Introduction
The impact of invasive mammalian predators on islands is well known (Atkinson Reference Atkinson1985, Jones et al. Reference Jones, Tershy, Zaveleta, Croll, Keitt, Finkelstein and Howald2008) and invasive species remain one of the key threats for the conservation of many bird species (Butchart Reference Butchart2008, Hilton & Cuthbert Reference Hilton and Cuthbert2010). On Gough Island in the South Atlantic Ocean, introduced house mice Mus musculus L. prey on the chicks of albatrosses and burrowing petrels. To date mice have been reported killing nestlings of three seabird species: the Critically Endangered Tristan albatross Diomedea dabbenena L., the Endangered Atlantic petrel Pterodroma incerta Schlegel, and the Least Concern great shearwater Puffinus gravis O'Reilly (Cuthbert & Hilton Reference Cuthbert and Hilton2004, Wanless et al. Reference Wanless, Angel, Cuthbert, Hilton and Ryan2007). Further evidence indicates that eggs and chicks of the endemic and Critically Endangered Gough bunting Rowettia goughensis Clarke are also preyed upon by mice on the island (Cuthbert & Hilton Reference Cuthbert and Hilton2004, Ryan & Cuthbert Reference Ryan and Cuthbert2008). For populations of winter-breeding Tristan albatross and Atlantic petrel the current levels of predation are unsustainable, with 65–80% of chicks killed each season (Cuthbert Reference Cuthbert2004, Cuthbert et al. Reference Cuthbert, Sommer, Ryan, Cooper and Hilton2004, Wanless et al. Reference Wanless, Ratcliffe, Angel, Bowie, Cita, Hilton, Kritzinger, Ryan and Slabber2012). There are further concerns and growing evidence that all winter breeding petrels on Gough are being affected by mice predation (Angel & Cooper Reference Angel and Cooper2006, Cuthbert et al. Reference Cuthbert, Louw, Lurling, Parker, Rexer-Huber, Sommer, Visser and Ryan2013). In this paper we describe observations of mice predation on chicks of the summer-breeding Atlantic yellow-nosed albatross Thalassarche chlororhynchos Gmelin and dark-mantled sooty albatross Phoebetria fusca Hilsenberg, the first records of such attacks on Gough Island for these two globally endangered species.
Study area and methods
Monitoring of Atlantic yellow-nosed albatrosses was undertaken at a long-term study colony (Cuthbert et al. Reference Cuthbert, Ryan, Cooper and Hilton2003) and three other sites that are part of a monitoring programme for the species (Cuthbert & Sommer Reference Cuthbert and Sommer2004). These four areas held 277–300 breeding pairs in the three seasons of study (2008/09, 2009/10 and 2010/11). Monitoring of chicks was made on at least weekly intervals at the long-term study colony from hatching until all chicks had fledged, with visits to the three other areas after the peak of egg laying, after hatching and shortly before fledging. Dark-mantled sooty albatross monitoring was conducted via scan counts of coastal and inland cliffs (Cuthbert & Sommer Reference Cuthbert and Sommer2004) that held 212–238 breeding pairs during the three seasons, with further counts in late December and mid-March of numbers of hatched and fledging chicks. Around 50 nests and chicks were checked during 2008/09 and 2009/10 from both coastal and inland areas with monthly nest checks in the 2009/10 season. Further incidental observations on sooty albatross chicks were made during ongoing plant eradication work on the island's sea-cliffs.
Results
The fresh carcass of a sooty albatross chick was discovered at a nest on 15 December 2008. The chick had round open wounds to the skin of its belly and the abdominal cavity was exposed with the intestines visible (Fig. 1). Other than the observed wounds, the chick was undamaged and appeared to be in good body-condition, with a mass of 0.45 kg and internal fat deposits. Directly underneath the chick and located within the chick's nest was the entrance to a mouse hole (Fig. 1). A second dead sooty albatross chick was found during the same period and on the same stretch of cliffs. The body of this chick was too decomposed to assess patterns of wounds, although it was clear that the skin had been removed from behind its neck. Fledging success (number of fledging chicks/number of hatched chicks) of sooty albatrosses during the three seasons was 56% (n = 87 hatched chicks), 69% (n = 78) and 46% (n = 82) in the 2008/09, 2009/10 and 2010/11 seasons, with an overall mean of 57.1 ± 11.5%.
a. The carcass of a sooty albatross chick with wounds to its belly and abdomen and exposed viscera, and b. location of the mouse hole within the nest and beneath the carcass of the chick.
An Atlantic yellow-nosed albatross chick was found dead on 15 January 2010. This chick had hatched on the 22 December and weighed 1.7 kg at the time of death. The carcass was fresh with very few fly maggots and other invertebrates present and no visible sign of decay. Wounds were found on the lower belly, exposing the viscera (Fig. 2). The nest mound had a mouse hole in the cup and a mouse hole at the base of the pedestal, with a well-used track leading to it (Fig. 2). This predated chick was one of 52 hatched chicks in the long-term study colony where regular (more than twice weekly) checks were made in the 2009/10 season, suggesting mice preyed upon a minimum of 2% of chicks in this season. Fledging success of this species was 53% (n = 194 hatched chicks), 64% (n = 201) and 82% (n = 175) in the 2008/09, 2009/10 and 2010/11 seasons, respectively. However, there was considerable variation between the four sites in rates of survival during the 2009/10 breeding season with a coefficient of variation (CV) of 0.631 in comparison to CVs of 0.101 and 0.115 in the other two years (Table I). The high variation in the 2009/10 season was due to one site with very low chick survival (Table I). Inspection of a sample of Atlantic yellow-nosed albatross nest mounds (n = 86 at two of the four sites) in the 2009/10 season indicated that 34% of nests had mouse holes in them, of which eight had holes into the nest cup.
a. The carcass of an Atlantic yellow-nosed albatross chick with wounds on the belly and exposed viscera, and b. location of the mouse hole within the nest mound and entrance of the mouse hole at the base of the pedestal.
Values of fledging success of Atlantic yellow-nosed albatrosses on Gough Island with sample size of hatched chicks in parentheses for the four monitored sites and three breeding seasons, with the overall mean, standard deviation (SD) and coefficient of variation (CV).
Discussion
This is the first time that mice have been reported to prey upon summer-breeding albatross species on Gough Island and the first records, of which we are aware, of mouse preying on Atlantic yellow-nosed albatross chicks and second observation of mice attacks on sooty albatross chicks (Jones & Ryan 2009). The wounds to the stomach and rump, and exposed viscera were identical in both location and character to wounds observed on Tristan albatross and Atlantic petrel chicks attacked by mice on Gough Island and to wounds on wandering albatross Diomedea exulans L. and dark-mantled sooty albatross chicks on Marion Island (Cuthbert & Hilton Reference Cuthbert and Hilton2004, Wanless et al. Reference Wanless, Angel, Cuthbert, Hilton and Ryan2007, Jones & Ryan Reference Jones and Ryan2010). The apparent rarity or absence of mice preying upon summer breeding species on Gough Island led Angel & Cooper (Reference Angel and Cooper2006) to conclude that the impact of mice was likely to be low for summer-breeders, in contrast to the situation for winter-breeders. This appears to be the case for both the yellow-nosed and sooty albatross on Gough Island, where, despite the observed predation events, overall rates of chick survival are relatively high and in the range previously recorded for the species on Gough and other islands (Tickell Reference Tickell2000, Cuthbert et al. Reference Cuthbert, Ryan, Cooper and Hilton2003). However, it is apparent that on a local scale mice may be detrimental, with only 7% of Atlantic yellow-nosed albatross chicks surviving to fledge at one site. It is also important to note that the scavenging behaviour of sub-Antarctic skuas Catharactica antarctica Lesson and southern giant petrels Macronectes giganteus Gmelin on Gough Island means that without frequent nest checks many chicks could have been consumed and “disappeared” leaving no evidence that they were originally preyed upon by mice. Consequently, levels of mouse depredation of sooty and yellow-nosed albatross chicks on Gough Island could be higher than suggested by our observations.
House mice have been reported to prey on chicks of passerines and small petrels in other studies (Fugler et al. Reference Fugler, Hunter, Newton and Steele1987, Ainley et al. Reference Ainley, Henderson and Strong1990, Campos & Granadeiro Reference Campos and Granadeiro1999, Michelsen-Heath & Gaze Reference Michelsen-Heath and Gaze2007), however it is only on Gough Island and on sub-Antarctic Marion Island (Jones & Ryan Reference Jones and Ryan2010) that house mice are known to prey upon chicks of albatross species. At both Marion Island and Gough Island predation events appear to be associated with house mice utilizing albatross nest mounds. Mice on sub-Antarctic islands are known to utilize burrows in order to provide a warmer microclimate and their burrows are disproportionally found in wandering albatross nesting mounds in comparison to other habitats (Avenant & Smith Reference Avenant and Smith2003). These authors attribute the high proportion of mice burrows around albatross nests to the high densities of soil macro-invertebrates in these manure enriched sites, as macro-invertebrates are a preferred food item for house mice (Gleeson & Van Rensburg Reference Gleeson and van Rensburg1982, Jones et al. Reference Jones, Chown and Gaston2003). An unfortunate by-product of this site selection by house mice is that it may have contributed to the predatory behaviour of mice on Gough and Marion islands, as such burrows will place mice directly underneath albatross chicks where chicks are most vulnerable.
Given the evidence for mice preying on two additional species, further monitoring of breeding success and causes of nesting failure are warranted for these and other species to determine if this behaviour is becoming more widespread and of greater severity on Gough Island. Proposed conservation management actions to eradicate house mice from Gough Island in order to remove a key threat to the Critically Endangered Gough bunting and Tristan albatross and Endangered Atlantic petrel will have further benefits in eliminating this cause of breeding failure for the Endangered Atlantic yellow-nosed albatross and dark-mantled sooty albatross.
Acknowledgements
We thank the South African National Antarctic Programme, University of Cape Town and Royal Society for the Protection of Birds for providing logistical support in the field and the Tristan da Cunha Conservation Department for granting permission to work on Gough Island. Funding was provided by the UK Government's Overseas Territories Environment Programme (OTEP). Thanks to M. Nysschens for assistance with albatross monitoring. The constructive comments of the reviewers are gratefully acknowledged.
