One of the most ubiquitous patterns in plant ecology is species loss following nutrient enrichment. A common explanation for this universal pattern is an increase in the size asymmetry of light partitioning (the degree to which large plants receive more light per unit biomass than smaller plants), which accelerates the rates of competitive exclusions. This 'light asymmetry hypothesis' has been confirmed by mathematical models, but has never been tested in natural communities due to the lack of appropriate methodology for measuring the size asymmetry of light partitioning in natural communities. Here, we use a novel approach for quantifying the asymmetry of light competition which is based on measurements of the vertical distribution of light below the canopy. Using our approach, we demonstrate that an increase in light asymmetry is the main mechanism behind the negative effect of nutrient enrichment on species richness. Our results provide a possible explanation for one of the main sources of contemporary species loss in terrestrial plant communities.

Summary Plant communities show two general responses to gradients of soil resources: a decrease in species richness at high levels of resource availability and an associated shift in species composition from small and slow‐growing species to large and fast‐growing species. Models attempting to explain these responses have usually focused on a single pattern and provided contradicting predictions concerning the underlying mechanisms. We use an extension of Tilman's resource competition model to investigate the hypothesis that both patterns may originate from the size‐asymmetric nature of light exploitation by competing plants. The only mechanism producing changes in species richness and species composition in our model is mortality due to competition. Under the framework of the model, asymmetric light exploitation is a necessary and sufficient condition to obtain the empirically observed responses of species richness and species composition to soil resource gradients. This theoretical result is robust to relaxing the simplifying assumptions of the model. Our model shows that the traits enhancing competitive superiority depend on the mode of resource exploitation: under symmetric exploitation, competitive superiority is achieved by tolerance of low resource levels, while under asymmetric exploitation, it is achieved by the ability to grow fast and attain a large size. This result indicates that a long‐standing debate concerning the traits that enhance competitive superiority in plant communities (the ‘Grime–Tilman debate’) can be reduced into a single parameter of our model – the degree of asymmetry in resource competition. The model also explains the observed shift from below‐ground to above‐ground competition with increasing productivity, the associated increase in the asymmetry of competitive interactions and the increasing likelihood of competitive exclusion under high levels of productivity. None of these patterns could be obtained under symmetric competition in our model. Synthesis. The ability of the model to explain a wide range of observed patterns and the robustness of these predictions to its simplifying assumptions suggest that the size asymmetry of competition for light is a fundamental factor in determining the structure and diversity of plant communities.

Aim Changes in global climate and land use are expected to alter water and nutrient availability. Various meta‐analyses and large‐scale experiments show that increasing nutrient availability is expected to decrease the diversity of ecological communities, but so far, no study has attempted to provide a global‐scale perspective of diversity responses to water manipulation. Location Global. Methods We conducted a meta‐analysis focusing on the effects of water and nutrient additions both on species richness and on biomass of herbaceous plant communities. We identified 41 water addition experiments, of which 19 experiments manipulated both water and nutrients. Results Although both water and nutrient additions increased biomass (by c. 15 and 34%, respectively), only the latter consistently decreased richness (by c. 23%). Biomass responses to water addition were mainly derived from an increase in forb biomass (by c. 37%), whereas corresponding responses to nutrient addition were derived from an increase in graminoid biomass (by c. 56%). Addition of both water and nutrients led to larger biomass responses compared with the addition of each resource alone (by c. 69%), but the negative effect on species richness was similar to nitrogen addition alone. None of these responses could be explained by general (resource‐independent) theories, such as the productivity–diversity hypothesis or the niche dimension hypothesis. Main conclusions While highlighting overlooked patterns, this meta‐analysis reveals a fundamental knowledge gap in our ability to predict biodiversity responses to global change and demonstrates that future theories attempting to explain and predict such changes must take into account the potential implications of resource‐specific and functional group‐specific responses.

Significance Nutrient enrichment of natural ecosystems is a primary characteristic of the Anthropocene and a known cause of biodiversity loss, particularly in grasslands. In a global meta-analysis of 630 resource addition experiments, we conduct a simultaneous test of the three most prominent explanations of this phenomenon. Our results conclusively indicate that nitrogen is the leading cause of species loss. This result is important because of the increase in nitrogen deposition and the frequent use of nitrogen-based fertilizers worldwide. Our findings provide global-scale, experimental evidence that minimizing nitrogen inputs to ecological systems may help to conserve the diversity of grassland ecosystems.

A fundamental notion in community ecology is that local species diversity reflects some balance between the contrasting forces of competitive exclusion and competitive release. Quantifying this balance is not trivial, and requires data on the magnitude of both processes in the same system, as well as appropriate methodology to integrate and interpret such data. Here we present a novel framework for empirical studies of the balance between competitive exclusion and competitive release and demonstrate its applicability using data from a Mediterranean annual grassland where grazing is a major mechanism of competitive release. Empirical data on the balance between competitive exclusion and competitive release are crucial for understanding observed patterns of variation in local species diversity and the proposed approach provides a simple framework for the collection, interpretation, and synthesis of such data.

Summary One of the most widely documented patterns in plant ecology is the decrease in species diversity following nutrient enrichment. A long‐standing explanation for this diversity decline is an increase in the relative importance of size‐asymmetric light competition which accelerates the rate of competitive exclusion (the ‘light asymmetry hypothesis’). Recently, an alternative hypothesis has been proposed which attributes the negative effect of nutrient enrichment on species diversity to a reduction in the number of limiting resources (i.e. a reduced niche ‘dimensionality’). A recent global‐scale experiment demonstrating that increasing the number of added resources leads to a decrease in species diversity was interpreted as a support for this ‘niche dimension hypothesis’. Here we highlight a number of theoretical considerations that question this interpretation and demonstrate that a deeper analysis of the new global‐scale dataset provides a stronger support for the light asymmetry hypothesis.

1. Ecological theory predicts that the soil seed bank stabilizes the composition of annual plant communities in the face of environmental variability. However, longterm data on the community dynamics in the seed bank and the standing vegetation are needed to test this prediction. 2. We tested the hypothesis that the composition of the seed bank undergoes lower temporal variability than the standing vegetation in a 9-year study in Mediterranean, semi-arid and arid ecosystems. The composition of the seed bank was estimated by collecting soil cores from the studied sites on an annual basis. Seedling emergence under optimal watering conditions was measured in each soil core for three consecutive years, to account for seed dormancy. 3. In all sites, the composition of the seed bank differed from the vegetation throughout the years. Small-seeded and dormant-seeded species had a higher frequency in the seed bank than in the standing vegetation. In contrast, functional group membership (grasses vs. forbs) did not explain differences in species frequency between the seed bank and the vegetation after controlling for differences between grasses and forbs in seed mass and seed dormancy. 4. Contrary to predictions, the magnitude of year-to-year variability (the mean compositional dissimilarity between consecutive years) was not lower in the seed bank than in the vegetation in all sites. However, long-term compositional trends in the seed bank were weaker than in the vegetation in the Mediterranean and semiarid sites. In the arid site where year-to-year variability was highest, no long-term trends were observed. 5. Synthesis. The effect of the seed bank on the temporal variability of the vegetation in annual communities depends on site conditions and timescale. While the yearto-year variability of the seed bank is similar to the vegetation, the soil seed bank can buffer long-term trends.

The enormous variation in seed mass along gradients of soil resources has fascinated plant ecologists for decades. However, so far, this research has focused on the description of such variation, rather than its underlying mechanisms. Here we experimentally test a recent model relating such variation to two fundamental properties of plant growth: allometry of biomass growth and size‐asymmetry of light competition. According to the model, mean seed mass should increase, and the variance of seed mass should show a unimodal response, to increasing soil resource availability (productivity). We test these predictions and their underlying assumptions using a combination of field observations, mesocosm experiments and greenhouse experiments focusing on Mediterranean annual plants. Our results support the predictions and assumptions of the model, and allow us to reject alternative models of seed mass variation. We conclude that growth‐allometry and size‐asymmetric light competition are key drivers of seed‐mass variation along soil resource gradients.