Feral or free-ranging pigs have been a problem in Australia since the first years of European settlement, and they now occur in a wide range of habitat types throughout much of the continent. Feral pigs impact environmental, agricultural, and cultural resources, but they can also have commercial value from harvesting and recreational or subsistence hunting. It has been difficult to quantify many of the adverse impacts of feral pigs. Consequently, most management programs aim to mitigate actual, potential, or perceived impacts that have been inferred from observational studies and anecdotes, untested retroductive hypotheses, or observations generated from outside of Australia. Lethal control to reduce population density is the most common form of management, and it is often applied as a form of insurance, rather than to mitigate specific, measurable impacts. We suggest that future management programs should aim to quantify the effects of management actions on pig populations and the vulnerable resources that are the basis of real management objectives. There are important ethical and practical reasons for this approach, which should in turn enhance the efficiency and efficacy of management programs and help to ensure continued public and government support for ongoing mitigation of the feral pig problem.

The risk of non-target species being harmed by pest management activities can be a major impediment to animal control being implemented. This problem is particularly acute in complex faunal communities, where the pest occurs with functionally-similar species. Here, we present a framework for developing target-specific pest management tools for complex communities that helps managers reduce the risk of non-target species impacts. We applied the framework to the problem of poison baiting for feral pig control in the tropical rainforests of north-eastern Australia, and identified several potential methods to reduce adverse baiting impacts on non-target species. We evaluated the target-specificity of different baits and bait presentation methods using camera traps and sandplots to monitor animal-bait interactions in the field. As predicted using the framework, making baits available only at night prevented bait take by diurnal species, and bait burial substantially reduced bait consumption by nonfossorial species. The illumination of bait sites also reduced the foraging intensity of small non-target mammals, without inhibiting bait take by feral pigs. Two putative auditory repellents for small mammals were ineffective. Temporal variation in bait take by pigs and other species was unpredictable, and we were unable to exploit seasonally-variable factors to further enhance targetspecificity. The use of a starch-rich vegetable bait, rather than an omnivore bait, prevented bait take by dingoes and improved bait take by feral pigs. The use of a light plastic container to cover this bait prevented bait take by small omnivores when an uncovered free-feed was also provided. This latter baiting method renders nocturnal distribution, bait burial, and site illumination unnecessary, and provides a simple, target-specific baiting protocol that will allow effective feral pig control in the region for the first time. Our field results demonstrate the usefulness of the framework in our study site. The framework is similarly suitable for developing target-specific methods for delivering other chemical or biological agents to wildlife or pest populations, or for developing other pest control tools such as traps. The framework awaits further evaluation in other complex faunal communities.