Can limiting similarity increase invasion resistance? A meta‐analysis of experimental studies

Synthesis We used meta‐analyses to examine experimental evidence that functional similarity between invaders and resident communities reduces invasion. We synthesized evidence from studies that experimentally added seed to resident communities in which the functional group composition had been manipulated. We found communities containing functionally similar resident species reduced invasion of forb but not grass invaders. However, experimental design dramatically influenced the results – with evidence for limiting similarity only found in artificially assembled communities, and not when studies used functional group removal from more ‘natural communities’. We suggest that functional group similarity plays a limited role in biotic resistance in established communities. The principle of limiting similarity suggests that species must be functionally different to coexist; based on the assumption that inter‐specific competition should be greatest between functionally similar species. There has been controversy over the generality of this assembly rule for plant communities with some studies finding evidence for limiting similarity and others not. One approach to testing this is to examine the ‘invasion’ success of species into communities in which the functional group composition has been manipulated. Using a meta‐analysis approach, we examined the generality of limiting similarity for plant communities based on published experimental studies. We asked – is establishment of an invading species less successful if it belongs to a functional group that is already present in the community compared to a community in which that functional group is absent? We explored separately colonisation (i.e. germination, establishment or seedling survival) and performance (i.e. biomass, cover or growth) of different functional groups (forbs and grasses) and experimental designs (removal experiments of more or less natural communities and synthetic‐assemblage experiments). We found that communities containing functionally similar resident species did reduce invader colonisation and performance of forb invaders, but did not reduce colonisation or performance of grass invaders. Evidence in support of limiting similarity was only detected in synthetic‐assemblage experiments and not when studies used functional group removal from ‘natural’ communities. Functional similarity is an important aspect of biotic resistance for forb invaders, but was only found in artificial communities. This has implications for restoration ecology especially when communities are built de novo. However, we suggest that limiting similarity plays a limited role in biotic resistance, because no evidence was detected in established communities.