Stoats (Mustela erminea) were removed from three islands in Fiordland with traps at densities of one trap per 3.6Á11 ha to determine whether stoats could be eradicated with traps at these densities, and how long eradication might take. Stoats were successfully eradicated from the two of the three islands within 4 months of the start of trapping. In a separate experiment, stoat traps were also set and checked for 4 years on 19 small islands at varying distances offshore, and the relationship between the number of stoats caught, the distance offshore and the length of adjacent mainland coast documented. From these data, we estimated the likely rate of stoat re-invasion onto large islands in Fiordland such as Resolution, Secretary and Coal, which are candidates for future stoat eradication programmes. We conclude that permanent trapping networks on large islands could keep stoat numbers at levels low enough to allow the recovery or re-introduction of endangered wildlife.

Introduced mammalian predators, in particular rats (Rattus spp.), are a major threat to New Zealand bat populations. Aerial application of the toxin sodium monofluoroacetate (1080) is currently the most costeffective method of controlling rats across large spatial extents. Lesser short-tailed bats (Mystacina tuberculata) may be vulnerable to secondary poisoning from 1080 because they feed on invertebrate prey on the ground that may have consumed toxic bait. We monitored individually marked bats before, during, and after an aerial 1080 operation in the Eglinton Valley, Fiordland, in December 2014 from a population that has been monitored since 2008. No symptoms of sub-lethal exposure in free ranging bats were detected and survivorship was high: 764 of the 771 marked bats (99.1%) recorded in the pre-monitoring period were still alive one week after toxin application and a record number of 1731 marked bats were recorded emerging from a single roost tree in January 2015. One bat pup was found dead under a roost tree and 1080 was detected in muscle tissue. Any immediate impact of 1080 was assessed as minimal because the calculated annual survival rates were high (91.5%). We conclude that survival of the population was likely enhanced by the large scale 1080 operation.

The Fiordland crested penguin Eudyptes pachyrhynchus is one of the least studied crested penguin species, with indications the species has a declining population, which would be in line with the historic and contemporary trends for most crested penguins. To determine the current population trend of the Fiordland crested penguin, a number of monitoring programmes using both abundance counts and demographic modelling approaches were carried out between 1990 and 2010 in the northern half of the species' range. A 2.6% ± 0.8% annual decline rate of active nests was detected at 14 monitoring plots, and the number of nests along two coastlines declined annually by 1.2% and 2.6%. Population matrix models using site-specific demographic rates for the species at two South Westland sites indicated contrasting population trends, with one site increasing by 1.6% annually and a second site decreasing at 0.3% annually. Due to concerns about the reproductive parameters used in the model, the trajectory indicated by the nest-chick data was deemed more robust and should be used to inform management. Six potential threats to Fiordland crested penguin were reviewed against the detected population trend and specifically adult survival, but it was determined that there is insufficient understanding about the species, particularly its foraging ecology and effects of fishing and terrestrial predation, to confidently identify the key threats. Therefore, the recommended management action is to address these knowledge gaps.

No abstract