LAUSR

The eradication of invasive species from islands yields significant conservation returns. However, novel challenges continue to arise as projects expand in their scope, complexity and scale. Prey-loss and secondary poisoning were historically considered to have limited impact on native top-order predators when planning eradications, but this has rarely been tested quantitatively.We used a 10-year timeseries of Brown Skua Stercorarius antarcticus lonnbergi breeding surveys and isotopic dietary analysis on Macquarie Island to investigate how prey-loss and secondary poisoning deaths resulting from the eradication of an abundant invasive prey species, European rabbits Oryctolagus cuniculus, affected a top-order predator.Skua nest density declined from 7.14 nests/km(2) (95% CI: 6.01-8.27) in the presence of rabbits (pre-eradication) to 3.73 nests/km(2) (95% CI: 2.96-4.51) in the first 3 years after the eradication of rabbits, before showing signs of recovery in the 4 years thereafter. However, breeding success dropped from 1.01 chicks/nest (95% CI: 0.76-1.26) to as low as 0.38 chicks/nest (95% CI: 0.23-0.53) with little evidence of recovery.Secondary poisoning affected a greater number of skuas than anticipated prior to the eradication, including skuas nesting in areas where rabbits were not typically hunted as prey. We highlight that invasive prey often replace native prey in the diet of native predators rather than provide an additional source of food, and rapid eradication of non-native prey can have long-term impacts for predators, particularly when recovery of native prey is slow.Synthesis and applications. Monitoring programmes that complement large-scale eradication projects and address (a) trophic-driven declines in predator populations and (b) population-level impacts of secondary poisoning are integral to ensuring bottom-up effects of eradications are anticipated and adequately quantified. If prey deficits caused by eradication of invasive prey are expected to be severe but short-lived, supplementary feeding programmes may buffer against increased predation pressure on native prey and reduced breeding success of native predators. Alternatively, if the rapid recovery of native prey is not expected to occur naturally, breeding programmes and translocation of native prey prior to assist recovery of native predators should be considered to support ecosystem restoration.