Population genetic structure and migration
Populations were highly differentiated, with a global FST value of 0.219. In comparison, population genetic structure of the codling moth (Cydia pomonella), which belongs to the same tribe (Grapholitini) as G. molesta and is also a major pest of fruit crops, varies from absent/low (France; Chile) to moderate (Switzerland) to high (South Africa; Italy) among different geographic areas. However, the present study was carried out at a global scale, while individual studies of C. pomonella were carried out at scales varying from a few to several hundred kilometers, impairing comparisons between the two species. Nonetheless, low levels of population structure among C. pomonella populations from some regions relative to G. molesta could be the combined result of higher dispersal abilities of the former species compared to the latter[19, 47–49] and the frequent use of pesticides against C. pomonella in some countries, which can stimulate flight in this species (reviewed by).
In the present study, clustering outcomes from two distinct Bayesian approaches were largely congruent, except that the analysis from Geneland suggested greater population substructure compared to the analysis from Structure. Evidence indicates that G. molesta was introduced to Australia from mainland Asia. The data also support a North American source of G. molesta in Argentina and Chile, The Azores and South Africa, since historical records indicate that G. molesta was present in North America from the beginning of the 20th century, but arrived much later in the latter three regions. However, a previous study suggested that South African populations were unlikely to be derived from Canadian populations, since the genetic distance estimates from these two regions were relatively high.
Unexpectedly, Brazilian populations clustered together with two European populations (Structure), and shared ancestry between Western European and Brazilian populations was also supported by results from Geneland. The two European populations that shared ancestry with populations from Brazil showed little evidence of admixture with other European populations. It is possible that these two populations from Europe were derived from the same source as other European populations, but were isolated from other populations on the same continent, and therefore became differentiated over time. If this is the case, Brazilian populations are likely derived from European origins. However, it is also possible that G. molesta was recently re-introduced to Europe from the South American part of its range. Chile is the primary source of off-season peach and nectarine imports to Europe, and Brazil exports nearly 90% of its apple crop to Europe[50, 51]. In contrast, stone and pome fruit exports from Europe to South America are negligible[50, 51]. This latter interpretation of our data is in line with recent findings of C. pomonella individuals with putative South American origin in France, suggesting that South America may be a common source of introduction for important stone and pome fruit pests to Western Europe. Moreover, other invertebrate pests of distantly related plant host species have recently been introduced from South America to Europe, including the pine-infesting woodwasp Sirex noctilio, which was introduced from Chile to Switzerland. Although most invasive species in Europe are assumed to originate from North America or Asia, there is growing evidence that a number of invasive invertebrates in Europe come from South America (e.g.,, this manuscript).
Although it has been suggested that G. molesta was introduced to North America from Japan[7, 12, 13, 53], we did not find evidence to support this hypothesis. The single Japanese population incorporated in this study clustered with other Asian populations, and did not show evidence of shared ancestry with North American populations. However, it is possible that limited sampling of populations in Japan prevented us from capturing a likely source population from that region, or that a historical bottleneck followed by high rates of mutation since the North American introduction have generated divergence levels that prevent the detection of shared ancestry. Our data did not provide an alternative hypothesis regarding the source of North American G. molesta populations.
We found an overall (albeit weak) pattern of isolation-by-distance, suggesting that regional introductions occur in a relatively stepwise manner. Also, we found little evidence for substantial admixture between populations sampled from different continents, which, combined with the clustering outcomes, implies that multiple cross-continental introductions are unlikely. We note however, that sampling from Australia and Africa was limited to one population each, and we cannot make definitive continent-wide conclusions for these regions based on these sample sizes.
It has earlier been suggested that G. molesta disperses primarily through the movement of fruit, bins and plant material between orchards[18, 21]. The natural dispersal ability of both males and females is limited, even though certain environmental factors can increase flight capacity. Flights between non-contiguous orchards are possible, however they are generally short[19, 47], and human-mediated dispersal is implicated as the main mode of dispersal. The poor natural dispersal ability of G. molesta likely accounts for the high levels of inbreeding observed in our study, since close relatives are probably geographically constrained to the same or neighboring orchards. Combined, these findings suggest that improved management of this invasive pest at regional and community levels (e.g. by area-wide pest management) and at the national level (e.g. by quarantine regulations) may be efficient strategies to limit its ongoing spread.
We did not find any evidence that G. molesta populations are structured according to host plant species. Although G. molesta has been shown to perform better on peach, its primary host, compared to apple, serial generations of a population sometimes infest different hosts as the growing season progresses. Such seasonal host shifts may preclude adaptation of populations to any specific host, particularly if several suitable hosts are available within close geographic proximity.
Genetic diversity and inbreeding
Estimates of genetic diversity among different geographic regions were high overall. A six times higher number of private alleles among Asian populations compared to populations from other continents provides, for the first time, convincing evidence that Asia is indeed the native range of this species, as is frequently suggested in the literature. We found that genetic diversity levels are surprisingly high across most of the invaded range, except for a few populations (Brazil, The Azores, and South Africa). It is therefore unlikely that founder effects impair the adaptive potential of G. molesta in most non-native areas, as is suggested to occur in other invasive species. Levels of genetic diversity were reported from the non-native range of 14 species of invasive insects, and of these only four species (Ceratitis rosa, Drosophila pseudoobscura, Polistes dominulus, and Rhagoletis completa) exhibited HE values above 0.3. In our study, HE exceeded 0.5 in most populations from both the native and invaded range. High levels of genetic diversity observed in this study could be a result of high microsatellite mutation rates in lepidopterans, and/or may result from the introduction of many individuals at each founding event.
High levels of genetic diversity may account for high levels of heritable phenotypic variation within G. molesta populations, and may contribute to the ability of G. molesta to adapt to pest management regimes. For example, there is considerable genetic variation within G. molesta populations with regard to olfactory response of female moths to their host plants. Similarly, intra- and inter-population variation in other traits may enable adaptation to new environments and/or pest control regimes, although heritable variation in such traits has not yet been quantified. Nonetheless, a number of authors have demonstrated that G. molesta has developed insecticide resistance in at least some parts of its invaded range[59, 60].