The spread of non-native species is one of the outcomes of global change, threatening many native communities through predation and competition. Freshwater ecosystems are particularly affected by species turnover with non-native species. One species that has been established in Central Europe for many decades – or even a few centuries – is the amphipod crustacean Gammarus roeselii. Although G. roeselii is nowadays widespread in major river systems, there have been recent reports of its spread into smaller streams that are typically inhabited by the native species Gammarus fossarum. Due to their leaf shredding ability, G. fossarum takes up a key position in headwater streams. This raises the important question, to what extent G. roeselii can equivalently take over this function. To answer this question, we collected both species from nine different sites in a mid-mountain river system (Kinzig catchment, Hesse, Germany) and investigated their functional similarity using a combination of stable isotope analysis, gut content and functional morphology. The species hardly differed in morphological characteristics, only females showed differences in some traits. Gut content analysis indicated a broad dietary overlap, while stable isotopes showed a higher trophic position of G. roeselii. The observed functional overlap could intensify interspecific competition and allow the larger and more predaceous G. roeselii to replace G. fossarum in the future as a headwater keystone species. However, the differentiation in the stable isotopes also shows that co-existence can occur by occupying different trophic niches. Moreover, the wide range of inhabited sites and exploited resources demonstrate the omnivorous lifestyle of G. roeselii, which is likely to help the species succeed in rapidly changing environments.

Freshwater ecosystems, functional morphology, Gammarus, gut content analysis, stable isotope analysis, trophic niche

The introduction and spread of non-native species pose a threat to native communities globally (Gallardo et al. 2016). Due to their interconnection, streams are particularly affected by invasive species (Bij de Vaate et al. 2002; Leuven et al. 2009), which can cause a decrease of native species through predation and competition (Van der Velde et al. 2000). This change is not only expressed on a taxonomic level (Van der Velde et al. 2009), but probably also leads to changes in functional characteristics of ecosystems (Rosenfeld 2002). Non-native species might possess traits and behaviour that enable them to fulfil similar ecological functions as the native species they displace. However, the non-native species did not evolve within the recipient ecosystem and might lack natural predators, competitors or parasites that would otherwise regulate their populations in their native range (Torchin et al. 2003; Sih et al. 2010). Consequently, this unregulated population growth can lead to changes in resource availability and disrupting ecological interactions. However, post-invasion transformations of community structure and ecosystem function can differ immensely, are context-dependent and, often, poorly understood (Parker et al. 1999; Strayer et al. 2006; Kenis et al. 2009; Jeschke et al. 2014; Kumschick et al. 2015; Bellard et al. 2016).

A key group that is currently enormously affected by species-turnover is the taxonomic group of amphipods (Jażdżewski 1980; Leuven et al. 2009). Native Gammarus spp. are key species in freshwater ecosystems, due to their role in the decomposition of organic matter – an essential process in the headwaters of stream ecosystems (Cummins and Klug 1979; Vannote et al. 1980; Gessner et al. 1999; Graça 2001). However, the species turnover that is currently attracting attention is taking place in larger rivers (Leuven et al. 2009; Jourdan et al. 2016), while small headwaters – which are often not monitored within the Water Framework Directive – receive far less attention. While Gammarus fossarum Koch, 1836, is a typical headwater species, Gammarus roeselii Gervais, 1835, mainly inhabits the downstream parts of the rivers (Pöckl and Humpesch 1990; Pöckl et al. 2003). The exact origin and classification (native or non-native) of G. roeselii is still under debate, but the reduced genetic diversity (Csapó et al. 2020) and the main distribution on the Balkan Peninsula (Grabowski et al. 2017; Kabus et al. 2023) suggest that G. roeselii is probably an older invader in northern and western river systems outside the Danube system (Jażdżewski and Roux 1988). The species has been described near Paris in 1835 (Jażdżewski and Roux 1988), probably from a non-native population and has been spreading in Western Europe in the past centuries (Jażdżewski 1980; Jażdżewski and Roux 1988; Csapó et al. 2020). Nowadays, we observe an increasing spread of G. roeselii into smaller tributaries and upstream regions (Jourdan et al. 2019). In the Kinzig catchment (Hesse, Germany), some first order streams are already colonised by G. roeselii, while others are still inhabited by G. fossarum. Interestingly, there are also streams where both occur syntopically. The similar phenotypic adaptation along environmental gradients in both species (Jourdan et al. 2019; Grethlein et al. 2022) now raises the question to which extent they share a similar trophic niche.

Investigating the trophic ecology of invasive species is necessary to gain a better understanding of the community-wide effects of invasions (Tillberg et al. 2007). Studies have shown that invaders across the animal kingdom often have a flexible and generalist diet (Garton et al. 2005; Tillberg et al. 2007; Caut et al. 2008; Zhang et al. 2010; Grey and Jackson 2012; Jackson et al. 2017; Mothapo and Wossler 2017). Especially in changing environments, omnivorous species or species with a broad trophic niche are more successful than those with a more specialised diet, so that omnivorous invasive species may prey on or compete with native species (Barbosa and Castellanos 2005; Simberloff 2010; Jackson et al. 2017).

The ecological niche of an organism is connected with its functional morphology (Bock and von Wahlert 1965). Therefore, morphological traits can provide additional important ecological insights (Premate et al. 2021). Characterising such traits – like body size and mouthparts – is suitable to evaluate the relationship between morphology and trophic ecology, which, in turn, can be tested against further results of stable isotope and gut content analyses (Premate et al. 2021). Morphological analysis combined with stable isotope analysis can be used to elucidate potential relationships between morphology and function (Hutchins et al. 2014), since trophic levels characterise the functional role of organisms (Hairston and Hairston 1993). So far, the relationship between morphology and ecology is not yet fully understood in many invertebrate groups and only now starting to become clearer in amphipods (Copilaș‐Ciocianu et al. 2021).

Stable isotope analyses are useful for answering general questions about trophic structure (Vander Zanden et al. 1999; Post et al. 2000). However, this analysis is not able to distinguish very well between closely-related food sources with a similar isotopic value (McCutchan et al. 2003; Hood-Nowotny and Knols 2007; Bowes and Thorp 2015). Therefore, to gain qualitative information necessary to interpret the isotopic results, stable isotopes should be used together with other information, such as the analysis of the gut content (Inger and Bearhop 2008). This analysis has, in previous studies, been demonstrated to have a strong correlation with stable isotope data and, thus, can be considered a reliable reflection of diet in amphipods (Bacela-Spychalska and Van Der Velde 2013; Aumack et al. 2017). Copilaș‐Ciocianu et al. (2021) identified a strong relationship between morphological traits and the proportion of food items found in the gut. These correlations can provide an important understanding of the functional morphology of freshwater amphipods (Copilaș‐Ciocianu et al. 2021).

Stable isotope analysis is a common method used to reveal the trophic position of organisms in the field (Peterson et al. 1986; Cabana and Rasmussen 1996; Vander Zanden and Rasmussen 1999; Vander Zanden et al. 1999; Post et al. 2000; Post 2002; Layman et al. 2007). This analysis is an important tool to study food webs and has already been successfully used to detect the impact of biological invasions on trophic structures (Vander Zanden et al. 1999; Van Riel et al. 2006; Mancinelli and Vizzini 2015; McCue et al. 2020). For nitrogen, isotope ratios (¹⁵N/¹⁴N; δ¹⁵N) of a consumer are on average 3 to 5‰ higher than of dietary items and increase with successive trophic levels (DeNiro and Epstein 1981; Minagawa and Wada 1984; Peterson and Fry 1987; Vander Zanden and Rasmussen 1999; Layman et al. 2007). On the contrary, the ratios of carbon isotopes (¹³C/¹²C; δ¹³C) show only minor changes since carbon moves through the food web with little alteration (DeNiro and Epstein 1978; Rounick and Winterbourn 1986; Peterson and Fry 1987; Layman et al. 2007).

We hypothesise that, in the focal study area, G. fossarum and G. roeselii share a similar trophic niche. This equivalence is suggested by a laboratory experiment where G. roeselii showed the same leaf consumption rate as G. fossarum (Jourdan et al. 2016). Moreover, we expect a high level of competition amongst ecologically similar species, which should lead to a niche shift in co-occurrence scenarios. Finally, G. roeselii is found both in large lowland rivers, but also occasionally in upper stretches, while G. fossarum is restricted to upper stretches (Pöckl and Humpesch 1990; Pöckl et al. 2003). Given that headwaters mainly receive energy input through terrestrial organic material, resulting in simpler food webs, while lower stretches exhibit more complexity due to increased nutrient availability and primary production (Vannote et al. 1980), we hypothesise that G. roeselii has a broader trophic niche compared to G. fossarum.

Our study revealed a strong overlap in morphology and gut content between the native G. fossarum and non-native G. roeselii in headwater streams. However, stable isotopes indicated a stronger dietary differentiation between the two species when occurring alone and a more similar trophic niche when occurring together, with G. roeselii, however, generally occupying a higher trophic level. In addition, the gut content analysis confirmed that G. fossarum appears to have a different diet when occurring alone. This indicates that, despite their apparent functional morphological equivalence, the two species exploit different food resources. Below, we expand on the significance of these findings.

Overall, morphological differentiation occurs between the sexes rather than between species when looking at the combined data, pooled over all sampling sites. Only in some traits of the females (gnathopods, pereiopods and coxae) could we see differences between the species. This differentiation between females of both species could be explained by different reproduction characteristics of the species, such as different thermal optima for maximum fecundity (Pöckl 1993). As reproduction involves a metabolic cost (Sutcliffe 1992), a different reproduction effort could mean that different energy levels are available for growth and, thus, translating into different sizes of morphological features. Besides, the size of males is significantly affected by sexual selection, while females are significantly affected by natural selection (Ward 1988). Sexual selection of a trait is induced by competition over mates (Andersson 1994), while natural selection is induced by trait variation amongst individuals (Endler 1986). The competition between males of both species might be similar throughout the sampling sites and, therefore, the males show similar morphological traits. On the other hand, females of both species experiencing natural selection might lead to morphological trait differentiation. Similar morphological features of males of both species point to an absence of morpho-functional differentiation, the occupation of a similar trophic niche and ecological similarity (Cothran et al. 2013; Fišer et al. 2015; Copilaș‐Ciocianu et al. 2021). However, morphological similarity does not always mean ecological equivalence. Fišer et al. (2015) found ecological differentiation between four morphologically similar Niphargus species, which could potentially lead to different ecological roles in the ecosystem. Moreover, Premate et al. (2023) discovered that morphological traits are influenced not only by trophic position, but also by the specific habitat type. This observation could also explain the similar morphology of our study species occurring in the same headwater habitat.

The gut content analysis showed a strong dietary overlap between the species when pooled over all sampling sites. Our results thus indicate that the foraging on the same food sources might lead to competition between the two species. In principle, a strategy adopted by different amphipods to reduce competition for limited resources could be to utilise resources in different ways, in different microhabitats or at different times (Piscart et al. 2010, 2011; Mauchart et al. 2017; Premate et al. 2021). Apart from that, our species differed significantly regarding the δ¹³C level, which has also been observed in other co-occurring freshwater amphipods (Premate et al. 2021). Such partitioning of food resources is expected amongst co-occurring species (Schoener 1974; Chesson 2000), facilitating a stable co-existence (Chesson 2000). Our analysis showed a significant differentiation in gut content for G. fossarum when occurring alone compared to co-occurring with G. roeselii and close to significance for G. roeselii when occurring alone compared to co-occurring with G. fossarum. This could be explained by their, in general, separated occurrence in different river sections (Pöckl and Humpesch 1990; Pöckl et al. 2003) and the different availability of resources in these sections (Vannote et al. 1980). On the other hand, the absence of dietary separation for co-occurring populations might be a consequence of high abundance of food during the study period and could indicate potential for interspecific competition (Piscart et al. 2011; Rothhaupt et al. 2014). Interspecific competition, in turn, can reduce intraspecific specialisation, therefore, reducing the niche breadth of individuals (Araújo et al. 2011), leading to a dietary overlap. In addition, dietary separations or overlaps can change frequently in populations depending on spatio-temporal availability of food items and on the abundance of competing species (Rothhaupt et al. 2014). Thus, stable isotope analysis is useful for studying the diet on a longer time-scale.

The stable isotope analysis revealed that both species had more similar trophic niches when they co-occurred. Specifically, G. fossarum underwent a trophic level increase while G. roeselii a decrease which was also accompanied by a shift in the carbon source. Although this result was unexpected and non-intuitive at a first glance, it could be explained by reciprocal predation on juveniles or recently moulted individuals of the other species, as observed in another native and non-native amphipod species pair (Dick et al. 1993; Dick 1996; Dick and Platvoet 1996). Moreover, other researchers have discovered niche segregation in co-occurring amphipods in order to facilitate co-existence and avoid competition (Piscart et al. 2011; Premate et al. 2021).

Nevertheless, our stable isotope analysis revealed that, even though both species experience a niche shift, G. roeselii still has generally a higher trophic position compared to G. fossarum when both species occur together, contradicting our first hypothesis of a shared trophic niche. One reason why we have not found this more carnivorous lifestyle in the gut content could be that animal material can be digested more quickly (Guerra-García et al. 2014), but the stable isotopes of animal prey are incorporated in the tissues of the amphipod and, thus, influence the stable isotope analysis. The higher trophic position of the non-native G. roeselii observed in our stable isotope analysis could be an advantage compared to the native G. fossarum. Animal material, a high quality and easily-assimilated food item, can improve the growth of Gammarus sp. (Anderson and Cummins 1979; Crenier et al. 2017). The acceleration of growth, maturation and reproduction progress when being fed animal matter has been observed for several Gammarus species (Vassallo and Steele 1980; Smith 1988; Delong et al. 1993). In addition, it has been observed for G. fasciatus that, with increasing size, the amount of animal matter found in the gut increased proportionally (Summers et al. 1997). The present study showed that larger individuals occupy a higher trophic niche, confirming an ontogenetic niche shift. Thus, since G. roeselii usually has a larger body length, it will presumably consume more animal material than the smaller G. fossarum (Delong et al. 1993). In return, individuals of G. fossarum have a lower trophic position than G. roeselii. Since G. fossarum also has a smaller body size and smaller gnathopods than G. roeselii, it may be mainly limited to detritus because their smaller gnathopods may not be large enough to handle larger food items, such as animal matter (Summers et al. 1997). This is supported by a laboratory experiment of Delong et al. (1993), which showed that small individuals of G. fasciatus showed a delayed growth response when fed animal matter, but a normal growth when fed only leaf litter. However, G. fossarum is generally also known to be predatory and able to exploit animal food resources (Stoffels et al. 2011; Georgievová et al. 2020; Syrovátka et al. 2020) resulting in gammarids generally being classified as rather omnivorous (MacNeil et al. 1997).

Overall, our results indicate that G. roeselii exhibits a broader trophic niche than G. fossarum confirming our second hypothesis. It has been shown that, when G. roeselii co-occurs with G. fossarum, it significantly affects their micro-distribution (Mauchart et al. 2017). G. roeselii chooses habitats with macrophytes, (dead) wood or deposits of CPOM (coarse particulate organic matter), whereas G. fossarum prefers coarse gravel and cobbles (Mauchart et al. 2017). This suggest different sources of algal or terrestrial δ¹³C (Finlay 2001, 2004) for the diet of G. roeselii and could be an explanation for its broader trophic niche. Moreover, in our stable isotope analysis, population from site 7 appears to be an outlier with lower δ¹³C values. Site 7, unlike the other sites, was highly morphologically degraded, canalised and without significant riparian vegetation. Unlike G. fossarum, G. roeselii is able to occur in many anthropogenically shaped habitats (Mauchart et al. 2014; Enns et al. 2023; Kochmann et al. 2023). Apparently, in such habitats, G. roeselii is able to exploit resources from different food chains, such as algae. On the other hand, the larger size of G. roeselii could be an explanation for its broader trophic niche, because the larger size can offer competitive advantage (Young 2004). This advantage arises from the potential for a wider dietary range as size increases, potentially resulting in accelerated growth rates and shorter reproduction times (Summers et al. 1997). Moreover, it enables the exploitation of seasonal changes of food items (Delong et al. 1993). High trophic diversity is one of the most important factors responsible for the dispersal success of amphipods (Legeżyńska et al. 2012), because invasive species must be able to gain food material in heterogeneous habitats (Rothhaupt et al. 2014). Nevertheless, in a future study, it needs to be elucidated whether G. roeselii has a similar or different niche in their native range on the Balkan Peninsula. Presumably, G. roeselii could become even more competitive through depending on leaf litter as constant adaptation, effectively exploiting leaf litter, and also higher quality food sources and continuing to colonise tributaries and headwaters (Jourdan et al. 2016).

The omnivorous diet of both amphipods, the higher trophic niche of G. roeselii seen in our stable isotope analysis and the observed microhabitat partitioning in the field (Mauchart et al. 2017) can facilitate and explain the co-existence of both species. However, unlike G. fossarum, G. roeselii possesses dorsal pleosomal spines. An experimental study identified these as an effective defence mechanism against predatory fish (Bollache et al. 2006). The protection through spines can probably not only facilitate its spatial distribution, but may also favour the colonisation of new niches, hence increasing its invasion potential (Copilaş-Ciocianu et al. 2020). The low predation of G. roeselii may promote its establishment in communities with native species, such as G. fossarum (Bollache et al. 2006). Moreover, headwater sections of streams are particularly impacted by anthropogenic pollution and a subsequent species turnover (Betz-Koch et al. 2023; Enns et al. 2023). This ongoing pollution can facilitate the spread and establishment of G. roeselii into headwaters (Mauchart et al. 2014; Jourdan et al. 2024) and can lead to the decline of the more sensitive G. fossarum (Enns et al. 2023). In addition, temperature records for the herein studied sites show water temperatures between 14.4 °C and 21.5 °C, which are congruent with an earlier study of some of the sites (Grethlein et al. 2022). G. roeselii reaches sexual maturity faster, has a shorter brood development and a higher reproductive success than G. fossarum at temperatures above 12 °C (Pöckl et al. 2003). Furthermore, these characteristics of G. roeselii are enhancing with increasing temperature (Pöckl 1992). These life-history traits could facilitate invasion success and alter the community structure of invaded habitats (Grabowski et al. 2007; Rothhaupt et al. 2014). Thus, due to increasing water temperatures associated with climate change and ongoing pollution through anthropogenic impact, G. roeselii might outcompete G. fossarum in headwaters in the future (Pöckl et al. 2003; Enns et al. 2023; Jourdan et al. 2024).

Our study revealed that the non-native G. roeselii is morphologically similar to the native G. fossarum in headwater streams. We also found similar food items in the gut content, which showed the generally omnivorous lifestyle of both species. However, stable isotopes indicated that the trophic niches of both species differ substantially, with G. roeselii being more predaceous and generally having a broader niche. This indicates that, despite their shared morphological characteristics and omnivorous tendencies, there is a noticeable niche differentiation in G. roeselii, consequently influencing the headwater food web. In situations of ample resource availability, co-existence between both species may be possible. However, in cases of resource scarcity, we expect G. roeselii to be competitively superior, particularly given its ability to exploit a broader range of food resources, regardless of their quality.

We thank Erik Aschenbrenner and Jana Kabus for fieldwork assistance. We thank an anonymous reviewer for constructive comments that helped to improve the manuscript. Lars Pelikan thanks the Laboratory of Evolutionary Ecology of Hydrobionts at the Nature Research Centre in Vilnius, Lithuania, for their kind hospitality during his stay.

Lars Pelikan was funded by an Erasmus+ scholarship. The open access publication of this article was funded by the Open Access Publication Fund of Goethe University Frankfurt am Main.