114 New Zealand Journal of Ecology (2016) 40(1): 114-120 © New Zealand Ecological Society. DOI: 10.20417/nzjecol.40.13 Flexibility of diet of stoats on Fiordland islands, New Zealand Elaine C. Murphy 1 *, Craig Gillies 2 , Fraser Maddigan 1 , Peter McMurtrie 3 , Kerri-Anne Edge 3 , Maheswaran Rohan 4 , B. Kay Clapperton 5 1 Department of Conservation, Private Bag 4715, Christchurch 8140, New Zealand 2 Department of Conservation, PO Box 516, Hamilton 3240, New Zealand 3 Department of Conservation, PO Box 29, Te Anau 0640, New Zealand 4 Department of Biostatistics and Epidemiology, Auckland University of Technology, Auckland 1142, New Zealand 5 56 Margaret Avenue, Havelock North 4130, New Zealand *Author for correspondence: (Email: emurphy@doc.govt.nz) Published online: 28 August 2015 Abstract: The eradication operations to remove stoats (Mustela erminea) from islands in Fiordland provided an opportunity to assess the diet of stoats in areas with no rodents or with only mice (Mus musculus) available as mammalian prey. The carcasses of stoats trapped on Chalky Island in 1999, Secretary Island and the adjacent mainland in 2005, and Resolution Island in 2008 were collected and their gut contents analysed. On rodent-free Chalky Island, most of the stoats had consumed birds, mostly passerines. Stoats on Secretary Island (rodent- free) and Resolution Island (mice present) preyed mostly on invertebrates, particularly wētā (Orthoptera). On Resolution Island, mice were probably at relatively low densities, and were consumed by only 12% of the stoats. While average consumption of birds and invertebrates was lower for stoats at the mainland site, the only signifcant differences amongst the sites were the high bird consumption and low invertebrate consumption on Chalky Island compared with the other sites. The diet of male stoats was similar to that of female stoats on both Secretary Island and Resolution Island. Chalky Island male stoats were heavier than those on the other islands, while the females on the various islands had similar body weights. The variability in diet of stoats from these islands may in part refect the temporal and spatial differences between the samples. However, it demonstrates the adaptability of stoats, and their ability to survive without mammalian prey in different ways. It supports the hypothesis that differences in body weights of stoats are at least partly driven by variation in prey size and/or availability. Keywords: bird predation; body weight; invertebrate; mouse; Mus musculus; Mustela erminea; weta Introduction Stoats (Mustela erminea) were introduced to New Zealand in the 1880s in an attempt to control rabbits (Oryctolagus cuniculus). At the time, conservationists warned that stoats would have a devastating impact on native birds but they were ignored (King & Murphy 2005). Stoats spread rapidly throughout the South Island and even remote islands in Fiordland had been invaded by the early 1900s (King & Murphy 2005). In their native habitats in the northern hemisphere, stoats evolved as specialist predators of small vertebrates (rodents, birds and lagomorphs) and eat insects only rarely (King & Powell 2007). The main prey items of stoats in mainland New Zealand forests are usually house mice (Mus musculus) and/or rats (ship rats Rattus rattus, Norway rats R. norvegicus) (King & Murphy 2005). In non-forested areas, rabbits can predominate in stoat diet (e.g. Alterio & Moller 1997). Birds can also make up a high proportion of stoat diet (King & Murphy 2005; Smith et al. 2008). However, invertebrates can be important prey where mammals are relatively uncommon (e.g. alpine or open tussock/riverbed areas). They are a good alternative source of protein and other nutrients (Banjo et al. 2006). Orthoptera: Anostostomatidae (wētā) and Coleoptera (beetles) are the two most common orders of invertebrates found in stoat guts (Murphy & Dowding 1995; Purdey et al. 2004; Smith et al. 2005, 2008; Murphy et al. 2008). In New Zealand forest habitats, the occurrence of birds and invertebrates in stoat diet can be linked with the changes in availability of their main mammalian prey. In beech forests (Fuscospora spp.) huge synchronous production of seeds occurs every few years (mast years) and mice, birds and invertebrates become very numerous (King 1983; Murphy & Dowding 1995; Alley et al. 2001). In mast years, mice are a major prey of stoats but in non-mast years when mice are scarce, stoats eat more invertebrates and lagomorphs (Murphy & Dowding 1995; Smith et al. 2005). This ability to change their diet is also seen in podocarp forests, where seasonal changes in rodent abundance induce functional responses in stoat feeding behaviour (Jones et al. 2011). Reductions in rat numbers following control operations also affect bird consumption by stoats (Murphy et al. 1998, 2008; Clapperton et al. 2011). So what happens in the complete absence of rodents and lagomorphs? No rats, mice or lagomorphs have been recorded from Chalky Island or Secretary Island, while the only rodent now found on Resolution Island is the house mouse. Veale et al. (2014) suggested that the presence of mice on Resolution Island was the cause of the larger body weight in the resident stoats compared with those on rodent-free Secretary Island. They recommended a study of the diet of these stoats, to test this hypothesis. Predator-free islands are widely used as sanctuaries for threatened species and New Zealand has become very profcient at eradicating pests from islands (Parkes & Murphy 2003). Because of advances in stoat control, it was thought feasible to eradicate resident stoats and control their potential re-invasion on some islands in Fiordland (Elliott et al. 2010; King et al.