Biological invasions cause organisms to face new predators, but also supply new anti-predator shelters provided by alien ecosystem engineers. We checked the level of anti-predator protection provided to three gammarid species by an invasive PontoCaspian zebra mussel Dreissena polymorpha, known for its habitat modification abilities. We used gammarids differing in their origin and level of association with mussels: Ponto-Caspian aliens Dikerogammarus villosus (commonly occurring in mussel beds) and Pontogammarus robustoides (not associated with mussels), as well as native European Gammarus fossarum (not co-occurring with dreissenids). The gammarids were exposed to predation of two fish species: the racer goby Babka gymnotrachelus (PontoCaspian) and Amur sleeper Perccottus glenii (Eastern Asian). This set of organisms allowed us to check whether the origin and level of association with mussels of both prey and predators affect the ability of gammarids to utilize zebra mussel beds as shelters. We tested gammarid survival in the presence of fish and one of five substrata: sand, macrophytes, stones, living mussels and empty mussel valves. D. villosus survived better than its congeners on all substrata, and its survival was highest in living dreissenids. The survival of the other gammarids was similar on all substrata. Both fish species exhibited similar predation efficiency. Thus, D. villosus, whose affinity to dreissenids has already been established, utilizes them as protection from fish predators, including allopatric predators, more efficiently than other amphipods. Therefore, the presence of dreissenids in areas invaded by D. villosus is likely to help the invader establish itself in a new place.

Predators shape the behavior and distribution of prey organisms, driving their evolution and environmental impact. We studied relationships between interspecific competition and predation risk in Ponto-Caspian invasive gammarid crustaceans: Pontogammarus robustoides and Dikerogammarus villosus. We hypothesized that a stronger competitor D. villosus would displace P. robustoides from a preferred habitat, but the frequency of this displacement would be reduced in the presence of a higher fish predator, the racer goby. We studied gammarid preferences for stone or sand substrata in 24-h pairwise-choice tests in single-species or mixed-species treatments, at 2 densities (12 or 24 individuals per tank), in the presence or absence of a predator. D. villosus displaced P. robustoides from the stone habitat, even at the low density. As P. robustoides tested separately preferred stones at both densities, its displacement resulted from active avoidance of a stronger competitor even when the substratum was not actually a limited resource. The presence of a predator reduced the number of instances of displacement of P. robustoides by D. villosus, allowing the weaker species to stay on its preferred substratum. Nevertheless, the presence of D. villosus still increased fish predation on P. robustoides. The presence of a predator modifies relationships between prey organisms, allowing a weaker competitor to stay in its preferred habitat. Thus, antipredator responses of prey organisms may have cascading effects on the functioning of the community, affecting habitat choice and competition among species.

Summary When an invasive predator encounters native and invasive prey, two scenarios are possible: the predator may benefit from the presence of naïve native prey or choose prey from its region of origin, reflecting their common evolutionary history. To determine interactions between an invasive predator and native and invasive prey, we used the Ponto‐Caspian racer goby Babka gymnotrachelus as predator and gammarids as prey: native Gammarus fossarum and Ponto‐Caspian Dikerogammarus villosus and Pontogammarus robustoides. We hypothesised that prey origin would affect fish preferences and growth rate and conducted a series of laboratory experiments on fish predation and growth and estimated profitability of prey of different origin. The goby preferred native prey to the Ponto‐Caspian gammarids, irrespective of prey motility, the presence of shelters or waterborne chemical cues. Moreover, fish grew better when fed native prey. Thus, we suggest that fish selectivity was based on the assessment of prey quality during direct contact with gammarids. A diet consisting of Ponto‐Caspian gammarids did not facilitate an invader originating from the same region, which benefited more from the presence of a local prey species. Ponto‐Caspian gammarids and gobies are successful invaders in inland waters, usually main rivers. The gobies, in contrast to the invasive gammarids, enter smaller tributaries that serve as refugia for native gammarids. We show that the gobies may benefit from the presence of native prey species in such locations.

Racer goby is one of several PontoCaspian gobiids spreading throughout European rivers and concurrent with recent declines in threatened populations of a native species of similar biology, the European bullhead. Although suggestive of competitive interactions, evidence thereof is scarce, so we examined behavioural interactions between racer goby and bullhead (single specimens of each species together, also pairs of each species) under experimental conditions (shared space with two shelters) to determine whether the invader displaces the native species when food resources are limited. Food (live chironomids) was added to a single feeder at rates below satiation levels twice over 24 h (once in light and once in darkness), with fish behaviour (aggressive interactions: attacks and threatening) and feeding activity (time spent near or inside the feeder) recorded using video cameras and infrared illumination. Racer goby exhibited aggressive behaviour towards bullhead (mean = 2.5 aggressive events h -1 ), but rarely the inverse (threatening only, mean = 0.05 events h -1 ), significantly limiting bullhead foraging time (by 62 %) and being faster to reach food in the feeding time in 76 % of cases. Gobies were more aggressive during daylight (77 % of all aggressive events occurring in light), and both species spent more time on feeding activities in darkness (88 and 66 % of all time spent in the feeder by bullheads and gobies, respectively). However, the adverse impact of goby on bullhead was independent of light conditions. Our results suggest that under natural conditions, racer goby are likely to displace bullhead during feeding, with potential consequences for foraging efficiency.

Detection of chemical cues of predators and food resources is a key for the behaviour of many species in aquatic ecosystems. We checked whether predator origin and diet, containing potential food and/ or alarm signals, affect the behaviour of omnivorous prey partly sharing their diet with a top predator. We conducted y-maze experiments to study the responses of invasive omnivorous and cannibalistic amphipods (Dikerogammarus villosus and Pontogammarus robustoides) to chemical cues of fish predators, injured amphipods and chironomid larvae (common amphipod food). As the predators, we used the goby Babka gymnotrachelus (sympatric with the amphipods) and piranha Pygocentrus nattereri (allopatric to the amphipods). The fish were either starved or fed amphipods or chironomids. D. villosus preferred predators fed chironomids and conspecifics as well as crushed conspecifics and chironomids, but avoided both hungry predators. Thus, D. villosus may perceive the presence of a top predator as an indication of both food resource and predation risk. P. robustoides avoided predators fed both amphipods (particularly conspecifics) and their alarm substances. The responses of both species depended more on the predator diet than on its species, which is likely to facilitate the recognition of allopatric predators and survival in newly invaded areas.