Lethal interactions among forest-grouse predators are numerous, motivated by hunger and carcasses, and their impacts determined by the demographic value of the victims

New vertebrate communities are emerging in Europe following the recovery of multiple native predators to highly anthropized landscapes where predator control is still prevalent. While the lack of reference points for these communities creates novel challenges for conservationists and wildlife managers, they also provide opportunities to further our understanding of species interactions. Despite a growing body of evidence, many aspects of interactions among predators remain poorly understood, impairing our ability to anticipate the effects of such changes in predator communities. Through a systematic literature review, we gathered all the available evidence concerning the existence, strength, and demographic impacts of lethal predator interactions among forest grouse predators in Europe. We found a highly interconnected predator community, with 44 pairwise lethal interactions among 12 taxa. Three of these resulted in some degree of population suppression of the victim, while another three did not. However, most interactions (38) have not been evaluated for population suppression. Additionally, we highlight how predators interact simultaneously with a large range of other predators and identified at least two further species likely suppressed through the combined impacts of multiple predators. We propose that interactions causing demographic suppression are characterized by impacts on individuals with high survival elasticity, and that they are motivated by food limitation and additionally, in mammals, by competition for carcasses. Predator interactions, and our still poor understanding of them, introduce large uncertainties to conservation actions based on the management of predator abundances, which should be carefully evaluated.We retrieved articles, book chapters and reports published up to July 2019 and available online that addressed interactions amongst forest grouse predators in Europe. We used electronic database search engines (Web of Science, Google Scholar, ResearchGate) for two searches. The first sought to review predator's diet and detect instances of intraguild predation, while the second sought to document the strength and impact of such interactions. The first was structured as "predator and diet" using the following search terms for predator: "lynx", "red fox", "golden eagle", "eagle owl", "goshawk", "buzzard" and "pine marten". The diet of the smallest predator taxa (diurnal raptors, owls, corvids, stoats and weasels) was not reviewed but instances of intraguild predation by these predators were included if available. The second search was structured as "predatorA AND predatorB AND type of interaction", where the type of interaction was a combination of: "intraguild", "predation", "killing", "suppression", "impact". Additionally, the terms: "stoat", "weasel", "mustelid", "corvid", "owl", "raptor" were added to the list of predators. Articles were first appraised from the title and abstract to include data on diet, killing or population impacts of our predators of interest and within the geographical boundaries of this study. Reference lists of retrieved articles were screened for additional studies. Where a study was cited by another but itself not available, the data was extracted if the study area, sample size and method were reported. This yielded a total of 240 studies. Eighty were later discarded multiple reasons. Seventy-nine were dietary studies that did not contain accounts of intraguild predation, focused on specific components of the diet (e.g. plants), were aggregated at a high taxonomic level (e.g. carnivora, passerines) or were already included in another study. The majority were of pine marten (36) and red fox (19) as well as three of both species. Eight were of common buzzard, the same number as golden eagle's, three of lynx and two of eagle owl. One other study by Björklund et al. (2016), which addressed the relationship of northern goshawks and common buzzards in Finland, was excluded because it considered a migratory population of buzzards (unlike in the rest of Europe). Hence, the study did not report demographic effects. CNW and LR performed the literature search. Results from diet studies of birds of prey are usually presented as the proportion of prey items belonging to a particular species relative to the total number of prey items across samples, while for mammalian predators these are generally presented as the proportion of samples (scats, kills, stomachs) with a given prey category in them. Occasionally avian diet was quantified in the latter manner. We refer to these as Frequency of Occurrence (% FO, hereafter) irrespective of the method used. Dietary studies often report data aggregated at different arbitrary levels (e.g. nests, territories, regions, seasons, multi-annual periods). Where available, the total frequency of occurrence per study was extracted from the papers or calculated from the data provided. To summarize the data, we averaged the frequency of occurrence of a given intraguild prey in the diet of an intraguild predator by method, weighting by the sample size of each study. For this, we only included papers with non-zero occurrence for a given species-pair. Note that dietary data does not allow discerning predation from scavenging. The data is presented per species pair and, where relevant, by method. Multiple species pairs were grouped in single sections were deemed reasonable (e.g. similarities in the species involved). For each species pair, dietary data is presented first, followed by killings, annual killing rates and any evidence of impacts on demographic parameters (i.e. breeding success, mortality). Evidence of population-level effects, direct or indirect, is presented last and qualitatively evaluated as the depression or, in their absence, release of a predator's abundance by another through lethal predator interactions. Interactions were deemed anecdotal when a single account was found across all the literature reviewed and empirical when there were two or more independent accounts