We hear the birds even before we reach the island: a raucous, unconducted symphony of screeches and whistles. Tui, bellbirds and saddlebacks contribute to the sound, but it is the kaka that stand out: their calls, like their name, loud, staccato and unmistakable. A burst of red flashes out among the foliage as one lands on a branch nearby. Then a babble of even louder squawks erupts as another kaka drops in and a playful squabble ensues. One bird departs and the other, perhaps a mate, soon flies off in pursuit. Although only 20 minutes by boat from Paraparaumu, the wildlife sanctuary of Kapiti Island seems more like a different planet—one where birds rule. This happy state of affairs exists because, of the various introduced predators present on the mainland, Kapiti has had but two representatives: both rat species and both now eradicated by poisoning. Birds exult and thrive in the pest-free environment. Surf hisses and rattles on to the stony beach, washing over our boots as we leave the boat. Peter Daniel, Kapiti's resident Department of Conservation ranger, welcomes us to the island. As we enjoy a picnic lunch, an inquisitive kaka hops on to a branch overhead, giving us a sly sideways glance. Without hesitation, the bird flaps down and tries to steal a sandwich. We laugh at its impertinence. For the past 30 years, kaka and other birds on Kapiti have received supplementary food supplies. Daniel regularly replenishes a large trough of sugar solution near his house. Tui and bellbirds jostle for position on the edges of the tray, dipping in, then lifting their heads with the sweet water dripping from their beaks. The birds show neither fear nor apprehension, despite the dozen or so people chattering and looking at them. But the whirr of larger wings behind makes the smaller birds scatter as a kaka lands on the tray. Kaka indisputably head this feathered nation. The bird then perches on a woman's head, the better to steal a piece of cheese. As its claws scrabble for a hold in her hair, she grimaces in a mixture of pain and delight. Kaka are ridiculously easy to photograph here, as they preen and prance unafraid. There is no need for the stealth and cunning necessary to get within cooee of mainland birds. On the few larger forested offshore island sanctuaries, kaka scream, chatter, clown around and almost fall over themselves in their eagerness to interact with human visitors. Yet on the mainland, the behaviour of kaka is entirely different. The birds are rare, more often heard than seen, and largely indifferent to humans. Daniel has tramped in Whirinaki Forest every year over the last decade. "I always hear kaka in Whirinaki, but not once have I got closer than 80 metres to one. It is such a contrast to Kapiti." One Kapiti kaka could even be picked up by his beak, says Daniel. "He'd let you lift him up, screeching, play-fighting with his claws and pretending to be annoyed." For mainland kaka, every day is a fight for survival. The birds can spend up to eight hours a day just gathering sufficient food to meet their requirements. But the greatest threat to the birds is from predators over the period when females incubating eggs or rearing chicks are confined to their nest holes. As a result of this mortality, males can outnumber females by six to one. Peter Wilson of Landcare Research in Nelson has examined kaka populations on all our forested islands and compared the numbers of birds with the range of predators present on each island. Where there are no stoats, kaka populations are healthy, even if cats and rats are present. Wherever there are stoats, kaka are in trouble. On Kapiti, there are over 1000 kaka in 1500 hectares of forest. On d'Urville Island—at the same latitude and with 8000 hectares of forest, but with stoats as well—there are only about a dozen kaka left. Unless the predator problem can be solved, kaka will eventually be lost from mainland forests and islands like d'Urville. By contrast, on stoat-free islands (including Little and Great Barrier, Kapiti, Hen, Stewart, Mayor, Codfish, Ulva and Nukuwaiata) kaka populations are growing. Plenty of food and less lethal predators mean the birds are able to breed successfully. However, when these islands reach their carrying capacity and run out of room and food, emigrating kaka will have to face the dangers associated with mainland life. Kaka are members of the 330-strong parrot family, also represented in New Zealand by the flightless kakapo, the alpine kea (see New Zealand Geographic, Issue 24) and several species of parakeet. Along with their indigenous relatives, kaka have been in decline since the onset of human occupation, due to predation and competition from introduced animals, forest destruction and hunting pressure. Kaka are now restricted to larger forested islands, and North Island kaka are rarely found outside the large forest tracts of Pureora, Whirinaki, Tongariro, Urewera and Kaimanawa. South Island kaka are seen in moderate numbers in Fiordland, Mt Aspiring and Westland National Parks and are still common on Stewart Island. The kaka is a large bird, standing about 46 cm tall and generally coloured dull olive-brown. Its most striking features are a large arched beak, dark, inquisitive eyes, scaly feet and the scarlet feathers on the underwings. North Island kaka have less grey in the feathers on their heads than their South Island counterparts do, are a centimetre shorter and weigh 100 grams less. Accordingly, they have been classified as a separate subspecies, Nestor meridionalis septenrrionalis, rather than the Nestor meridionalis meridionalis of the South Island. Although kaka are the same height as their close relative the kea, kea weigh close to twice as much. Male birds are larger than females (South Island range: males 525-640 g; females 430-550 g) and their upper mandible is longer than that of female birds. Museum specimens show that white, yellow and red varieties once occurred. Kaka may live for up to 20 years, though on the mainland few will die of old age. Kaka feed on nectar, fruit, berries, seeds, insects and sap. They concentrate on one food, then move on to others as each comes into season. Much of their feeding takes place in the canopy, bringing them into direct competition with possums. South Island kaka also use their semi-brushed tongue to lick honeydew, excreted by scale insects, from beech trees. Unfortunately for kaka, honeydew is also favoured by introduced wasps, which compete with kaka for this high-energy resource. Research suggests that kaka satisfy their energy needs from nectar and honeydew, while seeds provide them with protein and insects and grubs with fats. The kaka uses its powerful beak to tear long strips of bark from trees in search of insects and sap. This destructive approach to feeding earned kaka the nickname of "tomahawk" in parts of Otago last century. To the annoyance of foresters, the "tomahawk" is quite capable of ring-barking pine trees. When sap-feeding, kaka make lateral incisions on the trunks of trees such as totara, northern rata and pohutukawa, licking up the sap as it oozes out. This type of feeding happens during spring and autumn; the timing is thought to be related to changing sugar levels in tree sap. After kaka have stripped bark in search of sap and insects, the tree is open to infection, often leading to decay and death. It is thought that when kaka were abundant they may have played a significant role in the renewal of the forest, hastening the demise of older trees and thereby making room for the young and vigorous. As nectar feeders, kaka also perform the associated role of moving pollen around the forest. [Chapter Break] Peter Wilson and a team of Landcare researchers have been studying kaka for more than 10 years in a possum-chewed, wasp-infested beech forest near Nelson Lakes National Park. Wilson believes that competition from possums and wasps for the same high-energy foods kaka depend on is limiting kaka breeding: "If kaka are denied high-energy foods in late summer and autumn, they will go into winter in poor condition and not reach breeding condition by spring." In addition, when a female is nesting, she relies on the male to forage for her, but if he is unable to procure sufficient food, the female will leave the nest to forage for herself, causing incubation failure. Wilson monitored 31 kaka for five years. During this time only two pairs attempted to breed and only two fledglings were raised—by one female. A year later, this female was killed, probably by a stoat. "Even for a long-lived parrot, successful breeding by one out of 31 kaka in five years is a mighty low reproductive rate," comments Wilson. So in 1989 he set out to investigate whether supplementary feeding would improve breeding success for kaka. However, unlike the feeding free-for-alls common on Kapiti, getting mainland kaka to accept supplementary food proved difficult: "We tried trays of honey, honey-water and huhu grubs. Nothing worked until we hung silk fuchsia flowers—imitating mistletoe, a favourite kaka food—in beech trees." Kaka immediately visited the flowers, and Wilson then trialled a range of food types in what he calls a "cafeteria experiment." Grapes, sunflower and pumpkin seeds, avocados, peanuts, honey, cheddar cheese, pine nuts, cashew nuts and a mix of Complan and Farex were all placed near the flowers. "Pine nuts proved a great hit, but they're outrageously expensive!" says Wilson. He also attached hamster feeder bottles containing honey-water to trees, with the nozzles coming out of poinsettia flowers. Supplementary food was made available daily to kaka from three automatic feeders, and Wilson says the experiment was "a crashing success." Of 23 birds monitored, 39 per cent fed regularly, 22 per cent fed sometimes and 39 per cent sampled the supplementary food only once. Older birds seemed reluctant to try new foods, he says, while younger birds tended to be attracted to novel things. Have kaka been imprinting on humans, perhaps? Of particular interest to the researchers was the behaviour of a bird nicknamed Knuckle (see sidebar; page 104). Soon after the feeding programme began, the 15-year-old bird was taught to use the feeders by his mate. "Knuckle has bred with several different females in his life, but none of their eggs have hatched. Only since he has used the feeders have the eggs been successfully incubated," says Wilson. Although supplementary feeding does not trigger breeding, Wilson believes it may put birds in better health, so that when the unknown stimulus urges them to breed, they will have a greater chance of success. The breeding success of kaka varies from year to year, probably in response to annual variations in food sources. Without honeydew, mistletoe and other high-energy foods now taken by wasps and possums, South Island kaka fail to reproduce. North Island kaka may be more reliant on sap and heavy podocarp fruiting to trigger breeding. Mating occurs in spring, and the females make primitive nests in deep holes in tree trunks. The nest lining is a mixture of wood fragments chipped and chewed from the inside of the tree. Between two and five dull white eggs are laid in a hollow amongst the debris and are incubated for 24 days by the female. The male returns periodically to provide food for his mate. Kaka nests only become obvious once the grey, downy chicks reach 10 days old, when the smell of faeces may be detectable from 10 or more metres away. Chicks remain in the nest for 10 weeks, growing adult feathers before emerging flightless and noisy. They can spend up to a week on the ground before their feathers are ready for flight. From the moment the eggs are laid to the time that the chicks can fly is a staggeringly long three to four months—a time when female, eggs and chicks are extremely vulnerable to predation. [Chapter Break] It is midwinter and we are in the middle of the expansive Pureora Forest Park, a towering jungle of ancient podocarp trees that constitutes one of the few remaining homes for North Island kaka. Despite Swanndris and balaclavas, we are shivering as we huddle beneath ferns with a team of kaka researchers. Compared with the numbers of kaka this forest once contained, today's population is pitiful, yet Pureora is one of the best places on the mainland to see these elusive parrots. Terry Greene, the crew boss, sits with his head back, eyes straining to see a kaka he knows is perched near the top of a rimu tree. Strung high among these giant trees is an almost invisible mist net. It has taken a full two days to erect the net. First, a suitable site is selected, where kaka are known to occur, and then a space for the net is cut in the forest floor and all small shrubs and ferns that might entangle the net are removed. Next, Alan Jones, one of the taller team members, uses his long arms to advantage with a slingshot. After several unsuccessful attempts, a fishing sinker with nylon attached is shot over an appropriately elevated rimu branch at one end of the clearing. Since some of these podocarp giants are 50 metres tall, it is a feat requiring Robin Hood prowess. The same procedure is repeated at the opposite end of the clearing. A thick rope tied to the nylon is then hauled across. From this rope, two more ropes hang vertically, with the net suspended between. Once hoisted aloft, the net floats near the top of the canopy, and a pulley system enables it to be raised and lowered quickly. The researchers hope to catch at least 20 kaka, attach radio transmitters to the birds and follow them through an aerial 1080 possum-control operation to check that they survive. Possums threaten this majestic forest, and the Department of Conservation is using carrots laced with the poison 1080 to make a dent in their numbers. The use of 1080 is a successful, if controversial method of possum control. There are fears that the inquisitive kaka could be at risk from eating poisoned baits. Research shows that some kaka, particularly juveniles, will eat baits, but hunger might tempt older birds, too. Attaching radio transmitters to the birds is the only sure way of knowing if kaka will survive the possum blitz. For the crew, finding even two dozen birds is a needle-in-a-haystack challenge. The monitoring area is 24,600 hectares, and the birds are wary, unapproachable and can easily fly 50 km in a day. They also like to play hide-and-seek with would-be captors. "Trying to catch one of these cheeky, screeching, elusive birds is difficult at best, and downright foolhardy at worst," says Greene. "What's more, they seem to learn from their mistakes. I've seen one kaka we've caught telling a new group about the dangers of mist nests in no uncertain terms!" High in the rimu, a kaka calls. Greene presses a button on his CD player, and from a speaker hidden up in the trees a recorded kaka call scraarks out. Above, the real kaka calls again. The CD responds with a whistle. The kaka remains silent, watchful, suspicious. We sit below, motionless in the biting cold, anticipating. Nothing moves. Greene conjures another whistle from his artificial kaka, quieter this time. A softer, social "howdee doodee" contact call. Silence. Then suddenly the kaka swoops from its perch towards the clearing. It descends, but is still above the net. It clears the top and lands on the opposite side. We're disappointed. The thought of capturing a kaka had set the adrenalin flowing. Still, we can hardly expect to be so lucky on our first day when the crew has spent weeks doing this. The chess game begins again, with Greene playing canned calls on another speaker at the far side of the net. Curiosity rather than aggression brings kaka towards a net, Greene explains. We move, the kaka moves; we wait, it waits, neither sure of what will happen. Today it's a stalemate. The kaka is curious but never flies low enough to chance getting caught in the net. After several hours in the numbing cold we give up. The net is lowered and carefully packed away. The fine filaments can be easily caught and tangled, creating a mess that takes hours to undo. It's a far cry from the days when one Maori hunter sitting in a tree could snare dozens and sometimes hundreds of kaka in a day. Catching kaka in Pureora is a military-sized project requiring military-like commitment from the crew, who work in 10-day stints, mostly camping or staying in huts within the forest. They have 500 kg of gear including tents, climbing ropes, harnesses, cookers, slashers, bird-banding gear, mist nets, pulleys, slingshots, tape recording equipment, speakers, a CD player, transmitters, food and packs. The gear is flown in by helicopter, and a base set up for the crew. Winter daylight is short. Wet and miserable or cold and freezing are the only weather cycles. Greene says some days it is so cold that the bananas freeze inside the hut. "Temperatures drop below zero, and when you are hunkered down under fern fronds at dawn, waiting for a kaka to call, you need lots and lots of clothing. About three layers of trousers and six layers of tops just about keep out the chill!" But the numbing cold is quickly forgotten when a kaka is caught. The crew swings into practised action, lowering the net and quickly removing the bird. It is a stressful time for both bird and captor. "You need to grab the head and feet to immobilise them," says team member James Fraser. "The beak is like a can opener, and the claws are just as lethal." Once out of the net, the bird is put in a canvas bag—the darkness helps calm it. Over the next half hour, measurements are taken of wing length, weight, tail length and beak dimensions. These measurements help determine the gender of the bird, as males are larger and heavier than females. Of the 21 birds eventually captured, only three are females—confirmation of the dire state of the mainland kaka sex ratio. The birds are colour banded and a radio transmitter fitted, using a thin braided-nylon harness around the bird's chest and wings. The harnesses are designed not to restrict the kaka while flying, and a weak link is built into each harness so that it will eventually fall off. Each transmitter emits a different frequency, giving each bird its own signal which can be tracked on the ground or from the air. On one occasion the crew captured four birds in rapid succession. "We had these four birds in canvas bags, pegged to a tree, looking like wriggling Christmas stockings," says Greene. "The only problem was they were all poking their beaks through the bags, trying to bite their way out. There is nothing like the imminent loss of a bird to speed the measuring process up!" Given the impenetrability of the forest, it is impractical to try and follow the birds entirely on foot. Instead, an aircraft flown by Sid Marsh, a commercial pilot as well as a member of the kaka crew, is used to skim low over the tree tops. Two large aerials attached to the plane's wings allow kaka to be followed through the forest, using a receiver in the cockpit. The birds' positions need to be determined frequently to ensure they are moving—and therefore alive. After following the birds for six weeks (at which time the 1080 carrots were no longer toxic) the crew were certain that all monitored kaka had survived—the best possible result and an enormous relief to everyone concerned. As far as kaka are concerned, the way is clear for possum control by 1080. [Chapter Break] Although a kaka recovery programme as inaugurated in 1996, it will be at least three to four years before results are seen, because of infrequent kaka breeding. Even on pest-free, offshore islands, a good breeding season occurs only once every two to four years. In the first year, $250,000 was budgeted for the programme, and Terry Greene says most of this money was directed towards research. "Until we have answers to some critical questions, we won't know how to direct further money into management." Keeping populations of kaka on islands is regarded as an insurance policy in case kaka become extinct on the mainland. However, islands like Kapiti are not pristine environments, and even on islands the breeding success of kaka is cause for concern. "We can't afford to do nothing about kaka on the mainland and just rely on island birds to prevent extinction of the species," says Greene. "Kaka are an integral part of our mainland forests—without them they would be much poorer places." The forests are already in a parlous state, and it will be a long time before kaka are able to lose the suspicious nature necessary for survival on the mainland. Only with money, management and perhaps a miracle will kaka again reign in our treetops.