Plant species abundance and phylogeny explain the structure of recruitment networks

Abstract

Summary Established plants can affect the recruitment of young plants, filtering out some and allowing the recruitment of others, with profound effects on plant community dynamics. Recruitment networks (RNs) depict which species recruit under which others. We investigated whether species abundance and phylogenetic distance explain the structure of RNs across communities. We estimated the frequency of canopy–recruit interactions among woody plants in 10 forest assemblages to describe their RNs. For each RN, we determined the functional form (linear, power or exponential) best describing the relationship of interaction frequency with three predictors: canopy species abundance, recruit species abundance and phylogenetic distance. We fitted models with all combinations of predictor variables, from which we simulated RNs. The best functional form of each predictor was the same in most communities (linear for canopy species abundance, power for recruit species abundance and exponential for phylogenetic distance). The model including all predictor variables was consistently the best in explaining interaction frequency and showed the best performance in predicting RN structure. Our results suggest that mechanisms related to species abundance are necessary but insufficient to explain the assembly of RNs. Evolutionary processes affecting phylogenetic divergence are critical determinants of RN structure.

Phenotypic structure of plant facilitation networks

“…(2019) who found that the probability of germination of seeds from 10 species sown under 5 different nurses increased with the trait distance between the nurse and the facilitated species. Altogether, these results validate the premise and support the prediction made by previous studies using phylogenetic as a proxy of phenotypic distances between nurses and facilitated plants (Alcántara et al ., 2018, 2019; Marcilio‐Silva et al 2015; Valiente‐Banuet and Verdú, 2008; Verdú et al ., 2010).…”

“…Paradoxically, more than a decade after setting the initial assumption of divergent traits favouring facilitation and despite the vast amount of literature looking at trait patterns in facilitation (Butterfield and Callaway, 2013; Le Bagousse‐Pinguet et al 2015; Liancourt and Tielbörger, 2011; Michalet et al ., 2011; Schöb et al ., 2013; Soliveres et al ., 2014), phenotypic distances between species have never been used to explain the structure of plant facilitation networks. Instead, phylogenetic distances – taken as a proxy of trait divergence between species – have become the common procedure to test it (Alcántara et al ., 2018, 2019; Marcilio‐Silva et al 2015; Valiente‐Banuet and Verdú, 2008; Verdú et al ., 2010). Indeed, little empirical research addressing the link between traits and facilitation networks at the community level exists despite the broad consensus that predicting community and ecosystem processes from species traits is a ‘Holy Grail’ in ecology (Lavorel and Garnier, 2002; Funk et al ., 2017).…”

Plant–plant facilitation increases with reduced phylogenetic relatedness along an elevation gradient

“…It should be noted that phylogenetically clustered communities could be produced when competitive ability is evolutionarily conserved and therefore fitness differences instead of niche differences are the main community assembly mechanism (Mayfield and Levine 2007). However, facilitative interactions have been shown to be mainly mediated by niche differences (Valiente‐Banuet and Verdú 2007, Butterfield and Briggs 2011, Alcántara et al 2019a, Navarro‐Cano et al 2019). Including the phylogenetic specificity between interacting plants at the community level could thus shed light on SGH in natural communities.…”

Unifying facilitation and recruitment networks

“…Although the implications of these patterns for community dynamics are yet to be explored explicitly, they agree with two properties of ecological networks known to confer them stability: (a) that networks containing many weak and few strong interactions are more stable (McCann, Hastings, & Huxel, ; Wootton & Stouffer, ); and (b) that the coexistence of species that differ in fitness requires stronger intra‐ than interspecific limitation of population growth (Chesson, ). At least in the RNs analysed by Verdú and Valiente‐Banuet () and Alcántara, Garrido, and Rey () there is evidence suggesting that the second property may actually be acting in these communities, creating a community compensatory trend whereby more abundant species have lower rates of recruitment than rarer ones (Comita et al., ; Connell, Tracey, & Webb, ; Soliveres et al., ).…”