Discussion

Our genomic study of 50 global populations of a worldwide, selfing weed revealed several important findings. First, populations were found to form six unique genotype groups or ecotypes showing very little sign of admixture and are therefore considered reproductively isolated. Second, three of the most prevalent of these ecotypes present in Europe are found most commonly throughout the introduced ranges, indicating that more than one aggressive genotype was introduced and established. Third, the multiple independent introduction events we detected and the distribution of ecotypes have links to colonial history. Finally, the global distribution of ecotypes was found to be restricted based on latitude (and therefore environmental factors related to latitude such as climatic). We therefore propose that multiple successful genotypes and prior adaptation to climate in the native ranges are likely reasons for the success of this global weed, such that a the right (prior adapted) ecotype must have arrived in an introduced location with a corresponding environment in order to persist and thrive.

Global distribution of ecotypes

Our study is the first to sample populations of Plantago major so extensively across its global range and to provide an overview of the global distribution of the species’ genotypic variation. The low genetic differentiation we found within populations worldwide is in keeping with what is expected for a highly selfing plant, and with the findings of past genetic studies of P. major (Mølgaard, 1976; Wolff et al., 1994; Wolff & Morgan-Richards, 1998; Morgan-Richards & Wolff, 1999; Barrett, Colautti, and Eckert, 2008). One of our main findings was that based on genetic distances and shared ancestry tests, individuals from across the global range cluster into six genotype groups rather than by population despite extraordinary geographic distances between them. However, genotypes from a given population were homogenous, such that you can predict the genotype group of an individual from those of the other individuals in a given population, based on where in the world it is growing. A commensal species such as P. major that is known to flower continuously, to produce wind-borne pollen and thousands of propagules (seeds) from a single individual, has undoubtedly been moved around and is still being moved around especially given today’s global trade networks (McNeill et al, 2011; Early et al., 2016). However, the lack of admixture or limited evidence of geneflow among genotype groups, even where populations occur in close proximity, suggests very low rates of out-crossing (Wolff, 1991), and provides evidence that the groups we identified are largely reproductively isolated.
The six distinct genotype groups may represent different taxonomic subspecies or varieties of P. major and/or intermediates between taxa that may have arisen by historical introgression. Two well-accepted subspecies, P. major subsp. major and P. majorsubsp. intermedia , have been the focus of past genetic and ecological studies. For example, Wolff et al. (1994) demonstrated low levels of polymorphism between populations of both P. majorsubsp. major and P. major subsp. intermedia , but found notable genetic differentiation between subspecies. Furthermore, a comparison of Dutch and Scottish populations for the same two subspecies revealed that genetic divergence between subspecies was stronger than divergence within populations of the same subspecies between the two countries (Wolff & Morgan-Richards, 1998). These well-studied subspecies are sympatric and capable of interbreeding, yet genetic, ecological, and morphological differentiation between them was found to be pronounced enough for them to be resolved as separate species/subspecies likely because high selfing rates limits interbreeding (Mølgaard, 1976; Morgan-Richards & Wolff, 1999).
Even though our sampling was focused on populations that fit the morphological descriptions for subsp. major , it is possible that our genotype groups reflect the complexity of several intraspecific taxa—or hybrids of them resulting from historical introgression and being maintained due to reproductive isolation. El-Bakatoushi (2011) found populations in Egypt to be intermediate between two subspecies, and that some populations exhibited morphological characters resemblingP. major subsp. major despite evidence of introgression with P. major subsp. intermedia . Because the goals of our work were not taxonomic in nature, and because very little is known about the numbers and global distribution of intraspecific taxa and the extent of their introgression, we consider the six groups identified in our study to be ecotypes of P. major sensu lato .
Three of the ecotypes are found to occur in both native and introduced ranges, and one unique ecotype (from the populations in Yukon and Alaska [in part]) is found in the introduced ranges. Selfing or clonal plants have been found to exhibit low genetic variation due to founder effects and lack of admixture in introduced populations compared to native populations (Zhang et al, 2010; Ferraro et al., 2015). The three most prevalent ecotypes we identified in Europe (two in the northern latitudes, and one in the southern latitudes) are also found to be most widespread in the introduced ranges; very little genetic differentiation is seen between native and introduced populations of the same ecotype.
Plantago major is a prime example of a species that increased drastically in numbers and spread within its native range after the widespread clearing of land begun during the Neolithic, between 4900–2000 BCE as evidenced by palynological records (Godwin, 1944; Pysek, 1998; Preston et al., 2004; Kaplan et al., 2016). Weedy and aggressive ecotypes—which are commonly reported in successful introduced or invasive plants (i.e. Hawthorn, 1974; Taylor & Keller, 2007; Lambertini et al., 2012)—could have evolved and spread widely in Europe before being introduced to other continents. We show that not just one but at least three ecotypes have successfully colonized introduced ranges and may represent lineages that have evolved increased abilities to withstand human disturbance and human-altered habitats (Mølgaard, 1986; El-Bakatoushi et al., 2011; Hufbauer et al., 2012; Vigueira et al., 2013). The nature of our genomic data and lack of historical records make it impossible to determine when and where in the native range these ecotypes of P. major evolved and when their ranges expanded; however, as is hypothesized for other weedy and commensal European species, the rapid onset of agriculture and the profound modification of the native landscape in Europe around the Neolithic likely facilitated the spread (Godwin, 1944; Hicks, 1971; Preston et al., 2004, Brun, 2009).  Even though the historical herbarium specimens of P. major were collected up to centuries after it’s putative introduction to new ranges, DNA analyses of these specimens may shed light on historical introduction pathways (Dormontt et al., 2007; Martin et al., 2014).

Multiple introduction events with colonial ties

Multiple source origins from the native range are inferred for this global weed. The three most widespread ecotypes found in Europe and the unique ecotype found in the Middle East (Iran1) were all detected in the introduced ranges we sampled. Therefore, at least three separate introduction events occurred from Europe to North America, one event from the Middle East (Iran) to western North America (California), two introduction events to the Australasian continent, and at least one introduction event to each of South America, South Africa, Canary Islands, Hawaiʻi, Greenland and Iceland.
The introduction of more than one lineage from the native range has been reported for other introduced species. For example, Guo et al. (2017) found two introduced lineages in North America for the invasive grassPhragmites australis (Cav.) Trin. ex Steud. – one lineage native to north-central Europe and Asia occurs in the Great Lakes region of North America, and another lineage from the Mediterranean region is naturalized in the Gulf Coast (USA) and in Central and South America. These patterns conspicuously match patterns we see with ecotypes ofPlantago major from Europe and their distributions in North and South America.
Unfortunately, the very low genetic differentiation within eachPlantago major ecotype makes it difficult to infer more precise origins for the introduced populations, despite our extensive sampling across native and introduced ranges. Furthermore, we cannot rule out that multiple introduction events of the same ecotype occurred historically, particularly given the extraordinary dispersal abilities and commensal nature of the plant. Our data does however support hypotheses that P. major followed early European colonists to new ranges (Samuelsen, 2000). The distribution of ecotypes and shared ancestry inferred between plants in native and introduced ranges reveals links to early colonial human movements and/or European colonies. For example, ecotypes found in southern Europe (Spain, Portugal) are also found in former Spanish or Portuguese colonies in Florida (USA), Peru, Chile, Brazil, and the Canary Islands, and could have followed colonial Spanish and/or Portuguese voyagers and settlers between the 15th and 18th centuries (Mancall, 2006). Plants in southern Africa could have been transported and introduced by Portuguese voyagers along their voyages to southern and eastern Africa.
Similarly, the two ecotypes prevalent in central and northern Europe and western Asia likely gave rise to the majority of populations in North America, as well as Greenland, Iceland, and New Zealand. Plants could have travelled with early colonial explorers/settlers from France and England in the 16th and 17thcenturies where these countries had colonial power (Mancall, 2006). In spite of hypotheses that the Norse played a role in the movement ofPlantago major in their travels across the northern Atlantic (Samuelsen, 2000), our data do not reveal any specific shared ancestry between plants in Scandinavia, Greenland and northeastern Canada (Newfoundland). The earliest known record of the plant in New Zealand is from 1832; however, the plant likely was introduced well before that (Webb et al., 1988).
Some populations of Plantago major do not show obvious colonial ties. For example, although populations from both Melbourne and Perth (Australia) share ancestry with the southern European populations, there is no evidence of direct voyages between the Spanish or Portuguese during early colonial times, although Spanish and Portuguese were both in southern Africa and southeast Asia, and voyagers from the UK made stops in South Africa on route to Australia. Plants could have been introduced to southern Africa and then further dispersed to Australia. Alternatively, plants could have arrived with the Portuguese to Timor in the 16th century (Mancall, 2006), approximately 650 km from the Australian coast, and later dispersed by birds to Australia (Iwanycki Ahlstrand et al., 2019). The first record for the species in Australia is in 1770 (GBIF.org), around the same period the English made their first voyages there, which indicates that plant may have arrived with earlier explorers such as the Dutch, who arrived in Australia (and New Zealand) in the 17th century (Mancall, 2006). The plants we sampled in Hawaiʻi also belong to the southern European ecotype group. Given that English explorers were the first to reach Hawaiʻi (Captain James Cook in 1778), after stops in Tahiti and South Africa, it is also plausible that plant genotypes from more northern European latitudes also arrived in Australia, South Africa, and Hawaiʻi, but that the climate or environment did not allow their persistence. DNA from historical herbarium specimens could help to shed light on this.
The ecotypes found in eastern Asia (South Korea and Japan) were not found within any of the introduced ranges we sampled. This could be an artefact of poor sampling across Asia, or could represent a true biological scenario in which Plantago major ecotypes from eastern Asia are not as successful in colonizing and spreading in other regions. At least two ecotypes are recognized in Japan – one restricted to sandy shores and brackish waters (which is represented in our sampling), the other considered introduced and weedy (pers. comm. M. Amano, 2020). The population sampled in Yukon, and some of the individuals from Alaska (USA) share common ancestry with eastern Asian populations, yet are found to have a unique lineage. These findings for populations we sampled in Yukon and Alaska (in part) provide support to the hypothesis that the northerly latitudes in North America consist of native populations of P. major that predate the arrival of early Europeans (Hawthorn, 1974). Increasing sampling and further investigating of the link between Asian plants and those in northernmost North America is needed to unravel relationships and migration patterns, and clarify any possible links between the movement of eastern Asian plants to northern USA and Canada (for example by Russian colonists or by earlier human migrations from Beringia [Erlandson & Braje, 2011]).
One unexpected finding was that the population sampled in California was found to differ from plants elsewhere in North America, and shared ancestry with the population from Iran1. However, other weedy species with native distributions in the Middle East have been introduced to and have become well established in California (i.e. Ortiz et al., 2008; Meyers & Liston, 2008). The relationships among such populations is most likely due to similarities in climate. Pathways of migration between these biogeographic regions remain poorly studied but the dispersal of weedy species could be linked to the movement of plants for horticultural trade with which weeds are also moved (Chapman et al., 2017; Dullinger et al., 2017). We also cannot rule out non-anthropogenic introduction events, particularly since the genus Plantago is well adapted for long-distance dispersal (Meyers & Liston, 2008; Iwanycki Ahlstrand et al., 2019). In a study by Morgan-Richards & Wolff (1999), populations sampled in Los Angeles (California, USA) and Trinidad were genetically and morphologically determined to be P. major subsp. intermedia (syn. P. intermedia DC.). Although it was not the goal of their study to characterize the global distribution for each of the subspecies or retrace origins of introduced populations, their sampling demonstrates that P. major subsp.intermedia is found in some regions of the introduced range, such as the more southerly latitudes in North America and the Caribbean. More intensive population sampling around the Middle East and California would be needed to further narrow down the native origins and further resolve ancestry and distribution for this Mediterranean ecotype ofP. major and other Plantago species.

Prior adaptation is key to global success

Plantago major is a classic example of a widely-introduced species for which its successful colonisation and spread in new lands is not compromised by low genetic variation or the lack of sexual recombination (Beaumont et al., 2009; Dormontt et al, 2014). Based on the distribution of ecotypes that have spread to the introduced parts of the range, and the division seen along the 35–40o N latitude range, we propose that prior adaptation to climate in multiple successful ecotypes may explain the distribution and persistence of this global weed.
When introduced species successfully colonize new ranges, they either do so by quickly adapting to new conditions, or they arrive with prior adaptations that arose in their ancestral native range before introduction (Rey et al., 2012; Jackson et al., 2015; Barrett, 2015; Bock et al., 2015). The pattern we see in the global ecotype distribution where the majority of the global populations we sampled suggests that the successful colonisation of new ranges by P. major is dependent on plants having a prior adaptation to climate before arriving to new regions worldwide. Such prior adaptation to climate has been noted for other species with successful colonists, i.e. in plants (Sexton et al., 2002; Dlugosch and Parker, 2007; Schlaepfer et al., 2010; Rosche et al., 2018), birds (Jackson et al., 2015), and beetles (Vahsen et al, 2018). Although climatic conditions are quite different between regions occupied by populations belonging to the same ecotype (i.e., semi-tropical Florida and southern Spain), it is possible that further climatic niche shifts occurred after introduction, and although such shifts are thought to be rare in terrestrial plants (Petitpierre, 2012), it has been demonstrated for other species (i.e. Jackson et al., 2015). Alternatively, prior adapted lineages could have evolved tolerance to a large climatic amplitude, and therefore climatic and environmental variation deserves further investigation.
Although very little is known about the ecological habitats of P. major in pre-colonial times, the ecotypes in its native range may have evolved anthropogenically-induced adaptations to invade (AIAI; Theoharides et al., 2007; Rey et al. 2012; Hufbauer et al., 2012). Plants like P. major that occupy a broad range of habitat conditions in their native range, including those that are human-made and disturbed habitats, can naturalize more easily in new regions (Hufbauer et al., 2012). Human activity and disturbance have created conditions that are now more similar across broad geographic regions, and this homogenization of the environment likely facilitated the colonisation and successful spread of disturbance tolerant species likeP. major (van Dijk & van Delden, 1981; Estoup & Guillemaud, 2010; Kalusova et al, 2017).
Asexual plants have been found to change just as often and as fast as do sexual plants when introduced to a new range (Dalrymple et al, 2015). The extraordinary phenotypic plasticity known in P. major is likely important for ecotypes coping with a high level of environmental and climate heterogeneity (Warwick & Briggs, 1980; Samuelsen, 2000).Plantago major can vary extensively in morphology even within populations, and therefore high phenotypic variation for the species is not related to genetic diversity at least within a population level, though further studies are needed (Mølgaard, 1976; Lotz and Blom, 1986; Wolff, 1991; Iwanycki Ahlstrand et al., 2018). Wide environmental tolerance and an ability to grow in a multitude of climatic and edaphic conditions as a result of well-developed phenotypic plasticity, makesPlantago major a classic example of a weed that possesses the “general purpose genotype strategy” as coined by Baker (1965) – in this case, it appears that as few as three general-purpose genotypes or ecotypes may have been involved in the global success of the common plantain.

Conclusions

Population genomic approaches can provide useful data where historical records are lacking in helping to retrace migration routes of historical introductions of plants, such as Plantago major, and improve our understanding of the role of genomic variation in explaining success of globally distributed species. Our findings show that beyond human-mediated migration, factors such as prior adaptation to climate and anthropogenic disturbance may be key to success for a worldwide weed. That is, a genotype introduced into a new range with environmental conditions for which it is already adapted is more likely to succeed. This suggests that even if globalization creates more opportunities for species invasions, not all species with invasive potential will be as likely to succeed in a new area. The six ecotypes we identify serve as an excellent starting point for future ecological and genomic studies, and the distribution of genomic diversity across the globe provides a glimpse into the complex interactions between the environment and the genome that influence the distribution of plant species and mediate phenotypic adaptation to local conditions (Bragg et al. 2015). Introduced plants that exhibit high phenotypic plasticity are hypothesized to perform better under changing climatic conditions, despite low genetic variation and sexual recombination. Therefore, our findings can be used to advance our ecological and evolutionary understanding of successful invasions, and contribute to more accurately predicting species responses to global change (Chapman et al., 2016; Klonner et al., 2017).