Introduction
Anthropogenic inputs of nutrients, including nitrogen (N) and phosphorus (P), into the biosphere have greatly increased in recent decades and continue to rise (Sinha et al., 2017). This environmental eutrophication represents a major threat to biodiversity in many terrestrial, freshwater and marine ecosystems worldwide, as it is usually associated with biodiversity loss (Borer et al., 2014, Ren et al., 2017). In grasslands, nutrient enrichment, both deliberate (agricultural fertilization) and unintentional (atmospheric deposition), has been shown to have profound impacts on ecosystems (Erisman et al., 2008). Nutrient input usually increases primary productivity and reduces plant diversity and community stability (Midolo et al., 2018, Soons et al., 2017). This loss of plant diversity can then impact the functioning of ecosystems and their associated ecosystem services (Hautier et al., 2015, HautierIsbell et al., 2018, Hautier et al., 2014, Hector et al., 2010, Isbell et al., 2015). However, we do not have a complete understanding of the mechanisms by which nutrient inputs lead to the loss of plant diversity (Harpole et al., 2017) or the timing during the growing season when these mechanisms are most important.
In low-fertility grasslands, where soil resources are strongly limiting, diversity is often high. However, resource competition theory (R*theory) predicts dominance by the single species that can deplete soil resources to the lowest level (with the lowest value of R *) (Tilman, 1982, Tilman, 1980). We must therefore assume that low-fertility grasslands are either limited by more than one belowground resource (Fay et al., 2015, Hutchinson, 1957), or that additional mechanisms operate, such as negative soil feedbacks, that introduce frequency-dependence and hence stabilisation (Petermann et al., 2008). Coexistence might be made easier in such systems because competition for belowground resources is often assumed to be size-symmetric (Vojtech et al., 2007, HautierVojtech et al., 2018), thus leading to relatively small fitness differences between species, which can be offset by weak niche differentiation (Chesson, 2000).
Under fertilized conditions, when nutrient limitation is alleviated and light becomes the limiting resource, resource competition theory (I * theory) again predicts competitive dominance, this time by the species that is able to intercept light and reduce it to the lowest level (Dybzinski and Tilman, 2007, Vojtech et al., 2007). Because light is a directionally supplied resource, tall species can intercept and pre-empt light, making it unavailable to low-growing species. Competition for light is likely to be highly size-asymmetric and might therefore lead to very large fitness differences and hence the exclusion of smaller, slow-growing species (Hautier et al., 2009, DeMalach et al., 2017, Borer et al., 2014) even if the same stabilising niche differences still operate.
While direct measurements of mechanistic plant competition are extremely difficult, relative growth rate (RGR) is relatively easy to measure, and many plant species show striking differences in their relative growth rate, even when grown under similar environmental conditions (Grime and Hunt 1975). High RGR might confer a strong competitive advantage under highly fertile conditions, because it enables a species to quickly capture light and deny it to competitors. But under low-fertility conditions, we might expect high RGR to be a much poorer predictor of competitive outcomes, as other traits, reflecting niche differences, may play a greater role. The timing of growth might also be a key factor in determining competitive outcomes. For example, a species growing faster during the early stage of the growing season might reduce light availability and thus have a disproportionate competitive advantage relative to species that initially grow more slowly. RGR can be measured at different time points and thus be used to identify when during the growing season differences in RGR are particularly important.
We used two studies to test whether early differences in species growth rates better predict short-term competitive dominance under fertilised conditions: (1) a common garden experiment where species were grown in monoculture and in pairwise and five-species mixtures under low and high soil fertility and (2) an experiment in a natural grassland community that also included fertilizer treatments. Critically, both studies provide detailed measurements of aboveground biomass through the growing season. We focus mainly on competitive outcomes in fertilized conditions, where we expect competition to be primarily for light, hence species with high early-season RGR in monoculture should dominate mixtures. We contrast the fertile situation with less productive conditions but because the outcome of competition may be slower, the comparison is limited by the short-term nature of our study.