The Australian magpie (Gymnorhina tibicen; magpie hereafter) was introduced to New Zealand from 1864 in an attempt to help control pest invertebrates in pasture. Since the introductions, magpie populations have gradually increased and they are now widely distributed throughout most of the country. In New Zealand there are many reports of magpies attacking and even killing many other birds; however, the reasons for the attacks are largely unknown. As such, the aims of this thesis were to: (1) document and describe aggressive interactions between magpies and other birds to determine the extent and context of this behaviour, and (2) examine potential explanations for the interspecific aggression displayed by magpies. Accounts of magpie attacks on birds in New Zealand were collated from literature and a survey of the public, and then summarised to identify the frequency and characteristics of reported attacks on different species. Magpies were reported attacking 45 bird species. Species commonly found in rural habitats where magpies are abundant were attacked most; however, a directly proportional relationship between species abundance in rural habitats and reported attack frequency did not occur. Species consuming similar foods to magpies tended to be attacked more often, probably because these foods are more abundant in rural areas. Attacks on smaller birds regularly resulted in death, but deaths declined as victim weight increased. Non-contact and non-lethal contact attacks occurred throughout the year while attacks resulting in death occurred mainly during the magpie's breeding season (July to November). Regular observations on six territorial breeding groups and three non-breeding flocks of magpies over a year indicated that a range of bird species actively avoided foraging close to both these social groups cf. adjacent magpie-free control areas; fewer birds were also recorded flying near territorial breeding groups but not non-breeding flocks. Excluding harriers (Circus approximans: see below), only 8% of birds recorded close to territorial breeding groups were observed being attacked. Attacks were most frequent when numerous birds occurred near magpies and species recorded in the highest frequencies were generally attacked most. Territorial breeding groups attacked 39% of passing harriers. All attacks consisted of the victim being swooped or chased; no physical contact was ever observed. Both adult male and female breeding magpies were seen attacking other birds; juveniles in breeding groups sometimes supported adults but never initiated attacks. Magpies in non-breeding flocks were not seen attacking other birds, except harriers ( attacked in 17% of appearances). Thirty-eight nests of seven species were continuously monitored in rural areas over two successive breeding seasons using time-lapse video recording to determine the proportion that were preyed on by magpies compared to other animals in this ecosystem. Twentytwo lethal events were recorded over the two breeding seasons; magpies were responsible for only one of these. A further eight scavenging events were also recorded; magpies were not responsible for any of these. Harriers, ship rats (Rattus rattus), and cats (Fe/is catus) were the main predators and were responsible for 36%, 32% and 23% of lethal events respectively. A pukeko (Porphyrio porphyrio) was also recorded preying on nest contents on one occasion. These results imply that magpies are not important nest predators in rural areas. After territorial disputes with conspecifics magpies have been observed attacking other animals, suggesting that redirected aggression may be one reason for interspecific aggression. Residents from eight territorial groups were subjected to simulated territorial intrusions using live magpie and pigeon (Columbus livia) decoys. The magpie decoy was presented to subjects for 30min and then covered and aggressive responses toward the pigeon decoy by residents were recorded for a further 30min (Treatment l ). Two other experimental conditions were presented on each territory as controls; the two decoys were either both magpies (Treatment 2) or both pigeons (Treatment 3). No attacks on pigeon decoys were recorded during any test. In contrast, magpie decoys were always attacked by both male and female residents. In general, there was no difference in the percent of observations that residents were recorded at close proximity to the pigeon decoy between Treatment l and Treatment 3; however, both were lower than the percent of scans residents were recorded close to the magpie decoy in Treatment 2. From this experiment, there was no evidence that magpies redirect aggression onto other birds after territorial intrusions. It is possible, however, that the proportion of territorial disputes that result in redirected attacks are small, or only occur under highly specific circumstances, which were not recognised in this study. Finally, the hypothesis that magpies currently attack birds in New Zealand partly because they evolved in an environment that rewarded such behaviour was investigated. Data were presented that suggest magpies had evolved in the presence of many predators, numerous dangerous resource competitors, and in areas where resource availability was limited. Therefore, individual magpies that displayed high levels of aggression towards a wide variety of species would have been more successful at obtaining and defending quality nesting and foraging sites against both conspecifics and other birds. Consequently, more aggressive magpies would have had higher levels of fitness and ultimately had a selective advantage over time. In New Zealand many of the environmental constraints from their evolutionary past no longer appear to be present, yet attacks on other birds still occur. It is suggested that the generalised, apparently unnecessary, attack regime is maintained by magpies in New Zealand because (I) interspecific aggression is still adaptive for magpies, (2) magpies have not been in New Zealand long enough for high levels of interspecific agonistic behaviour to be lost from the population, (3) magpies currently face no selection pressure to stop attacking other birds, or ( 4) interspecific aggression is a pleiotropic by-product of conspecific agonistic behaviour. This research has determined that there are a range of mutually non-exclusive mechanisms that may explain interspecific aggression by magpies in New Zealand; however, circumstances under which attacks occur appear to be highly variable. Widely focused interspecific aggression by magpies may be best explained as a behaviour that was previously effective at securing resources or protection from threats during their evolution, which may no longer confer improved fitness in New Zealand magpie populations.