1 | INTRODUCTION
The transport of species beyond their native ranges by human actions is breaking down biogeographical barriers and causing global reorganization of biota (Capinha et al. 2015, van Kleunen et al. 2015), with the ensuing invasions posing a serious threat to biodiversity, agriculture and human health (Simberloff et al. 2013). Successful invaders must disperse into a geographically distant area, establish a viable and fertile population, and spread throughout this new environment, where the biotic and abiotic pressures may differ from those they faced in their native range (Kolar and Lodge 2001). This invasion pathway occurs despite the reduction of genetic diversity that typically follows introductions of invasive species, which is usually associated with inbreeding costs and a loss of adaptive potential. For these reasons, invasions are often seen as paradoxical, since invaders are able to overcome these costs to become ecologically dominant in their novel environment, outcompeting native species adapted to local ecological conditions (Sax and Brown 2000, Facon et al. 2006).
Several life-history traits of invaders may favor them over the course of their invasion. Specific breeding systems, modes of dispersal or physiological characteristics may influence their ability to spread and their establishment success, which thus enhance their colonization rate. Exploring invasion mechanisms relies on unraveling whether these traits differ between introduced and native populations. These differences may result from evolutionary events occurring in the invaded population during the initial phase of introduction (Wares et al. 2005), evolving after the introduction due to new ecological pressures (Keller and Taylor 2008), or may already be present within native populations, thereby pre-adapting the source population for invasion success. Therefore, determining the source population of invasive species is critical to conduct comparative studies of life-history traits between introduced and native ranges to understand how they evolved under distinct biotic and abiotic pressures.
Exploring invasion mechanisms also requires assessment of the invasion history, in which a series of demographic events may influence the invasion process and patterns of genetic diversity. A simple invasion history may be the result of a single introduction from the native range; however, an invasive population may also stem from multiple introductions out of the native range, either from the same or different source populations. Similarly, distinct invasive populations may stem from different introduction events from one source population, or from different source populations from the native range. Finally, a successful invasive population itself may become a source for subsequent invasions -a phenomenon coined the ‘bridgehead effect’(Lombaert et al. 2010, Bertelsmeier and Keller 2018). Therefore, distinct invasion histories have different outcomes in terms of patterns of genetic diversity and life-history trait evolution. The bottleneck event following an introduction usually results in a loss of genetic diversity in the introduced population (Dlugosch and Parker 2008), but the amount of genetic diversity lost may vary under different invasion histories. The reduction of genetic diversity may be limited when the initial colonizing force is large, when the introduced population is subsequently re-invaded by additional individuals during multiple introduction events, or when the introduced population is invaded by individuals from several genetically distinct source populations (Facon et al. 2006). In rare cases, when there are several introductions from different source populations and these interbreed within an invasive population, genetic diversity may even be higher within this population than its native source populations (Facon et al. 2008). In contrast, the bridgehead effect may result in a severe loss of diversity, as subsequent introductions arise from an already depauperate introduced population. However, the bridgehead effect may promote the spread of phenotypic traits enhancing invasion success in secondary invasive populations, as these traits are already selected for and widespread in the initial introduced population. Investigating patterns of genetic diversity in native and introduced populations can therefore provide information on the invasion history of invasive species.
Reticulitermes flavipes is a subterranean termite species native from Texas to Massachusetts in the eastern USA. The termite has become invasive in localities both near to and distant from the eastern USA. This includes the western USA (Austin et al. 2005, McKern et al. 2006), the Province of Ontario in Canada (Kirby 1965), the Bahamas (Scheffrahn et al. 1999), Chile (Clément et al. 2001) and Uruguay in South America (Austin et al. 2005, Su et al. 2006) and France, Germany, Austria and Italy in western Europe (Kollar 1837, Weidner 1937, Clément et al. 2001, Ghesini et al. 2010). Previous genetic analysis based on microsatellite markers and mtDNA haplotypes have shown that the introduced French population exhibits an average decrease in genetic diversity of 60-80% compared to native USA populations (Perdereau et al. 2013). The analysis also revealed the occurrence of three main genetic clusters within the native USA range -the ‘Eastern cluster’ (West Virginia, Virginia, Delaware, North and South Carolina), the ‘Gulf Coast cluster’ (Florida and Eastern Mississippi–Louisiana) and the ‘Southern Louisiana cluster’ (the New Orleans and Baton Rouge regions in Louisiana) (Perdereau et al. 2013). Notably, some microsatellite and mtDNA haplotypes found in France were unique to the Southern Louisiana cluster (Perdereau et al. 2013). This finding, together with a similarity in chemical profiles and breeding structures found between France and Louisiana (Perdereau et al. 2010b, Perdereau et al. 2015), suggested thatR. flavipes was introduced to France from Louisiana, most likely during the 17th and 18th centuries via wood trade between New Orleans and the major French ports on the Atlantic coast (Dronnet et al. 2005, Perdereau et al. 2010a, Perdereau et al. 2013). Although the Louisiana origin of the invasive French population appears well supported, several points remain unclear. First, Perdereau et al. (2019) recently identified a French haplotype more closely related to the ‘Eastern cluster’ than Louisiana, suggesting multiple native populations from the USA may have invaded France. Additionally, the source(s) of the Canadian and Chilean invasions remain unidentified. Although several populations of R. flavipes occur in the Northeastern and Midwestern USA (i.e.,adjacent to Ontario), the only haplotype found in Canada was shared with Louisiana and France (Perdereau et al. 2013). Therefore, it is unclear whether the Canadian population arose from a primary introduction from Louisiana or from a secondary introduction through France (i.e., bridgehead introduction), as Canada and France share a long common history. Similarly, Chile’s unique haplotype was closest to one shared between Louisiana and France (Perdereau et al. 2013), raising the same question regarding primary versus secondary introduction. Overall, these findings suggest a complex invasion history for R. flavipes and raise the question of how many native populations may have served as sources of the introduced populations and what the role of bridgeheads might be in the global distribution of this species.
Substantial variability in breeding structure is present among the native USA populations of R. flavipes . Most populations are comprised of colonies headed by a monogamous pair of primary (alate-derived) reproductives (simple family). However, some populations are mainly comprised of colonies headed by a few secondary reproductives (i.e., nymph-derived neotenics; extended-family), and other populations comprised of fused colonies (mixed-family colonies) (Vargo and Husseneder 2009, Vargo et al. 2013, Aguero et al. 2020). Interestingly, the French introduced population of R. flavipes differs from most native populations, exhibiting a set of ‘invasive’ traits that enable colonies to form amplified versions of the extended and mixed-family forms. French colonies contain several hundred secondary reproductives and are usually spatially expansive (Dronnet et al. 2005, Vargo and Husseneder 2009, Perdereau et al. 2010a). They display highly similar chemical signatures, which reduces intraspecific antagonism between non-nestmate workers and allows for frequent fusions between colonies (Bagneres et al. 1990, Clément et al. 2001, Perdereau et al. 2010a, Perdereau et al. 2010b). Interestingly, this set of phenotypic traits occurs to a lesser extent in a population from Louisiana (Vargo 2019), which may have enhanced the invasion success of this introduced population in France (Perdereau et al. 2010a, Perdereau et al. 2010c, Perdereau et al. 2015). However, whether this set of traits is common among all introduced populations ofR. flavipes remains unknown.
Here, we used population genetic analyses and approximate Bayesian computation (ABC) to investigate the invasion history of R. flavipes . Using ddRadSeq, we first generated a SNP dataset sequencing 23 native populations in the USA and six introduced populations of this species in France, Germany, Chile, Uruguay, the Bahamas and Canada. We then assessed patterns of genetic structure within the entire native range of the species, and within each of the introduced populations. Finally, in order to elucidate the invasion history of R. flavipes , we compared support for different invasion scenarios modeling the number, size, and origin of each introduction event and their admixture using ABC.