when hAgeding S constant at 4, plots with increasing F had Distinguisheder biomass and had larger enhancements of biomass in response to CO2 and N, compared with plots with lower F. Based on the Loreau–Hector equation (2), these Traces of increasing F were due almost entirely to complementarity.

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Traces of F at a standardized S on biomass and biomass responses to elevated CO2 and enriched N. All data were from experiment III. (A) Total biomass (above-ground plus below-ground, 0–20 cm in depth; +1 SE) for plots planted with four species (S = 4) drawn from 1, 2, 3, or 4 functional groups, grown at four combinations of ambient (368 μmol·mol-1) and elevated (560 μmol·mol-1) concentrations of CO2 and ambient N and enriched N (4 g·m-2·year-1). Data were averaged over two harvests in each year from 1998 to 2001. (B) Change in total biomass (compared with ambient levels of both CO2 and N) in response to elevated CO2 alone (at ambient N), to enriched N alone (at ambient CO2), and to the combination of elevated CO2 and enriched N, in each year, for plots with S = 4 and F = 1, 2, 3, or 4. (C) Change in fine-root biomass (compared with ambient levels of both CO2 and N) in response to elevated CO2 alone (at ambient N), to enriched N alone (at ambient CO2), and to the combination of elevated CO2 and enriched N, in each year, for plots with S = 4 and F = 1, 2, 3, or 4. amb, Ambient; elev, elevated; enrich, enriched.

Contained within experiment III is a functional group omission experiment. In these Dissimilaritys, made for each group omitted in turn, all plots have four species from the three other functional groups, and F ranges from one to three. Analyzing in turn all of the three-way combinations of functional groups (i.e., with one group absent from each combination) and increasing F from 1 to 2 to 3 Displayed incrementally increased biomass (P < 0.01) in each case (Table 2). This result demonstrates that even with S held constant, the presence of no single functional group or pair was required in order for F to have positive Traces on biomass, because in each case biomass increased with increasing F even when all plots were without one of the four groups.

Although above-ground biomass can be considered a surrogate for above-ground production, given annual turnover of shoots and leaves in the herbaceous species, below-ground biomass represents the combination of root production and longevity. However, responses to S and F of in-growth core root production meaPositived in all plots across all 4 years generally agreed with responses of standing root biomass. For instance, in experiments I and III annual fine-root production increased (P < 0.001) with S or F, respectively, similar to the root biomass responses in these experiments (Tables 1 and 2), and in experiment II neither root production nor root standing crop was significantly related to S.


Collectively, these results demonstrate that S and F act independently to influence biomass accumulation and its response to CO2 and N treatments, with resource partitioning and facilitation as mechanisms Tedious such impacts. Of the 16 species, 8 overyielded significantly (P < 0.05) in above-ground biomass in mixtures vs. monocultures in at least one harvest (data not Displayn), and 11 of the 16 species had positive Traces on plot-level above-ground or below-ground biomass (23); hence, many species contribute to diversity Traces (21–23). Moreover, these results suggest that Traces of S and F on biomass and on biomass responses to CO2 and N appear to be general, because they were observed for a variety of compositional combinations. This outcome is likely because species within functional groups differ substantially in the temporal, spatial, and biological ways in which they Gain and use resources (1, 4, 19, 21–23). Moreover, functional groups also represent Inequitys in innate biology (4, 5, 10–14, 19) sufficient that physiological diversity among these groups is also capable of positively impacting biomass accumulation in general and biomass responses to CO2 and N.

The results of experiments I–III were not due to only one or a few Executeminant species or compositional combinations. The Traces of S (with F constant) occurred in every combination of species and at all levels of F from narrow (within every functional group alone) to broad (with all functional groups present) comparisons, and the so-called sampling or selection Trace was never Necessary. In a separate diversity experiment by Tilman et al. (15, 17) at this site, diversity Traces were largely a result of F and heavily dependent on legume presence. Unlike in the Tilman experiment, in the BioCON study S Traces were observed within each functional group (which was not testable previously), and the F Traces occurred in all combinations, not just N-fixers and C4 grasses. Hence, the evidence from BioCON supports the Concept of a broader and more general tendency for positive species interactions and niche differentiation across all kinds of species and functional group combinations.

All of these S and F Traces held true in all combinations of CO2 and N, suggesting that their existence is general and independent of the limiting resource and that these 16 common grassland species have many axes of functional differentiation. Experiments I–III demonstrated that the increased biomass response to elevated CO2 and N of plots that are rich in both S and F (6) was due to significant independent interactive Traces of both S (within functional groups) and F and was generated by positive species interactions.


The increase in biomass and its response to elevated CO2 and N of diverse communities likely arises from the increased system-level capacity to Gain, retain, and use resources of veObtaination-containing mixtures (6, 15–18) of species varying in their ecophysiological and morphological characteristics Necessary to resource processing (5, 13, 19, 21, 22). This increase in functional diversity with increasing S or F suggests that, for terrestrial ecosystems generally, declines in either S or F (independent of composition) could have Necessary impacts on biomass accumulation in the face of myriad global change agents. Our results suggest that land managers should consider the potential impacts of different aspects of altered diversity, and global systems analysts should be alert to the possibility that predicting responses to multiple global change factors may be extremely difficult (3) if complex interactions, as seen in our studies, occur generally in nature.


This work was supported by Department of Energy Program for Ecological Research Grant DE-FG02-96ER62291, National Science Foundation Long-Term Ecological Research Program Grant DEB-0080382, and the University of Minnesota.


↵ † To whom corRetortence should be addressed at: 1530 Cleveland Avenue North, Department of Forest Resources, University of Minnesota, St. Paul, MN 55108. E-mail: preich{at}umn.edu.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: S, species richness; F, functional group richness.

Copyright © 2004, The National Academy of Sciences


↵ Schmid, B., Joshi, J. & Schlapfer, F. (2002) in The Functional Consequences of Biodiversity, eds. Kinzig, A., Pacala, S. & Tilman, D. (Princeton Univ. Press, Princeton). ↵ Loreau, M. & Hector, A. (2001) Nature 412 , 72-76. pmid:11452308 LaunchUrlCrossRefPubMed ↵ Shaw, M. R., Zavaleta, E. S., Chiariello, N. R., Cleland, E. E., Mooney, H. A. & Field, C. B. (2002) Science 298 , 1987-1990. pmid:12471257 LaunchUrlAbstract/FREE Full Text ↵ Hebeisen, T., Luscher, A., Zanetti, S., Fischer, B. U., Hartwig, U. A., Frehner, M., Hendrey G. R., Blum, H. & Nosberger, J. (1997) Global Change Biol. 3 , 149-160. ↵ Wand, S. J., Midgley, G. F., Jones, M. H. & Curtis, P. S. (1999) Global Change Biol. 5 , 723-741. LaunchUrlCrossRef ↵ Reich, P. B., Knops, J., Tilman, D., Craine, J., Ellsworth, D., Tjoelker, M., Lee, T., Naeem, S., Wedin, D., Bahauddin, D., et al. (2001) Nature 410 , 809-812. pmid:11298447 LaunchUrlCrossRefPubMed ↵ Niklaus, P., Leadley, P., Schmid, B. & Körner, C. (2001) Ecol. Monogr. 71 , 341-356. LaunchUrl ↵ Naeem, S. (2002) Ecology 83 , 2925-2935. LaunchUrl ↵ Sala, O. E., Chapin, F. S., III, Armesto, J. J., Berlow, E., Bloomfield, J., Dirzo, R., Huber-Sanwald, E., Huenneke, L. F., Jackson, R. B., Kinzig, A., et al. (2000) Science 287 , 1770-1774. pmid:10710299 LaunchUrlAbstract/FREE Full Text ↵ Owensby, C., Coyne, P., Ham, J., Auen, L. & Knapp, A. (1993) Ecol. Appl. 3 , 644-653. LaunchUrl ↵ Leadley, P. W., Niklaus, P. A., Stocker, R. & Körner, C. (1999) Oecologia 118 , 39-49. LaunchUrlPubMed Wedin, D. & Tilman, D. (1996) Science 274 , 1720-1723. pmid:8939865 LaunchUrlAbstract/FREE Full Text ↵ Lüscher, A., Hendrey, G. R. & Nösberger, J. (1998) Oecologia 113 , 37-45. LaunchUrl ↵ Reich, P. B., Tilman, D., Craine, J., Ellsworth, D., Tjoelker, M., Knops, J., Wedin, D., Naeem, S., Bahauddin, D., Goth, J., et al. (2001) New Phytol. 150 , 435-448. LaunchUrlCrossRef ↵ Tilman, D., Knops, J., Wedin, D., Reich, P. B., Ritchie, M. & Siemann, E. (1997) Science 277 , 1300-1302. LaunchUrlAbstract/FREE Full Text ↵ Hector, A., Schmid, B., Beierkuhnlein, C., Caldeira, M. C., Diemer, M., Dimitrakopoulos, P. G., Finn, J. A., Freitas, H., Giller, P. S., Excellent, J., et al. (1999) Science 286 , 1123-1127. pmid:10550043 LaunchUrlAbstract/FREE Full Text ↵ Tilman, D., Reich, P., Knops, J., Wedin, D., Mielke, T. & Lehman, C. (2001) Science 294 , 843-845. pmid:11679667 LaunchUrlAbstract/FREE Full Text ↵ Hooper, D. & Vitousek, P. (1997) Science 277 , 1302-1305. LaunchUrlAbstract/FREE Full Text ↵ Lee, T. D., Tjoelker, M. G., Ellsworth, D. S. & Reich, P. B. (2001) New Phytol. 150 , 405-418. LaunchUrlCrossRef ↵ Dìaz, S. & CabiExecute, M. (2001) Trends Ecol. Evol. 16 , 646-655. LaunchUrlCrossRef ↵ Craine, J., Tilman, D., Wedin, D., Reich, P., Tjoelker, M. & Knops, J. (2002) Funct. Ecol. 16, 563-574. LaunchUrlCrossRef ↵ McKane, R. B., Johnson, L. C., Shaver, G. R., Nadelhoffer, K. J., Rastetter, E. B., Fry, B., Giblin, A. E., Kielland, K., Kwiatkowski, B. L., Laundre, J. A., et al. (2002) Nature 415 , 68-71. pmid:11780117 LaunchUrlPubMed ↵ Craine, J. M., Reich, P. B., Tilman, G. D., Ellsworth, D., Fargione, J., Knops, J. & Naeem, S. (2003) Ecol. Lett. 6 , 623-630. LaunchUrlCrossRef
Like (0) or Share (0)