In the native forests of the North and South Islands some 150 kaka are playing a key role in the conservation of their species. These birds are the focus of a nationwide study designed to determine if large-scale pest control operations can reverse the decline of kaka on mainland New Zealand. Over the past five years their movements have been followed and their every nest found and monitored, sometimes to the extent of being placed under 24-hour video surveillance. It is well known that introduced predators are the nemesis of many native species, and that controlling these pests is, in principle, a good thing to do. However, proving that a particular method of predator control benefits a given native species is usually very difficult. This is primarily because species which are vulnerable to introduced predators have often become so rare that scientists struggle to find sufficient individuals to study. Evaluating the effectiveness of predator-control methods is crucial because of the continuing threat predators pose to native species and because of the amount of time and money tied up in their control. In an effort to protect populations of threatened species on the mainland, the Department of Conservation has initiated a series of "mainland island" projects. The aim is to create "islands" of native biodiversity similar to those found offshore (such as Little Barrier, Codfish, Kapiti, etc), but on the main islands of New Zealand. The concept is simple: if pest numbers can be reduced sufficiently over a large enough area, populations of native species should recover. Unlike true islands, though, mainland islands are not immune from reinvasion, so pest management must be ongoing to keep the numbers of rats, stoats, possums and other pests below the level at which they have a significant impact on native species. Such management is a formidable drain on resources, but, short of installing predator-proof fences, it is the only way of conserving many native species on the main islands of New Zealand. The question is, does it work? In 1996 a group of DoC scientists set out to answer this question with respect to the kaka, a threatened native parrot that is now rare in most of the North and South Islands. Their rarity, combined with high mobility and cryptic coloration, make them difficult birds to study—in many respects as difficult to study as marine mammals. While the latter spend most of their time underwater, kaka spend most of theirs equally outside the range of easy observation, high in the forest canopy. Because of this elusiveness, radio transmitters have become a vital research tool. Attached to the birds with a backpack-style nylon harness, they make it possible to locate individual kaka in hundreds of hectares of forest. Modern transmitters can run for as long as three years, allowing a considerable amount of information to be collected on the wearer's movements, breeding and, ultimately, survival. Kaka are especially vulnerable to predation when nesting, so the primary objective of researchers has been to locate and monitor the breeding outcome of as many nests as possible. Finding a nest is in many respects the high point of this kind of study. Each nest contributes data which will eventually reveal if predator control benefits kaka or not. Kaka nest inside the hollow trunks and major branches of big trees, anywhere from 8 to 25 m above ground. To avoid the problem of having to climb the trees repeatedly to inspect the nests, small video cameras are installed, allowing the nest contents to be checked from the ground. Kaka have proved to be remarkably tolerant of such intrusions, and there hasn't been a single case of a female deserting a nest because of them. So far, the fates of 93 nests have been documented. Nesting success at sites with predator control has been 80 per cent or higher, compared to 38 per cent at three sites without predator control. Predation on nesting females has been 5 per cent at sites with predator control compared to 65 per cent at unprotected sites. There seems to be little doubt that pest control makes a difference. However, for mainland island-style pest control to be useful in conserving kaka we need to know if improvements in nesting success and female survival are large enough to actually reverse the decline of kaka populations. Female survival is the key to the health of kaka populations. Because female kaka spend long periods inside the nest chamber they are much more vulnerable to predators, such as stoats, than males are. Consequently, females are often in short supply, and it is this shortage that is driving populations to extinction. For a kaka population to increase, the numbers of young females surviving to adulthood must exceed the number of adult females lost to predation or other causes. This has been the case at all sites with predator control, but not at a single site without it. In other words, mainland island-style pest control does appear capable of actually reversing the decline of kaka populations. Numbers aside, perhaps the most tangible evidence of the difference predator control can make for a kaka population was the sighting of a group of young kaka in the village of St Arnaud, adjacent to Nelson Lakes National Park, in the autumn of 1998. These birds had fledged within the neighbouring Rotoiti Nature Recovery Project area, a mainland island and a kaka study area. It was the first time in many years that locals had seen a group of young kaka in the village. Since then, young kaka have become regular visitors to St Arnaud. Now that there is hard evidence that predator control can reverse the decline of kaka, perhaps flocks of the boisterous birds will start to be seen in other parts of the country, too.