Results and Discussion
Our analyses do not support that the probability of observing the signature of environmental filtering and competition is scale-dependent (Park et al. 2020). We found that the phylogenetic distance of exotic species to the closest coexisting native relative was, on average, three times larger at the local community level than at the continental level (Fig. 2A ), consistent with previous studies (Park et al. 2020). Yet a similar pattern was found when we instead estimated the minimum phylogenetic distance between each exotic species and the native species from communities where those exotic species are absent (Fig. 2B ). Thus, it is unlikely that the pattern is a signature of environmental filtering. A more plausible alternative is that the increase in species richness at larger spatial scales increases the probability of including a closely related exotic and native species (Fig. 2C ).
Given the risk that results are biased by sampling artifacts, we opted by addressing the conundrum at lower spatial scales and ask whether the success of the invader in a community is influenced by the degree of phylogenetic relatedness with native neighbors. Under the naturalization hypothesis, for example, we should expect that within any region exotic species will be restricted to those communities where they do not find close relatives. Instead, the preadaptation hypothesis predicts a preference for communities with close relatives. As some exotic species were absent from certain communities despite being well-established in near-by communities (505 out of 1,359 cases), we tested whether the probability that an exotic species occurred in a community was related to the phylogenetic distance with the native species from the community. Despite previous studies suggesting that the naturalization hypothesis should be favored at smaller spatial scales, we found no evidence that the presence of close relatives competitively exclude exotic species. On the contrary, we found that exotic species were more likely to be present in communities containing native species that are on average more closely related (Fig. 3A, B ).
One possible explanation for the failure to support the naturalization hypothesis is that native species may reduce the success of exotic species but rarely prevent invasion (Levine et al. 2004). Thus, the relative abundance of an exotic species in the community may be a more informative measure of invasion success to address the naturalization hypothesis. Based on estimations of exotic species abundance in communities where they were established (N = 792), we again found no evidence that exotic species more distantly related to native residents reach higher relative abundances than exotic species with closer relatives (Fig 3C, D ). Rather, exotic species abundances marginally decreased with the mean and nearest phylogenetic distance to native species. Although most native species in the studied communities had low relative abundances (Sol et al. 2020b), and thus may be unlikely to inhibit invasions even if the exotic species have similar resource needs, the conclusion does not change when using mean and nearest phylogenetic distances weighted by their relative abundance (Fig 3E-H ).
The lack of support for the naturalization hypothesis may reflect that in birds competition plays a minor role in influencing invasion success, relative to other intrinsic or environmental factors (Duncan et al. 2003). One reason is that exotic species tend to attain higher success in disturbed habitats, mostly urbanized habitats (Sol et al. 2020b), where competitive regimes have frequently been altered by the extinction of native species (Bartomeus et al. 2012). These human-altered habitats often impose strong environmental filtering, as only a few species have the adaptations needed to live in artificial environments frequently disturbed by human activities (Sol et al.2014b). Adding the interaction between phylogenetic distance and degree of habitat alteration improves the fit of previous models of invasion success (Table S1 ). When considering this interaction, we find support for the preadaptation hypothesis, with a negative relationship between invasion success and phylogenetic distance but only for human-altered environments, such as urbanized habitats (Fig. 4 ). Importantly, this pattern is not confounded by species richness (Tables S2-S3 ) because urbanized environments tend to contain fewer species than the surrounding non-urbanized environments (Solet al. 2020b).
While our results support the pre-adaptation hypothesis, they are not exempt of potential biases. An issue that may make the patterns difficult to interpret is that the subset of species that have been introduced and successfully established in regions outside their native range is a non-random sample of the entire avian phylogeny, a possibility well-documented in the literature (Duncan et al.2003). The clustering of introduced species within the avian phylogeny should increase the probability that a native species randomly selected from the avian phylogeny belongs to a different clade. This effect may be further accentuated by the fact that introduced species often come from distant regions. For example, if we compare the phylogenetic distances of pairs of species randomly sampled within and across the main bioregions involved as donors and recipients of invaders, we find that by chance the average distances across bioregions are higher than within regions for a same sample size (Fig 2D, E ). To better interpret the results, we assessed to what extent phylogenetic patterns reflect ecological differences.
The notion that phylogenetic patterns reflect ecological differences is supported by two lines of evidence. First, we found that the main axes of niche variation across all species—measured in terms of morphology, dietary use and foraging behavior (see methods)— showed clear evidence of phylogenetic effects (Fig. S1) , supporting the existence of substantial niche conservatism. A Mantel test further confirmed that phylogenetic distance co-varied with niche differentiation (P = 0.001), although the analysis also showed that it only accounted for 31% of variation in the niche. Second, similar to phylogenetic distance, we found a negative relationship between niche distance and invasion success measured in terms of abundance (Fig. 4; Table S4 ). The relationship between invasion success and niche distance appears to primarily occur in human-altered environments (Fig. 4K,L,O,P ).
Instead, we found a different pattern when we examined the probability that the invader occurred in a community within a given region. In communities occupying less human-altered areas, this probability tended to increase with functional distance, suggesting that invaders prefer to settle in communities that lack species with similar niches (Fig. 4I-N, Table S5 ). The alternative that the pattern results from extinction of close-relatives by the invader is less likely because invaders rarely achieved higher densities in little disturbed environments. Moreover, native species are expected to be more competitive than invaders in these environments because they have had more opportunities to adapt (Sax & Brown 2000; Sol et al.2012a).
The discrepancy between patterns may indicate that functional and phylogenetic distances contain different information. Phylogenetic distances among species reflect differences in the niche, but they also contain important information on traits related to the ability of the species to persist in novel environments (Fig. 5) . While niche availability is arguably essential to proliferate in a new environment, response traits like behavioral plasticity and a future returns life history strategy can also be important in novel environments to buffer individuals against the demographic consequences of maladaptation (Solet al. 2012b). As response traits are particularly relevant to live in human-altered environments (Sol et al. 2014a), it might be that phylogenetic distance better captures differences among species that are relevant for the pre-adaptation hypothesis than for the naturalization hypothesis. This finding further highlights the importance of using functional traits to interpret phylogenetic patterns. It also suggests that the accuracy of predictive models should increase when including functional information, a possibility supported by our data (Table S1 ).
Our comprehensive analysis of phylogenetic relatedness between invaders and natives yielded no support for naturalization hypothesis. Previous works suggest that the failure to support the hypothesis reflects flaws in the analyses, like the adoption of inappropriate spatial scales or the use of taxonomic instead of phylogenetic information (Thuilleret al. 2010; Li et al. 2015). However, we analyzed well-surveyed communities within a phylogenetic framework, and considered possible sampling artifacts and a number of confounds such as the body size of the invader and the relative abundance of native close relatives. We also considered the possibility that native competitors do not exclude invaders but merely reduce their abundance. Despite all these improvements, we failed to find evidence that nonindigenous organisms are more successful in communities in which their close relatives are absent. While the role of competition in invasions remains debatable, our results fit better with growing evidence that, at least in birds (Duncan et al. 2003; Sol et al. 2012a), competition only plays a minor role in explaining invasion success. However, our finding that the probability that the invader occurred in a community increased with functional distance may suggest that invaders tend to avoid natural communities where they find species with similar niche requirements. While human-altered environments are novel for both invaders and native species, in natural environments the opportunity for adaptation is substantially higher for native species. These communities should thus exert stronger biotic resistance, which may in part explain why exotic species tend to avoid them (Lovell et al. 2021). However, our analyses indicate that the signatures of these effects are difficult to detect with phylogenetic relatedness.
Our analyses provide more credence for the pre-adaptation hypothesis, showing that phylogenetic distance is a good surrogate for both niche and functional response traits and that lower functional distances are also associated with enhanced invasion success. This results aligns with growing evidence showing that invasion success in birds critically depends on the possession of appropriate adaptations to flourish in human-altered communities, where biotic resistance has been reduced by the extinction of specialized competitors (Sol et al. 2012b). Because these adaptations are phylogenetically conserved (Sol et al. 2014a, 2017a, 2020b), it follows that the existence of phylogenetically close species in the communities can inform us on the likelihood that an invader establishes itself and spread in a new region. Our findings thus extend those of Redding et al. (Reddinget al. 2019) by showing that the presence of closely related species in a region not only increases the probability of establishment but also the abundance a species attains, a key metric of invasion success and a factor determining the impact of exotic species. This provides some of the clearest evidence to date that the presence of close relatives enhances —rather than inhibits— biological invasions. It remains to be shown whether this conclusion can be extrapolated to organisms with less ability than birds to sample habitats and select those that best match their phenotypes.
We therefore conclude that phylogenies can improve our ability to accurately predict the likelihood that a species proliferates when introduced to a new region. However, phylogenetic information is only useful to make predictions when considering the environmental context in which invasions take place, particularly the degree of human-related disturbances. Moreover, phylogenetic information should be combined with functional data to accurately predict invasiveness, as phylogenetic distances are not always an accurate surrogate of ecological distances. The rapid development of new statistical techniques to integrate phylogenetic and phenotypic information (Penone et al. 2014) provides new, largely unexplored opportunities to improve our ability to predict and prevent the success and impact of biological invasions.