Exotic species that invade remote islands, usually following human settlement, have had catastrophic effects on native biota. However, on islands it is increasingly feasible to eradicate key exotic species allowing extant native species to recover in situ or to return naturally. The practice of marooning threatened species on islands where the threat is absent, irrespective of whether the threatened species once occurred on the island, is well established. However, less focus has been given to the ‘island’ as the management unit on which to return extirpated species or related surrogates for extinct species. We use the example of Lord Howe Island as a case study to explore options for island restoration should the remaining critical exotic pests (rodents and perhaps owls in this case) be eradicated as planned. Lord Howe Island, in the south-west Pacific Ocean, is remote, biologically diverse, has a high degree of endemism, and was only discovered by humans in 1778. Consequently, the original and exotic biota and their interactions are all better known than for many islands with a more ancient human history. Two species of plants, nine terrestrial birds, one bat and at least four invertebrates have been lost from the island since 1778. One plant and two invertebrates could be returned as conspecifics. One plant and all the terrestrial birds that are extinct could be replaced by closely related species from elsewhere in the Pacific Ocean. Decisions on replacing extinct species with surrogates should be based on the taxonomic relatedness of the candidates for reintroduction: the same species before subspecies before genera, with functional replacement being a further filter on candidates that are not the same species. In our opinion, taxa with functional equivalence but without taxonomic relatedness would not be acceptable candidates for reintroduction.