Documenting the Damage

Stoats are probably the most abundant carnivores in the remaining forests of New Zealand, yet the question of the extent to which they threaten the surviving birds was for a long time hard to answer. Early ornithologists often commented on the devastating effects of stoats on native birds (O'Regan 1966), but could not provide the systematic observations needed to balance estimates of predation by stoats against multiple other losses (O'Donnell 1996).

The first and most obvious approach to the question, analysis of what stoats eat, had been done on a large scale by the end of the 1970s (see Figure 5.4). This work confirmed that stoats in New Zealand do indeed eat a great many birds, but that information was no help in trying to assess the effect of the removal of so many birds from a given population.

The only thing that could be said from studying the stoats themselves was that, at certain places and times, stoats eat so many birds that they must be potentially a real threat. New Zealand does not have enough small rodents to "buffer" the birds from the attentions of predators. In the northern hemisphere, good seasons for rodents tend to be easy ones for the birds (Figure 13.3a, b). But in New Zealand beech forests in summer, extra supplies of mice do not significantly reduce the proportion of birds eaten by stoats until the mice reach plague proportions (Figure 13.3c), and that has been observed only twice in 20 years (White & King in press).

Voles/ha Rodents/ha

Mouse Abundance Index (C/100 TN)

Figure 13.3 In England, as in Europe and North America generally, abundant rodents "buffer" birds against predation by weasels, both in open farmland (a) and in woodland (b). The graphs show that when voles are abundant, birds constitute a small part of weasel diets. The same is usually not true in New Zealand beech forests (c), where the rate at which stoats eat birds is unaffected by modest increases in mice. There, the buffering effect is possible only at extremely high densities of mice, and then only for a short period, because the periods when mice are abundant enough to distract stoats from taking birds are rare and brief (without the three right-hand points, the relationship is not significant). (Redrawn from Dunn 1977, Tapper 1979, and King & White, in press.)

Mouse Abundance Index (C/100 TN)

Figure 13.3 In England, as in Europe and North America generally, abundant rodents "buffer" birds against predation by weasels, both in open farmland (a) and in woodland (b). The graphs show that when voles are abundant, birds constitute a small part of weasel diets. The same is usually not true in New Zealand beech forests (c), where the rate at which stoats eat birds is unaffected by modest increases in mice. There, the buffering effect is possible only at extremely high densities of mice, and then only for a short period, because the periods when mice are abundant enough to distract stoats from taking birds are rare and brief (without the three right-hand points, the relationship is not significant). (Redrawn from Dunn 1977, Tapper 1979, and King & White, in press.)

Throughout the 1970s a great effort was made to document the biology of stoats, the necessary prerequisite for developing a workable management policy for them (King & Moody 1982). At that time it was impossible to disentangle the effects on birds of predation versus habitat destruction, and no hard data from New Zealand habitats were available to compare with the international predation literature that was still strongly influenced by the ideas of Paul Errington. King (1983b) suggested that the periodic summer irruptions of stoats she observed in 1976 and 1979 in the southern beech forests of Fiordland (see Figure 10.4) could be periods of greatly increased predation on birds, but confirming evidence derived from measurements of the productivity of the birds did not appear for another decade.

All the surviving native birds of mainland New Zealand live in habitats modified and invaded by a whole range of introduced alien mammals, and reduced and dissected to various degrees by human activities (especially deforestation). From the little that was known in the 1970s about the population dynamics of the remaining mainland bush birds, they appeared either to have accommodated themselves to the new conditions or to be already gone, to extinction or to permanent exile on safe refuges offshore (King 1984b). Doubt remained about the extent to which stoats could still be causing further losses, and it was not clear how we could protect the contemporary species from predation or, indeed, whether protection might now be too late.

Part of the reason for the doubt was that the only way to estimate the impact of predation on birds is to observe the interaction from the birds' point of view, and this is not easy to do. One of the earliest attempts was made in an intensively studied patch of bush near Kaikoura, in the South Island. There, stoats and common weasels together accounted for 77% of known nest losses of bush birds over three seasons, affecting native (101 nests) and introduced (48 nests) species equally (Moors 1983a, 1983b).

It was something of a surprise to learn that the introduced birds, which had evolved with predators and were supposedly better able to deal with them, suffered the same devastating losses as the native birds living in the same forest. Nonetheless, that information on its own does not tell us whether predation affects the numbers of either native or introduced species. Potts' partridge model (Chapter 12) is an example of how much information is needed to understand predator-prey interactions, and nothing like it was available for any New Zealand species until the 1990s.

A pioneering study of the South Island robin, done over several years at Kaikoura and on three mammal-free offshore islands, showed that the mainland robins survived by making a huge effort to compensate for their losses (Flack & Lloyd 1978; Powlesland 1983). By comparison with the island birds, which still live, presumably, in more or less their natural state, the mainland birds studied at Kaikoura for seven summers during the 1970s had much larger breeding territories (1 to 5 ha, instead of 0.2 to 0.6 ha); they started breeding much earlier and finished much later; their productivity was higher (clutch size three instead of two, three broods reared per season instead of one, and three juveniles fledged per pair per season instead of 0.14 to 1.1); nest failure due to predators averaged 55% per season instead of less than 10%; and annual adult mortality ranged from 23% to 37% instead of 17%.

Despite this enormous extra effort, the mainland birds lived at a much lower density (5 to 9 ha per pair, instead of 0.4 ha per pair) and not all available breeding habitat was used every season. The Kaikoura robins were able to compensate at that time, but elsewhere they and many other species could not. Most of those populations, including the one originally studied with such care and intensity at Kaikoura, are now at or beyond the threshold of extinction.

Under New Zealand law it has been illegal to kill or possess any of the threatened native birds since the Wildlife Act 1953, but that did not help them while forests (especially on the lowlands) were still being felled. Therefore, the first and clearest target for the young environmental movement of the 1970s was the fact that many important lowland forests and wetlands still remained unprotected (King 1984b). Some birds are acutely vulnerable to predation, others somewhat less so, but all are vulnerable to the loss of their habitat.

Throughout the 1980s, attention was focused on fierce arguments against large-scale logging of the remaining lowland forests, and the draining of wetlands. The main issue was that the system of reserves and protected lands then established did not adequately represent these lowland habitats, yet they supported the highest diversity of endemic birds. Many native species could not survive the loss or fragmentation of these habitats, with or without stoat control. At that time there was still a chance to defend, by radical action if necessary, lowland forests that were then still at risk of clear-felling (Wright 1980), so protesting against habitat destruction was, rightly, top of the conservation agenda.

Much progress was made during the 1980s in establishing legal protection for the lowland forests. Nevertheless, birds continued to disappear even from protected forests. The first edition of this book (1989:227) concluded cautiously that if "habitat protection can be assured, especially in areas where threatened birds still survive, then by analogy with the partridge work we may hope that some carefully planned stoat control work could be worthwhile." On what we know now, that must count as one of the understatements of the century (Wilson 2005).

Over the last few years, intensive field studies allied with computer models have begun to distinguish between the effects of predation and habitat loss. Understanding predation requires rigorous analyses and numerical models based on field-based, quantitative measurements of the mortality definitely due to each species of predator relative to all forms of natural mortality or failures in recruitment. Such measurements have been possible only since the development in the late 1980s of techniques able to measure productivity and mortality of eggs and chicks. The general conclusion is that the prime requirements for protecting native birds in New Zealand are similar to those for protecting partridges in England (p. 317): Habitat protection is absolutely necessary, but not by itself sufficient. Now that the need for protection of lowland forests is widely accepted, attention has turned back to the problem of predation.

Time after time, catastrophic failures in recruitment of young chicks of many of New Zealand's endemic birds have been traced to persistent predation. For some species, stoats are the main threat, and for other species, only an additional one (Table 13.1). For example, young kiwi are especially vulnerable to stoats, but uncontrolled dogs kill many adult kiwi. In beech forests in recent years, heavy seedfalls have been followed by irruptions of ship rats as well as of stoats, doubling the danger for birds that nest in tree holes. The yellowhead (an endemic forest passerine) was once very abundant but has now disappeared from more than 75% of its former range. Yellowheads are at particular risk from agile, tree-climbing stoats and rats (see Figure 6.4), and their productivity and mortality are both significantly affected in years of stoat irruptions (Elliott 1996a). As with the Wytham tits that nested in boxes accessible to common weasels (see Figure 7.1), the sitting females suffer much higher mortality than the males in years when the rates of nest predation are high. In New Zealand beech forests the increased danger can be predicted well in advance (O'Donnell & Phillipson 1996) and can be averted, at least locally, by intensive stoat trapping, except in years when rats are also numerous (O'Donnell et al. 1996; Dilks 1999; Dilks et al. 2003).

Likewise, the survival and nesting success of kaka (an endemic parrot) are seriously affected by predation on eggs, chicks, and nesting females by stoats (Moorhouse et al. 2003). In one forested study area monitored for 11 years, only four young kaka survived to independence out of 20 breeding attempts, and four adult females were killed on their nests (Wilson et al. 1998). The continuing slow decline of the endemic New Zealand long-tailed bat (one of only two surviving native land mammals) is also linked with post-seedfall irruptions of rats, probably accompanied by stoats (Pryde et al. 2005).

Table 13.1 Native Birds in New Zealand Potentially or Actually at Risk of Stoat Predation1, 2


Study area



North Island brown kiwi

Lake Waikaremoana


(McLennan et al. 1996;

Basse et al. 1999)

Haast kiwi, Okarito brown



(Pyle 2003)


South Island kaka

Nelson Lakes N P


(Wilson et al. 1998)




(Elliott 1996b)

Black stilt



(Dowding & Murphy 2001)

New Zealand dotterel


(Dowding & Murphy 2001)




(Lee & Jamieson 2001)

Yellow-crowned and orange-

South Island


(Elliott et al. 1996)

fronted parakeets

North Island kokako

North Island


(Basse et al. 2003)

Yellow-eyed penguin

South Island


(Alterio et al. 1998)

1. Levels of vulnerability refer to stoats only (+++ very high; ++ high; + other risks higher); all species listed are at serious risk, but for some, factors other than stoats are more important.

1. Levels of vulnerability refer to stoats only (+++ very high; ++ high; + other risks higher); all species listed are at serious risk, but for some, factors other than stoats are more important.

2. For a summary of the current status of New Zealand birds, see Wilson 2005.

All species of New Zealand's national icon, the flightless kiwi, are very vulnerable to predation by stoats on chicks up to the age of about 6 months (McLennan et al. 1996, 2004). A model developed from systematic field observations predicted that "the persistence of kiwi on the mainland is now largely dependent on the development of new technology for controlling stoats" (Basse et al. 1999). In the summer of a good mouse year, stoat densities in kiwi areas can reach up to 10 per km2; the model predicted that the maximum critical threshold density tolerable by kiwi is about two per km2. More recent work suggests that this figure is, if anything, too high (Barlow & Choquenot 2002), since it takes only a few stoats to do a lot of damage (McDonald & Murphy 2000; Gillies et al. 2003).

Monitoring of kiwi populations has confirmed that effective local reduction in the numbers of stoats can produce increases in counts of kiwi calls (Pierce & Westbrooke 2003) and real improvements in the productivity and survival rates of juveniles. On one study area beside Lake Waikaremoana, in Te Urewera National Park in the eastern North Island (Blackwell et al. 2003), the proportion of young kiwi to reach stoat-resistant size (800 g) increased from 4% to 58% by the third year of stoat trapping. In another area, Trounson Kauri Park north of Auckland, intensive predator control for 6 years, at a cost of more than NZ $1 million, raised the survival of kiwi chicks from (probably) less than 10% to an average of 38% (range 25% to 69%) (Gillies et al. 2003).

Most estimates of stoat densities in New Zealand forests are well above the two per km2maximum critical threshold density signaling danger to kiwi (Chapter 10). Where kiwi can be rigorously protected, they can hold their own provided immigration of stoats from surrounding areas can be minimized and dogs can be kept under control. Elsewhere, they are still rapidly disappearing; in some years and places, even intensive trapping cannot always reduce the numbers of stoats enough to save the kiwi chicks. For example, after a good fruiting season for native forest trees (especially rimu) in the winter of2002, ship rats and stoats became hugely abundant in the Okarito kiwi sanctuary (100 km2), on the west coast of the South Island. The removal of 353 stoats and 577 rats in the summer after the rimu mast (compared with 124 stoats and 61 rats the year before; J. Crofton unpublished) was not enough to prevent all 14 monitored kiwi chicks from being killed by stoats (Pyle 2003).

Even if they survive, young kiwi have a strong tendency to disperse out of all but the very largest protected areas. This means that the minimum viable area for a kiwi population must be large, at least 100 km2 (Basse & McLennan 2003)—and controlling stoats over such large areas is difficult. The best prospects are in places protected by natural barriers to immigration, such as in the Moehau kiwi sanctuary at the tip of the Coromandel peninsula. Elsewhere, the ultimate goal of controlling stoats to help the surviving, often aging breeding kiwi replace themselves has been hard to achieve (Gibbs & Clout 2003).

Species that have been long resident in New Zealand are especially vulnerable to predation because they have too few, or inappropriate, antipredator reactions compared with related species that have arrived more recently. New Zealand birds have always had avian predators that hunt during the day by sight, but mammalian hunters such as stoats and rats hunting at night by scent are something entirely new to them. The endemic black stilt and the recently arrived pied stilt have a common Australian ancestor, but whereas the black stilt has lived in New Zealand for millennia, the pied stilt appeared only about 200 years ago. Both nest in the same open gravel riverbed habitats, but the black stilt has lost several antipredator behaviors such as the broken-wing display, and its incubation period has become longer, extending the most vulnerable period of the brood (Pierce 1986).

Likewise, the South Island takahe (Figure 13.2) and the pukeko also have a common Australian ancestor, but the takahe has been in New Zealand since the Pleistocene and is now flightless, while the pukeko arrived within the last 1,000 years. In an experiment reported by Bunin and Jamieson (1996), pukeko responded much more strongly than takahe to a model stoat, and one cross-fostered takahe chick learned this behavior from its pukeko guardians while a parent-reared takahe chick did not.

Young New Zealand robins do not instinctively recognize stoats as a threat, but have to learn to associate the sight of a stoat with the alarm calls made by their parents. Fortunately, they can learn quickly. Wild young robins trained for only 5 minutes with stoat models and robin alarm or distress calls responded with more feather and body displays and stayed away from the nest for longer than other young robins (Maloney & McLean 1995).

Some species of small burrowing seabirds were extinguished from the main islands by predators; others have survived by nesting on predator-free offshore islands. Paradoxically, the Hutton's shearwater still nests in the alpine zone of the Seaward Kaikoura Mountains, in the northwest of the South Island. At this particular site, breeding seabirds are so abundant, and available over such a short period of the year, that the locally resident stoats can make no serious impact on their productivity. Remains of shearwaters were found in 785 of 788 stoat scats collected over three seasons (Cuthbert et al. 2000), but predation by stoats removed on average 0.25% of adults and 12% of chicks per year (Cuthbert & Davis 2002).

The main reason that the future of this particular colony is in no danger from stoats (Cuthbert 2002) is it is so large. At smaller colonies, the toll taken by predators can easily exceed the birds' production, just as colony size is important for Scottish seabirds visited by introduced American minks (Craik 1997). No doubt many small breeding groups of seabirds around the coasts of New Zealand have been wiped out in the past, unrecorded. For example, nesting colonies of New Zealand dotterels and yellow-eyed penguins are often threatened by predators including stoats (Moller & Alterio 1999; Dowding & Murphy 2001).

The effects of stoat predation can become very complex, and can extend across more than one trophic level. At Craigieburn, an isolated patch of mountain forest, low pollination rates and reduced fruit set of the native mistletoe is attributed to low densities of bellbirds, induced by stoat predation. Removal of stoats benefited the bellbirds: Nests in the trapped area in 2000-2001 were four times more likely to succeed than in the untrapped area (64% compared with 16%), although the increase in bellbird density was not enough to make any difference to the mistletoes (Kelly et al. 2005).

Was this article helpful?

0 0

Post a comment