The ground-based eradication of Norway rats (Rattus ... · solutions and results of the operation on the permanently inhabited island are covered. Bait stations with cereal-based

269

INTRODUCTION

The Isle of Canna is located off the west coast of Scotland in the Inner Hebrides (6o30’W, 57o03’N), and consists of two semi-connected main islands; Canna (1126 ha) and Sanday (191 ha), and several small offshore stacks and islets (Fig. 1). The Highland Ringing Group, which has monitored the seabird colonies of the Isle of Canna for over 40 years had recorded that seabird populations (in particular razorbills (Alca torda), European shags (Phalacrocorax aristotelis) and Manx shearwater (Puffinus puffinus)) had been declining since the early 1990s (Swann 2002). Brown (Norway) rats (Rattus norvegicus) were identified as the most likely factor influencing this decline, from the observed increased predation on eggs and chicks (Swann 2002). Rats are known to have devastating effects on seabird populations, causing extinctions of birds on numerous islands throughout the world (Moors and Atkinson 1984; Atkinson 1985; Jones et al. 2008). Many islands have been successfully cleared of rats (Thomas and Taylor 2002; Howald et al. 2007) with a subsequent increase in bird populations (Towns and Broome 2003; Jones et al. 2008).

The National Trust for Scotland (NTS) commissioned a feasibility study into the potential for the eradication of rats from Canna (Bell and Bell 2004), based on an earlier proposal (Patterson 2003). The Canna Steering Group, a partnership of NTS, Edinburgh Zoo, Scottish Natural Heritage (SNH), the Joint Nature Conservation Committee (JNCC) and Royal Society for the Protection of Birds (RSPB), decided that eradication of rats using a ground-based eradication technique was to proceed. Wildlife Management International Limited (WMIL) won the tender to direct the eradication with the assistance of NTS volunteers and staff. The three-phase Canna Seabird Recovery Project (Phase I eradication of Norway rats; Phase II monitoring for surviving rats and implementation of quarantine and contingency procedures; Phase III long-term monitoring of seabirds) began in August 2005 (Bell et al. 2006). Complete details of the project are available on the project website (www.ntsseabirds.org.uk).

STUDY AREA AND METHODS

Canna and Sanday are naturally joined at low tide and are now linked by a road bridge. Canna is approximately 8 km long, east to west, and 2 km across at its widest point.

The ground-based eradication of Norway rats (Rattus norvegicus) from the Isle of Canna, Inner Hebrides, Scotland

E. Bell1, D. Boyle1, K. Floyd1, P. Garner-Richards1, B. Swann2, R. Luxmoore3, A. Patterson4, and R. Thomas5

1Wildlife Management International Limited, 35 Selmes Road, Rapaura, RD3, Blenheim, 7273, New Zealand. <[email protected]>. 214 St Vincent Road, Tain, Ross-shire, IV19 1JR, Scotland. 3Scotland. National Trust for Scotland,

28 Charlotte Square, Edinburgh, EH2 4ET, Scotland. 4ABI Wildlife Consultancy, 4 Carnbee End, Edinburgh, EH16 6GJ, Scotland. 5The Royal Zoological Society of Scotland, 134 Corstorphine Road, Edinburgh, EH12 6TS, Scotland.

Abstract: Seabird populations, particularly European shags (Phalacrocorax aristotelis) and Manx shearwater (Puffinus puffinus), on the Isle of Canna have been in decline for several years. Norway rats (Rattus norvegicus) were identified as the most likely factor. The Canna Seabird Recovery Project, developed as a result of this information, was a three year project incorporating the ground-based eradication of Norway rats (Phase I) followed by long-term monitoring, quarantine and contingency for rodents (Phase II) and continued long-term monitoring of the seabird populations (Phase III). The National Trust for Scotland (NTS) contracted Wildlife Management International Limited to direct the first and second phases of the project (August 2005-June 2008), with the assistance of NTS staff and volunteers. Techniques, problems, non-target species, (particularly wood mice (Apodemus sylvaticus) and white-tailed eagles (Haliaeetus albicilla)) issues, solutions and results of the operation on the permanently inhabited island are covered. Bait stations with cereal-based wax blocks containing diphacinone at 0.005% w/w were established on a fifty to one hundred metre grid over the island. Some offshore islets harboured rats, although some of the more sheer stacks did not. Interference with bait stations by non-target species was moderate to high, and bait stations required extra strengthening or protection to prevent damage or disturbance by cattle (Bos taurus), ponies (Equus caballus), sheep (Ovis aries), rabbits (Oryctolagus cuniculus) and hooded crows (Corvus corone cornix). Monitoring confirmed the successful eradication of rats from the Isle of Canna in June 2008. This provides another example of the effectiveness of ground-based rodent eradication techniques and provides an opportunity to restore the seabirds.

Keywords: Wood mouse, white-tailed eagle, diphacinone, monitoring, quarantine and contingency

Pages 269-274 In: Veitch, C. R.; Clout, M. N. and Towns, D. R. (eds.). 2011. Island invasives: eradication and management. IUCN, Gland, Switzerland.

Fig. 1 Location of Isle of Canna, Inner Hebrides, Scotland.

Bell, E.; D. Boyle, K. Floyd, P. Garner-Richards, B. Swann, R. Luxmoore, A. Patterson, and R. Thomas. The ground-based eradication of Norway rats (Rattus norvegicus) from the Isle of Canna, Inner Hebrides, Scotland

Island invasives: eradication and management

270

With the exception of the in-bye land close to the farmhouse and buildings, Canna rises steeply from beach platforms on all sides to a rolling plateau with the highest point at 210 m above sea level. Sanday is approximately 3 km long, east to west, and just over 1 km at its widest point. Sanday reaches 59 m a.s.l., but is similar to Canna with steep coastal cliffs, beach platforms and a low rolling plateau.

Owned and managed by NTS, the Isle of Canna has 15 permanent residents who maintain the farm and crofts, or manage the tourism ventures. The islands (i.e. both Canna and Sanday) are popular with visitors interested in the seabirds, raptors, flora and history. There are several houses, crofts, farm buildings, churches, a lighthouse and school. Canna has a jetty and is regularly serviced by the Caledonian MacBrayne ferry from Mallaig.

The islands are covered by maritime heath, coastal pasture and heather moorland, apart from Tarbert and in-bye areas, which are improved pasture. There are also small areas of mixed woodland adjacent to the in-bye land. The island is grazed by domestic livestock including horses (Equus caballus), sheep (Ovis aries), and cattle (Bos taurus); three feral goats (Capra hircus), and rabbits (Oryctolagus cuniculus). Canna also has three other small mammals; the pygmy shrew (Sorex minutus), house mouse (Mus musculus) and wood mouse (Apodemus sylvaticus). Hedgehogs (Erinaceus europaeus) are also present. A small number of domestic cats and dogs are kept by the residents.

The Isle of Canna (excluding all the in-bye land) was designated a Site of Special Scientific Interest (SSSI) in 1987 for its biological and geological features. Following this, Canna was also designated in 1997 as a Special Protected Area (as part of the European Union NATURA 2000 network of important bird sites) for its internationally important concentrations of breeding seabird species. Shags, razorbills, kittiwakes (Rissa tridactyla), puffins (Fratercula arctica), guillemots (Uria aalge), black guillemots (Cepphus grylle) and fulmars (Fulmarus glacialis) all breed on the island. However, many of these species are now in decline (Swann 2001, 2002). Historically, Manx shearwaters were also recorded to breed on the island, but have declined to almost zero (Swann 2002). Two pairs of white-tailed eagles (Haliaeetus albicilla) and a pair of golden eagle (Aquila chrysaetos) breed on the Isle of Canna, along with buzzards (Buteo buteo), peregrines (Falco peregrinus) and kestrels (Falco tinnunculus).

It is not known when rats became established on Canna; but this is likely to have occurred more than two hundred years ago, either as an accidental introduction with supplies

or from an early shipwreck. Only the Norway rat is known from the island and previous surveys recorded them in all habitat types (Patterson and Brough 1999; Patterson and Lloyd 2000; Patterson and Quinn 2001; Patterson 2003). Distribution, however, was not uniform with the highest densities occurring around the in-bye land, the shoreline and coastal slopes; a pattern common to most island situations where rats are dependent on foraging for food in inter-tidal zones and at seabird colonies. Other notable rat presence on the island was generally related to farming activity, watercourses and rabbit habitat (Patterson 2003).

The eradication option adopted for this project was a ground-based poison programme using protective bait stations to reduce risk to non-target species, particularly the white-tailed eagle and other raptor populations. The programme ran from 25 August 2005 to June 2008 and included bait station establishment, capture of wood mice, poisoning, monitoring, quarantine and contingency, and a final check and rat-free declaration (Table 1). Each operational task was undertaken and completed as follows:

Bait station gridThe bait station grid was established between 3

September and 27 October 2005. Bait stations were made from 750 mm lengths of 100 mm diameter corrugated plastic drainage pipe, pegged to the ground with wire “legs” to prevent movement by wind and/or stabilised with rocks or other material to reduce interference by sheep, cattle, and ponies. Additional wires pushed through both entrances reduced the entrance size to exclude smaller non-target species such as rabbits, hooded crows (Corvus corone cornix) and gulls and to help secure the station to the ground. Both entrances were raised slightly off the ground to deter entry by insects.

Bait was placed in the centre of the station through a small access hole cut in the top which was covered by an additional short clip-on section of pipe as a lid. “Crow clips” (a piece of wire across the station), as used during the Lundy Island rat eradication (Bell 2004), were also used to prevent crows from removing lids to access bait.

Bait stations were placed on a 50-metre grid on the coastal slopes and cliffs, the in-bye area and on Sanday (Fig. 2). On the higher plateau areas on Canna, stations were placed more widely at 100 m (Fig. 2). All areas, except steep or sheer cliffs with no vegetation had bait stations. All offshore rock stacks and islets had bait stations, as did areas with sizeable vegetation below steep cliffs with difficult access. Ropes and a boat were used to access these areas.

Fig. 2 Bait station grid on Canna and Sanday (bait station positions are marked by a black dot).

271

In outdoor areas, each station was marked with a cane and flagging tape to ensure visibility in thicker vegetation or during foggy conditions.

Tube bait stations were also positioned along the Beach Road and at the pier. Philproof and/or Protecta lockable stations were used inside all buildings.

Canna has numerous archaeological sites. WMIL, NTS and Historic Scotland worked together to identify important sites and,whenever possible, bait stations were placed outside recognisable structures (e.g., stone walls and remnant houses). If this was not possible, one or two stations were positioned in the best possible way to minimise disturbance or damage to the site. These sites were identified on maps for the field team and access to all archaeological sites was limited to work purposes only.

The entire grid of 4388 stations was positioned before being individually numbered and mapped using GIS (Manifold). Since Norway rats are reputedly neophobic and can be wary of new items placed in their environment, the grid was left for a period of two to six weeks to allow the rat population time to become familiar with it.

Capture of wood miceIt is thought that the race of wood mouse found on

Canna is distinct, possibly a sub-species of the wood mouse found on the Scottish mainland (Berry et al 1967; Lloyd 2000; Patterson 2003) as it is larger and more golden than the mainland population (Patterson 2003). Eradication programmes can have an associated risk that non-target species will be poisoned either by direct consumption or through secondary poisoning. Principle preventative methods can include the design of the bait station, but the Canna wood mouse was small enough to gain direct access to the bait as well as being at risk from secondary poisoning by eating invertebrates that have eaten the bait. The spacing of the bait station grid meant the chance of accidentally eradicating the Canna wood mouse was unlikely (due to their small home range), but it was decided that a small, but viable ‘assurance’ population would be live-captured and held as two captive sub-populations at Edinburgh Zoo and the Highland Wildlife Park for the duration of the baiting period.

The translocation of wood mice was undertaken by staff from The Royal Zoological Society of Scotland during the bait station establishment period from 8 September to 3 November 2005. Longworth live traps were deployed in a range of habitat types and locations over Canna and Sanday. Traps were run for three nights at each site before being moved to alternative sites. Traps containing bedding materials were baited with grain and invertebrates and checked every four hours. A total of 158 wood mice were captured. All individuals were maintained and transported in North Kent Plastic MB1 laboratory rat cages which

minimised handling, and transferred without loss to Edinburgh Zoo and the Highland Wildlife Park.

PoisoningFirst generation rodenticides were chosen for the

eradication campaign to minimise the risk of secondary poisoning, particularly to birds. The main toxicant used was a 28 g, cereal-based wax block bait with 0.005% active ingredient diphacinone (Ditrac, manufactured by Bell Laboratories). The other rodenticide used was also a 28 g cereal-based wax block bait but with 0.005% active ingredient bromadiolone (Contrac, also manufactured by Bell Laboratories). However, only three blocks of Contrac bait were deployed at one location. Both types of bait were dyed blue (or green/blue), which makes them less attractive to birds.

The bait was delivered to Canna on 28 October 2005 and was transported to depots around the island by tractor and trailer and/or all terrain vehicle (ATV) and trailer.

The poisoning programme (Phase I) commenced on 1 November 2005 and continued through to 6 March 2006.

Baits were present in each station throughout the poisoning programme and replaced as required, when eaten by rats or non-target species and/or damaged by weather. Ten bait blocks were available in each bait station for most of the programme but this was reduced to three blocks when rat activity waned towards the end of the poisoning phase (6 January 2006). By mid February these bait blocks were wired into the stations to ensure missing baits were being taken by surviving rats rather than non-targets as crows and cattle sometimes shook the stations to displace and consume the bait.

The majority of stations (n = 4229) were checked and serviced every three to six days (November and December 2005) or every 15 to 20 days (January to March 2006). However, difficult to access bait stations (n = 66) in tide, weather, rope or boat dependent areas) were loaded with thirty blocks per station and checked whenever possible. Permanent bait stations (n = 93) established in the farmyard and buildings around the island were regularly inspected and maintained as required with ten blocks into each station.

To present the data on bait take gained from these varied bait station checks we grouped the data into 12 periods or checks (mean (±SEM) = 9.6 ± 1.8 days between checks, range 3-22 days) shown as days from baiting (Figs. 3 and 4).

Towards the end of the poisoning phase (15 February - 6 March 2006), when isolated incidents of rat activity, such as teeth marks or droppings, were detected in a monitoring or bait station, an additional bait block was staked inside the entrance of an identified rat hole in the vicinity and/

Table 1 Timetable of activity on the Canna Seabird Recovery Project

Dates Phase Activity27 August 2005

PHASE I

Team arrive on Canna3 September to 27 October 2005 Bait station grid established8 September to 3 November 2005 Capture of wood mice28 October 2005 Poison arrives1 November 2005 to 6 March 2006 Poisoning operation13 December 2005 to 27 March 2006

PHASE II

Intensive monitoring (M1)28 March to 23 September 2006 Long-term monitoring (M2)24 September to 19 December 2006 Intensive monitoring (M3)20 December 2006 to 10 March 2008 Long-term monitoring (M4)11 to 28 March 2008 Final check and quarantine and contingency audit (M5)2 to 9 May 20087 June 2008 Declaration of rat-free status

Bell et al.: Eradication of rats, Canna, Scotland

Island invasives: eradication and management

272

or inside a purpose made ‘natural’ bait station such as a tunnel made from rocks. Both types of bait were used on these occasions, placed out of sight and reach of non-target species by covering entrances with rocks, vegetation or wire netting. These natural sites were marked with flagging tape, numbered and added to the bait station grid.

Bait take was recorded in field notebooks by bait station number and the species believed to have consumed or removed the bait. These data were entered into a data base and large-scale maps showing active stations were produced in real-time to enable the team to effectively monitor bait take activity and target any “hot spots”. All rat corpses found were collected and returned to base for safe disposal to reduce risk for non-target scavengers.

MonitoringFive distinct periods of monitoring were undertaken

as the project progressed (Table 1). Intensive monitoring (M1) using 5296 stations at 50 m spacing was carried out from 13 December to 27 March 2006 to detect rats surviving through the poisoning phase. This was followed by a six-month period of long-term monitoring (M2), from 28 March to 23 September 2006. A second intensive monitoring period (M3), utilising 7608 stations, was completed from 24 September to 18 December 2006 followed by a period of long-term monitoring (M4) using 801 stations from 20 December 2006 to 10 March 2008. These were established at high risk areas on the island; around the coastal seabird breeding sites, Beach Road, at the pier, around the farmyard, in out-buildings, in all properties and around the coast of Sanday (Bell et al 2006, 2007, Table 1). The final check (M5), using 1610 stations, was carried out between 11 and 28 March 2008. WMIL staff and NTS volunteers carried out the intensive and final checks and NTS staff maintained the long-term monthly monitoring over summer. There were two types of monitoring stations using rat attractive food items; one was secured to the ground by a wire and the other was secured inside a tube station. Both were individually numbered and any evidence of activity (i.e. teeth marks) was recorded in field notebooks by station number and the species believed to have consumed or marked the monitoring item.

Monitoring items such as soap, chocolate, chocolate wax, and candles (but most frequently chocolate wax) were placed inside and outside each station. Mud traps of mud smoothed out to detect rat foot prints were established on stock feeding sites, Tarbert Barn, Beach Road and at the pier. Checking for active rat burrows and rat runs, along with trapping at Tarbet Barn, was also undertaken.

Each monitoring site was checked regularly, either separately or together with the poisoning bait station grid. Both rat and non-target species sign found on detection devices was recorded and added to the database. If rat sign (usually tooth marks) was detected, an intensive poisoning (bait stations at 25-m) and trapping (snap traps) programme was established around the site.

RESULTS

Bait acceptance and takeGreen/blue rat droppings appeared within five days

of baiting and rats accounted for 540 kg of bait taken (estimated 3000-5000 rats). The bait take pattern was typical of other Norway rat eradication campaigns (Thomas and Taylor 2002). It was very high six days after original baiting (1st check) and dropped to a relatively low level 28 days after original baiting (5th check). A small increase was recorded at day 32 after the original baiting (6th check), but dropped away to a low level throughout the rest of the poison programme, reaching zero bait take on day 64 after the original baiting (9th check) (Fig. 3).

Throughout the poisoning phase, 62% of bait stations were visited by rats, with 50% active within the nine days of the original baiting. The low percentage of active stations shows that rats were not distributed evenly across the island nor were they in high numbers. This was reflected in bait take levels on the slopes. Ten percent of the bait stations had more than 12 blocks taken, and 3% had more than 21 baits taken by rats. On the plateau, 12 blocks were taken from 4% of the stations and 21 blocks from 1% of stations (Fig.5).

The coastal cliff areas, where breeding seabird colonies are established during summer, also had high bait take by rats, as did sites at Geugasgor, Lamasgor, Iolasgor, and the Nunnery where shag colonies are present during summer. There were few stations on the cliffs or slopes that had no bait take by rats (Fig. 5).

Bait take was also high on the offshore stacks and littoral areas of the main island accessed by boat (Fig. 5). Every bait station on the rock stacks had at least ten bait blocks taken, as did many of the shoreline stations on Canna.

The average number of blocks taken by rats was 8.06 (± 1.01) blocks per active station (n = 2732). The average number of blocks taken per station (n = 4388) was 4.4

Fig. 3 Amount (kg) of bait consumed by rats at each bait check (marked by black dot) during the Norway rat eradication on the Isle of Canna, Inner Hebrides, Scotland.

Fig. 4 Amount (kg) of bait consumed by non-target species at each bait check (marked by black dot) during the Norway rat eradication on the Isle of Canna, Inner Hebrides, Scotland.

273

(± 0.1). However, 38% of stations had no bait taken by rats and 54% had six or fewer blocks taken by rats.

There were low to moderate levels of interference by non-target species (Fig. 4). Cattle trampled bait stations, ate bait (<200 kg) and removed numbered tags. Ponies, sheep and goats had minor impact that was generally related to removing numbered tags or knocking over poles. Stock were not affected by the poison during the eradication. Crows (and other birds) ate moderate amounts of bait (<120 kg) and green regurgitation pellets were seen at known crow roost sites. No birds were found killed by poison during the eradication operation. Wood mice (< 15 kg), pygmy shrews (<2 kg) and insects and slugs (<25 kg) consumed small amounts of bait. Five wood mice and two pygmy shrew carcasses were found during the eradication operation. More than three tonnes of bait was lost to weather events, particularly on the coastal cliffs during storms.

MonitoringMonitoring for rat presence continued island wide

(Canna and Sanday) for two years after the end of the poisoning operation.

Three rats were detected on Canna during the four-month period when monitoring overlapped the end of the poison operation (last rat detected on 20/2/06). These were caught using traps and the alternative Contrac bait. No rats or sign were detected during monitoring after the end of the poisoning operation. Rat-free status for the Isle of Canna was declared in June 2008.

The wood mouse population on both Canna and Sanday recovered quickly after the eradication. Wood mice tooth marks were recorded at more than 75% of the monitoring points during sessions M1 to M5. Rabbits left tooth marks on devices at 23% of the monitoring points and pygmy shrew tooth marks were recorded at 17% of monitoring sites.

DISCUSSION

The success of the Isle of Canna rat eradication campaign shows that a well-planned, adequately resourced, well-executed programme, supported by the landowner and residents and directed by experienced operators can eradicate brown (Norway) rats from a large inhabited island using a ground-based poisoning technique.

Once set up, the island was cleared of rats within six weeks (42 days from original baiting; 9th check) with very few secondary and primary non-target species affected

(and these mainly restricted to wood mice and pygmy shrews). Bait-take showed that the rat population was low to moderate and not evenly distributed across the island. High concentrations on the coastal slopes meant rats would have had an effect on nesting seabirds.

Problems encountered were few and mainly limited to weather and interference with bait and monitoring stations by non-target species. Although wood mice were recorded taking bait and a small number of losses did occur, the population quickly recovered in numbers and range after the removal of rats. Since the wood mouse population was recovering naturally there was no requirement to reintroduce individuals taken into captivity. Ten of the captive Canna wood mice have been used in an unrelated mark-recapture study and the remainder held at the Highland Wildlife Park as a permanent display.

There is no doubt that the eradication of Norway rats from Canna will benefit the recovery of breeding seabirds. Manx shearwaters were presumed to be extinct on the island, but a few individuals were still present (Swann 2008) and the first chick to be recorded on the island in ten years was found and banded in September 2006 (A. Ramsay, Caledonian Ornithological Services pers. comm.). This increases the possibility for successful recovery of the Manx shearwater population. There are also increases in productivity and/or numbers of puffins, razorbills and European shags (Swann 2008; Bob Swann pers. obs.). The seabird populations will continue to be monitored by the Highland Ringing Group.

With rats gone from Canna, it is important that they are never provided with an opportunity to re-establish on the island. As a permanently inhabited island, the greatest risks of rats reaching Canna comes from infested fishing boats mooring overnight, from equipment and food being brought to the island (via the Caledonian MacBrayne ferry or other vessels); and with visitors to the island. A rodent quarantine and contingency plan was developed which minimises the risk of rats being re-introduced, without being too onerous for island residents, ongoing projects, and visitor programmes (Bell et al. 2007, 2008; Bell and Garner-Richards 2006).

Bait stations have been established on the mainland (on Mallaig pier) and on the neighbouring islands of Rum, Eigg, and Muck which are not rat-free, but have some level of rodent control. Bait stations are maintained on the Caledonian MacBrayne ferry and landing-craft that service the island. Bait stations and trapping points have been established on Canna pier, Beach Road, farm buildings,

Fig. 5 Distribution of total bait take by rats (as bait blocks consumed per station) during the entire Norway rat eradication on the Isle of Canna, Inner Hebrides, Scotland.

Bell et al.: Eradication of rats, Canna, Scotland

Island invasives: eradication and management

274

tearooms, food storage areas, residents’ homes, and guest accommodation. Rodent ‘motels’ (i.e. large wooden boxes that act as an attractive shelter and nesting area for rodents into which traps and monitoring items can be placed) have been placed in all high risk areas. All staff and residents on the island have been trained in quarantine methods, rodent sign and detection. One resident NTS staff member has been made responsible for enforcement of quarantine and implementation of any contingency action. Rodent-proof areas have been identified for unpacking suspicious or high-risk containers. All visitors to the island and boat owners mooring offshore are informed of the rat-free status of the island and are asked to be vigilant for rats and rat sign.

A contingency protocol was developed for Canna that details procedures for interviewing persons who report a rat sighting, inspecting the location of the sighting, determining if this is a likely rat event, establishing and maintaining monitoring, trapping and/or baiting grids, identifying tooth marks (or other sign) and reporting and recording all incidents (Bell and Garner-Richards 2006).

Rodents have now been successfully eradicated from islands ranging in size from 1 to 11,200 ha throughout the world. The successful eradication of rats from Ailsa Craig (100 ha; Zonfrillo 2001, 2002), Handa Island (Stoneman and Zonfrillo 2005), Ramsey Island (256 ha; Bell et al. 2000), Lundy Island (500 ha; Bell 2004) and now the Isle of Canna (1300 ha), demonstrates how ground-based poisoning operations can be effectively applied on islands around the UK and Europe. The success on Canna builds on the efforts of many projects that have gone before and lessons learnt will be invaluable for future eradication programmes, particularly those with important non-target species. It also shows that ground-based eradication techniques can be adapted for, and undertaken on, permanently inhabited islands of various sizes, and serves as a good example of the significant long-term benefits that can be achieved through short-term investment.

ACKNOWLEDGEMENTS

This project was funded by the European Union (EU) LIFE Programme, The National Trust for Scotland (NTS), the Royal Zoological Society for Scotland (RZSS) and Scottish Nature Heritage (SNH). We would like to thank all the WMIL staff and NTS volunteers who worked on Canna; all the Islanders (in particular Patrick and Wendy Mackinnon, Kathryn Mackinnon, Gerry Mackinnon and Murdo Jack); Canna Seabird Recovery Project Steering Group; NTS staff (in particular John Hollingsworth, Les Smith and Euan Turner); Graham Tucker (EU); RZSS staff (in particular Gareth Bennett, Edwin Blake, Fraser Dodds and Gavin Harrison); Martin Carty; RJ McLeod’s; Don Merton; Bill Simmons; Bill Cash; Steve Ebbert, Peter Dunlevy and Lisa Sharf (Alaska Maritime National Wildlife Refuge); Will Golland and May-Ellen Spoerke (Bell Laboratories Ltd); Mark Lambert (Central Science Laboratories); Sanitarium (NZ) Ltd; Hubbards (NZ) Ltd; Tunnocks Ltd; Shipton Mill; Caledonian MacBrayne Ltd; Fisher Scientific Supplies; Carberry Candles; Prices Patented Candles Limited; and Rentokil (UK) Ltd.

REFERENCESAtkinson, I.A.E. 1985. The spread of commensal species of Rattus to

oceanic islands and their effects on island avifaunas. In: Moors, P.J. (ed.). Conservation of island birds, pp. 35-81. ICBP Technical Publication 3.

Bell, E.A. 2004. Seabird Recovery Project, Lundy Island, Bristol Channel, UK: 2002-2004: Final Report. Unpublished report to Lundy Seabird Recovery Project Partnership co-ordinated by Royal Society for the Protection of Birds. South-west Regional Office, Devon, UK.

Bell, E.A. and Bell, B. 2004. An assessment of the proposal to eradicate brown rats from the Isle of Canna, Scottish Hebrides. Unpublished report to the National Trust for Scotland, Edinburgh, Scotland.

Bell, E.A.; Boyle, D.; Garner-Richards, P.E. and Floyd, K. 2006. Canna Seabird Recovery Project: Phase 1: eradication of brown rats 2005-06. Unpublished report to the National Trust for Scotland, Edinburgh, Scotland.

Bell, E.A.; Boyle, D.; Garner-Richards, P.E. and Floyd, K. 2007. Canna Seabird Recovery Project: Phase 2: intensive monitoring 2006. Unpublished report to the National Trust for Scotland, Edinburgh, Scotland.

Bell, E.A.; Garner-Richards, P.E. 2006. Isle of Canna: Quarantine and Contingency Procedures. Unpublished report to the National Trust for Scotland, Edinburgh, Scotland.

Bell, E.A.; Tayton, J.; Garner-Richards, P.E.; Floyd, K. and Boyle, D. 2008. Eradication of brown rats from Isle of Canna: final monitoring check and quarantine and contingency audit; March to May 2008. Unpublished report to the National Trust for Scotland, Edinburgh, Scotland.

Bell, M.D.; Bullock, I. and Humpridge, R. 2000. The eradication of rats from Ramsey Island. Wales. Unpublished report prepared for Royal Society for the Protection of Birds, UK, by Wildlife Management International Limited. New Zealand.

Berry, R.J.; Evans, I.M. and Sennitt, B.F.C. 1967. The relationship and ecology of Apodemus sylvaticus from the small isles of the Inner Hebrides, Scotland. Journal of Zoology (London) 152: 333-346.

Howald, G.; Donlan, C.J.; Galvan, J.P.; Russell, J.C.; Parkes, J.; Samaniego, A.; Wang, Y.; Veitch, D.; Genovesi, P.; Pascal, M.; Saunders, A. and Tershey, B. 2007. Invasive rodent eradication on islands. Conservation Biology 21: 1258-1268.

Jones, H. P.; Tershy, B. R.; Zavaleta, E. S.; Croll, D. A.; Keitt, B. S.; Finkelstein, M. E. and Howald, G. R. 2008. Severity of the effects of invasive rats on seabirds: A global review. Conservation Biology. 22: 16-26.

Lloyd, L. 2000. Small mammal populations on the Inner Hebridean Island of Canna, Scotland: implications for rat control. Unpublished Master of Science thesis. University of Edinburgh, Scotland.

Moors, P.J. and Atkinson, I.A.E. 1984. Predation on seabirds by introduced animals and factors affecting its severity. In: Croxall, J.P, Evans, P.G.H. and Schreiber, R.W. (eds.). Status and conservation of the world’s seabirds, pp. 667-690. ICBP Technical Publication 2.

Patterson, A. 2003. Seabird conservation on the islands of Canna and Sanday. Unpublished report to The National Trust for Scotland, 15 pp.

Patterson, A. and Brough, C. 1999. Small mammal survey on the island of Sanday. Unpublished report to The National Trust for Scotland.

Patterson, A. and Lloyd, L. 2000. Small mammal survey on the island of Canna. Unpublished report to The National Trust for Scotland, Edinburgh, UK.

Patterson, A. and Quinn, S. 2001. Rat survey on the island of Canna and Sanday. Unpublished report to The National Trust for Scotland, Edinburgh, UK.

Stoneman, J. and Zonfrillo, B. 2005. The eradication of brown rats from Handa Island, Sutherland, Scotland. Scottish Birds 25: 17-23.

Swann, R.L. 2001. Canna Seabird Studies 2000. JNCC Report 314.

Swann, R.L. 2002. Canna Seabird Studies 2001. JNCC Report 327.

Swann, R.L. 2008. Canna Seabird Studies 2007. JNCC Report 376.

Thomas, B.W. and Taylor, R.H. 2002. A history of ground-based rodent eradication techniques developed in New Zealand, 1959-1993. In: Veitch, C.R. and Clout, M.N. (eds.). Turning the Tide: the eradication of invasive species, pp. 301-310. IUCN, Gland, Switzerland and Cambridge, UK.

Towns, D. R. and Broome, K. G. 2003. From small Maria to massive Campbell: Forty years of rat eradications from New Zealand islands. New Zealand Journal of Zoology, 30: 377-398.

Zonfrillo, B. 2001. Wildlife Conservation on Ailsa Craig. The Thomas Duncan Memorial Lectures. Friends of the McKechnie Institute, 22 pp.

Zonfrillo, B. 2002. Puffins return to Ailsa Craig. Scottish Bird News 66: 1-2.