We explored the social and ecological outcomes associated with emergence of a management panacea designed to govern a stochastic renewable natural resource. To that end, we constructed a model of a coupled social-ecological system of recreational fisheries in which a manager supports naturally fluctuating stocks by stocking fish in response to harvest-driven satisfaction of resource users. The realistic assumption of users remembering past harvest experiences when exploiting a stochastically fluctuating fish population facilitates the emergence of a stocking-based management panacea over time. The social benefits of panacea formation involve dampening natural population fluctuations and generating stability of user satisfaction. It also maintains the resource but promotes the eventual replacement of wild fish by hatchery-descended fish. Our analyses show this outcome is particularly likely when hatcherydescended fish are reasonably fit (e.g., characterized by similar survival relative to wild fish) and/or when natural recruitment of the wild population is low (e.g., attributable to habitat deterioration), which leaves the wild population with little buffer against competition by stocked fish. The potential for release-based panacea formation is particularly likely under user-based management regimes and should be common in a range of social-ecological systems (e.g., fisheries, forestry), whenever user groups are entitled to engage in release or replanting strategies. The net result will be the preservation of a renewable resource through user-based incentives, but the once natural populations are likely to be altered and to host nonnative genotypes. This risks other ecosystem services and the future of wild populations.stocking panacea | wild population collapse | angler satisfaction | social-ecological model | cultural ecosystem service

Although the effects of introduced predators on prey populations in aquatic ecosystems have been studied frequently, less is known about the interactions between predators. We performed a whole-lake experiment by stocking a non-native top predator (pikeperch (Sander lucioperca)) to two residential piscivores (Eurasian perch (Perca fluviatilis) and northern pike (Esox lucius)). By analyzing spatial distribution, diet composition, growth, and consumption rates of the piscivores before and after pikeperch introduction, we tested how both density-dependent and trait-mediated responses affected interactions between the three predators. Total piscivore biomass increased 1.5 times and annual consumption by the piscivores increased 1.7 times after stocking, attributable to the stocked pikeperch and increased northern pike abundances. Abundance, distribution, and consumption data indicated that northern pike was hardly affected by pikeperch stocking and even increased its biomass, whereas piscivorous perch shifted its habitat use towards the littoral lake areas in response to competition with pikeperch. Furthermore, all piscivores increasingly fed on small perch. The forced habitat shift of piscivorous perch in combination with increased predation on small perch led to a decreased abundance of large perch, attributable to the compensatory effects of intraguild predation and cannibalism.Résumé : Alors que les effets des prédateurs introduits sur les populations de proies dans les écosystèmes aquatiques ont souvent été étudiés, les interactions entre les prédateurs eux-mêmes sont moins bien connues. Dans une expérience à l'échelle d'un lac, nous avons ajouté un prédateur supérieur non indigène (« pikeperch » (Sander lucioperca)) à un peuplement de poissons ichtyophages résidents (la perche (Perca fluviatilis) et le grand brochet (Esox lucius)).

1. Antipredator behaviour by the facultative planktivorous fish species roach (Rutilus rutilus), perch (Perca fluviatilis) and rudd (Scardinius erythrophthalmus) was studied in a multi-year whole-lake experiment to evaluate species-specific behavioural and numerical responses to the stocking of pikeperch (Sander lucioperca), a predator with different foraging behaviour than the resident predators large perch (P. fluviatilis) and pike (Esox lucius). 2. Behavioural responses to pikeperch varied greatly during the night, ranging from reduced activity (roach and small perch) and a shift in habitat (roach), to no change in the habitat use and activity of rudd. The differing responses of the different planktivorous prey species highlight the potential variation in behavioural response to predation risk from species of similar vulnerability. 3. These differences had profound effects on fitness; the density of species that exhibited an antipredator response declined only slightly (roach) or even increased (small perch), whereas the density of the species that did not exhibit an antipredator response (rudd) decreased dramatically (by more than 80%). 4. The maladaptive behaviour of rudd can be explained by a 'behavioural syndrome', i.e. the interdependence of behaviours expressed in different contexts (feeding activity, antipredator) across different situations (different densities of predators). 5. Our study extends previous studies, that have typically been limited to more controlled situations, by illustrating the variability in intensity of phenotypic responses to predators, and the consequences for population density, in a large whole-lake setting.