= Solution
\b[Variation in <plant> strategies helps explain why different <species> contribute different functions, and why a diverse community can sustain performance across conditions. The evidence must connect <phenotypic trait> differences to demographic consequences and then to measured ecosystem processes.] A <life-history trade-off> is a limitation on jointly improving components such as growth, survival and reproduction under specified conditions. A <plant functional trait> is a measurable characteristic affecting or indicating performance; it is not necessarily a direct measure of lifetime <genotype fitness>.
Resource allocation supplies a mechanistic starting point. If an individual acquires an amount $A$ and allocates fraction $a$ to reproduction and the remainder to growth, then $R=aA$ and $G=(1-a)A$. At fixed $A$, increasing one allocation decreases the other. But individuals that acquire more can grow more and reproduce more simultaneously, so a positive observed growth–reproduction correlation does not disprove an allocation constraint. Conversely, a negative correlation can be caused by different environments rather than a genetically fixed constraint. Common-garden comparisons, manipulated resource supply and experimental alteration of reproductive demand can distinguish acquisition, allocation and environmental effects. Lifetime costs should include later survival and reproduction, rather than only current mass.
One strong comparative pattern is the <leaf economics spectrum>. Relatively inexpensive, short-lived <leaves> tend to have high mass-based <photosynthesis> and nutrient concentrations, whereas costly, persistent <leaves> provide slower returns over longer periods. <specific leaf area> and <leaf mass per area> quantify opposite sides of the area–construction-mass relationship. A large global dataset found coordinated variation in <leaf> construction, physiology and longevity across habitats, while also showing overlap among broad <plant> types.<The worldwide leaf economics spectrum by Ian J. Wright and others>[Nature 428, pp. 821–827; global leaf-trait relationships and overlap among functional types]{ref} This supports a recurring acquisition–conservation axis, not a rule that every high-photosynthesis <species> must grow fastest in every habitat. Mass-based and area-based measurements must be distinguished, and leaf-level rates cannot by themselves determine whole-plant performance.
Demographic evidence is equally important. Comparisons of tropical trees link differences in growth and mortality to <phenotypic traits> including <wood density>, but a single <leaf> or wood measurement does not explain every <species>' strategy.<Functional traits and the growth-mortality trade-off in tropical trees by S. Joseph Wright and others>[Ecology 91, pp. 3664–3674; trait associations with demographic variation in a tropical forest]{ref} Low construction costs can permit rapid acquisition in favorable conditions while providing less persistence under damage or resource scarcity. Direct demographic measurements are needed because a <phenotypic trait> may be correlated with several different mechanisms. Likewise, at a fixed reproductive budget, a large <seed> can carry more reserves but fewer <seeds> can be produced; its possible establishment advantage must be weighed against dispersal and the number of sites reached.
There is no single universal fast–slow line. <plant> height, <seed> size, <leaf> construction, belowground deployment and phenology can vary along partly independent dimensions. A global analysis of <plant> form and function identified broad size-related and leaf-economics dimensions rather than one <phenotypic trait> that orders all <plant> strategies.<The global spectrum of plant form and function by Sandra Díaz and others>[Nature 529, pp. 167–171; multidimensional variation in major plant traits]{ref} Patterns also vary within <species> through <phenotypic plasticity> and inherited differences. Closely related <species> are not independent experimental replicates, and environmental sorting can generate <phenotypic trait> correlations. Analyses should incorporate shared ancestry, sampling across environments and uncertainty in the relation between an easily measured <phenotypic trait> and the underlying physiological mechanism.
These differences can affect coexistence. A <species> efficient at acquiring a limiting resource when abundant may be less able to persist when it is scarce, and <species> can differ in rooting depth, season of activity or response to disturbance. Such differences can reduce direct competition and allow complementary resource use. But a trade-off alone does not establish stable coexistence: the relevant population dynamics must allow each <species> to recover when rare under the proposed mechanism. <niche differentiation> can stabilize coexistence, while large differences in competitive ability may still exclude a weaker <species>. <phenotypic trait> variation is therefore a source of testable mechanisms, not automatic proof that greater diversity must arise.
At the ecosystem level, <phenotypic trait> effects can operate through dominance and through interactions. The <mass-ratio hypothesis> emphasizes that abundant <species>' <phenotypic traits> strongly influence a community's processes, so high productivity may reflect dominance by a productive <species>. <complementarity effects> can arise when <species> use resources differently or facilitate one another, increasing their combined performance. A <species richness> experiment can separate an arithmetic selection component from an average relative-yield component. Let $M_i>0$ be <species> $i$'s monoculture yield under comparable conditions, $p_i$ its initial mixture proportion, $Y_i$ its mixture yield and $\Delta RY_i=Y_i/M_i-p_i$, with $\sum_i p_i=1$. Then
$$
\Delta Y=\sum_iY_i-\sum_ip_iM_i
=N\,\overline{\Delta RY}\,\overline M+N\,\operatorname{Cov}(\Delta RY,M),
$$
where the <covariance> uses denominator $N$. The first term is the <complementarity effect>; the second is the <selection effect in biodiversity experiments>.<Partitioning selection and complementarity in biodiversity experiments by Michel Loreau and Andy Hector>[Nature 412, pp. 72–76; additive partition of net biodiversity effects]{ref} A positive first term does not uniquely identify belowground resource partitioning: facilitation, changed enemies or other processes can also raise relative yield. Nor does this “selection effect” mean evolutionary <natural selection>.
A compelling experiment manipulates <species richness> and composition in replicated, randomized plots, holds initial total density or planting design comparable, and includes the appropriate monocultures. Multiple mixtures at each <species richness> help separate <species> identity from <species richness>. Measuring soil resources, seasonal production and root deployment can then test the proposed interaction, rather than merely report an effect on final standing biomass. <transgressive overyielding>, in which a mixture exceeds its best monoculture comparator, is stronger evidence against a simple best-species sampling explanation, although it does not itself identify the physiological cause. Field diversity–productivity correlations alone can be confounded by soil fertility, management or reverse causation.
<phenotypic trait> variation also matters through time. <species> with different responses to drought, temperature or seasonal conditions need not fluctuate together. For total community biomass $B=\sum_iB_i$,
$$
\operatorname{Var}(B)=\sum_i\operatorname{Var}(B_i)+2\sum_{i<j}\operatorname{Cov}(B_i,B_j).
$$
Less synchronized fluctuations can stabilize total production even if individual populations remain variable. A decade-long grassland <species richness> experiment found greater community-level productivity and temporal stability in more diverse treatments.<Biodiversity and ecosystem stability in a decade-long grassland experiment by David Tilman and others>[Nature 441, pp. 629–632; manipulated prairie diversity and temporal variation in productivity]{ref} A stability metric such as $\mu_B/\sigma_B$ can increase through a higher mean as well as a lower variability, so these contributions should be examined separately. <response diversity> is especially relevant when <species> with similar effects on a function differ in their vulnerability to change.
Finally, one function cannot stand for all ecosystem functioning. Rapid biomass production, long-term <carbon storage>, nutrient retention, decomposition and resistance to disturbance can depend on different <phenotypic traits> or involve conflicts. <ecosystem multifunctionality> can therefore require <species> that contribute differently across functions, places and years; apparent redundancy for one measured function need not persist under a new climate or when another function is considered. \b[Trade-off evidence supports a mechanistic explanation for the roles of <biodiversity>, but the inference is strongest when trait–performance relationships, causal diversity manipulations and the particular ecosystem functions are all measured.]
Back to article page
