Abstract
BackgroundDespite numerous tests of Darwin’s naturalization hypothesis (DNH) evidence for its support or rejection is still contradictory. We tested a DNH derived prediction stating that nonnative species (NNS) without native congeneric relatives (NCR) will spread to a greater number of localities than species with close relatives in the new range. This test controlled the effect of residence time (Rt) on the spread of NNS and used naturalized species beyond their lag phase to avoid the effect of stochastic events in the establishment and the lag phases that could obscure the NCR effects on NNS.MethodsWe compared the number of localities (spread) occupied by NNS with and without NCR using 13,977 herbarium records for 305 NNS of weeds. We regressed the number of localities occupied by NNS versus Rt to determine the effect of time on the spread of NNS. Then, we selected the species with Rt greater than the expected span of the lag phase, whose residuals were above and below the regression confidence limits; these NNS were classified as widespread (those occupying more localities than expected by Rt) and limited-spread (those occupying fewer localities than expected). These sets were again subclassified into two groups: NNS with and without NCR at the genus level. The number of NNS with and without NCR was compared using χ2 tests and Spearman correlations between the residuals and the number of relatives. Then, we grouped the NNS using 34 biological attributes and five usages to identify the groups’ possible associations with spread and to test DNH. To identify species groups, we performed a nonmetric multidimensional scaling (NMDS) analysis and evaluated the influences of the number of relatives, localities, herbarium specimens, Rt, and residuals of regression. The Spearman correlation and the Mann–Whitney U test were used to determine if the DNH prediction was met. Additionally, we used the clustering objects on subsets of attributes (COSA) method to identify possible syndromes (sets of biological attributes and usages) associated to four groups of NNS useful to test DNH (those with and without NCR and those in more and fewer localities than expected by Rt).ResultsResidence time explained 33% of the variation in localities occupied by nonnative trees and shrubs and 46% of the variation for herbs and subshrubs. The residuals of the regression for NNS were not associated with the number or presence of NCR. In each of the NMDS groups, the number of localities occupied by NNS with and without NCR did not significantly differ. The COSA analysis detected that only NNS with NCR in more and fewer localities than expected share biological attributes and usages, but they differ in their relative importance.DiscussionOur results suggest that DNH does not explain the spread of naturalized species in a highly heterogeneous country. Thus, the presence of NCR is not a useful characteristic in risk analyses for naturalized NNS.
Highlights
After we selected species that have overcome their lag phase, we considered the nonnative species (NNS) occupying more or fewer localities than expected based on their residence time, which was calculated using a regression between the number of localities and residence time
We obtained a significant regression model (R2 = 0.4601, P < 2.2e−16) between the number of localities occupied by nonnative species (N ) and their adjusted residence time (Rt ) (N = −2.549 + 1.8213 ∗ Rt )
Our results indicate that the presence of native congeneric relatives (NCR) in a highly heterogeneous country does not hinder the spread of naturalized plant species; these results agree with the general trend found for regions showing that Darwin’s naturalization hypothesis (DNH) was not supported at the regional scale (Ma et al, 2016)
Summary
Darwin’s naturalization hypothesis (DNH), known as the ‘‘phylogenetic repulsion’’ hypothesis (Mack, 1996b; Daehler, 2001; Duncan & Williams, 2002; Strauss, Lau & Carroll, 2006; Proches et al, 2008), proposes that the phylogenetic closeness between two species impedes their coexistence within the same community because they share similar resource requirements, resulting in intense competition between them; in addition, the natural enemies of the native species would attack the nonnative, again hindering species coexistence.DNH has been tested several times, the evidence for its support (Mack, 1996a; Rejmánek, 1996b; Rejmánek, 1999; Ricciardi & Atkinson, 2004; Strauss, Webb & Salamin, 2006; Schaefer et al, 2011; Bezeng et al, 2015) or rejection (Duncan & Williams, 2002; Lambdon & Hulme, 2006; Küster et al, 2008; Diez et al, 2009; Park & Potter, 2013; Li et al, 2015; Park & Potter, 2015) is still contradictory. A high likelihood of becoming invasive among NNS with close NCR was confirmed in the Asteraceae family in countries with Mediterranean climates (Park & Potter, 2015) These opposing results for Poaceae and Asteraceae might be explained if the functionality of DNH depends on taxonomic groups and/or species groups with shared functional attributes. We tested a DNH derived prediction stating that nonnative species (NNS) without native congeneric relatives (NCR) will spread to a greater number of localities than species with close relatives in the new range. We selected the species with Rt greater than the expected span of the lag phase, whose residuals were above and below the regression confidence limits; these NNS were classified as widespread (those occupying more localities than expected by Rt ) and limited-spread (those occupying fewer localities than expected) These sets were again subclassified into two groups: NNS with and without NCR at the genus level.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
