4.1. Three North American species of Girardia colonise the world
Our phylogenetic analysis showed that all non-American Girardia samples belonged to either G. tigrina, G. sinensis or G. dorotocephala. The close phylogenetic relationship between G. tigrina and G. dorotocephala and the fact that they are of North American origin, suggest that they diversified only relatively recently on the North American continent. However, G. sinensis, which is phylogenetically closely related to G. tigrina and G. dorotocephala (Fig. 2), was described on the basis of specimens from China, but the phylogenetic tree shows the presence of this species on the island of Cuba, which lies on the North American Plate (Fig. 2 and S1). Therefore, and on the basis of the phylogenetic relationships and the geographic distribution of these three species, G. sinensis is actually a North American species that was introduced into China, but thus far is undetected in North America, as discussed in Benítez-Álvarez et al (2022b).
Thus, our results corroborate (1) the earlier hypothesis that G. tigrina was introduced from North America, and (2) the introduction of G. dorotocephala in Japan and Hawaii. In addition, our results revealed the introduction of G. dorotocephala in Europe and Brazil and point to a third introduced species, G. sinensis, that has been introduced from North America into China, Europe, and Australia (Figs. 1 and 2).
Our present results indicate multiple introductions of Girardia into Europe, corroborating previous results of Ribas et al. (1989) and Stocchino et al. (2019), who had already described the presence of four morphological biotypes of G. tigrina named classes A to DOn the basis of a protein electrophoretic analysis, Ribas et al. (1989) detected genetic differentiation between animals of class A and those of classes B + C and suggested that this may signal the presence of at least two separate taxa, either at the subspecies or species level. Moreover, both Ribas et al. (1989) and Stocchino et al. (2019) suggested that presence of the four morphological groups in Europe may be the result of multiple independent introductions. Our present results corroborate this notion of multiple introductions, albeit that now we can establish molecularly that two species are involved in the areas where Ribas et al. (1998) and Stocchino et al. (2019) analysed the presence of G. tigrina, viz., G. tigrina and G. sinensis.
With respect to their external appearance and internal morphology, G. sinensis and G. tigrina are highly similar. This paucity of morphological differences between both species may be the reason why G. sinensis has gone undetected in its presumed original area of distribution, North America. In this sense, it is noteworthy that the copulatory apparatus of the sexually reproducing animals on Menorca analysed by Ribas et al. (1989) conforms to that of G. tigrina, and their habitus and pharynx pigmentation pattern to class C (Ribas et al. 1989, Fig. 2). On the other hand, the specimens from Menorca analysed in our study, and which came from a locality close to the B-type locality of Ribas et al. (1989), belong to G. sinensis. All of this provides ample evidence that G. tigrina and G. sinensis have colonised the same areas, where they may even occur at exactly the same site, exemplified also by the sampling locality #138 in mainland Spain (Table S1).
The morphological similarity of G. tigrina and G. sinensis and their co-occurrence in Western Europe, implies that records of presumed G. tigrina simply based on external appearance can no longer be considered to be reliable. In such cases where the animals only show asexual reproduction through the process of fission, as frequently being the case with introduced populations, discrimination between specimens of G. tigrina and G. sinensis can be achieved only through molecular markers. In that respect, it is noteworthy that earlier reports mentioned presumed G. tigrina for Australia (see Sluys et al. 1995 and references therein), while the individuals that we analysed from Queensland and Tasmania unequivocally ranked as representatives of G. sinensis.
The widely separated geographic localities where all three species were introduced, in combination with the fact that planarians are poor dispersers, strongly suggest that there were multiple introductions and that these were due to human-induced dispersal events. Most likely, the latter were effectuated through the international trade in aquatic plants and the activity of aquarists or through the importation and subsequent culturing of North American strains of G. tigrina and G. dorotocephala as model organisms in education and scientific studies all around the world. The fact that G. tigrina is very sticky, clearly enhances its attachment to all kinds of surfaces and thereby facilitates its passive dispersal (Stocchino et al. 2019). Wright (1987) suggested that multiple occasions of disposal of the contents of aquaria and of small ornamental garden ponds into local rivers might underlie the pattern of distribution of G. tigrina in Great Britain. In addition, disposal of such contents into waste-water pipes might facilitate entrance of flatworms into a river via surface water drains (Wright 1987). The present data does not allow us to infer whether introductions into different countries came directly from North America or may have occurred in a stepping-stone model from one introduction site to the next. Most probably, both types of events have contributed to the spread of alien species of Girardia. Future analyses of more variable markers may be necessary, in order to determine the various routes that may have been followed by the three North American species of Girardia in their conquest of the world.
4.2. What makes Girardia species successful colonisers?
Our niche modelling indicated that G. tigrina, G. dorotocephala and G. sinensis do not only have the ecological potential to successfully survive at the localities where they have been introduced but are also capable of expanding their distributional range. Interestingly, in the Maxent model the responses of the three species to different environmental variables indicated a key feature explaining their high capacity for colonization, viz., shared tolerance for anthropogenic habitats, such as artificial freshwater habitats (like canals) or freshwater habitats in cultivated and managed areas (Fig. 3). The importance of this variable has been reported also for many other introduced plant and animal species (Johnston et al. 2017; González-Ortegón and Moreno-Andrés, 2021; Rickart et al. 2011; Salomidi et al. 2013; Simkanin et al. 2013), including terrestrial planarians (Álvarez-Presas et al. 2014) and is also one of the main factors explaining the distribution of another introduced freshwater planarian, Dugesia sicula (Leria et al. 2022). In that respect, it is noteworthy that the factor “tolerance for anthropogenic habitats” played no role in the explanation of the autochthonous distribution of D. subtentaculata in the Iberian Peninsula (Leria et al. 2022). Anthropogenic habitats generally are characterised by low autochthonous species diversity, thus offering empty niches for invasive species that can use these as a springboard to natural habitats (Dietz and Edwards 2006).
Apart from association with humans and anthropogenic habitats, other traits of successful colonizers that may become invaders, are, for example, high abundance in the native range, short generation time, high genetic variability, wide physiological tolerance, phenotypic plasticity, and asexual reproduction (Kolar and Lodge 2001). These are general features that may vary or combine in different ways, depending on the intrinsic characteristics of each group of organisms and of the receiver area. Our Maxent analysis allowed us to determine that the introduced species of Girardia present some of these traits.
A case of wide physiological tolerance concerns temperature, with the three Girardia species presenting a range of highly suitable temperatures that is much broader than in other freshwater planarian species (Vila-Farré and Rink 2018; Leria et al. 2022). For example, Polycelis felina exhibits an optimality maximum at 4ºC, with the optimality rapidly decreasing at temperatures above 10ºC, while Dugesia subtentaculata shows a narrow temperature range between 12o and 15ºC with a high suitability (Leria et al. 2022). In contrast, in Girardia high suitability was found for temperatures ranging from around 4ºC to roughly 20ºC. This minimum value of about 4ºC found for Girardia in our niche modelling, coincides with ecological studies that found temperature to be a limiting factor in the expansion of G. tigrina in the United Kingdom and, curiously, catalogued it as a warm-water species, with a lower tolerance limit of 6ºC for feeding and 10ºC for asexual reproduction (Wright 1987 and references therein). On the other hand, in the present study, G. dorotocephala has been found in Mexico and G. sinensis in Cuba in waters with temperatures ranging around 19.5–20.5ºC and 24ºC, respectively.
Our niche modelling also revealed that the three species are suitable to a wide range of lithological classes and water discharge regimes, as well as precipitation regimes (Fig. 3). However, for slope not all of the introduced species presented a flexible suitability, as only G. tigrina would be highly suited to an extremely wide range of terrain slopes. This is surprising, since it has been reported that G. tigrina generally occurs in the lower reaches of lowland rivers with low slopes (Wright 1987). All of these results suggest that each of the three introduced Girardia species exhibits high tolerances to most of these key environmental variables.
As remarked above, asexual reproduction is considered to be a trait enhancing successful colonization and invasion. For G. tigrina, and G. dorotocephala, it has been established that they show both sexual and asexual reproduction in their native areas, with populations being either strictly sexual or asexual or showing both reproductive modalities (Kenk 1937, 1972; Puccinelli and Deri 1991; Knakievicz et al. 2006). For G. sinensis, in the laboratory it was established that the species can reproduce both sexually and asexually by fission (Chen et al. 2015). However, most of the introduced populations are constituted by fissiparous individuals. Thus far, only five immigrant sexual populations are known from Western Europe, located in Great Britain, Spain (Menorca Island), Italy, and France (Stocchino et al. 2019 and references therein), while all other known populations reproduce by fission. This prevalence of asexual populations might be advantageous for dispersion and population growth (Gee et al. 1998; Wright 1987) and has already been proposed as a major factor explaining the successful recent colonization of Dugesia sicula in the Mediterranean region (Lázaro and Riutort 2013).
Another important characteristic of planarians possibly contributing to their colonization capability concerns their top position in the trophic web. Although there is large number of species known to predate on freshwater planarians, the impact of predation on natural populations of flatworms generally is rather low (Vila-Farré and Rink 2018). Evidently, this may promote rapid population growth, which, in turn, fulfils a basic requirement for high rates of spread.
4.3. Potential ecological impact of introduced Girardia
Since freshwater communities are highly sensitive to composition changes through the introduction of invasive species (Havel et al. 2015; Gallardo et al. 2016), prediction and risk assessment of potential invader species are vitally important for the protection of these habitats. However, risk assessment is distinct from prediction whether a certain species will become an invader, and even more difficult to achieve than the prediction. With respect to freshwater planarians, ecological studies on the specific impact of introduced species on their receiver freshwater communities are (1) limited, (2) focused on only small geographic regions, or (3) are even contradictory. For example, while some studies concluded that Girardia tigrina may be a competitive invader (Reynoldson 1985; Wright 1987; Gee and Young 1993), its ecological impact on native communities was downplayed in a recent paper (Ilić et al. 2018).
For making predictions about whether certain non-native species will become invasive, it is unfortunate that there is no suite of characteristics that can unequivocally signal successful invaders (Kolar and Lodge 2001). Nevertheless, a non-native species must go through five phases to become an invader: (1) transport to a new region, (2) release or escapement to the wild, (3) establishment, (4) dispersal or spread, and (5) integration or impact (García-Berthou 2007). From that perspective, all three species of introduced Girardia have already reached phases four and five, albeit that their ecological impact has not been properly analysed.
We have now run a preliminary study to highlight the potential impact of introduced species of Girardia on autochthonous freshwater planarians restricted to the Iberian Peninsula, for which there is a good knowledge about its autochthonous species diversity. Most native species in this region occur at high altitudes (Crenobia alpina (Dana, 1766), Polycelis felina (Dalyell, 1814), various species of Phagocata Leidy, 1847), while others, such as Phagocata ullala Sluys, 1995, are endemic to very restricted areas in which Girardia has never been detected (Baguñà, Saló, and Romero 1980, 1981; Sluys, Ribas, and Baguñà 1995). Nonetheless, Dugesia subtentaculata (Draparnaud, 1801), the most widely distributed autochthonous species, has been found together with Girardia representatives. At locality #247, D. subtentaculata coexists with G. sinensis, while in the North of the Iberian Peninsula it co-occurs with unidentified Girardia individuals (M. Vila-Farré, pers. comm.) and with G. dorotocephala (recently identified by DNA barcoding in our laboratory). Moreover, optimal environmental conditions for D. subtentaculata were recently estimated in a niche modeling analysis (Leria et al. 2022). For this species optimal conditions are present along the northern and western coasts of the Iberian Peninsula, a region with principally deciduous tree cover, an annual average temperature around 14ºC, and relatively high precipitation regimes. These environmental conditions partially overlap with the optimal ecological niche for G. sinensis, as identified in the present study (Fig. 4). Moreover, niche modeling showed that the Iberian Peninsula presents extensive unsuitable areas for D. subtentaculata (zero suitability values) (Leria et al. 2022), while in our modeling analyses of Girardia, almost all regions of the entire Peninsula received high suitability values (> 0.6), especially for G. tigrina and G. sinensis. In that light, it is important that it has been shown that introduced species of Girardia may exert a negative impact on native planarian species, for instance by competing for the same food. Gee & Young (1993) showed that introduced G. tigrina had considerable dietary overlap with the native planarian species Polycelis tenuis Ijima, 1884 and P. nigra (Müller 1774).
Our data also evidences that the genetic diversity within and among the three introduced species is quite low, in fact, much lower than, for example, in freshwater planarian species of the genus Dugesia in Western Europe. There, more than 10 species of Dugesia have been characterized, representing a broad diversification that evolved during the last 25 My (Dols-Serrate et al. 2020; Leria et al. 2020; Leria et al 2022; Benítez-Àlvarez et al 2022c). If changes in ecosystems induced by human activities favour the alien over the autochthonous species, a highly diverse and locally adapted fauna will be dangerously menaced by a generalist and low-diverse group with the evolutionary consequences this may have. Evidently, the ecological impact of introduced Girardia may not be restricted to other planarians but may concern also other freshwater organisms that they may predate or compete with.
4.4. Future expansion trends of introduced Girardia
Our niche modelling indicated that at present the three introduced Girardia species have the ecological potential to expand their distributional range (online resources Fig. S1), while modelling of future situations, regardless of the particular climatic scenario, pointed to an extension of their potential areas of suitability (Fig. 5 and S2). Their potential for range expansion was mostly restricted to suitable areas in the Northern Hemisphere, in particular to regions throughout Europe without diminishing high suitability of regions in the south. In this light, and also considering the generalist characteristics of the species and their demonstrated suitability for human-modified habitats, it is only to be expected that in our current, increasingly anthropogenic world they may have an advantage over native species.
In this aspect, it is especially worrisome that G. sinensis has recently been detected for the first time in Africa (Benítez-Álvarez et al. 2022a) in regions where our niche modeling indicated its suitability, also a region presenting a high diversity of Dugesia autochthonous species (Leria et al 2022; Benítez-Álvarez et al. 2022c). In this region, it will be especially important to avoid a future introduction of G. dorotocephala with a wider suitable region than G. sinensis, covering all the Atlas Mountains even in future suitability predictions (Fig. 5).
Final remarks
In the present work we do a preliminary dissection of a long-held invasion. We corroborate that three species of Girardia are colonizing the world (Benítez-Àlvarez et al 2022b). We characterize the features that may explain such successful colonization, especially as compared to some autochthonous species, concluding that there exists an overlap between autochthonous and Girardia suitability areas, with an advantage in area extension and physiological plasticity for the introduced species. Moreover, we demonstrate the usefulness of two molecular markers to rapidly and cheaply identify the introduced species at any locality, which may facilitate a continuous survey of freshwater habitats for new introductions and expansions. Also, niche modelling demonstrates its value to predict potential areas with a higher risk of being invaded, which may help focus future projects to monitor the expansion and potential detrimental effects of the three species. To avoid future expansions and reintroductions it will be extremely important to ensure importers, aquarists, researchers, and schools keeping alien freshwater organisms in their facilities refrain from disposing of their culture waters into waste-water pipes without killing first any rest of organisms.