Random processes and phylogenetic loss caused by plant invasions.
Aim: Although biological invasions represent a major cause of biodiversity loss, the actual mechanisms driving species extinctions remain insufficiently understood. Here we investigate the role of three processes as drivers of phylogenetic loss in invaded local plant communities, namely the 'biotic resistance', 'environmental filtering' and 'functional equivalence' hypotheses. Location: Balearic Islands (western Mediterranean). Methods: We quantified the phylogenetic diversity and structure of 109 pairs of invaded and non-invaded local plant communities from two Mediterranean islands. Each pair contained one control plot and one plot invaded either by the deciduous tree Ailanthus altissima, the succulent subshrubs Carpobrotus spp. or the pseudoannual geophyte Oxalis pes-caprae. We combined generalized linear models, analyses of phylogenetic community structure and generalized linear mixed models using a Markov chain Monte Carlo technique (MCMCglmm) to contrast the 'biotic resistance', 'environmental filtering' and 'functional equivalence' hypotheses. Results: While the phylogenetic structure of the non-invaded communities was not more clustered or overdispersed than expected by chance, minimum phylogenetic distance to the invasive species increased in invaded assemblages, in which the magnitude of phylogenetic diversity loss ranged from 6 to 37% depending on the invader's identity. Invader or island identity did not explain the probabilities of native species becoming locally extinct. Rather, the likelihood of extinction was mainly explained by species abundance, with scarcer species exhibiting a higher chance of becoming locally extinct. Species identity explained a small fraction of the variation in extinction risk (12%), independently of each species' evolutionary history. Main conclusions: The most relevant driver of local extinction is a stochastic process where less abundant species tend to disappear more frequently irrespective of their evolutionary history. This has strong implications for conservation because it suggests that in the study region the invaders are unlikely to drive regional and global extinctions except in cases where the native species is already rare.