Summary1. Understanding how environmental factors drive plant community assembly remains a major challenge in community ecology. The strength of different assembly processes along environmental gradients, such as environmental filtering and functional niche differentiation, can be quantified by analysing trait distributions in communities. While environmental filtering affects species occurrence among communities, functional divergence or convergence is strongly related to species abundances within communities, which few studies have taken into account. We examine the trait-mediated effect of these two processes along a stress-resource gradient. 2. We measured species abundances and the distributions of eight traits related to vegetative and regenerative phases in plant communities along a gradient of soil depth and resource availability in Mediterranean rangelands. We quantified environmental filtering, defined as a local restriction of trait range, and trait divergence, based on abundance-weighted trait variance, using a two-step approach with specifically designed null models. 3. Communities presented a clear functional response to the soil gradient, as evidenced by strong trends in community-weighted trait means. We detected environmental filtering of different traits at both ends of the gradient, suggesting that, contrary to widespread expectations, trait filtering may not necessarily be the result of abiotic filtering under harsh conditions but could likely also result from biotic interactions in productive habitats. 4. We found marked shifts in trait abundance distributions within communities along the gradient. Vegetative traits (e.g. leaf dry matter content) diverged on shallow soils, reflecting the coexistence of distinct water-and nutrient-use strategies in these constrained habitats and converged with increasing soil resource availability. By contrast, regenerative traits (e.g. seed mass) tended to diverge towards deeper soils, while plant reproductive heights diverged all along the gradient. 5. Synthesis: Our study highlights how the combination of abundance data with traits capturing different functional niches is critical to the detection of complex functional responses of plant communities to environmental gradients. We demonstrate that patterns of trait divergence and filtering are strongly contingent on both trait and environment such that there can be no expectation of a simple trend of increasing or decreasing functional divergence along a gradient of resource availability.

1. CC-BY-NC 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/024315 doi: bioRxiv preprint first posted online Aug. 10, 2015; Species interactions, ranging from antagonisms to mutualisms, form the architecture of biodiversity and determine ecosystem functioning. Understanding the rules responsible for who interacts with whom, as well as the functional consequences of these interspecific interactions, is central to predicting community dynamics and stability. Species traitssensu lato may affect different ecological processes determining species interactions through a twostep process. First, ecological and lifehistory traits govern species distributions and abundance, and hence determine species cooccurrence, which is a prerequisite for them to interact. Second, morphological traits between cooccurring potential interaction partners should match for the realization of an interaction. Moreover, inferring functioning from a network of interactions may require the incorporation of interaction efficiency. This efficiency may be also traitmediated, and can depend on the extent of matching, or on morphological, physiological or behavioural traits. It has been shown that both neutral and traitbased models can predict the general structure of networks, but they rarely accurately predict individual interactions, suggesting that these models may be predicting the right structure for the wrong reason. We propose to move away from testing null models with a framework that explicitly models the probability of interaction among individuals given their traits.The proposed models integrate both neutral and traitmatching constraints while using only information about known interactions, thereby overcoming problems originating from undersampling of rare interactions (i.e. missing links). They can easily accommodate qualitative or quantitative data, and can incorporate trait variation within species, such as values that vary along developmental stages or environmental 2 . CC-BY-NC 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/024315 doi: bioRxiv preprint first posted online Aug. 10, 2015; gradients. We use three case studies to show that they can detect strong trait matching (e.g. predatorprey system), relaxed trait matching (e.g. herbivoreplant system) and barrier trait matching (e.g. plantpollinator systems). Only by elucidating which species traits are important in each process, i.e. in determining interaction establishment, frequency, and efficiency, can we advance in explaining how species interact and the consequences for ecosystem functioning.

Summary1. Species interactions, ranging from antagonisms to mutualisms, form the architecture of biodiversity and determine ecosystem functioning. Understanding the rules responsible for who interacts with whom, as well as the functional consequences of these interspecific interactions, is central to predict community dynamics and stability. 2. Species traits sensu lato may affect different ecological processes by determining species interactions through a two-step process. First, ecological and life-history traits govern species distributions and abundance, and hence determine species co-occurrence and the potential for species to interact. Secondly, morphological or physiological traits between co-occurring potential interaction partners should match for the realization of an interaction. Here, we review recent advances on predicting interactions from species co-occurrence and develop a probabilistic model for inferring trait matching. 3. The models proposed here integrate both neutral and trait-matching constraints, while using only information about known interactions, thereby overcoming problems originating from undersampling of rare interactions (i.e. missing links). They can easily accommodate qualitative or quantitative data and can incorporate trait variation within species, such as values that vary along developmental stages or environmental gradients. 4. We use three case studies to show that the proposed models can detect strong trait matching (e.g. predator-prey system), relaxed trait matching (e.g. herbivore-plant system) and barrier trait matching (e.g. plant-pollinator systems). 5. Only by elucidating which species traits are important in each process (i.e. in determining interaction establishment and frequency), we can advance in explaining how species interact and the consequences of these interactions for ecosystem functioning.

Background and aims Since its emergence in the mid‐20th century, invasion biology has matured into a productive research field addressing questions of fundamental and applied importance. Not only has the number of empirical studies increased through time, but also has the number of competing, overlapping and, in some cases, contradictory hypotheses about biological invasions. To make these contradictions and redundancies explicit, and to gain insight into the field’s current theoretical structure, we developed and applied a Delphi approach to create a consensus network of 39 existing invasion hypotheses. Results The resulting network was analysed with a link‐clustering algorithm that revealed five concept clusters (resource availability, biotic interaction, propagule, trait and Darwin’s clusters) representing complementary areas in the theory of invasion biology. The network also displays hypotheses that link two or more clusters, called connecting hypotheses , which are important in determining network structure. The network indicates hypotheses that are logically linked either positively (77 connections of support) or negatively (that is, they contradict each other; 6 connections). Significance The network visually synthesizes how invasion biology’s predominant hypotheses are conceptually related to each other, and thus, reveals an emergent structure – a conceptual map – that can serve as a navigation tool for scholars, practitioners and students, both inside and outside of the field of invasion biology, and guide the development of a more coherent foundation of theory. Additionally, the outlined approach can be more widely applied to create a conceptual map for the larger fields of ecology and biogeography.