Eradication of Island Invasives: Practical Actions and Results Achieved

Abstracts

Cat eradication on the Montebello Islands

Algar, D., A. A. Burbidge and G. J Angus.

Montebello Renewal (part of ‘Western Shield’) aims to eradicate feral animals from and reintroduce and introduce threatened animals to the Montebello Islands, an archipelago off the Pilbara coast of Western Australia. Feral cats were eliminated from Hermite Island, at 1 020 ha the largest of the Montebellos, in 1999. Eradication comprised aerial baiting, using a recently developed cat bait, followed by intensive trapping utilising an innovative technique. Eradication was achieved over a six-week period. Intensive searches for evidence of cat activity one year later confirmed earlier conclusions that cats had been successfully eradicated.
 
 

Removing a diverse suite of invasive threats to recover an endangered Hawaiian bird species and its dry forest habitat

Banko, P. C., S. Dougill, L. Gold, D. Goltz, L. Johnson, P. Oboyski and J. Slotterback.

Recovery of the Hawaiian forest bird community that includes the endangered palila (Loxioides bailleui) requires removing a wide array of invasive pests and weeds from subalpine dry forest habitat on Mauna Kea volcano. Palila are threatened by predators and food competitors, and their habitat is threatened by aliens that browse native vegetation, increase fire fuel levels, and suppress forest regeneration. Due to these and other factors, most palila are concentrated in only 30 km² of habitat. In addition, the palila is a seed specialist that obtains most of its food resources from mamane (Sophora chrysophylla), an endemic leguminous tree that is sensitive to browsing by feral sheep, mouflon sheep, and cattle. Episodically during the past two decades, sheep and mouflon populations have been reduced, resulting in mamane regeneration in many areas. However, annual counts suggest that the palila population may not benefit from these habitat improvements until saplings have grown larger. Invasive annual grasses suppress mamane regeneration and accumulate as fire fuel, and an alien vine overgrows trees. We are mapping the distribution of these and other weeds to facilitate control strategies. We are investigating the ecology of alien mammals to develop control priorities and strategies. Feral cats and black rats destroy many palila nests, and the tendency of birds to roost repeatedly in the same trees may increase their vulnerability to mammalian predation. Cats are readily trapped, but tracking studies indicate that immigrants will arrive from far outside control areas. Cats on Mauna Kea seem to prey more on birds than on house mice; therefore, reducing mouse populations may have little impact on cat numbers. Mice are abundant in palila habitat and are active in tree canopies, but their potential threat to palila or the forest is unclear. Threats to insect food resources of palila include alien wasps that parasitise and prey on caterpillars. Ants also are spreading into palila habitat and threaten the entire insect community. Protecting and enhancing the main palila population and re-establishing another population elsewhere on Mauna Kea depend on the effectiveness of reducing the impacts of this complex suite of invasive aliens.
 
 

Man-made marinas as sheltered islands for alien marine organisms: Establishment and world-first eradication of a marine pest

Bax, N., C. Hewitt, M. Campbell, and R. Thresher

The typical tidal range in the Northern Territories, Australia is 7m. Sheltered marinas with double lock gates have been developed from estuaries, or dug from the shoreline, to provide regulated environments with no tidal range. These sheltered marinas are novel environments in northern Australia and provide islands of opportunity for colonisation by invasive marine pests. In March 1999 a fouling mussel, Mytilopsis sp., closely related to the damaging and costly freshwater zebra mussel in Europe and the US, was discovered in one of the marinas at densities up to 10,000 m^-2. It had reached those densities in less than 6 months. In this talk we describe the colonisation of this and other marinas by the mussel, and the approaches taken to quarantine and eventually eradicate the mussel from those marinas. Lastly we compare the environment fauna in the regulated environments to adjacent unregulated natural environments and determine those attributes that could have led to increased invasion success.
 
 

The eradication of alien mammals from five offshore islands, Mauritius, Indian Ocean

Brian D. Bell

Following the removal of rabbits from Round Island (1979) and the publication of a management plan (1989), the Mauritius Government contracted Wildlife Management International Limited in 1993 to fulfil one of the plans recommendations to survey the offshore islands of Mauritius and Rodrigues and prepare an offshore islands management plan. This plan made a number of recommendations and priorities in relation to the removal of alien species. In 1995 work on the priorities began with the removal of rats (Rattus norvegicus) and hares (Lepus nigricollis) from Gunners Quoin, rats (R. rattus) from Gabriel Island and mice (Mus musculus) from Cocos and Sables Islands. In 1998 cats (Felis catus), rats (R. rattus) and mice were removed from Flat Island and rabbits (Oryctolagus sp.), which had been illegally released following the earlier eradications, from Gunners Quoin. These programmes were hand-laid operations. In all cases 0.02gm/kg brodifacoum was used in grain based pellets. The bait was set out at half the maximum grid recommended for the rodent species targeted. The exception was cats which were trapped in leg-hold traps. Plans are being considered for the re-introduction of reptiles and birds. Some planting of native trees has begun. This paper covers the eradication sector of the management.
 
 

The eradication of possums from Kapiti Island

Kerry P. Brown

Kapiti Island is a 1965 ha nature reserve lying 5 kilometres off the south-west coast of the North Island of New Zealand. The Australian marsupial the brushtail possum Trichosurus vulpecula was introduced to the island in 1893. Possums have negative impacts on New Zealand forest ecosystems - they can kill individual trees, potentially alter forest succession and regeneration processes, suppress flowering and fruiting, and prey upon native birds and other native animals. Various attempts were made to control possums on Kapiti by trapping between 1920 and 1968. A moratorium was placed on trapping in 1969 and research commenced to better quantify the importance and nature of possum impacts. Possum control was initiated again in 1980 using commercial trappers following recommendations from research. Intensive control using trappers on wages and including an aerial application of 1080 occurred during 1983 -1985. Eradication using traps, dogs and guns commenced in March 1985 and the last possum was killed in October 1986. Approximately 21 000 possums were removed during 1980 - 1985 by trapping and poisoning. Dogs located 32 of the last 80 possums that were removed during 1985 - 1986. The methods used, reasons for success, and lessons for future eradication attempts are described.
 
 

The impact of rabbit and goat eradication on the ecology of Round Island, Mauritius

Bullock, D. J., S. G. North, M. E. Dulloo, and M. Thorsen.

Round Island (151 ha) is an important refuge for several endemic reptiles and plants that formerly occurred on the mainland of Mauritius. Rabbits (Oryctolagus cuniculus) and goats (Capra hircus) introduced in the early 19th century greatly modified the island's main habitats and several species were threatened with extinction. By late 1986 rabbits and goats had been eradicated. In this paper we describe changes in the extent and composition of key elements of the biota (vegetation, reptiles and invertebrates) before and after eradication. Primary objectives of our work are to measure and interpret changes in the populations of threatened species which inform their conservation management. Repeatable sampling methods have allowed the effects of eradication on trends and population sizes to be estimated and the identification of ecological responses, some of which were unexpected. Ten years after eradication the following changes had occurred: Abundant regeneration of the three main tree species and the first substantial cohorts of newly mature individuals for two centuries; a slight increase in the extent of ground vegetation; substantial increases in ground vegetation height, perennial herbs and shrubs; major declines in plants of open habitats; colonisation and rapid spread of several non-native "weedy" species. Round Island's ground layer vegetation is becoming dominated by non-native species. Future regeneration of several native species (including the trees) may now be threatened by the increasing impact of invasive plants. Amongst reptiles, we have detected no population increases for the threatened Gunther's gecko (Phelsuma guentheri) and Round Island boa (Casarea dussumieri). Fluctuations in populations of Vinson's gecko (P ornata) and Durrell's night gecko (Nactus serpensinsula) include apparent declines but Telfair's skink (Leiolopisma telfairi) and Bojer's skink (Gongylomorphus bojeri) have increased, tracking increases in the number of mature trees. These changes, and those of some invertebrate taxa sampled, are discussed in terms of the removal of rabbits and goats and the persistence of key threatened species. We also discuss the unpredictable responses of modified island ecosystems to eradication of introduced herbivores and consider the desirability of the observed changes in relation to the future management of Round Island and similar islands.
 
 

Introduced animal eradications for nature conservation on WA islands: a review

Burbidge, A. A. and K. D. Morris

There are about 2500 islands off the Western Australian coast, many of which have high nature conservation values. Eight species of introduced mammals occur or occurred on about 100 islands. In addition, Aborigines have introduced dingoes to at least four islands. Three domestic animals – horse, camel and sheep – have also been recorded. Six exotics – fox, cat, goat, rabbit, black rat and house mouse – have now been eradicated from about 45 islands in a series of projects since the 1960s. Most effort has been directed at black rats with at least 31 islands now clear of this species. Pindone, vacuum-impregnated into oats, was used until the 1990s, when bran pellets with brodifacoum were used in the Montebellos. Rabbits have been eradicated using carrots soaked in 1080, foxes with 1080 dried meat baits and cats with a combination of baiting and trapping. Goats were eradicated from Bernier Island using an experienced shooter operating from a helicopter. The house mouse has been eradicated from Barrow Island twice after introductions in food and equipment, and from Varanus and adjacent islands after introduction in food. Both islands are utilised by the oil industry. Difficulties and how they were overcome, and future eradication priorities will be discussed.
 
 

Introduced neotropical tree frogs in the Hawaiian Islands: Control technique development and population status

Campbell, E. W., F. Kraus, S. Joe, L. Oberhofer, R. Sugihara, D. Leaseand P. Krushelnycky

Two species of neotropical tree frog, Eleutherodactylus coqui and E. planirostris have been introduced into the Hawaii Islands via the horticulture trade. Since 1997 frog colonies within the state have rapidly spread accidentally and intentionally and frog abundance within colonies has grown rapidly. Colonies of these frogs are currently known from 150+ locations on the island of Hawaii, 35+ on Maui, 5+ on Oahu, and 1 on Kauai. Although these frogs were originally restricted to horticulture sites, they are now found in residential areas, resorts and hotels, and public lands. Individual frogs or frog colonies have been verified at sites ranging from sea level to over 3500 ft. Within their native range, where their populations may be restrained via predation and other natural checks, they may reach densities of 20,000 frogs/ha and consume an estimated 140,000 prey items/night. Given the current population eruptions of these frogs in Hawaii, similar densities could be reached or exceeded. Given the high potential biomass of introduced frogs there are realistic ecological and anthropogenic concerns associated with the spread of these frogs. Currently, there are limited techniques to control these animals. Research has been conducted to evaluate the efficacy of various mechanical and chemical techniques for frog control. Thus far, hand-capture and trapping have proven labour intensive for frog control in sites with locally moderate to high frog densities. In collaboration with state pesticide regulatory and wildlife management agencies, we tested 30+ compounds (registered insecticides, surfactants, human pharmaceutical compounds and food additives) in the laboratory to determine their efficacy for tree frog control. In these trials, caffeine and water solutions proved to be the only compounds that could effectively be used for tree frog control. Currently, field trials are being conducted to evaluate the efficacy of a direct spray application of a concentrated caffeine and water solution for tree frog control on 0.1 - 0.5 ha tree frog infested plots. If these trials are successful, it is hoped that management agencies in the State of Hawaii will be able to reduce the spread and potential impact of these pest species on a landscape scale.
 
 

Exotic trees as refuges from macaque predation on endangered Mauritian birds

Carter, S. P. and P. B. Bright.

Introduced mammals on Mauritius are threatening the survival of a suite of highly endangered endemic birds. Crab-eating macaques and black rats are the major nest predators, as revealed by camera traps and surrogate nests. Rats are controlled locally by poisoning in the most critical areas, and other predators such as cats and mongooses are trapped and destroyed. There are, however, no currently acceptable options for controlling macaques, which may have a significant impact during the bird-breeding season. A previous study of the Mauritius fody and other forest birds suggested that plantations of non-invasive Japanese red cedar provide a refuge from nest predation. Fodies increased their breeding success eight-fold by nesting in these plantations, and prior to subsequent releases the last remaining wild pink pigeon population nested entirely within one such plantation. Further research using surrogate nests and quail eggs confirmed that nest predation by macaques is significantly lower in cedar than native forest, apparently because the macaques are avoiding these areas. The careful planting and management of these beneficial exotics could therefore be carried out to create more refuge areas. A spatially explicit model using GIS shows how the creation of a small number of additional nest refuges might significantly increase bird populations. Thus providing a cost effective and sustainable solution to the problematic macaques, which are currently outside the reach of conventional predator control.
 
 

Tackling tussock moths: strategies, timelines and outcomes of two programs for eradicating tussock moths from Auckland suburbs

Clearwater J. R.

Suburb-wide, aerial sprays of an organic insecticide (Bacillus thuringiensis) were applied to an infestation of the white-spotted tussock moth (Orgyia thyellina) followed by intensive monitoring for the remnant of the population. Caged females in sticky traps caught small numbers of wild males from a tightly localised area. This area was sprayed from the ground and with helicopters. Ground searches for eggs and caterpillars found nothing. The catch in the female-baited traps decreased as the summer program of targeted sprays continued. A synthetic pheromone was identified from an international effort and deployed in a large number of sticky traps in the second year of the program. No males were caught. The lures were subject to an independent quality assurance test for attractiveness. The moth was declared to have been eradicated and was not found again during the following two years. A second species of tussock moth (Teia anartoides) was found in two Auckland suburbs the following year. Localised ground spraying of infested areas with synthetic insecticides was followed by ground searches for caterpillars. These searches have yielded a steady number of caterpillars for a 12-month period. No attempt to use natural pheromone sources was made and an attempt to identify the sex pheromone by a local group has failed. The moth continues to be present at the time of writing. The use of pheromones is concluded to be the key tool in any attempt to eradicate an invading moth species.
 
 

Invasive weed control on Poor Knights and Hen and Chickens Islands

Coulston G. J.

In 1995 the Department of Conservation initiated an intensive weed control programme on the Poor Knights Islands, 16km offshore from Tutukaka, Northland, New Zealand, with the intention of eradicating all existing environmental weed infestations to the point where windborn reinvasion from the mainland was the only threat. We are now in a position to evaluate the progress/successes of this programme. The weeds targeted are Ageratina adenophora, A. riparia, Araujia sericifera, Cortaderia selloana and C. jubata Control methodologies will be explained. In essence all known weed sites are visited twice a year and all weeds found destroyed. Visits are timed to coincide with peak germination periods and pre to early flowering to prevent further seed set. Aerial surveys are completed during early flowering to locate any plants on cliff faces or in the canopy of trees. Areas of the island prone to reinvasion are fully ground searched every year in spring and the weed free areas are searched every second year. A summary of the results to date will be graphically presented. Key points are the success with A. riparia and Cortaderia spp. A. adenophora numbers reduced from several thousands to less than fifty. A. sericifera has continued to have high germination of seedlings, but are now in a process of decline. It appears the seedbank is starting to expire. A. sericifera has posed the greatest difficulty to physically locate. An Excel spreadsheet of weed control was developed that provides useful field data for control purposes and the raw statistical information for management and monitoring purposes. Once refined this database could be beneficial for, analysing weed site dynamics, potential seedbank viability under various site conditions and an indication of the rate of reinvasion from outside sources. The model has been replicated on the Chickens group since 1997 and is showing promising results.
 
 

Eradication planning for animal pests – where we have come from and where we are going

Cromarty, P. L., K. G. Broome, A. Cox, R. A. Empson, W. M. Hutchinson, and I. McFadden.

The Department of Conservation is now a world leader in the field of animal pest eradication on islands, particularly rodent eradication. Organisational features which have lead to this include: Capacity building - Maximising research opportunities, particularly improving techniques for future eradication operations; Building on information and experience to minimise risk of failure and re-invention of the wheel; Forward planning skills development - Identifying training opportunities for new teams to gain experience by participating in eradication operations elsewhere; A team approach - Setting up project teams to oversee each major pest eradication operation; Peer review - The formation of an eradication advisory group and associated network to provide advice on major pest eradication operations. Focus is on planning and readiness before an operation takes place; Review - Operations are routinely debriefed to effectively transfer lessons learnt with each operation through to future projects. Debriefs are inclusive of all contributors and members of future project teams. Eradication operations rely entirely on existing technology used in control operations. The difference is the mind set. The right "mind set" is important for everyone involved in an eradication operation and is different from that for a "control" operation. An eradication operation requires 100% focus and effort! 1. The design and planning of the eradication operation must be as robust as possible to prevent the operation failing. 2. All pest animals must be put at risk for eradication to be achieved. 3. Failure of an eradication operations is frequently linked to a casual approach or a "can’t be done" attitude. The approach outlined above has application where animal pest eradication on islands is planned elsewhere.
 
 

Control of buffel grass (Cenchrus ciliaris L.) on Airlie Island off the Pilbara Coast of Western Australia

Dixon, I. R., K. W. Dixon, and M. Barrett.

The aim of this project, now into the second year of the implementation phase, is to eradicate buffel grass, develop and implement methods to restore the indigenous vegetation, collect and store seed for future restoration works. The most effective herbicides trialed under these conditions are Roundup Biactive® 8 litres/ha and Verdict® 6 litres/ha. Extensive field trials indicated the main perennial shrubby species on the island Acacia bivenosa, Acacia coriacea and Rhagodia preissii are very resistant to Roundup Biactive®, Verdict®, with the exception of native grasses, had no adverse effect on any indigenous species. Initial blanket and spot spraying with Roundup to kill the parent plants followed by blanket spraying, avoiding native grasses, with Verdict is the best and most cost effective method of control. A temporary (three years) water pipe for filling battery operated 250 litre spraying units was laid through the centre of the buffel populations. Hoses 60m long with hand held lances were pulled out to their extremities as spraying took place. Four operators with two units can spray about 2 hectares each day. The best time for spraying is six weeks after heavy rain when the parent plants are actively growing and the new seedlings are large enough to spray, too early spraying misses the seedlings, too late and the seedlings as well as the parent plants are dropping seed or drying off. The window of opportunity for spraying is only two weeks. About 98% of the buffel has been controlled. Replanting with greenstock is preferable after heavy rainfall, the main shrubby species planted after spraying with Roundup can then be oversprayed, when required, with Verdict®. Greenstock survival rates, 5 to 90%, are entirely dependent on follow up rainfall. Two to four sprays a year, depending on rainfall events, are required for a period of at least three years (estimated age of soil seed bank) to control this weed with follow up monitoring and backpack spot spraying or hand removal. Eulalia aurea a perennial dominant native grass is best planted after the three year spraying programme to avoid spray damage and for easier operations.
 
 

Management of island invasives to restore biodiversity on Alaska Maritime National Wildlife Refuge

Ebbert, S. E. and G. V. Byrd

Alaska Maritime National Wildlife Refuge encompasses 3.5 million acres and more than 2500 islands around the coast of Alaska. Many large islands have had one or more non-endemic vertebrate species (eg., Arctic and red fox, ground squirrel, Norway rat, house mouse, caribou, reindeer, cattle, Arctic and European hare) introduced accidentally or purposely by man. Introductions drastically altered these island ecosystems. Although new introductions are against U.S. Fish and Wildlife Service policy and State law, accidental introductions, especially with non-native rodents, is still a threat. To restore native biodiversity, refuge personnel have inventoried islands for introduction of species, measured the impact of invasive wildlife on native species, adapted predator control methods for eradication on refuge islands, and assessed the benefits of successful eradication projects. This paper reviews the history of these projects, particularly the effort which has resulted in the eradication of introduced foxes from 38 islands totalling more than 1 million acres. Special emphasis is placed on successful exotic fox removal on 38 refuge islands. History of rodent introduction and prevention on the refuge are presented in another paper.
 
 

Control and eradication of the introduced grass, Cenchrus echinatus at Laysan Island, Central Pacific Ocean

Flint, E. N.

The sandbur, Cenchrus echinatus, an annual grass native to Central America, was first documented at Laysan Island, Hawaiian Islands National Wildlife Refuge, in 1961. It was probably brought there inadvertently by the military or researchers visiting the island. By 1989 it had spread to dominate the vegetation of nearly 30% of the 212 vegetated hectares of Laysan Island. By displacing the native bunchgrass Eragrostis variabilis, it diminished important breeding habitat for two endemic endangered landbirds, the Laysan finch (Telespyza cantans) and the Laysan duck (Anas laysanensis), as well as several species of indigenous seabirds and terrestrial arthropods. In 1991 Refuge staff started a year-round control program designed to remove all plants before they went to seed and continue to do so until the seedbank was depleted. After experimenting with techniques including heat and saltwater application, application of a herbicide (glyphosate) and mechanical control (hand pulling) were chosen. A system of marking locations of all plants found and regular visitation of these sites allowed the staff to anticipate new sprouts and remove them before they went to seed. Concurrent studies of the life history of the plant allowed continual adjustment and refinement of the control regime. Time from sprouting to seeding ranged from 8 to 12 weeks allowing us to increase the rotation interval for visiting plots. Decline in the rate of finding new plants in a previously cleared plot from as high as 84.7 plants per hour in Autumn 1994 to 0.043 plants per hour in Autumn 1999 is evidence that the seedbank is being depleted. Cenchrus is now so rare that it is no longer having ecosystem effects at Laysan Island. This project demonstrates that in a controlled-access situation it is possible to reduce and eventually eradicate an introduced annual grass. Costs associated with this undertaking include a monetary investment averaging $150,000 US per year for staff and volunteers, supplies, and vessel charter to this extremely remote site (5 days by boat from Honolulu); disturbance to nesting seabirds, particularly burrow nesting species such as Bonin petrels, Tristram's storm-petrels, and wedge-tailed shearwaters; and the risks of introducing new island pests despite stringent quarantine procedures.
 
 

Sustained control of feral goats in Egmont National Park, New Zealand

Forsyth, D. M., J. P. Parkes; D. Choquenot, G. Reid and D. Stronge

Egmont National Park (33 540 ha) is a forested mountain ‘island’ surrounded by a ‘sea’ of farmland. Feral goats have been present in the Park since c. 1910. Control efforts have been ongoing since 1925, making it one of the longest-running sustained vertebrate pest control operations in the world. Although helicopter-based hunting has proven effective at reducing goats above timberline, most of the Park is forested and the primary method of control in this habitat has been ground-based hunting with dogs. We used indices of hunting effort (days hunted) and goat population density (goats killed/days hunted), to investigate trends in the goat population in response to management during the period 1961-1999. Annual hunting effort generally increased over the period 1961-1986 but, following a change in the management organisation in 1987, has since declined. Goat density was highest in the earliest years of control (c. 7 goats killed/day) and steadily declined until 1987 (0.8 goats killed/day). Post-1987 the population has been maintained at low densities (<2 goats killed/day). The likely consequences of alternative strategies for allocating hunting effort on goat densities will be discussed.
 
 

The rat eradication program of Monito Island, West Indies

García, M. A., and P. J. Tolson

Monito Island (15 ha) is located between Puerto Rico and the Hispaniola (West Indies). On this island the endemic Monito Island gecko (Sphaerodactylus micropithecus) has been considered scarce, and rat predation was postulated as the most likely explanation. Therefore, the Puerto Rico Department of Natural and Environmental Resources (PRDNER) started a rat eradication program in October 1992, using Maki mini blocks Ò . This rodenticide was spread at 10 meter intervals covering the entire island. This initiative decreased rapidly the number of rats from a relative estimate of 0.63 to 0.1 RHT (rat/hour trap). However, in April 1993 this project was stopped by the U.S Fish and Wildlife claiming that PRDNER had not satisfied all the requirements of the Federal Insecticide, Fungicide, and Rodenticide Act (FRIFRA). The main concern was the possibility of poisoning geckos with the rodenticide. Thus, we proved experimentally that the geckos were not attracted to the paraffinized rodenticide blocks. Meanwhile, this delay caused a fast increase (0.38 RHT) in rat’s relative abundance. Finally, we started this project again in September 1998, and a surprisingly low rat population was barely detected using chewing sticks. No rats were caught on snap traps. Nevertheless, we went ahead and completed the eradication using Talon-G Ò . We put the Talon-G Ò blocks inside plastic baiting stations, at 20 meters intervals. Current surveys have not detected rats on Monito Island.
 
 

Changes in bird numbers on Tiritiri Matangi Island over the period of rat removal

Graham, M. F. and C. R. Veitch

Tiritiri Matangi is a 220 ha island 25 km north of Auckland City in the Hauraki Gulf. Most of its forest cover was removed during many centuries of Maori and European occupation and farming. Some areas of extant forest canopy remained. Since 1984 farming has ceased and some 300 000 native trees have been planted to restore the natural ecosystem. Twenty-seven species of native bird are naturally present and breeding on the island. Twenty-two exotic species introduced to mainland New Zealand have found their way to the island. Nine species of native bird have been translocated to the island. Data from bird counting transects within extant forest areas in spring of each year are considered. The data from the six year period before eradication of Pacific rats (Rattus exulans) in September 1993 are compared to the six year period following rat eradication. A number of significant changes are recorded with both increases and decreases in bird numbers. These are attributed to either the direct impact of the rats or changes in forest composition following rat removal or are confused by other factors.
 
 

Recovery of invertebrate populations on Tiritiri Matangi Island following eradication of kiore (Rattus exulans)

Green, C. J.

The effects of kiore, or Pacific rat (Rattus exulans) on indigenous species has historically been based on anecdotal accounts and circumstantial comparisons. The eradication of kiore from 220ha Tiritiri Matangi Island in 1993 provided an opportunity to obtain empirical data on the effects of this rodent on invertebrates. Long term monitoring of ground invertebrates began three months before the removal of kiore and continued for five years following removal. Pitfall traps were set in a mature broadleaf forest remnant and in a younger regenerating forest. Larger numbers of invertebrates were caught in the mature forest and these also increased to a greater degree after rat removal. Capture rates of several large (>10 mm) species increased during the study, including ground weta (Orthoptera: Anostostomatidae) and several species of prowling spider (Araneae: Miturgidae). Capture rates of other species that changed over time appear to be correlated with weather that varied dramatically during the period of monitoring. Seasonal changes are reported and the life histories of large flightless, nocturnal ground dwelling invertebrates are correlated with kiore eradication.
 
 

Spartina eradication and inter-tidal recovery in Northern Ireland estuaries

Hammond, M. and A. Cooper

Spartina anglica and Spartina townsendii are invasive plants that were introduced into Northern Ireland between the 1930s and 1950s. They spread to develop mono-dominant swards. These swards reduce the feeding area for wildfowl and waders and exclude native plants such as Zostera spp. (a major food source for overwintering Brantabrenicla). In 1998 an experiment was initiated to study the effectiveness of eradication methods at two sites, one a S. anglica sward, the other a S. townsendii sward. The treatments studied included, smothering with black plastic sheeting, applying the herbicides Dalapon and Glyphosate, and cutting followed by herbicide application. The effects of the eradication treatments on live Spartina stem density and associated flora was examined. Dalapon (applied at a rate of 57kg ha^-1) and smothering were the most effective eradication methods, reducing live Spartina stem density by over 95%. Glyphosate was relatively ineffective. Live Spartina stems re-established over the two years following treatment applications, suggesting that Spartina eradication would require the long-term re-application of treatments. Other species established after eradication treatments. They were, Salicornia spp., Puccinellia maritima, Aster tripolium and Plantago maritima. This suggests that treated areas will probably not revert into Zostera beds but will be colonised by other spp. benefiting from a changed habitat of accreted sediment. Preventing the further expansion of Spartina swards and infestation of new areas is thus suggested as a practical conservation strategy.
 
 

Control of cats on mountain "islands", Stewart Island, New Zealand

Harper, G. A. and M. Dobbins

The southern subspecies of the New Zealand dotterel is currently restricted to nesting areas on the bleak alpine mountaintops of Stewart Island. By the early 1990s the species had declined to a total population of 65 individuals. The principal cause of the decline was attributed to predation by feral cats. A cat-control programme was initiated in 1992. The programme involved a perimeter of bait stations set up at the bush-line. Poison baits for cats were presented in these stations during spring and summer, when dotterels were nesting. Research suggests that cats are not resident year-round in the sub-alpine scrub. They generally stray into the sub-alpine scrub and above the bush-line during the summer "low" in abundances of rats, their principal prey on Stewart Island. Little alternative prey is available during summer. The cat control appears to have been successful, as the population of dotterels has expanded to 170 individuals by April 2000. Research into the habitat preference of cats is continuing with a view to more efficient use of resources for ongoing cat control.
 
 

Eradication of feral goats and pigs from Sarigan Island, Commonwealth of the Northern Mariana Islands; methods and results

Kessler, C. C.

Sarigan Island is one of the Mariana Islands, is tropical in climate, an covers about 500 ha. During January and February of 1998 the field operations for eradicating feral goats and pigs was undertaken and completed. This was with the idea of stopping and reversing the loss of forest and accompanying erosion and thus improving the habitat for the endangered Micronesian megapode (Megapodius laperouse laperouse) and other native species. Methods used to remove ungulates were based on shooting and included ground shooters, helicopter shooting, dogs, and trapping. By March 1 of 1998 the project was thought to have been successfully completed with the removal of 68 pigs and 904 goats. Subsequent follow-up surveys in February and July of 1999 and July of 2000 removed an additional five goats. Monitoring of vegetation before and after has shown an increase in ground cover, and increase in tree seedlings, and the rapid expansion of the introduced vine Operculina ventricosa. Surveys of Mariana fruit bats (Pteropus marianus marianus), birds, lizards, and rats also show changes. It appears that the project has been successful in stopping and reversing the trend of forest loss. Habitat seems to have improved not only for the megapode but for fruit bats as well. It is still undetermined as to what affect the vine Operculina ventricosa will have on the regeneration and expansion of the native forest.
 
 

The response of herbaceous vegetation and endemic plant species to the removal of feral sheep from Santa Cruz Island

Klinger, R. C., P. Schuyler, and J. D. Sterner.

From 1984-1998 we monitored the response of herbaceous vegetation and endemic plant species to the eradication of feral sheep (Ovis aries) from the western 90% of Santa Cruz Island. Species diversity of herbaceous vegetation did not increase significantly between 1984-1996, but total herbaceous cover increased and bare ground decreased after sheep were eradicated from the island. The number and relative frequency of native herbaceous species was inversely related to increased herbaceous cover and the relative frequency of alien species. There was no evidence that native species were being displaced by new invading alien species, but by an increase in abundance from the alien species which already occurred in an area. Sixty percent of the 42 endemic plant species on the island showed an increase in distribution and/or abundance following the eradication. New populations of two of the five rarest species on the island were discovered within seven years of the end of the eradication program, and abundance within most of the populations of these two species increased. Of the other three species, the distribution and abundance of one remained unchanged while the two other species showed alarming declines. The declines of these two species were attributed to a proliferation of non-native grasses and impacts from feral pigs. The response of the herbaceous vegetation was relatively undesirable from a conservation perspective, especially in regards to some of the rare plants and the fact that grasslands occupy almost 50% of the area of the island. But many of the other endemic species showed positive responses, and other studies on Santa Cruz Island indicate that communities on the island that are dominated by shrubs and trees appear to be showing rapid rates of recovery from sheep impacts. A variety of outcomes can be expected to occur as a result of eradication of large numbers of grazing ungulates from islands, and eradication programs should only be considered the first step in a long process of restoration rather than an end in themselves.
 
 

The status of invasive ant control in the conservation of island systems

Krushelnycky, P. D., E. Van Gelder, L. L. Loope and R. Gillespie.

Introduced ants have often been responsible for significant ecological disruption in both continental and insular systems. In the oceanic islands of the Pacific in particular, ants have long been implicated in the wholesale extirpation of the native lowland arthropod fauna. While removal of the most detrimental of these invasive ant species would benefit many natural areas, the currently available techniques utilizing baits with chemical toxicants are appropriate only in certain situations. Eradication of invasive ants for conservation goals is a fairly new endeavor, and recent attempts have met with mixed results. At least one effort in the Galapagos employing the toxicant hydramethylnon may have been successful against the little fire ant (Wasmannia auropunctata), while experimental plots in Haleakala National Park, Hawaii, have failed to eradicate the Argentine ant (Linepithema humile). The present focus at Haleakala has therefore temporarily shifted to strategies for suppressing further invasion, and this has been moderately successful. Annual treatments of the expanding edges of the park's two ant populations with Maxforce bait have limited rates of spread to 5%-50% of historical rates. As a new campaign against the Argentine ant is being undertaken in New Zealand, and as problematic tramp species in general continue to expand their ranges, the need for renewed investigations into ant eradication techniques is critical. We propose that further experimentation with bait carriers, employed at biologically opportune times, may lead to a higher rate of bait sharing among nestmate queens and brood. Trial and error may be necessary to achieve maximal bait acceptance in different situations.
 
 

Ecological restoration at Wenderholm Regional Park and results of the first bird release

Lovegrove, T. G., B. S. Greene, B. W. Green and C. H. Zeiler

Ecological restoration has been undertaken at Wenderholm Regional Park, a 134 hectare peninsula on the mainland north of Auckland. Early efforts included fencing 60 hectares of coastal forest from livestock, brushtail possum (Trichosurus vulpecula), rabbit (Oryctolagus cuniculus) and plant pest control and revegetation of retired pastureland. More intensive animal pest control began in October 1992, following a study which showed that native pigeons (Hemiphaga novaeseelandiae) were failing to breed, mainly because of nest predation by possums and ship rats (Rattus rattus). A bait station grid (100 x 50 metre spacings) was established to control rats using anticoagulant poison. Baits have been laid annually between September and March since 1992, and annual brushtail possum control has been undertaken. Feral cats (Felis catus) have also been shot and trapped, and since March 1999, mustelids have also been trapped. Possums and rats have now been reduced to very low numbers, resulting in enhanced forest health, significantly improved pigeon breeding success and more abundant large invertebrates. As a result of successful restoration, and the presence of linkages to other forest areas, Wenderholm was deemed suitable for experimental releases of birds, which had become locally extinct on the northern New Zealand mainland. The objective was to select forest birds known to persist with introduced predators on the mainland elsewhere in the North Island, because despite predator control at Wenderholm, small numbers of some pests could be expected to occur. The first release was of 21 North Island robins (Petroica australis), translocated from nearby Tiritiri Matangi Island in March 1999. In September 1999 at least 62% of the robins were still alive, and during the 1999-2000 summer six pairs fledged 23 young. In September 2000, there were five pairs of robins at Wenderholm and two single males in a nearby contiguous forest area with no rodent control. If robins establish territories at Wenderholm in similar densities as reported elsewhere, the reserve could potentially support about 60 pairs. However, linkages with nearby forest areas mean that robins disperse easily, and more than one translocation may be required to establish a viable population inside the reserve.
 
 

Controlling invasive plants: hard-won lessons from continents and islands

Mack, R. N. and W. M. Lonsdale

So far, the record of controlling invasive plant species whether on islands or continents consists of few clear victories, some stalemates and many defeats. Instructive, if expensive, lessons have nevertheless been learned. (1) The advantages are recognised intuitively for destroying immediately all individuals of a potentially invasive species upon their arrival in a new range. Yet few immigrants meet this fate (cf. the detailed epidemiology of Centaurea tricocephala and Crupinavulgaris in the U.S. Pacific Northwest [USA]), illustrating a universal failure to act until damage has been inflicted by the invader. (2) Failing the total destruction of immigrants upon their entry, then maximum effort should be lodged against the immigrants’ small, isolated foci, even at the expense of attacking the most conspicuous centres of the invasion. As with (1), implementing this sound advice has often however proven difficult. However, the radical reduction of the range of Striga asiatica in North Carolina (USA) represents the clearest application of this principal; control of Pennisteumsetaceum in Hawaii has shown a similar result. (3) Effective control of invasive species requires continual effort: only by repeatedly canvassing the same area for survivors can an invader be controlled. Control of Schinus terebinthifolius and other invasive species in Bermuda has clearly succeeded through such diligence; re-emergence of Berberis vulgaris in the USA, once surviving only in isolated relictual populations, illustrates the hazards of relaxing a once diligent control program. (4) Control exerted upon a single invasive species may be undone ultimately, if its removal only sparks another non-indigenous species to fill the ecological void, e.g. the rise of Centauraeasolstitialis coincident with the control of Hypericum perforatum in California and role reversals of invasive aquatic macrophytes in New Guinea following biological control. Although the areas so treated are very small, success has been achieved by this scorch-earth strategy within distinct, isolated habitats in Hawaii and on some riverine islands in the continental U.S. Islands with their intrinsic borders and geographic isolation provide excellent locations for experimentation within which these lessons can be honed, identifying both the effective and ineffective components of any control effort.
 
 

The effectiveness of weeded and fenced ‘Conservation Management Areas’ as a means of maintaining the threatened biodiversity of mainland Mauritius

Mauremootoo, J. R, C. G. Jones, W. A. Strahm, M. E. Dulloo and Y. Mungroo.

Mauritian native ecosystems continue to be degraded by the action of alien plants and animals. Lack of management is not an option if Mauritius’ unique biodiversity is to be maintained. Weeding of alien plants and the fencing out of deer and pigs was first recommended for conservation in the 1930s. Ten weeded and fenced ‘Conservation Management Areas’ (CMAs) have been set up in a variety of ecosystem types. Predator control is also practised in some CMAs. The effects of management on native flora and fauna have been quantified in several upland CMAs. Consistent weeding and maintenance of fences appears to result in a spectacular regeneration of native flora. In the Brise Fer ‘Old Plot’, first weeded and fenced in 1987, a minimum of between 53% and 68% of native tree taxa are regenerating. Regeneration of some taxa was probably prevented by mammals that cannot be excluded by conventional fences. The diversity of native seedlings and saplings is relatively low in a more recently managed part of Brise Fer and in the nearby Mare Longue CMA respectively. In the former this may be due to the fact that several deer were fenced into the CMA for over two years. In the latter, rocks were not placed at the foot of the fence, thus allowing pigs to burrow into the plot. Native butterflies were on average nineteen times more abundant in the surveyed CMAs than in non-managed areas while results for native birds were equivocal. In contrast densities of some native snail groups were lower in CMAs. This may be due to the effect of persistent rat poisoning and the change in habitat after initial weeding. Current CMAs can be highly effective if the fencing is of a consistently suitable standard and if any incursions of deer and pig are dealt with rapidly. Weeding methods may have to be modified to minimise non-target damage. Non-regenerating or negatively impacted species may have to be managed individually. An alternative or complement to this would be the use of predator-proof fences.
 
 

Eradication of kiore (Rattus exulans) from Whenua Hou, Putauhinu and Rarotoka islands

McClelland P. J.

In August 1998 a programme to eradicate kiore (Rattus exulans) from Whenua Hou Nature Reserve (Codfish Island) was carried out. While not the largest eradication of this kind, the number and vulnerability of important non-target species involved made it one of the most complicated. Numerous trials were carried out over 5 years to gauge risk to non-target species and to determine methods of minimising this risk. All Kakapo (Strigops habroptilus) (48 out of a world population of 62) were evacuated for the duration of the programme. Kiore were eradicated from the smaller nearby island (Putauhinu) in order to establish a reserve population of an endemic subspecies of fernbirds (Bowdleria punctata). Additionally 37 ha of ground baiting was undertaken in the core fernbird area on Whenua Hou. 380 southern short-tailed bats (Mystacina tuberculata tuberculata) were caught and held in captivity for over 3 months. Also a quicker-weathering bait type was selected to reduce the period that non-targets would have access to bait. The only species which suffered significantly at a population level, other than the rats, was the fernbirds. However, it appears that sufficient have survived to repopulate Whenua Hou. Following monitoring, the Whenua Hou was declared rat free in December 2000. Putauhinu and Rarotoka are effectively privately owned by Maori for the purpose of muttonbirding. As kiore eradications for islands of these smaller sizes are relatively straightforward, it was decided to remove the kiore from nearby Rarotoka (Centre Island) during the Putauhinu preparation. Eradications were carried out in July 1997 in conjunction with the "owners" of the 2 islands and both were declared successful in July 2000. The Putauhinu eradication has enabled robins (Petroica australis) to be reintroduced along with fernbirds and has lead to a dramatic increase in the populations of other native birds and lizards and significant increase in the regrowth of many of the plants. Rarotoka was heavily grazed for many years prior to the eradication, meaning that the removal of rats is simply the second step in the restoration of the island, which will involve ongoing weed control and a planting programme.
 
 

Preparation for the eradication of Norway rats (Rattus norvegicus) from Campbell Island

McClelland P. J.

Norway rats have populated Campbell Island for nearly 200 years. During this time they, in combination feral cats which have since died out, have had a devastating effect on the island’s fauna, marooning several species of bird to the small rat-free islands around the coast and probably causing the extinction of several other undiscovered species. At 11300 ha the attempt to eradicate rats from Campbell Island will be the largest ever undertaken. The island’s size coupled with its location in the furious fifties (700 km south of the New Zealand Mainland) renown for their strong winds and frequent rainstorms, means the attempt will be stretching the boundaries of current technology. In order to make the eradication logistically feasible the margin for error that has been build into all previous eradications has had to be significantly reduced. Instead of two bait drops totalling 12kg/ha as is usually used, Campbell will be done with a single drop, but with a 50% overlap to eliminate the risk of gaps, totalling 6 kg/ha. This technique was tested in 1999 with a 600ha field trial carried out on the island. Rhodomine dye showed that all the rats in the baited area ate bait and would therefore have been killed. While there is only 70 hours of bait dropping required and 3 helicopters will be used, the short daylight hours on Campbell during the winter and the predictably bad weather, mean that the project team must plan to be on the island for up to 3 months. Non target issues are minimal, with the only priority species at direct risk being southern skua (Catharacta skua), which fortunately are absent from the island until mid August. However if the drop is delayed not only will skua be affected but so will the large colonies of Mollymawks which nest at the North end of the island and present a real risk to the helicopters. The drop will be carried out in July - September 2001 with no follow up until 2003. There will only be one attempt and we either succeed or fail, there is no halfway.
 
 

Island Quarantine – prevention is better than cure

McClelland P. J.

While there is currently an international focus on the eradication of introduced invasives particularly rodents, it is easy to forget that usually the easiest, cheapest and often the only way to avoid the significant impacts that invasives can have on an island ecosystem, is to prevent them getting to the island in the first place. There are numerous examples of "near misses", where rodents in particular have made it on to islands or have only failed to do so by shear luck. That these incidents did not result in the establishment of the predators can only be put down to good fortune. It is vital that in future precautions are put into place that mean that we don't rely on luck to keep islands free of introduced species. To date quarantine precautions have focused largely on rodents, and they have been the flagship of the quarantine battle. This has been very productive as everyone hates rats and can understand/relate to the damage they can do on an island. However with increased awareness and knowledge it is apparent that other species can pose a greater albeit not so obvious risk, this includes invertebrates, plants and even micro-organisms. Standard precautions such as sealed containers work well for rodents but do little to prevent the introduction of a vast and ever increasing number of weed species and invertebrates. This is at a new level and relies heavily on the individual taking responsibility. Basic precautions such as scrubbing footwear and checking pockets are simple and can significantly reduce the risk, but are often overlooked. All visitors to islands, tourists, researchers and managers pose a quarantine risk, and a quarantine plan must be practical for all the situations relevant to the specific islands. Currently there are significant resources being dedicated to trying to stop invasives, rodents in particular, getting established once they make it to an island. However these contingencies are often expensive and are at best unreliable. Further research is required on the effectiveness of the various options eg traps v's toxin, and also on the behaviour of rats reaching a new island.
 
 

Alien animal eradication and quarantine on inhabited islands in the Seychelles

Don Merton, Gideon Climo, Victorin Laboudallon and Serge Robert

The biological and conservational benefits of eradicating alien pest animals from uninhabited or sparsely inhabited islands has long been recognised. However, the practicability of permanently removing such pests (especially rodents) from oceanic islands supporting human settlement and/or major development, together with information on any enduring ecological benefits, appear largely unknown. Eradication, and in particular effective rodent quarantine on such islands has generally been considered impractical – if not impossible. During the period 1996 – 2000 eradication of five mammal (including three rodent) species, and local control of an voracious invertebrate predator of small animals, the crazy ant, was attempted on four inhabited islands, including three resort islands, ranging in size from 100 – 300ha in the Seychelles (tropical Indian Ocean). Objectives were to avert extinction and to restore urgently needed habitat for localised threatened endemic animals, and to facilitate ecological restoration in line with a national biodiversity strategy. Although too soon to gauge the outcome on three of the islands, one (Bird Island) has been maintained free of rats and rabbits since their eradication in 1996. Strategies and techniques developed, and some of the many challenges encountered in conducting eradication and quarantine programmes on remote, inhabited, tropical islands are outlined, together with progress to date. The apparent success of the projects will, we hope, inspire similar effort on other inhabited islands with high biological values or potential.
 
 

The role of parasitoids in eradication or area-wide control of tephritid fruit flies in the Hawaiian Islands

Messing, R. H.

Koinobiont larval endoparasitoids (Hymenoptera:Bracondae:Opiinae) of tephritid fruit flies have been used with variable success rates in classical biological control, but have not previously been considered useful adjuncts to eradication programs. Successful eradication of tephritid flies from other Pacific Islands has been accomplished by utilisation of semiochemical-toxicant combinations (male annihilation) and/or the sterile insect technique (SIT). However, pilot projects designed to show feasibility of medfly (Ceratitis capitata) eradication on the Hawaiian island of Kauai failed because of the inability to achieve adequate overflooding ratios of sterile:wild males. Medfly population reduction sufficient to achieve suitable overflooding ratios is not possible with insecticides due to the location of populations in remote and environmentally sensitive areas. Population dynamics models demonstrate the synergistic effect of combined augmentative parasitoid releases with SIT. More cost effective would be the release of new species or strains of parasitoids that are self-perpetuating and dispersing. The potential is demonstrated by Fopius arisanus from Asia, which causes over 95% egg mortality of Oriental fruit fly (Bactrocera dorsalis) in guava. Comparable levels of parasitism for melon fly (B. cucurbitae) in cucurbits might make eradication in Hawaii feasible.
 
 

Eradication of rats and rabbits from Saint-Paul Island, French Southern Territories

Micol, T. and P. Jouventin

Introduced Rattus rattus (black rats) appear to have decimated the seabird colonies on Saint-Paul Island in the Southern Indian Ocean (approximately 800 ha). Six of the 13 seabird species originally breeding on Saint-Paul are now represented by only a few individuals confined to a 2 ha islet (Roche Quille) located 150 m from the main island. The persistence of these remnant populations led us to believe that recolonisation was possible on Saint-Paul Island if all rats were removed from it. French Southern Territories therefore decided to eradicate rats from Saint-Paul Island. Part of the funding was provided by the European Commission. Because rabbits disturbed some burrowed nesting species and had a non-negligible impact on vegetation, we decided eradicate rabbits also. Aerial dropping was chosen in view of the large size of the island and the inaccessibility of cliffs. Brodifacoum bait (Talon 20P) was used as its efficiency was already confirmed in operations in New Zealand. Two preliminary campaigns were conducted, in 1995 and 1996, in order to assess (a) the palatability of baits to Saint-Paul’s target species, (b) the distribution of rats and rabbits on the island, (c) the risk of killing non target species (subantarctic skuas), and (d) the colonisation potential from Roche Quille. In January 1997, 13.5 tonnes of bait were spread by helicopter from a monsoon bucket. Saint-Paul Island was intensively checked for rats during three months after the dropping. During the same time 48 rabbits were killed using dogs, shooting and traps. More bait was spread by hand in some areas. Two follow-up operations were carried out, in November-December 1997 and December 1998-February 1999. Snap trapping and searches failed to detect the presence of rats. Eighteen and five rabbits were killed in 1997 and 1999 respectively, and no sign of rabbit was found when we left the island the 22th February 1999. We are now confident that black rats were eradicated following the single dispersal of bait. Eradication of rabbits still needs to be confirmed. The Saint-Paul Island project demonstrates the efficiency of the aerial technique against rats (black rats here) but that rabbit eradication needs a more sustained effort. Breeding of endemic Macgillivrayi’s prion and great-winged petrels has already begun on Saint-Paul Island.
 
 

Direct and indirect effects of house mice on declining populations of a small seabird, the ashy storm-petrel (Oceanodroma homochroa), on Southeast Farallon Island, California

Mills, K. L; P. Pyle, W. J. Sydeman, J. Buffa, and M. J. Rauzon

There is concern over severe population decline of the ashy storm-petrel (Oceanodroma homochroa) on Southeast Farallon Island (SEFI), California (37^oN 123^oW). Evidence from a mark-recapture analysis suggests that a primary cause of this decline is increased predation on this species, whose main predators include expanding populations of western gulls (Larus occidentalis) and migrant burrowing owls (Athene cunicularia). There is evidence that the introduced house mouse (Mus musculus) may occasionally prey upon ashy eggs and chicks, although the extent of this is unknown. Owl arrival in the fall coincides with the peak mouse population, but with decreasing food supplies in the late winter, the mouse population reaches a low point. When this occurs, the wintering owls lose a primary food source and may shift their diet from house mice to ashy storm-petrels, which arrive to SEFI in early spring to begin their breeding cycle. Thus, the indirect effect of mouse presence on ashy storm-petrel populations, through burrowing owls, is perhaps more severe than the direct effects. The U.S. Fish and Wildlife Service is currently considering a proposal to remove house mice from the island. Justification of this action will rest heavily on documentation of adverse affects of mouse presence to the natural ecology of SEFI. Before eradication plans are implemented, all factors, both direct and indirect, must be considered.
 
 

Response of forest birds to rat eradication on Kapiti Island

Miskelly, Colin and Hugh Robertson

Five-minute bird counts were used to determine whether the eradication of kiore Rattus exulans and Norway rats R. norvegicus from Kapiti Island in 1996 had any measurable impact on the diurnal forest bird community. Counts undertaken quarterly from April 1999 to January 2001 were compared with counts undertaken using the same methodology from April 1991 to January 1994. At least four species appear to have increased since rat eradication (red-crowned parakeet Cyanoramphus novaezelandiae, robin Petroica australis, saddleback Philesturnus carunculatus and bellbird Anthornis melanura). None of the 15 species investigated showed evidence of a consistent decline since rat eradication, although two (tui Prosthemadera novaeseelandiae and tomtit Petroica macrocephala) were less conspicuous than in 1991-94 in four of the eight count sessions completed to date. Weka Gallirallus australis were adversely affected by the rat poisoning operation, but had recovered to pre-eradication levels by 1999. The present series of counts will be completed in January 2002.
 
 

Cat eradication and restoration of endangered iguanas (Cyclura carinata carinata) on Long Cay, Caicos Bank, Turks and Caicos Islands, British West Indies

Mitchell, N. C., R. Haeffner, V. Veer, M. Fulford-Gardner, W. Clerveaux, C. R. Veitch

Endangered Turks and Caicos rock iguanas, Cyclura c. carinata, are being displaced on Big Ambergris Cay by an expansive development project. We chose Long Cay, Caicos Bank, as a relocation site for some iguanas because it: 1) is a large (111 ha), uninhabited, protected reserve, 2) previously supported iguanas but did not have a current population, 3) could support thousands of iguanas, 4) had no native mammals, few scavenging birds, and no nesting colonies of scavengers. There was a small population of feral cats, well-known iguana predators. To restore the island, we conducted an intensive cat poisoning campaign using 1080, in July 1999. 460-500 bait stations allowed us to distribute poison baits thoroughly over the 3.5 km long island. Bait stations were flagged with surveyor’s tape, numbered, and spaced 25 m apart in parallel lines a maximum of 50 m apart. Fish chunks (Clupeidae) or whole minnows (Atherinidae) were injected with 0.009 ml of 1080 in a 22% solution and placed at a height of 15 cm on branches overhanging clearings and trails or, in areas without vegetation, on inverted plastic cups filled with sand to prevent them from being displaced by wind. Bait was laid at the stations between 1600–1900 h to minimise exposure to heat and scavenging birds. Old baits were collected when fresh bait was deposited daily for 5-6 days. At the end of the week, leftover toxin and contaminated items were diluted to non-toxic levels and disposed of or burned, respectively. In November 1999, a three-day survey revealed no evidence of cats on Long Cay, and we proceeded with iguana relocation. A test-group of 25 iguanas, was taken from Big Ambergris Cay to Long Cay. Since their successful establishment we have relocated almost 400. Occasional trapping may be necessary, to maintain Long Cay free of cats. We have begun patrols and courtesy visits to vessels cruising the area, installed informational/cautionary signs, and have produced public service announcements for television, to reinforce the importance of keeping domestic animals away from uninhabited islands.
 
 

Evaluation and cost-benefits of controlling house mice Mus domesticus on islands: an example from Thevenard Island, Western Australia

Dorian Moro

Feral mammals occur on many offshore islands around Australia, and their eradication by poison-baiting requires a careful assessment of the associated costs and long-term benefits. Past studies to eradicate or control house mice, Mus domesticus, have rarely been designed to reduce the impact on non-target and native species of rodents. Indiscriminate poison-baiting on an island reserve off the Western Australian coast required management actions to control or eradicate house mice in the presence of a native short-tailed mouse, Leggadina lakedownensis. I describe a trial program to poison house mice selectively on Thevenard Island, and evaluate whether the eradication of house mice from the island is an achievable and cost-effective goal. Laboratory trials were first conducted to ascertain an effective bait station that would allow house mice greater access than native mice to a poison-bait, yet be simple and inexpensive to construct. Three bait stations were tested for their effectiveness at selectively capturing house mice, or for the selective delivery of bait, and one showed promising results; field trials were then conducted to test this design and identify an optimal spacing for the bait-delivery stations. It was found the densities and survivorship of house mice declined significantly more on grids with bait stations spaced every 10m than on grids baited every 20m when compared to unbaited (control) grids. On one grid baited every 10m, the abundance of house mice declined by 83% 22 days after baiting commenced. This decline was correlated with an increase in the abundance of short-tailed mice L. lakedownensis on the grid. A total of 33 person days was required to conduct the present poison-baiting trials, and expenses incurred were approximately $AUS1110 ha^-1. Baiting islands for house mice can prove a costly exercise, and management authorities need to assess whether ground-based baiting is an efficacious and cost-effective management option if complete eradication fails.
 
 

The control of the introduced black rat on Barrow and adjacent islands off the north west coast of Western Australia

Morris, K. D.

The black rat (Rattus rattus) has been introduced to many islands around the world and has been shown to have a detrimental impact on a wide range of fauna. They are known from about 10 % of Australian Islands, many of these are off the Western Australian Pilbara or Kimberley coasts, where rats were accidentally introduced in the late 1800s by the pearling industry. Barrow and adjacent islands are nature reserves with significant conservation value, particularly for threatened mammals. Rats were known to inhabit the six smaller adjacent islands, but it was not until 1990 that they were located on the south end of Barrow Island. Eradication programs on North and South Double, Boomerang, Pasco and Boodie Islands in 1983-6 have been successful, but most of these islands had no non target mammals. Seven mammals were considered to be at risk from an oat based baiting program on Barrow Island. Barrow Island was also considerably larger than other islands where successful eradication had occurred (23 000 ha vs 5-1 000ha). The rats on the smaller islands, without non target mammals, were successfully eradicated using oats impregnated with the anticoagulant Pindone. Baits were laid on the ground in a 25m grid. On Boodie Island unsuccessful attempts were made at covering the oat baits to prevent access by the threatened burrowing bettong (Bettongia lesueur). While the rats were eradicated, the bettongs also disappeared. They have since been successfully reintroduced. Fortunately on Barrow Island, the rats were found to be present only in 270 ha at the south end of the large island. A bait station was designed that allowed climbing access by the black rats (and native rodents) but prevented access by other native mammals. These bait stations were set on a 25m grid throughout the area where the rats occurred. This eradication program has been successful and the native rodents have since reinvaded the area. These bait stations were also used to eradicate rats on Middle Island where the golden bandicoot (Isoodon auratus) also occurs. Ongoing monitoring of these reserves is continuing.
 
 

Eradication of possums and wallabies from Rangitoto and Motutapu Islands

Mowbray, S. C.

The graceful symmetrical form and triple peak of Rangitoto Island is recognisable from many places in Auckland and in turn provides from its summit a panoramic view of the Hauraki Gulf and Auckland metropolitan area. So unique is this small area of New Zealand that it has been proposed as a separate ecological district. It is a relatively young volcanic landform (about 700 years old) with unusual native plant communities having a high level of endemism. Rangitoto Island has considerable scenic, recreational and scientific significance in New Zealand and internationally. Motutapu Island is an older landform and its landscape - rolling green pastures and coastal Metrosideros trees - differs markedly from its immediate neighbour, a young lava field with developing Metrosideros canopy. The islands are linked by a narrow causeway and bridge which meant any eradication operation had to cover the total area of 3850 hectares. In 1990 the operation began with a 1080 aerial drop on Rangitoto Island, achieving an estimated 93 percent kill of possum and wallaby. This was followed from 1990 to 1997 by ground work on both islands to complete the eradication of both species. Methods used were trapping, cyanide poisoning, dogs and spotlighting. This was followed by several years of ground monitoring and mop up operations. Aerial surveillance (FLIR) was conducted on two occasions to survey the islands for surviving animals. A DGPS navigational system logged flight lines and animals sighted. It was then interfaced on video footage so the survey information could be displayed in real time. Rangitoto Islands provided an extreme and rugged terrain for the hunting team, which included dogs. On the bare rock the summer heat was intense and lava terrain made walking very difficult. There were successes and failures with the various field methods employed in this operation. From recent survey it appears eradication of an estimated 21,000 possum and 12,500 wallaby was achieved in the eight years of the operation. This eradication operation has been successful in restoring the battered Metrosideros tree canopy to both islands.
 
 

Eradication of feral goats from Lord Howe Island

Parkes, J. P. and N. L. Macdonald

Lord Howe Island is a 1455-ha World Heritage site in the south Pacific Ocean that was one of the few Pacific islands never settled by humans until European discovery in 1788. It’s unique biota is a mix of mainly New Zealand and Australian elements, with the usual extinctions of animals since the arrival of exotic pests. Feral cats and feral pigs were eradicated in the 1980s, and mice and ship rats remain. Feral goats were eradicated by the islanders from the lower-altitude northern half of the island by the early 1970s, but remained on the more-intact habitats (about 700 ha) on the southern mountains despite frequent control. This paper reports on the planning and execution of an eradication campaign completed in 1999. We estimated, from changes in kill-rates and guesses at rate of increase, that there must have been about 200 goats on the island, and that eradication was possible using a combination of ground hunters and dogs and helicopters at a cost of $NZ 100 000. The ground hunters shot 106 goats in 168 hunter-days and 189 goats were shot from the helicopter. There are always more pests than first estimated! Techniques used, monitoring procedures, risks to non-target animals (the endemic woodhens), and some of the social risks of failure (the presence of domestic goats and some opposition from some islanders) will be described.
 
 

Kiore: their impacts on two small seabird species in the Hen and Chickens Islands

Pierce, R. J.

The Hen and Chicken Islands support introduced kiore (Rattus exulans) populations and also remnant populations of two small burrow-nesting seabirds, the summer breeding, Pycroft’s petrel (Pterodroma pycrofti) and the winter-nesting little shearwater (Puffinus assimilis haurakiensis). The sequential eradication of kiore from the larger Chickens in the 1990s provided an opportunity to measure the responses of these seabirds to kiore removal. The following hypotheses were tested: (1) Breeding success of the two seabird species is not limited by kiore presence, (2) Breeding success is not limited by the presence of tuatara (Sphenodon punctatus) an endemic predatory reptile, (3) The two seabird species are not in competition with each other. Two study islands were used, Coppermine and Lady Alice Island. Study burrows were checked early and late in the seasons to determine breeding success. Success was significantly lower when the study islands contained kiore: little shearwaters averaged a 16% breeding success in the presence of kiore and 61% in the absence of kiore, while Pycrofts petrels averaged a 33% breeding success in the presence of kiore and 57% in the absence of kiore. Contemporaneous data for the two islands enabled other factors such as food supply and heavy rainfall to be eliminated as compounding variables. For example, the lowest breeding successes of little shearwater (5%) occurred in a kiore-present scenario for two years on Coppermine, but in the same two years productivity was high on kiore-free Lady Alice Island. Similarly, for Pycroft’s petrels, the lowest years of breeding success were in kiore-present scenarios, but in the same years there was significantly higher productivity in the kiore-free scenarios. The presence of tuatara in burrows did not significantly influence the breeding success of these seabirds, at least in a post-kiore scenario. However, in burrows used by both species of seabirds, late-fledging little shearwaters disrupted the nesting of Pycroft’s petrels, causing some pairs to be displaced to other burrows or abandon nesting for the season. Currently, the effects of this competition on Pycroft’s petrel are small and more than compensated by their increased productivity following kiore removal. In conclusion, these findings demonstrate clear negative impacts of kiore on small seabird productivity. They are consistent with the recorded decline in seabird populations on the Hen and Chickens Islands during the 20^th century. Count data from Lady Alice Island from 1992 to 2000 indicate that this population decline has been halted and apparently reversed.
 
 

Invasive red mangrove eradication in a Hawaiian wetland, waterbird responses, lessons learned

Rauzon, Mark J. and Diane Drigot

Introduced red mangrove are a major pest in Hawaii's wetlands, including Nu'upia Ponds, a wildlife reserve and ancient Hawaiian fish pond on US Marine Corps Base Hawaii. These ponds are home to approximately 10% of Hawaii's endemic black-necked stilt population. After 20 years, thousands of volunteers, and over $US 2 million dollar investment, approximately 20 acres of mangrove were eradicated here. Monitoring showed that stilts readily colonise cleared mudflats near established breeding areas. Challenges included: complex permit requirements; difficult vehicle manoeuvres in muddy terrain; need for labour-intensive hand removal in archaeological areas; and unexpected biotic developments. For example, black-crowned night herons and cattle egrets colonised the mangrove, causing work schedule alterations and hazing permits. An opportunity was forgone to sustain three self-introduced great blue herons, who prefer mangrove roosts. Lessons learned regarding waterbird conservation and guidelines for other agencies with similar invasive species challenges are discussed.
 
 

Seabird recolonisation after cat eradication on equatorial Jarvis, Howland and Baker Islands

Rauzon, Mark J.; Douglas J. Forsell, and Elizabeth N. Flint

In 1990, the last cat was removed from Jarvis Island National Wildlife Refuge in the Central Pacific Ocean. Cats were removed from two other equatorial islands: Baker Island in the 1960s and Howland Island in 1986. Introduced during the 1930s, cats had extirpated some species of terns, small procellarids, Polynesian and Norway rats from Jarvis, Howland and Baker Islands. With cat and rat eradication, previously extirpated seabirds; blue-gray noddies, Christmas and Audubon’s shearwaters and white-throated storm-petrels have recolonised Jarvis Island. Baker Island has been recolonised by wedge-tailed shearwaters and hundreds of thousands of birds that moved from Howland Island. Small tern populations are returning to Howland Island as well as increased numbers of wintering shorebirds. Even though the predation-free period is longer for Howland and Baker than Jarvis Island, fewer new species have recolonised probably because of the great distance to other colonies that could serve as a source.
 
 

Exotic pest plants: when is complete eradication a realistic goal?

Rejmanek, M. and M. J. Pitcairn

There are several encouraging examples where widespread alien animals have been eradicated. Can we eradicate equally widespread and difficult alien plants? For both plants and animals, the extent of an infestation, its configuration in the landscape, and the amount of resources available to managers are key issues. Using a unique data set on eradication attempts by the California Department of Food and Agriculture (16 species and 50 separate infestations targeted for eradication in the period 1972-1999), we can show that professional eradication of exotic plant infestations smaller than one hectare is almost always possible. Also, about 1/3 of all infestations between 1 and 100 ha have been eradicated. Costs, however, increase dramatically. With a realistic amount of resources, it is very unlikely that infestations larger than 100 ha can be eradicated. Early detection of the presence of an invasive taxon can make the difference between being able to employ feasible offensive strategies (eradication) and the necessity of retreating to a defensive strategy that usually means an infinite financial commitment. Obviously, a substantial increase in resources for exclusion and early detection of exotic weeds would be the most profitable investment. Nevertheless, depending on the potential impact of individual exotic species, even infestations larger than 100 hectares should be targeted for eradication effort or, at least, substantial reduction and containment. However, if an exotic species is already widespread, then species-specific biological control is usually the only practical way of switching between defensive and offensive strategies.
 
 

Management of Malvaceae on islands off Perth, Western Australia

Rippey, M. E., J. J. Rippey, and J. N. Dunlop

The number of native plant species on small seabird islands off Perth, Western Australia has declined significantly over the past 40 years (eg Penguin Island 12% reduction, Bird Island 82% reduction, Seal Island 67% reduction). This has coincided with the invasion by European Tree Mallow (Malva dendromorpha) that developed dense stands up to 3 metres in height to the exclusion of other species. This appears to interfere with seabird nesting by physically excluding some species and altering the breeding habitat for others. The objectives of this study were to reduce or eradicate the growth of M. dendromorpha and revegetate with perennial species previously native to the island. Methods: Seal Island was chosen for experimental management. Much of the plateau of this 1.2 ha island was dominated by M. dendromorpha. The Malva was either uprooted or cut and the stumps poisoned with glyphosate (15%). Germination was suppressed in test plots by hand tilling to bury the seed or with weed mat. Revegetation with native perennials during winter rains was carried out in untreated areas and in weed mat. Results: Initial eradication attempts were unsuccessful. Uprooted plants often rerooted, weeding was followed by new germination and new growth reached 1.5m and flowered in four months unless eradication was repeated. At the end of the second growing season and four eradication attempts, the germination of M dendromorpha is thinning, to be replaced by alien annual grasses and herbs. Replanting with native perennials at the time of cutting was swamped by this growth. Weed mat was successful in preventing regrowth of weeds and allowing revegetation to become established but is expensive and difficult on an island site. Tilling inhibited the growth of weeds and rotivation may be of use in weed control during revegetation. In conclusion, eradication of established M dendromorpha populations on seabird islands and restoration of the original habitat is probably impossible. Islands on which Maustraliana grows seem to be at risk of invasion and monitoring and early removal of M dendromorpha is the best defence. Research is continuing.
 
 

Practical concerns in the eradication of island snakes

Rodda, G. H., T. H. Fritts, E. W. Campbell, G. Perry, and C. P. Qualls.

Successes trapping the introduced and invasive brown treesnake (Boiga irregularis) from 1-4 ha areas on Guam led us to suggest that the snakes could be eradicated from large nature reserves if emigration of snakes from adjoining areas could be eliminated or greatly reduced with perimeter snake barriers. Practical problems in the design of snake barriers on Guam include the extraordinary abilities of brown treesnakes to climb, high levels of rat damage to chewable barrier surfaces in snake-reduced areas, and frequent and destructive cyclonic storms. A variety of snake barrier designs have been developed to overcome these problems, and one 23-ha site on Guam has been effectively trapped out following erection of a snake-fence around the perimeter. Unfortunately, the snake-fence used in this situation is relatively vulnerable to storm damage, but authorities are reluctant to commit the substantial funds needed for a more permanent structure. Another problem with the design is that it does not exclude cats, which have preyed heavily on endangered Guam rails (Gallirallus owstoni) repatriated to the snake exclosure. One emerging concern about our fence-and-trap approach is that the snake traps currently in use are relatively ineffective against small snakes. Retrapping in subsequent years may be needed to eliminate snakes that were too small to be trapped during an initial eradication effort. Eradication efforts on newly colonised islands such as Saipan are probably also subject to this problem with snake trap size selectivity. New colonies may also be refractory to trap-based eradication as a consequence of reduced capture success in a prey-rich environment associated with an incipient or sparse predator population. The problem with trapping in a prey-rich environment is likely to crop up in all eradication efforts that occur over a period of time long enough for prey species to irrupt following initial predator density reduction. Both small-snake size selectivity of trapping and problems trapping in a prey-rich environment may have had an impact on our unsuccessful preliminary effort to trap Wolf Snakes (Lycodon aulicus), an invasive introduced species on Ile aux Aigrettes, a nature reserve near the coast of Mauritius.
 
 

An ecological basis for improving the management of introduced mongooses on Mauritius

Roy S., S. Harris, and C. G. Jones

The small Indian mongoose Herpestes auropunctatus is a highly successful animal in its native range. Due to its success it was introduced to many locations around the world to control vertebrate pests in agricultural habitats. Many of these locations are oceanic islands, where the mongoose became an ecological pest and began to decimate populations of endemic species. It has been blamed for the extinction of many endemic vertebrate species on islands. On Mauritius, many endemic species still survive, largely due to intensive predator trapping regimes. From 1997 to early 1999, the population and spatial ecology of the species was studied with a view to improving the current control efforts for mongooses. Using trapping, capture-mark-recapture, indirect census techniques, and radio-telemetry, the population densities and habitat preferences of the small Indian mongoose was assessed at both the population and individual level. Densities were found to vary considerably from season to season, and in different habitats, ranging from five to thirty animals per square kilometre. Mongooses significantly preferred low-grade woodlands and riparian habits to pristine forests, sugar cane and coastal areas. Population sizes were estimated for the National Park and the mainland of Mauritius. Using these estimates, the populations of mongooses within conservation areas have been modelled, and the potential impacts on native species have been estimated. Field based population parameters of mongooses and endemic species have been used to model the control efforts required to maintain the species to within acceptable ecological limits. A management plan for the species has been drawn up, and data on the habitat preferences and spatial ecology of the species have been incorporated into the management recommendations to optimise allocation of control resources in space and time. The methods, results and models from the study of the mongoose on Mauritius can be applied to other oceanic islands where the mongoose is a problem. This study concludes by investigating the feasibility of eradicating the mongoose from areas of ecological importance.
 
 

Managing pest mammals at near-zero densities at sites on the New Zealand mainland

Alan Saunders

Six "Mainland Island" projects are being undertaken by the Department of Conservation at sites on the North and South Islands of New Zealand. A feature of these projects is the range of pests targeted for control and the relatively low pest densities achieved as a result of pest control operations. These early results are significant in that they suggest that effectively managing the impacts of mammalian carnivores and herbivores is achievable on the mainland as well as on off-shore islands. The challenge now is to develop more efficient pest control regimes so that conservation outcomes may be sustained. Better targeting and timing of pest control and more effectively controlling pest re-invasion rates will result in further advances in our capacity to conserve native biodiversity. In view of the urgency in implementing more effective conservation programmes to arrest further declines, and recognising our inability to predict ecological outcomes from intensive, multi-pest control programmes, an adaptive experimental management approach has been proposed in order that our understanding of pest impacts and ecological responses is enhanced as part of the management process.
 
 

Control of feral goats (Capra hircus) on Santa Catalina Island, California, USA

Schuyler, Peter T., Dave Garcelon and Stan Escover

Santa Catalina Island, a mountainous, 194 kilometer² island is the third largest of the eight California Channel islands. In addition to numerous endemic species, it also has a resident human population of approximately 4,000 people and nearly 1,000,000 visitors per year. The Santa Catalina Island Conservancy owns and manages 88% of the island with a primary goal of natural resource protection while still allowing appropriate public access. Among Catalina’s non-native mammal species are feral goats (Capra hircus) which were well established by the mid 1800s and may have reached a population high of 30,000 in 1930s. Impacts by goats on natural resources has been severe, including destruction of endemic plant species and island plant communities, increased erosion, and soil compaction. Although sport hunting and island resource managers removed large numbers of goats throughout the years, sizeable populations remained until the 1990s. From 1990 to 1994, ground and aerial hunting removed over 7700 goats from the island, but due to lack of funding the program was stopped after approximately 95% of the goats were removed. In 1996, a new effort was initiated to remove all goats from the west end of the island. By early 1998, the only goats known to be in this area had telemetry tracking collars attached and the program was expanded to include the rest of the island. During the next six months over 600 goats were removed by hunting. Following a community forum in January 1999, a outside animal welfare organisation submitted a live capture proposal. The Conservancy Board elected to suspend hunting to try the proposal. In the fall of 1999, 121 goats were captured and shipped off the island. In January 2000, permission to resume hunting was granted and 66 goats were removed. Shortly thereafter, another live capture proposal was submitted and the Board elected to follow a live capture program for all remaining goats on the island (estimated 25-30). By the start of 2001, all uncollared goats should be removed and we hope to remove the collared goats by the end of 2001, thus reaching the objective of zero goats.
 
 

Control of feral pigs (Sus scrofa) on Santa Catalina Island, California, USA

Schuyler, Peter T., Dave Garcelon and Stan Escover

Santa Catalina Island, a mountainous, 194 kilometer² island is the third largest of the eight California Channel islands. In addition to numerous endemic plant, bird, mammal, and invertebrate species, it also has a resident human population of approximately 4000 people and nearly 1,000,000 visitors per year. The Santa Catalina Island Conservancy owns and manages 88% of the island with a primary goal of natural resource protection while still allowing appropriate public access. Among Catalina’s non-native mammal species are feral pigs (Sus scrofa) which were introduced in 1932. Direct and indirect impacts by pigs on natural resources has been severe, including destruction, as well as reduced reproduction of both flora and fauna, increased erosion, and impaired water quality. Although sport hunting removed pigs throughout the years, systematic island wide control efforts were not attempted until 1990. From 1990 to 1996, over 4350 pigs were removed, but the program was stopped before 100% removal was reached. In 1996, a new effort was initiated to remove all pigs from the west end of the island (20% of total island and isolated by fencing). Baited live traps, spotlighting, opportunistic shooting, and/or location/capture by dogs, removed nearly 700 pigs in a 18 month period. By early 1999, no pigs were known to be in the area although periodic monitoring continues to take place. Based on this removal success, the Conservancy’s Board of Directors approved a 5-6 year program to remove all pigs from the remainder of the island. Funding was identified and work began in July 1998. Fifteen miles of new pig proof fence was constructed to create three additional management zones. A zone by zone removal approach is currently under way. Work in the most challenging zone that includes the island’s main town should be completed by May 2001. In addition to the logistical and technical concerns of working so close to town, a public educational component was added to the program. By 2002, close to 99% of pigs should be removed from the island and we expect to reach our objective of complete removal shortly thereafter through intensive monitoring and spot hunting.
 
 

Today Tiritiri Matangi, tomorrow the world! – Are we aiming too low in invasives control?

Simberloff, Dan

Eradication of invasive non-indigenous species is often viewed as impossible, and there have been several high-profile failures. However, many animals and plants have been successfully eradicated. Although the majority of successes concerned quite circumscribed invasions (eg., on small islands), others rid substantial continental areas of invaders (eg., Anopheles gambiae from northeastern Brazil). Successful eradications usually share four features: (1) Sufficient economic resources must exist for the project to be completed. (2) Clear lines of authority must exist; someone has to be in charge and to be able to compel cooperation. (3) The biology of the target organism must be appropriate. (4) Probability of rapid reinvasion must be low. Further, even when the above criteria are met, a successful eradication can lead to unforeseen problems, such as mesopredator release or a proliferation of exotic weeds at the expense of native plants. Finally, not only can attempted eradication of widely distributed invaders be costly, but it can generate nontarget impacts (eg., on human health or species of conservation concern), the importance of which will be weighed differently by different stakeholders. Thus, successful eradication may be as much a function of political skill and public education as of technology. To date, eradication has been a rather idiosyncratic matter, often resting on the drive and ingenuity of a few people. Partly this has been a problem of generally insufficient attention paid to invasions. Other improvements in management of invasions should increase the appeal of eradication attempts. The development of more comprehensive monitoring and reporting systems, as well as of rapid response procedures, should lead to the more frequent eradication of invasions before they become metastatic. However, even invasions that escape initial elimination and spread widely may be susceptible to eradication. Many invasions that appear suitable for eradication by the above criteria have not been attacked because no one mustered the enthusiasm to try it or generated the requisite political support. In general, we do not know the geographic limits of current technologies. For example, just how great an investment would be needed to rid a large island or substantial continental region of a pestiferous mammal? As with many other aspects of the invasion problem, eradication may largely be a victim of an unwarranted fatalism – we are not doomed to the biotic homogenisation of the earth, but we will surely lose this war if we aim too low.
 
 

Control of the invasive exotic yellow crazy ant (Anoplolepis gracilipes) on Christmas Island

Slip, D. J.

The exotic invasive ant Anoplolepis gracilipes was accidentally introduced to Christmas Island between 1915 and 1934. It remained in relatively low numbers until 1988 when an isolated infestation of very high densities of ants was discovered. In 1998 several more infestations were found. Infestations range in size from less than 1 ha to over 100 ha. Currently about 1400 ha or 14% of forest is infested. In areas of infestation A. gracilipes forms extensive multi-queened supercolonies where high densities of workers are sustained on the forest floor and on most plant surfaces including rainforest canopy species. These infestations have serious impacts on the integrity of the rainforest ecosystem of Christmas Island by eliminating the dominant red crab Gecarcoidea natalis from infested areas. The red crab imposes regulatory controls on seedling recruitment and litter breakdown, and their removal results in a rapid transformation of the rainforest ecosystem in terms of habitat structure, species composition and ecosystem processes. The ant infestations also pose a serious threat to endangered species of birds and reptiles. Parks Australia has developed a methodology for a chemical control program. The key requirements of this program are that the bait be (a) highly attractive to ants such that they monopolise the baits, (b) slow acting so that maximum transfer of bait occurs among individuals, (c) effective over a wide range of concentrations, and (d) not detrimental to non-target species. A number of chemicals and attractants were tested and the most effective bait was fipronil in a fish protein base. Broad acre distribution at a rate of 0.5 grams active ingredient per hectare on half hectare plots demonstrated that this bait was effective in reducing ant densities, wiped out ant nests, and had no detectable non-target impacts. Larger scale baiting is currently underway. While complete eradication of A. gracilipes from Christmas Island is probably an impossible task initial baiting trials have shown it may be possible to reduce ant densities to levels where red crabs and ants can coexist. A concurrent research program is being undertaken along with the control program in order to provide better information for the management of this issue.
 
 

Eradication of potentially invasive plants with limited distributions in the Galapagos Islands

Soria, M., M. Gardener and A. Tye

A bold proactive bi-institutional project between the Charles Darwin Foundation and the Galapagos National Park Service has been initiated to attempt to completely eradicate several populations of potentially invasive species from the Galapagos. More than 500 introduced plant species have been recorded in Galapagos- these can be divided into non reproductive, naturalised and invasive. Many of the species in the first two categories are potentially invasive but still have limited distributions and can be eradicated. The aim of this paper is to discuss our attempts at plant eradication and illustrate them with three examples with differing degrees of invasiveness. We use the following protocol. A list of priorities of species to be eradicated and their potential invasiveness is developed with a combination of information from the existing herbarium database, literature, local knowledge and ongoing surveys. The populations of the target plants are then mapped using GIS and a survey of landholders is made to further our knowledge. If an effective control treatment is known for a particular species the team of field workers perform the eradication work, if not, a series of chemical trials are conducted to determine the best technique. Once initial work has been carried out, populations are monitored at three month intervals until plant has not been recorded for three years. Pueraria phaseoloides, a known invasive vine, was recently introduced and has a limited distribution. It has not been seen again since it was treated in 1997. Rubus glaucus, a potentially invasive scrambler, was introduced more than 25 years ago and is sparsely distributed over about 100 ha. It has had two initial treatments and monitoring will continue. The timber tree Citharexylum gentryi is thought to have been introduced more than 40 years ago but was only recorded in 1999. It has many invasive characteristics, has mature reproductive stands and is distributed over about 200 ha. All known reproductive individuals have now been removed. This species will require long term monitoring to eradicate it.
 
 

Preventing rat introductions to the Pribilof Islands, Alaska, USA

Sowls, A. L. and G. V. Byrd

The Pribilof Islands have about three million nesting seabirds, a million northern fur seals, an endemic shrew, and other wildlife. Rat introduction would greatly reduce bird and shrew populations and might transfer diseases to humans and wildlife. The islands have been inhabited since 1786, and although the lack of harbours impeded rodent introduction, house mice became established on St. Paul in 1872. In the early 1990s harbours were constructed on both St. George and St. Paul Islands. A boom of commercial fisheries soon followed and eventual rat introductions seemed a certainty. With the objective of keeping the Pribilofs rat free, a prevention program was begun in 1993 based on cooperation with local communities, government agencies, and industry. The program consists of maintaining trap and poison stations, community education, local shipwreck response capabilities, outreach to vessels to make them rat free, and regulations. Over 450,000 trap nights have passed and several rats have been killed on the St. Paul docks, but there is no evidence of rats becoming established anywhere in the Pribilof Islands. Improved design of preventive stations has decreased maintenance needs. Snap traps have been more effective than poisons, but have caused minor loss to non-target species. Both techniques are recommended. The local communities are taking increasing ownership in the program and it appears fewer ships using the Pribilofs carry rats. Unless there is a major advancement in rodent removal technology, the prevention program will have to be maintained forever. It is too early to be certain that the program is adequate to protect the Pribilof Islands, but as each rat free year passes, hopes are rising. Technical advice from Rowley Taylor, Joe Brooks, and Paul O’Neil was instrumental in the initiation of this program.
 
 

Island conservation in Northwest Mexico

Bernie Tershy, Donald Croll, Jose Angel Sanchez, C. Josh Donlan, Brad Keitt, Bill Wood, and Miguel Angel Hermosillo

The 230+ islands of northwestern Mexico have over 210 species and subspecies of endemic vertebrates. Isolation and aridity have protected these islands from many human impacts, and they have relatively intact biotas. Nevertheless, they have a much higher apparent extinction rate than the adjacent mainland (12% vs. 2% for bird species, 1% vs. 20% for mammals species and subspecies). Introduced mammals are responsible for 93% of these apparent extinctions. Most of the islands are government owned and legally protected. They are managed by the Natural Protected Areas Department under >8 jurisdictions and receive dramatically different levels of actual protection. Islands in the Gulf of California are well protected, but others including the biodiverse Guadalupe, Revillagigedo, and Tres Marias islands are not. The Island Conservation & Ecology Group (ICEG) developed a regional database of native species, introduced species, and extinctions to prioritize conservation action. ICEG, the National Autonomous University of Mexico, Center for Biological Investigations, and National Protected Areas Department collaborated with local people and NGO’s to remove one or more introduced mammals from each of 18 islands and will soon complete eradication on five more. This work has protected >50 endemic vertebrates. Science based, collaborative conservation action has removed introduced mammals from most islands under 4,000 ha in this biologically important region. This experience will make eradications possible on larger, more difficult islands. Once introduced mammals are removed, long-term protection of these islands depends on active management by Mexico’s Natural Protected Areas Department.
 
 

Ecological restoration of islands in Breaksea Sound, Fiordland, New Zealand

Thomas, B. W.

New Zealand has been long renowned internationally for its innovative conservation strategies. Translocation as a conservation management technique was pioneered in Fiordland between 1894 and 1900 when Richard Henry undertook 700 transfers of a range of vulnerable birds such as kakapo (Strigops habroptilus) and kiwi (Apteryx australis) to Resolution Island (and smaller adjacent islands) from nearby mainland sites in Breaksea and Dusky Sounds. This far-sighted project was abandoned when introduced stoats reached the area. Biological surveys of islands in Breaksea Sound in the 1970s resulted in an ambitious island restoration project in which Norway rats (Rattus norvegicus) were eradicated from bush-clad Hawea Island (9 ha) in 1986 and rugged Breaksea Island (170 ha) in 1988. Before poisoning on Breaksea Island, South Island robins (Petroica australis) were transferred to Hawea Island as a precautionary measure. The resultant population is the densest recorded, and they have even dispersed across 300 m of open water to neighbouring Wairaki Island (3 ha). South Island saddlebacks (Philesturnus carunculatus) were released onto Breaksea Island in 1992 with similar success and yellowheads (Mohoua ochrocephala) are confirmed breeding following an experimental transfer from the mainland in 1995. We undertook some of the first experimental translocations of lizards and invertebrates: Fiordland skinks (Oligosoma acrinasum) being released onto Hawea Island in 1988, and knobbled weevils (Hadramphus stilbocarpae) and flax weevils (Anagotus fairburnii) transferred to Breaksea Island in 1991. A programme to monitor ecological change, using several key, indicator species of flora and fauna, was set up before poisoning to document the benefits of eradicating rodent pests. Natural dispersal of Fiordland skinks onto Breaksea Island from a nearby rock-stack was confirmed within two years, and recovery and/or natural dispersal of vulnerable flightless mega-weevils on islands in Breaksea Sound provide further examples of the dynamic ecomarine interface. Despite a lack of funding, monitoring has been maintained, albeit at a reduced level, by using eco-tourism to provide the essential but expensive logistic support and field assistance needed to undertake research in such a remote location.
 
 

The development of a ground-based rat eradication technique for large islands

Thomas, B. W., and R. H. Taylor

For decades, eradicating rats from islands was deemed highly desirable but impossible. In his summing up of The Ecology and Control of Rodents in New Zealand Nature Reserves Conference in 1976, Dr John Yaldwyn (then Director of the National Museum of New Zealand) concluded that eradication of rats on islands remained unachievable. Yet, unbeknown to the Wildlife Service personnel who had undertaken a poison programme to reduce rat numbers at seabird colonies, extermination of Norway rats (Rattus norvegicus) from 32 ha Titi Island in the Marlborough Sounds had unwittingly already been achieved. The 1976 conference provided impetus for several eradication studies on small islands, and the possibility of the unthinkable was also being seriously contemplated - to take Norway rats off the remote and rugged 170 ha Breaksea Island in Fiordland. Amid widespread scepticism and resistance from administrators, we attained this important milestone in 1988 during a 21-day, ground-based, poison campaign in which anti-coagulant brodifacoum baits (Talon 50WB) were distributed in bait-stations on a grid over the whole island. The confidence engendered by this major success led to further cooperation between researchers and conservation managers in an operation on 269 ha Ulva Island in 1992, where significantly wider spacings were used between bait stations. Opportunity was also provided to trial the "rolling front" technique, in which poisoning was undertaken sequentially in sections, a third of the island at a time. Radio telemetry was used to investigate Norway rat behaviour at the interface between poisoned and non-poisoned blocks. This research culminated in the successful operation to eradicate Norway rats from 3100 ha Langara Island in the Queen Charlotte Islands, Canada, in 1995.
 
 

Early detection of new invasive weeds on islands

Timmins, Susan M. and Helen Braithwaite

The New Zealand Department of Conservation’s Weed Surveillance Plan facilitates the early detection and control of invasive weeds across all ecosystems. New incursions on islands are of particular concern because islands often have high conservation value but low visitor (spotter) numbers. On some islands the terrain and vegetation makes it difficult to see a new incursion (eg., hen and chickens islands). Visitors to infrequently visited islands are primed to look out for new weeds but in some cases the new species is not noticeable at the time of year the island is visited. Often weeds escape from old gardens on the islands (eg., Cuvier I, Rangitoto I) or are readily wind or bird-dispersed from the mainland to an adjacent island (eg., Mana Island). There is always a dilemma on finding a new, unidentified species; is it a weed or a threatened plant? The dilemma is greater on island sites where threatened plant status is more likely than on the mainland. But the opportunities to get back to control the weed, should it prove so, are fewer on an isolated island. When white briony was first found it was thought to be a threatened species but once it flowered it was recognised as a weed. Theoretically, islands could have stringent quarantine standards that are not possible between different areas on the mainland. In practice, it is difficult to control landings on islands even those requiring permits, eg., Little Barrier Island. Examples will illustrate the conservation benefits that have resulted from early detection of weeds on islands. The presentation will cover true islands, eg., selaginella found on Raoul and surveillance on Auckland and Waikato islands, plus islands of habitat in a sea of degraded landscape. One such is the regular, six-monthly surveys of Southland’s lakes and rivers for the oxygen weed lagarosiphon. As soon as the weed turns up in any Southland water body, it is detected and controlled, before control becomes impossible as it has in adjacent Otago. New funding, which recognises the importance of weed surveillance, has facilitated increased action against new finds on islands eg, broom on Chathams.
 
 

Eradication of rabbits and mice from subantarctic Enderby and Rose Islands

Torr, N.

In 1993 rabbits (Oryctolagus cuniculus cuniculus) were eradicated from Enderby (700ha) and Rose (80ha) islands in the New Zealand subantarctic Auckland Island group. This was achieved by a widespread poison campaign followed by an intensive second phase which included hunting with a dog, spotlighting and trapping. During the poison campaign a helicopter was used to apply a cereal based pelleted bait incorporating the anticoagulant toxin, brodifacoum, to both islands. Mice (Mus musculus), which were present on Enderby, disappeared during the poison campaign and appear to have been eradicated during this phase. Rabbits and mice were the last of several introduced mammal species present on Enderby and Rose. Without them the unique ecological values of these islands have a chance to recover.
 
 

Unpredicted rebound responses following eradication of introduced rats and rabbits from islands in northern New Zealand

Towns, D. R.

Invasive species that reach islands can have effects that ripple through communities. As a corollary, once invasive species are removed, rebound responses by resident species may also have ripple effects, sometimes with outcomes that are unpredicted. One such unpredicted response is reported on islands off north-eastern New Zealand following the removal of rabbits (Oryctolagus cuniculus) and Pacific rats or kiore (Rattus exulans). As composition of the vegetation changed and geckos became increasingly abundant, a hitherto unknown source of energy for the geckos was revealed: honeydew produced by the scale insect Coelostomidia zealandica (Hemiptera: Margarodidae). Honeydew may have significant effects on the carrying capacity of invertebrates and birds in mainland forests of New Zealand. However, its importance for geckos on islands was apparently masked by reduced gecko abundance in the presence of introduced predators, and suppression of host plants by introduced herbivores. Possible mechanisms of spread and new hosts of C. zealandica are described and the vulnerability of the scale insect on islands with introduced mammals that suppress recruitment of selected host species is emphasised.
 
 

A strategy for Galapagos weeds

Tye, A., M. Soria and M. Gardener

Introduced species are the most serious problem facing the native biota of Galapagos. In addition to 500 native vascular plant species, more than 500 introduced species have been recorded. The worst invasives are woody species, vines and grasses, most having been introduced deliberately. Many have invaded both the National Park and agricultural zones. Invasive plants have only recently been widely recognised as a high priority problem. A strategy for tackling them includes prevention, control, eradication and restoration, and the research required to develop these management actions. A first step is to prioritise the problems, given limited resources for control. A risk assessment system will evaluate both species that are already present in the islands and proposed introductions. The latter forms part of prevention: a quarantine system for Galapagos has been designed, and implementation commenced. Quarantine can slow but never stop introductions, but is essential if the balance between introduction and eradication is to be tipped towards the latter. Control and eradication have two components: research and management. Investigations of the ecology and distribution of introduced plants are helping to determine factors essential to the design of successful control and eradication programmes, such as reproduction and dispersal rates, and seed-bank longevity. Research on effective control techniques that cause minimal damage to native biota is essential, since many Galapagos invasives have not been subject to control elsewhere. Restoration research is focussing on methods of control combined with promotion of regeneration of native communities, such as seeding with native species. Attempts to eradicate potentially invasive species with limited distributions have begun, as detailed in another paper at this meeting. A pilot project will assess the feasibility of eradication of a well-established invasive tree. These measures also act as confidence-builders, changing the climate of opinion about the dangers of introduced species and the feasibility of their control.
 
 

Eradicating Indian Musk Shrews (Suncus murinus) from offshore islands

Varnham, K. J., S. S. Roy, A. S. Seymour, S. Harris, J. Mauremootoo and C. G. Jones

Through human agency, the Indian Musk Shrew, Suncus murinus, has spread from its original ho