The complex pest: interaction webs between pests and native species

Transcription

1 The complex pest: interaction webs between pests and native species Chris R. Dickman Institute of Wildlife Research, School of Biological Sciences, University of Sydney, NSW 2006, Australia ABSTRACT Pests are often viewed as having consistently strong and negative effects on biodiversity values and agricultural productivity, especially if they have been introduced from elsewhere. However, pests do not necessarily have just negative impacts, and there is evidence that in some situations their effects can be beneficial. The positive effects of pests arise when they become deeply embedded in ecological communities and are involved in webs of direct and indirect interactions with other species. In this paper, I first outline the concepts of direct and indirect interactions, and then describe two case studies that illustrate how these interactions develop between pests and native animals. In the first case study, house mice (Mus domesticus) introduced to Boullanger Island in Western Australia have direct effects on a small dasyurid marsupial (Sminthopsis griseoventer boullangerensis), but also exert indirect negative and indirect positive effects on four species of insular skinks. In the second case study, high levels of activity of domestic house cats (Felis catus) in suburban bushland in Sydney are associated with reduced richness of bird species. However, high cat activity also appears to depress the rate of egg predation in above-ground bird nests, apparently by suppressing the activities of small egg predators such as rats. In these and other examples, control of the putative pest would have unintended negative consequences on some native species, making decisions about management more difficult. I propose a preliminary framework to predict the likelihood of indirect interactions occurring at different times after a pest has been introduced, at different places, and at different pest densities, and use this to suggest options for management of the pest species. Key words: pest, management, direct interactions, indirect interactions, house mice, cat Introduction Pest animals come in many shapes and sizes, from invertebrates such as crazy ants (Anopholepis gracilipes) and tiny plant thrips, to larger and more conspicuous taxa such as the cane toad (Bufo marinus) and European rabbit (Oryctolagus cuniculus). In Australia, different species of pest occupy virtually all of the continent s land area, its fresh and marine waters and, in the case of the 32 or so introduced species of birds, much of the aerial environment too. The economic and environmental costs of pests are enormous. Taking just nine well-known species of introduced mammals, plus the cane toad and European carp (Cyprinus carpio), McLeod (2004) estimated the overall cost of their impact in Australia to be some $720 million a year. This estimate was based on the lowest cost figures that were available, suggesting that actual costs to both the economy and environment may be higher still. Indeed, another 45 species of introduced vertebrates have been identified as pests in Australia and its dependent territories (Olsen 1998), and at least an order of magnitude more species of invertebrates can be considered pests at different times throughout the continent (CSIRO 1991). More species of pests arrive on the shores of Australia each year (Low 1999). These observations suggest that pests are easy to identify and represent clear targets for management. However, pests are not always what they seem, and controlling their impacts can often be problematic. For example, feral horses or brumbies (Equus caballus) are seen by some as being highly destructive to parts of the Australian high country environment, causing damage directly by consuming and trampling native vegetation and by assisting in the spread of weed species (NSW National Parks and Wildlife Service 2002). To others, wild horses are icons of the bush, with any attempt at controlling their numbers leading to hysterical over-reactions in the public and political domains (Chapple 2005). Similarly exquisite tensions exist over the use and management of deer, and also of native species such as kangaroos and wombats. Large kangaroos are considered to be pests by many pastoralists but are seen as valuable resources by the kangaroo industry; they are usually viewed favourably also by the broader public (Grigg 2002). Many other taxa enjoy an ambiguous status in the Australian community, including several introduced species of trout, European finches, native galahs (Cacatua roseicapilla), magpie geese (Anseranas semipalmata) and water-rats (Hydromys chrysogaster). In some local areas, even the widely-despised European rabbit and feral house cat (Felis catus) are seen as desirable newcomers (Low 2004). To make matters more complicated, some species that have achieved pest status after their establishment in a new environment have become threatened in their original habitat. In this situation, pest control could have terminal consequences for the species as a whole. The one-humped camel (Camelus dromedarius), for example, has declined in its native Saharan homelands and now occurs in free-living state only in central Australia. Should outback camels be targeted for pest control, their future as a wild species could not be guaranteed. Pp in Pest or Guest: the zoology of overabundance, edited by Daniel Lunney, Peggy Eby, Pat Hutchings and Shelley Burgin Royal Zoological Society of New South Wales, Mosman, NSW, Australia.

2 The complex pest These potential problems in identifying and managing pests are well-known and have been subject to considerable debate (e.g., Braysher 1993; Olsen 1998; Dickman and Lunney 2001). However, there is another problem emerging for pest managers that has been subject to much less scrutiny. This is the realisation that pest species are often deeply embedded within ecological communities, and are involved in webs of interactions with other species that may be affected by attempts at pest control. Such interaction webs are usually most obvious for native pest species because these have co-evolved for long periods with other members of their communities. However, introduced pest species might also engage extensively with native species, especially if they reach high numbers or substitute the ecological roles of native species that have been displaced. If pests interact strongly with other species, is management of their populations always likely to be beneficial for remaining members of the community? Will their management necessarily benefit the environment, pastoral production or other goals that may be sought? Although we can often gauge the obvious consequences of controlling a pest, such as by measuring the rise in yield of a crop or by documenting the increase in population size of a target native species, we usually do not know the broader consequences of our management actions (Caughley and Gunn 1996). Nor do we have a framework for predicting what will happen. Interactions between species are so pervasive and complex that we have very limited ability at present to make robust predictions about what happens when a pest is controlled. In this paper, I first discuss different kinds of interactions that occur between species, emphasising indirect interactions that occur when three or more species are involved. Two case studies are presented that highlight the problems of pest control if indirect interactions are not considered. I then propose a preliminary framework for categorising and predicting the likely importance of indirect effects for managers. Kinds of interactions Direct interactions The simplest types of interactions occur directly between two species, and result in positive (+), negative (-) or neutral (0) effects on each other s population size or performance. This yields six pairwise types of interaction, although the situation where each species has a neutral effect on the other is trivial and is not pursued here. Excluding this, five kinds of interaction are left, and are defined as competitive (-,-), mutualistic (+,+), commensal (+,0), amensal (-,0) and contramensal (+,-) (Dickman 2006a). This scheme describes the effects that two interacting species have on each other, and not the mechanisms that produce the effects. For the first four interactions the effect also describes the mechanism that achieves it, but for contramensal interactions (Arthur and Mitchell 1989) at least four mechanisms can be identified. These are predation, disease, parasitism and Batesian mimicry (Dickman 2006a), all of which result in one species obtaining a net benefit from the interaction (the predator, disease organism, parasite or mimic) at the expense of the other (the prey, disease/ parasite host or model). The rabbit is often targeted for control because of the damage that it does to soil, pasture, crops, plantation seedlings and native vegetation (Williams et al. 1995). However, the potential effects of culling rabbits on other species are more diverse, as depicted in Figure 1. In this Figure 1. Direct interactions that could potentially occur between the rabbit (Oryctolagus cuniculus) and other species in Australia. Arrows point towards the species affected, while the +, - and 0 symbols show the direction of the effect. Only some of the potential interactions are shown and, of these, not all would be expected to occur in one geographical area. These and other direct interactions are discussed in Newsome et al. (1997). Pest or Guest 209

3 Dickman Figure 2. Indirect interactions between (a) the dingo (Canis lupus dingo) and rabbit (Oryctolagus cuniculus) and green plants; (b) the dingo, red fox (Vulpes vulpes) and any species of quoll (Dasyurus sp.); and (c) the feral house cat (Felis catus), any species of honeyeater (Lichenostomus, Meliphaga, Melithreptus spp. and others) and the plants that they pollinate. Arrows point towards the species affected, while the +, - and 0 symbols show the direction of the effect. Solid arrows indicate direct effects between species, while broken arrows indicate indirect effects between species that do not interact in any direct manner. These and other examples are discussed in Newsome et al. (1997) and Glen and Dickman (2005). hypothetical example, rabbit control could be expected to benefit macropods via reduction in competition, to benefit arboreal species by reducing damage to shrubs and trees, and to improve conditions for soil crust organisms by reducing soil disturbance (amensal) effects. However, rabbit control could lead to declines in populations of native raptors and other carnivores if rabbits form a staple part of their diets, and to reduced activity of commensal burrow dwellers such as goannas when their rabbit-built shelters decay. If rabbits are associated in a mutualistic manner with weeds such as blackberry (Rubus fruticosus) - gaining benefit from being sheltered by the thorny clumps while providing the soil disturbance needed, in turn, for weed germination - control of rabbits could also lead to reduced weed cover. Of course, not all the relationships shown in Figure 1 would be expected in all situations. The figure simply highlights some of the less obvious consequences of rabbit control that might arise due to their direct interactions with other species. Indirect interactions These types of interactions take place only when three or more species are present, and can be recognised when one species affects the interaction between two or more others. Because indirect interactions include more of the species that co-occur in ecological communities, they are likely to be more representative of reality than simple systems comprising just two species. However, they are also more complicated to depict, and the numbers of potential interactions increase greatly with even small increases in numbers of species. Among the best-studied indirect interactions are trophic cascades, indirect commensalisms and indirect amensalisms; all have relevance to the management of pests. Trophic cascades occur when predators indirectly affect the population sizes or activity of species in lower trophic levels. In the example in Figure 2a, the dingo (Canis lupus dingo) suppresses rabbit numbers and has an indirect but positive effect on the biomass of forbs and other plants that are eaten by the herbivore. Effects on the lower trophic levels are not always positive, and may be neutral or negative depending on the numbers of trophic levels in the system and whether interactions are confined to species in adjacent trophic levels (Flecker and Townsend 1994; Sinclair et al. 2000). Indirect commensalisms occur when one species has an unreciprocated and indirect positive effect on another. In Figure 2b, for example, the red fox (Vulpes vulpes) is assumed to reduce quoll (Dasyurus sp.) populations directly by competition for resources. In the presence of the dingo, however, fox population numbers may be reduced, this interaction indirectly releasing quoll populations from competition. Indirect amensalisms are unreciprocated and indirect negative effects of one species on another. In Figure 2c, the feral cat is assumed to reduce the abundance of honeyeaters or other pollen vectors by consuming them, thus indirectly reducing pollination and seed set in plants. These kinds of interactions and other, more complex ones, are turning out to be extremely prevalent and influential in ecological systems (Newsome 1990; Morin 1999; Glen and Dickman 2005; Dickman 2006b; Dickman and Murray 2006). This in turn has important consequences for how we manage pests: rather than viewing them as undesirable entities to be controlled under all circumstances, we need to appreciate the flow-on effects on other species in the system and attempt control only if there is net benefit to biodiversity in doing so. 210 Pest or Guest

4 The complex pest Case studies House mice (Mus domesticus) on Boullanger Island In 1985, Fuller and Burbidge made the exciting discovery of dibblers (Parantechinus apicalis) on Boullanger Island, a small (25.9 ha) and mostly sandy outcrop not far off the coast of Jurien in Western Australia (Fuller and Burbidge 1987). Not only this, Boullanger Island was found to support another dasyurid marsupial, a form of Sminthopsis, that appeared to be taxonomically distinct from related dunnart species (S. griseoventer boullangerensis; Crowther et al. 1999, but see also Start et al. 2006), as well as insular and geographically restricted subspecies of the skinks Egernia pulchra and E. multiscutata. The island was also home to a very dense population of the introduced house mouse (Mus domesticus), with densities of this species achieving peaks in summer of 700 animals per hectare. Experimental manipulations of mouse numbers by Dickman (1992) showed that the dunnart population was strongly depressed in the presence of mice, and that fourfold increases in numbers could be achieved by mouse control. In view of the ecological similarity of the two species of dasyurids, it would be reasonable to postulate that dibblers also would be affected by mice. Given these observations and the generally destructive impacts of mice when at high density (Caughley et al. 1994), the prima facie case for management would be to attempt to reduce mouse numbers or even eradicate the population. However, preliminary monitoring of the responses of the island s entire terrestrial vertebrate fauna gave pause for thought. Detailed results of experiments carried out from have been presented by Dickman (1988, 1999, 2006b) and are summarised below; more recent monitoring has been described by Bencini et al. (2001); Miller et al. (2003), and Wolfe et al. (2004). Although removal of mice from parts of the island allowed a sustained increase in the numbers of dunnarts, it unexpectedly led to a short-term depression in the numbers of dibblers. Analysis of the diet of this species showed that it frequently ate mice, so depletion of the mouse population during the removal experiment deprived dibblers of part of their food base. Not only this, natural depression of the mouse population by dibblers and by another native predator, the barn owl (Tyto alba), appeared to partially release dunnarts from competition with mice, thus providing an indirect benefit to the dunnarts. As neither dibblers nor owls included dunnarts in their diets to any extent, the interaction between these species can be classified as an indirect commensalism. The removal of mice had other, unexpected effects. In the months following the start of the experiment, the depth of leaf litter increased by up to 24% in the mouse removal areas compared to control areas, presumably because mice were no longer eating newly-fallen leaves. The deeper litter in turn provided habitat for richer and more abundant populations of invertebrates. Capture rates of two small species of litter-dwelling skinks (Ctenotus fallens and Morethia lineoocellata) increased concomitantly by up to 35% in the mouse removal areas, reflecting both the increased shelter in the litter and access to richer sources of food there. By contrast, capture rates of two further species of skinks (Egernia pulchra and E. multiscutata) declined by up to 50% on the mouse removal plots, reflecting the aversion that these species show for foraging in dense litter. As both the island s marsupial species and owls seldom ate skinks, the depression of mouse numbers by these predators appears to have an indirect commensal effect on C. fallens and M. lineoocellata, but an indirect amensal effect on the two species of Egernia. In these circumstances, a program of mouse removal on Boullanger Island could be expected to have manifold consequences for the native terrestrial vertebrates and for invertebrate assemblages. It would be advisable to proceed slowly, and to monitor the numbers of both dibblers and the two species of Egernia to ensure that these species populations did not fall to low levels. House cats (Felis catus) in suburban Sydney Many studies show that feral house cats eat a broad range of terrestrial vertebrates, and suggest further that they can have strong predatory impacts on the population sizes of preferred species (reviewed in Dickman 1996). Evidence for effects of owned cats in suburban environments is less compelling, in part because it is difficult to separate catimpacts from the many other kinds of disturbances that influence populations of prey species when living near humans, and in part because impacts at the population level cannot be deduced simply from observations of diet (Barratt 1998; Gillies and Clout 2003). There is also some evidence that owned domestic cats eat many introduced species of vertebrates such as black rats (Rattus rattus), Indian mynas (Acridotheres tristis) and other small passerines (Reark 1994), so reducing potential competitive pressure between these species and ecologically similar natives. Nonetheless, there is no doubt that large numbers of native species are killed by owned domestic cats (Fitzgerald 1988; Meek 1998). There is also a groundswell among local government authorities, urban and rural residents that sensitive suburban areas near remnant bushland should be designated cat-free zones, that cat ownership should be regulated, and that domestic cats should be subjected to strict controls to reduce their predatory activities outside the home (Grayson et al. 2002; Baldock et al. 2003; Grayson and Calver 2004; Lilith et al. 2006). Although these considerations suggest that there is a strong prima facie case for cat management, there are, again, some contrary observations that warrant a closer look at how and where management should be imposed. These studies have been described by Matthews et al. (1999) and Dickman (2006c), and a commentary is given below. Matthews et al. (1999) constructed artificial nests with clay eggs moulded to resemble those of the eastern yellow robin (Eopsaltria australis), and placed these in trees in 24 patches of bushland throughout the Sydney region to identify nest predators and the intensity of any predatory attacks. The predation rate, calculated as the percentage of nests attacked by predators, ranged from 45% to 100% per patch. Birds were responsible for attacks in all patches, and black rats in 10; no nests were attacked by cats, even Pest or Guest 211

5 Dickman though cats were virtually ubiquitous. Although there was no evident relationship between predation rate and size of patch or position of the artificial nests within patches, there was a strong negative relationship between cat activity (an index derived by walking along tracks in each patch and counting cat faeces) and predation rate. Together with evidence that cats were eating all the identified nest predators, these findings imply that cats might actually reduce predation on nests located in trees. If so, this would be an example of a trophic cascade, with cats suppressing the numbers of smaller nest predators and thus indirectly reducing predation pressure on populations of tree-nesting bird species. Soulé et al. (1988) described a similar situation in which coyotes (Canis latrans) indirectly protected scrub-breeding birds in urban habitat remnants by controlling smaller bird-eating predators such as foxes and cats. Loss of coyotes from some patches allowed populations of the smaller predators to increase and cause local extinctions of scrub-breeding birds; this phenomenon was termed mesopredator release (Soulé et al. 1988). Of course, the situation is more complex than this because, in Sydney s bushland, cats will usually eat birds if they get access to them. Indeed, surveys of all bird species within patches showed that their numbers were correlated inversely with cat activity, and suggest that cats have a negative effect overall on avian species richness. What might we expect if cats were removed entirely from Sydney s bushland patches? Although the above studies reported only correlations, a first expectation is that bird species richness would increase. Ground-active species in particular may return, as these occur presently in patches where cat activity is minimal. However, another expectation is that the activity of above-ground nest predators would increase, potentially depleting populations of tree-nesting birds. Examples of such cascading effects have been documented elsewhere (e.g., Soulé et al. 1988; Crooks and Soulé 1999; Glen and Dickman 2005), with several studies suggesting that cats specifically have beneficial effects on island bird faunas by suppressing populations of introduced rats (Disney and Smithers 1972; Fitzgerald et al. 1991; Tidemann et al. 1994; Courchamp et al. 1999). In these circumstances, cat control programs would be best advised to proceed circumspectly, with attempts made at the same time to prevent the release of black rats and other nest predators. Other examples Many further studies indicate that pest management, and our view of pest species generally, would benefit from being more nuanced; I list a small number of relevant examples here that concern the red fox (Vulpes vulpes). In Western Australia, Risbey et al. (2000) showed that cat activity increased three-fold after foxes were controlled, and that this led in turn to a striking (80%) depression in the numbers of small mammals captured. Increases in cat activity post-fox control were observed also by Burrows et al. (2003) in the Gibson Desert, while post-control shifts in cat habitat use were reported by Molsher (1999) in New South Wales. In another instructive study, Kuchling (1997, 2000) reviewed efforts that had been made to recover the critically endangered Western Australian swamp turtle (Pseudemydura umbrina) in reserves near Perth. Foxes and dogs had been diagnosed as major predators of the turtle, and this led to the construction of a fence around each of two swamps to keep the predators at bay. While this was successful, at one swamp the absence of predation led to an explosion in the numbers of Pacific black ducks (Anas superciliosa); the ducks consumed much of the food that would have been used by the turtles, polluted the water, and attacked and damaged many hatchlings directly. At the other swamp, ravens (Corvus coronoides), black and brown rats (R. norvegicus) increased and killed many juvenile and sub-adult turtles that were in aestivation or on the move to aestivation sites (Kuchling 2000). In the Australian Capital Territory, Banks et al. (1998) showed that removal of foxes led to outbreaks of rabbits, with potentially deleterious consequences for native vegetation. Many further examples of indirect effects impacting on management decisions overseas have been documented by Zavaleta et al. (2001) and Ewel and Putz (2004), while a review of analogous indirect effects caused by host-specific biological control agents has been made by Pearson and Callaway (2003). A framework for prediction and management The two case studies and several brief examples above show that introduced pest species can become involved in complex webs of interactions with native species, and suggest that we should be cautious before concluding that such pests are irredeemably bad. Of course, this is not to say that pests should not be managed they should be. It is just that we should try to be more thoughtful in how we approach the management of pests, and try to predict the consequences of management actions before they are implemented. In many cases this will not be easy because the system has been extensively modified and its natural state, or states, may not be known. From the case studies and examples above and recent theoretical explorations (e.g., MacDougall and Turkington 2005), however, a preliminary framework can be constructed that predicts the likelihood of occurrence of indirect interactions between introduced pests and native species, and the management actions that should be taken (Table 1). In the first instance, the time since introduction of the pest species is an important consideration. In the early stages of invasion, a pest may impact directly on native biota but there will be a small risk that it will have embedded itself in more complex webs of direct and indirect interactions with native species. If there are grounds to suspect that the pest may have strongly deleterious consequences in its new environment, such as with the recent introduction of the red fox into Tasmania, every effort should be made to remove it. If the pest species has been established for a long time in its new environment there is a much stronger chance that its removal will have effects that reverberate through the ecological community. Here, judicious use should be made of small-scale removal trials to evaluate the net benefit that broad-scale control might have. 212 Pest or Guest

6 The complex pest Table 1. A preliminary framework for prediction and management of the effects of pest control in different contexts: time, place and pest density. 1. Time Context Likelihood of indirect interactions Management Early invasion Low Remove pest Pest established Low-high Small-scale removal trials 2. Place Context Likelihood of indirect interactions Management Low species diversity, inc. islands Low, if system is in a new stable state ; moderate if system in original state Remove pest with caution High species diversity High, irrespective of system state Small-scale removal trials 3. Pest density Context Likelihood of indirect interactions Management Low density Low Remove or monitor pest High density High Small-scale removal trials The spatial context is also important. If a pest has been introduced to a place where there are few native species, it will be unlikely to engage in many indirect interactions simply because there are few opportunities to do so. This should be especially so in systems that have been pushed into a new stable state due to anthropogenic or other disturbances, because co-evolutionary linkages will have been already disrupted. If the pest s new environment has a low diversity of native species but is still in its original state, there may be a greater chance of establishment of indirect interactions. Examples here include small islands such as Boullanger. Management in these low-diversity situations should focus on pest removal, but proceed cautiously to reduce the chance that any component of the native biota experiences inadvertent negative effects. In places where the diversity of native species is high there will be a greater chance of a pest species establishing complex interactions. Here, the indirect effects of removal will be much less certain, and pest control should proceed only after small-scale pilot removal trials have revealed the broad effects of this management option. Finally, pest density should be considered. At low density there will be less likelihood that complex interaction webs have developed with native species. Management Acknowledgements I thank A. Dickman, A. Glen and C. McKechnie for helpful discussions, P. Banks and B. Murray for review comments on the manuscript, A. Greenville for References Arthur, W. and Mitchell, P A revised scheme for the classification of population interactions. Oikos 56: Baldock, F. C., Alexander, L. and More, S. J Estimated and predicted changes in the cat population of Australian households from 1979 to Australian Veterinary Journal 81: Banks, P. B., Dickman, C. R. and Newsome, A. E Ecological costs of feral predator control: foxes and rabbits. Journal of Wildlife Management 62: could probably proceed safely by removing or controlling the pest. If pest density is very low even this course of action may be unwarranted as the pest is unlikely to be having strongly negative effects; instead, monitoring may be sufficient, provided that triggers are in place to allow management intervention above some density threshold. If the density of the pest is high, its direct and indirect effects on native species are likely to be stronger. Control measures would again proceed most advisedly after smallscale removal trials had confirmed that they were safe to implement. The framework in Table 1 is preliminary and summarises only some of the factors that should be considered prior to the implementation of a large or costly pest control program. Other factors that need to be incorporated are pest behaviour, the trophic position of the pest, and the feasibility of implementing a control program that would maintain pest damage at acceptably low levels. These are challenges for the future. Progress will likely depend on forging novel approaches that combine good ecological understanding, development of theory such as network analysis (Proulx et al. 2005), and application of costeffectiveness analyses from the economic domain (Gentle 2005). assistance with the figures, and the Australian Research Council for funding the primary research underpinning this chapter. Barratt, D. G Predation by house cats, Felis catus (L.), in Canberra, Australia. II. Factors affecting the amount of prey caught and estimates of the impact on wildlife. Wildlife Research 25: Bencini, R., McCulloch, C., Mills, H. R. and Start, A. N Habitat and diet of the dibbler (Parantechinus apicalis) on two islands in Jurien Bay, Western Australia. Wildlife Research 28: Pest or Guest 213

7 Dickman Braysher, M Managing Vertebrate Pests: Principles and Strategies. Australian Government Publishing Service: Canberra. Burrows, N. D., Algar, D., Robinson, A. D., Sinagra, J., Ward, B. and Liddelow, G Controlling introduced predators in the Gibson Desert of Western Australia. Journal of Arid Environments 55: Caughley, G. and Gunn, A Conservation Biology in Theory and Practice. Blackwell Science: Oxford. Caughley, J., Monamy, V. and Heiden, K Impact of the 1993 Mouse Plague. Occasional Paper Series No. 7, Grains Research and Development Corporation: Canberra. Chapple, R The politics of feral horse management in Guy Fawkes River National Park, NSW. Australian Zoologist 33: Courchamp, F., Langlais, M. and Sugihara, G Cats protecting birds: modelling the mesopredator release effect. Journal of Animal Ecology 68: Crooks, K. R. and Soulé, M. E Mesopredator release and avifaunal extinctions in a fragmented system. Nature 400: Crowther, M. S., Dickman, C. R. and Lynam, A. J Sminthopsis griseoventer boullangerensis (Marsupialia: Dasyuridae), a new subspecies in the S. murina complex from Boullanger Island, Western Australia. Australian Journal of Zoology 47: CSIRO The Insects of Australia, 2 nd edition. Melbourne University Press: Melbourne. Dickman, C. R Body size, prey size, and community structure in insectivorous mammals. Ecology 69: Dickman, C. R Conservation of mammals in the Australasian region: the importance of islands. Pp in Australia and the Global Environmental Crisis, edited by J. N. Coles and J. M. Drew. Academy Press: Canberra. Dickman, C. R Overview of the Impacts of Feral Cats on Australian Native Fauna. Australian Nature Conservation Agency: Canberra. Dickman, C. R Rodent-ecosystem relationships: a review. Pp in Ecologically-based Management of Rodent Pests, edited by G. R. Singleton, L. A. Hinds, H. Leirs and Z. Zhang. Australian Centre for International Agricultural Research: Canberra. Dickman, C. R. 2006a. Species interactions: direct effects. In press in Ecology: an Australian Perspective, 2 nd edition, edited by P. Attiwill and B. Wilson. Oxford University Press: Melbourne. Dickman, C. R. 2006b. Species interactions: indirect effects. In press in Ecology: an Australian Perspective, 2 nd edition, edited by P. Attiwill and B. Wilson. Oxford University Press: Melbourne. Dickman, C. R. 2006c. House cats in the Australian environment: behavior, impacts, and management. Berryman Bulletin. Dickman, C. R. and Lunney, D. (eds) A Symposium on the Dingo. Royal Zoological Society of New South Wales: Mosman. Dickman, C. R. and Murray, B. R Species interactions: complex effects. In press in Ecology: an Australian Perspective, 2 nd edition, edited by P. Attiwill and B. Wilson. Oxford University Press: Melbourne. Disney, H. J. de S. and Smithers, C. N The distribution of terrestrial and freshwater birds on Lord Howe Island, in comparison with Norfolk Island. Australian Zoologist 17: Ewel, J. J. and Putz, F. E A place for alien species in ecosystem restoration. Frontiers in Ecology and the Environment 2: Fitzgerald, B. M Diet of domestic cats and their impact on prey populations. Pp in The Domestic Cat: the Biology of its Behaviour, edited by D. C. Turner and P. Bateson. Cambridge University Press: Cambridge. Fitzgerald, B. M., Karl, B. J. and Veitch, C. R The diet of feral cats (Felis catus) on Raoul Island, Kermadec group. New Zealand Journal of Ecology 15: Flecker, A. S. and Townsend, C. R Community-wide consequences of trout introduction in New Zealand streams. Ecological Applications 4: Fuller, P. J. and Burbidge, A. A Discovery of the dibbler, Parantechinus apicalis, on islands at Jurien Bay. Western Australian Naturalist 16: Gentle, M. N Factors affecting the efficiency of fox (Vulpes vulpes) baiting practices on the Central Tablelands of New South Wales. PhD thesis, University of Sydney: Sydney. Gillies, C. and Clout, M The prey of domestic cats (Felis catus) in two suburbs of Auckland City, New Zealand. Journal of Zoology 259: Glen, A. S. and Dickman, C. R Complex interactions among mammalian carnivores in Australia, and their implications for wildlife management. Biological Reviews 80: Grayson, J. and Calver, M Regulation of domestic cat ownership to protect urban wildlife: a justification based on the precautionary principle. Pp in Urban Wildlife: More than meets the Eye, edited by D. Lunney and S. Burgin. Royal Zoological Society of New South Wales: Mosman. Grayson, J., Calver, M. C. and Styles, I Attitudes of suburban Western Australians to proposed cat control legislation. Australian Veterinary Journal 80: Grigg, G Conservation benefit from harvesting kangaroos: status report at the start of a new millennium. A paper to stimulate discussion and research. Pp in A Zoological Revolution: Using Native Fauna to Assist in its own Survival, edited by D. Lunney and C. R. Dickman. Royal Zoological Society of New South Wales: Mosman. Kuchling, G Managing the last survivors: integration of in situ and ex situ conservation of Pseudemydura umbrina. Pp in Proceedings: Conservation, Management and Restoration of Tortoises and Turtles, July 1993, edited by J. van Abbema. New York Turtle and Tortoise Society: New York. Kuchling, G Conservation strategies for remnant turtle populations: the Western Australian swamp turtle Pseudemydura umbrina recovery programme. Stapfia 69, Zugleich Kataloge des O. Ö. Landesmuseums N.F. 149: Lilith, M., Calver, M., Styles, I. and Garkaklis M Protecting wildlife from predation by owned domestic cats: application of a precautionary approach to the acceptability of proposed cat regulations. Austral Ecology 31: Low, T Feral Future. Penguin Books: Melbourne. Low, T The New Nature. Penguin Books: Melbourne. MacDougall, A. S. and Turkington, R Are invasive species the drivers or passengers of change in degraded ecosystems? Ecology 86: Matthews, A., Dickman, C. R. and Major, R. E The influence of fragment size and edge on nest predation in urban bushland. Ecography 22: McLeod, R Counting the Cost: Impact of Invasive Animals in Australia, Cooperative Research Centre for Pest Animal Control: Canberra. 214 Pest or Guest

8 The complex pest Meek, P. D Food items brought home by domestic cats Felis catus (L.) living in Booderee National Park, Jervis Bay. Proceedings of the Linnean Society of New South Wales 120: Miller, S., Bencini, R., Mills, H. and Moro, D Food availability for the dibbler (Parantechinus apicalis) on Boullanger and Whitlock Islands, Western Australia. Wildlife Research 30: Molsher, R The ecology of feral cats (Felis catus) in open forest in New South Wales: interactions with food resources and foxes. PhD thesis, University of Sydney: Sydney Morin, P. J Community Ecology. Blackwell Science: Oxford. Newsome, A The control of vertebrate pests by vertebrate predators. Trends in Ecology and Evolution 5: Newsome, A., Pech, R., Smyth, R., Banks, P. and Dickman, C Potential Impacts on Australian Native Fauna of Rabbit Calicivirus Disease. Environment Australia: Canberra. NSW National Parks and Wildlife Service Draft Wild Horse Management Plan for the Alpine Area of Kosciuszko National Park. NSW National Parks and Wildlife Service: Sydney. Olsen, P Australia s Pest Animals: New Solutions to Old Problems. Kangaroo Press: Sydney. Pearson, D. E. and Callaway, R. M Indirect effects of host-specific biological control agents. Trends in Ecology and Evolution 18: Proulx, S. R., Promislow, D. E. L. and Phillips, P. C Network thinking in ecology and evolution. Trends in Ecology and Evolution 20: Reark The Metropolitan Domestic Cat: a Survey of the Population Characteristics and Hunting Behaviour of the Domestic Cat in Australia. Petcare Information and Advisory Service: Melbourne. Risbey, D. A., Calver, M. C., Short, J., Bradley, J. S. and Wright, I. W The impact of cats and foxes on the small vertebrate fauna of Heirisson Prong, Western Australia. II. A field experiment. Wildlife Research 27: Sinclair, A. R. E., Krebs, C. J., Fryxell, J. M., Turkington, R., Boutin, S., Boonstra, R., Seccombe-Hett, P., Lundberg, P. and Oksanen, L Testing hypotheses of trophic level interactions: a boreal forest ecosystem. Oikos 89: Soulé, M. E., Bolger, D. T., Alberts, A. C., Wright, J., Sorice, M. and Hill, S Reconstructed dynamics of rapid extinctions of chaparral-requiring birds in urban habitat islands. Conservation Biology 2: Start, A. N., Moro, D., Adams, M. and Bencini, R Dunnarts from Boullanger Island: new evidence and reassessment of a taxonomic issue with resource implications. Australian Mammalogy 28: Tidemann, C. R., Yorkston, H. D. and Russack, A. J The diet of cats, Felis catus, on Christmas Island, Indian Ocean. Wildlife Research 21: Williams, K., Parer, I., Coman, B., Burley, J. and Braysher, M Managing Vertebrate Pests: Rabbits. Australian Government Publishing Service: Canberra. Wolfe, K. M., Mills, H. R., Garkaklis, M. J. and Bencini, R Post-mating survival in a small marsupial is associated with nutrient inputs from seabirds. Ecology 85: Zavaleta, E. S., Hobbs, R. J. and Mooney, H. A Viewing invasive species removal in a whole-ecosystem context. Trends in Ecology and Evolution 16: Pest or Guest 215