Phil Bell and AI Bramley

*,.*;1&ffi.r4g.,r ' - 1,'- i:-ur* .-c

l)cl)1r rt ll lrllt ()l'( irrscn'aliort'la I'it l)tt -.11(t tt'1.10 i

. Ib;,r',i,

't'"

newzealand .govt.nz

Cover image: Sculpture at the enEance to Rakiura National Park, Stewa.t IslandPhoto. Sue Wilkins (2oo8)

o Copyright May2o13, New Zealand Depa ment of Conseryation

Prepared by Phil Bell and Al BramleyFuture of Predator Control ProsrammeTechnical DevelopnentScience and Capability GroupWellinston

In the intercst of fo€st conseffation, we support paperless electronic publishinq-

EXECUTIVE SUMMARY

Stewart lsland represents a significantly larger biodiversity restoration project than wehave ever attempted. At over 169,0oo ha any attempt to remove rats, feral cats, andpossums would be r5 times larger than the largest successlul island eradicationconducted in New Zealand (Norway rats lrom Campbell Island).

Using our current suite oftools and best practice for island eradications, we do notconsider that it is technically feasible to eradicate rats, cats, and possums from StewartIsland at this time.

Pivotal to the success of this project is the ownership of it by rAe Srewaft Islandcomm&nicy. Through seeing and understanding what can be achieved, the communitywill become project champions and the guardians ofthe natural capital that is restored.By signilicantly enhancing ihe biodiversity around the Halfmoon Bay area thatunderstanding will occur, and the aspiraiions of the local communitywill includecapturing the environmental AND economic benefits that a predator-free Stewart Islandcan bring. To achieve this, we propose starting with a project to eliminate predators froman area encompassing the township (somewhere between loooha and soooha). The finalextent and methods used to eliminate the predators would be for the community todecide. We recommend that, to the extent it is financially possible, the communitybeallowed to make the biodiversity gains 'on their terms'.

We have also identified the critical technical innovations that would'change the game'{or this project (and most large scale predator control projects in New Zealand). If theseinnovation challenges were overcome, theywould go a long way towards making aproject of this magnitude feasible.

The critical teehnical innovations are;. large scale detection of key predators. targeted application of toxin 1th;s ;nnovorion would fLow on from the detection

innovationl. improved efficiency olbait spreading. lure development. deer repelient

In addiiion to the critical technical innovations, a number of important knowledge gapshave been identified during this analysis. These gaps need to be better understood beforrwe can determine the feasibility of this project. Theyare:

. predator interactions e.g. will cats be driven towards local extinction if rats areremoved?

. will mice establish and erupt if rats and/or cats are removed?r reline technologies forpredator proof fencing. new toxin application methods. ongoing biosecurity, especially against rats and miceo invasion behaviour

In order to drive these innovations forward a minimum investment ofgr million perannum is requircd for the next 5 years.

SETTING THE SCENE

Dr Gareth Morgan (of the Morgan Foundation) has approached the Department ofConservation proposing a partnership approach to investigating the elimination of rats,possums, and feral cats from Stewart Island. This report scopes the main issues and areasfor consideration belore such a venture can be undertaken'.

Taroet areaStewart Island/Rakiura is ihe third largest island of New Zealand. It is locatedapproximately 3okm south ofthe South Island.It is a large island surrounded by anumber of smaller islands and rock stacks (some of which are already predator {ree, suchas Codlish Island (Whenua Hou) and Big South Cape Island (Taukihepa)). The total landarea that would need to be targeted in an elimination attempt would be 169,464 ha. Thisrepresents a project that is 15 times largerthan the largest successful rodent eradication(Norway rats off Campbell lsland); and 6 times larger than the largest successful cateradication (2g,oooha Marion Island).

Approximately 9oz olthe island is public conservation land managed by the Departmentof Conservation, ofwhich Bo% is ihe Rakiura National Park. The remaining 10% ofiheisland is managed by the Rakiura Maori Land Trust (approx. 8z); or in freehold title (zz)centred on the town. The population on the island is approx. 38o people, the majority ofwhich are concentrated in the township area on the northern side of Paterson Inlet.

Conservation priodtiesWithin DOC's ecosystem prioritisation system of Public ConseNation Land, StewartIsland has twelve eeosystem management units present in the top 7OO ranked units. Thehighest ranked ecosystem management unit is the Rakeahua River wetlands (664ha),ranked at 47. The other ecosystem management units (and associated ranls) are:

. Freshwater River (!o,748ha) - zs!

. Tin Ranse (7482ha) 257

. Toitoi flat wetland (9a8ha) - 337

. Ruggedy dunes/saline (rl8ha) ' 352

. Port Pegasus (1452rha) 38o

. Mason Bay dunes (L193ha) ' 986r Smokeybeach dunes (65ha) 498. Mount Anglem (14s83ha) - sso. Rakeahua (7,o68ha) 619

. Douqhboy Bay dunes (84ha) - 67so Three Lesged Woodhen (2ha) - 693

While it is clear that Stewart lsland contarns some very important conservation sites,only the top 4oo ecosystem management units across the countryhave been funded formanagement ftom within DOC's current budget. Therefore, the restoration ol {urthersites will not happen without the help of the community. Accordingly, the Departmenthas embarked on a new operating model, whereby we seek to achieve more conservationacross the country by engaging with others in ourwork. Our engagement with Dr GarethMorgan and our serious consideration of the proposal to eliminate predators fromStewart Island is an example of this new operating model in practice.

l This report draws heavily lrom the work done in 2oo8 by Brent Beaven for the Stewa.t Island Rakiura

Taroet speciesStewart Island is fortunately free ftom mustelids, rabbits, feral pigs, and feral goats - all <

which have caused signilicant environmental damage across New Zealand.

This investigation is centred on the following target species present on Stewart Island:. Norway rat (Rotlus norvegicus). Ship rat (Rotrus rotrus). Kiore/Pacific rat (Rartus erulans). Possum (Irichosurus rulpecula)o Feral cat (Felis carus)

However, we also consider that any {uture work should include hedgehogs (Ennaceuseuropaeus), which are known to be present in small numbers near the township of Oban

Deer (red deer Cervus elopAus; and white tailed deer Odocoileus virginionus) havebeen excluded from this project; largely due to the high value placed on these animals b'the local and extended community, and the expectation that inclusion of deer will resultin significant public opposition to any proposal to undertake this predaior eliminatronproject.

It is necessaryto mention that the retention ofdeer on the island will have seriousecological consequences, namely through the inhibition of lorest regeneration. This lev€of impact will need to be monitored and the exclusion of deer from this project does notremove the Department of Conservations responsibility to control deer ifthe necessityarises in the future-

Current approach to eradicationsThe current agreed best practice for island eradications of rodents' in New Zealand is thaerial application frcm a spreader bucket beneath a GPS guided helicopter ofbrodifacoum (second generation anti coagulant toxin) cereal baits. The bait is spread intwo complete applications over an entire island (first being 8kg/ha; and second being4.skg/ha) 7'1o days apart. Projects targeting multiple species at the same time (e.g. theRangitoto Motutapu Islands multi species eradication) have used larger application rateand/or a third application to ensure that the target animals gain access to the necessaryvolume of toxin.

Given the large scale ol this project and the number oftarget species involved, iftheproject were to proceed in the short term in accordance with current best practice, a

higher total bait application rate ol z5kg/ha (ikely spread over z or 3 applications) woullikely be required. As such, a minimum ol4240 tonnes ofbait would be needed. AnimalControl Products (the company that produces brodifacoum bait in New Zealand) cancurrently produce approx. 10 tonnes of bait per day. Therefore to produce the amount ofbait being suggested, Animal Control Products would take 424 days or over 14 months. Iaddition, the bait currently has a recommended storage life of 3 to 4 months.

Success in a multi species eradication project relies heavily on possums eating the baitdirectly, and cats succumbing to secondary poisoning through scavenging and eatingpoisoned rats and possums. However it is unlikely that all individuals will be killed as a

'zhoome, KGjBrown, DrCox,A.; Croma.ty, Pj Mcclelland, ?., Goldins, C.; Giffiths, R, Bell, P.2orr CurretriAsreed Best Pm.ti.e for Rodenl l.adi.arion Ae.ial broadcastins poisonbait (Ve6ion r.2). New Zealand

result of the toxin, and follow-up work, such as traps, hunting with dogs, etc, would beneeded to eliminate those individuals that do not succumb to the poisoning regime.

One predator dog team can effectively hunt approximately so 6oha per day (K. Vincengpers. comm). Currentlythere are 4 fully certified predator dog handlers in New Zealand(with z additional handlers currently holding intedm certification), and not all ofthemare certified for all ofthe target species. Given the scale ol Stewart Island, at the rate the)hunt, ii would take t5 dog teams 19o days to cover the entire island.

Therefore, using current best practise techniques, methods, and technology, it is notIeasible to eradicate rats, possums, and feral eats from Stewart Island at this time. Thisproject is not simply a matter of'scaling up' our existing methods we consider there is tsignilicant risk offailure using our current best practise on an island ofthis magnitude.

However, we have identified several critical technical innovations and knowledge gapsthat would'change the game'for this project (and most large scale predator controi workin New Zealand). Ilthese innovation challenges were solved, we consider theywould golong way towards deiermining the success of a project ofthis scale.

Steppinq stone approachThe Department of Conse ation has a history of incremental improvements in islanderadications - taking on islands of increasing size or complexity, learning all the while todo even bigger islands. An example ol this being the combined rat and cat eradicatron orRaoul Island (2938ha) in 2oo2 led onto the multispecies eradications on Rangitoto andMotutapu Islands (combined 3842ha) in 2oo9. Given the two orders of maqnitude sizedifference involved and the requirement for significant changes in how we undertakeeradications ofthis scale (outlined throughout this report), the value in this steppingstone approach appears limited. Where a stepping stone approach is likely to be ofbenefit. is as trial sites lor some of the technieal innovations described further in thisrepon.

ISSUES IDENTIFICATION

Sociel engegemeflt end ownership

There are very few eradications that have taken place on islands with residentpopulations,let alone ofthe size of Stewart Islands community. There are approx.3Soresidenis on Stewart Island at present. There is a growing concern over the viability ofthe isiand community (with a sh nking population, reduced school pupil numbers, andexpensive diesel generated electricity identified as the key drivers ofthese concerns;D. Belworthy, pers. comm). This entne project hinges on its social acceptance by thiscommunity - every aspect has a'social dimension' underpinning it. Even if all of thetechnical feasibility challenges are meq the social aspect has the ability to prevent atryoperation ever taking place (e.g. ifthe loeal community does not want it, it will nothappen).

Genuine engagement with the community is critical, but its success will be dependent or

the community developing a sense oI ownership ofthe project. The community needs tofeel that this project delivers some ofthe aspects for their tision'ofthe island. What tha'vision' for the island is is not well understood, but improved economic prospects are on(result we would expect from this proposed major gain in ecological capital andbiodiversity (e.9. via increased tourism).

Likewise, in advancing this project, we need to understand and respect the community'sattitudes to eradication, biodivercity, the various methodologies and technologies, theirsupport and their opposition.

Starting the journq/:To develop this sense ol ownership from within the Siewart Island community ofthisproject, we consider that the immediate operational goal should be to build on, andenhance, the biodiversity gains around the Halfmoon Bay area. There are already anumber of small scale community-driven projects happening in that area:

. Stewart Island Rakiura Community and Environment Trust (SIRCET) predarorcontrol and restoration work around Halfmoon Bay

. Daneing Stars Foundation predator prool fenced sanctuaryr Partnership between DOC and Air New Zealand to increase biodiversity around

the Rakiura Tracko Ulva Island open sanctuary (with the UIva Island Trust)

An appetite for conservation and biodiversity restoration already exists in some parts ofthe community - after all the community 'demanded' that the invading mt population orUIva Island was re eradicated in 2ou.

We propose building on ihat work bytargeting an area encompassing the to\anship(somewhere between roooha and Soooha) for elimination of all the target species(including hedgehogs). The exact size ofthe area and methods used to eliminate thesepredators would be for the communityto decide.In essence, we recommend thecommunity be allowed to make the biodiversity gains'on their terms'.

This smaller-scale project would enable us to enhance our relationship with thecommunity and understand theii motivations, attitudes to eradication and the techniqueinvolved, and their willingness to engage with or own a Iarge scale project. At the sametime, the project will 'bring more biodiversity to the community's backdoor' buildingenthusiasm and understanding ofthe connection betvreen eliminating predators and

Target Species

Hedsehoss:We consider that any future work should include hedgehogs, which are knoe,n predatorsof ground nesting birds, as well as having impacts on invertebraie and reptilepopulations. The current population ofhedgehogs is believed to be small and Ioealisednear the towaship oI Oban (B. Beaven, pers. comm) however this needs to beconfirmed.

Hedgehogs offer the possibility of an 'early win' for eradication, iftheir limiteddistribution is confirmed. However, a small eradication targeting the hedgehogs alonemay have Iimited visibility, as there appears to be a laek ofcommunity recognition olthepresence or damage ofthe hedgehogs.

Mice:There is no current evidence to suggest that mice are present on the island. The presenc€

of cats and rais may have prevented mice from establishing; conversely, the mice may besuppressed to below currently detectable levels on the island.

Miee have been eaught on occasion at the main wharfin Oban @. Beaven, perc. comm),so they are a known invasion risk. No mice have ever been detected within the DancingStars Foundation management area in the six years it has been managed to zero densityof all predators @ehind a predator proof fence), suggesting that they have not beeomeestablished on the island (or at Ieast not in that area ofthe Island).

The presence of mice on the salmon farm in Big Glory Bay needs cladfication. It isunderstood that mice have arrived there on occasion with salmon feed (A. Roberts, pers.

comm).lfthey are prcsent on the farm, discussions should be held with the salmon farmcompany with the goal o{eradicating them and reducing the risk of reinvasion. This iscdtical, given the farm presents a high risk ofintrodueing mice onto Stewart Island.

Ifthe elimination of rats and cats proceedg it could greatly increase the likelihood ofmice becoming established on the island (due to the lack of natwal enemies andcompetitors). Br.rilding understanding about the likelihood of mouse establishment anderuption, and the ecological consequences of those events, will be critical to many futureaspects ofthis project - choice of target species, biosecurity requirements, etc.

Rots:Generally, where rat species co-exist, habitat partitioning tends to occur. Harper et al3

{ound on Stewart island that ship rats dominated in podocarp broadleaf forest andriparian shrub land; Nor_way rats were most common in subalpine shrub land; and kioredominated in manuka shrub land. However, ship rats (and Norway rats in smallernumbers) were found in all vegetation/habitat types- It will be important to understandthe distdbution and habitat usage by each rat species to determine lvhether there is anycompetitive interplay for resources (and therefore how that may influence bait or deviceaccess lor each species).

3 Harpe., G.A., Dickinsoa K.J.M., and Seddon, P.J. (2oo 5) Hdbitdt use by thtee rut spec;es @anus spp) on

In addition, understanding the relationship between the rats and the feral cats will beimportant. It has been suggested4 that removing all of the rats could place significantpressure on the feral cats and potentially drive them to extinction by starvation. Thistheory needs to be robusdy tested.

Cots:Discussions with local DOC staff on Stewart Island has suggested that the disrribution ocats is limited, possibly to habitat that provides best ptotection from wet weather. Thisneeds invesiigatio& however it may open up the possibility oftreating sections oftheisland with cat-speci{ic techniques (rather than the whole island).

There is a strong possibility that a number ol cats would be kiiled via secondarypoisoning in any bait operation (involving 1o8o or brodifacoum) targeting rats and/orpossums. In a small expe ment on Stewart Island g out of to radio collared cats werekilled during a roSo operation targeting possums (the cats were found to be scavengingdead possums). As mentioned above, the relationship between cats and rats needs to bestudied to identi& ilthere are dependences to be exploited in an elimination operation.

However, secondary poisoning can not be relied on to account for all ofthe {eral cats onthe island - cats survived the aedal eradication of rats on Raoul Island; andRangitoto/Motutapu Islands. As such, we need to determine howto conduct the follow uwork to detect and kill the remaining animals. The eurrent use of dogs and hunters isIogistically and technically infeasible on an island the size and scaie of Steurart Island.We need new tools of sufficient high sensitivityto detect low numbers of animals acrosslarge landscapes.

Some Stewart lslanders own cats as pets. There is the potential for some pets to be killedas a result of an elimination operation. This maydamage any public support for theproject. Mitigation measures need to be investigated (and tested with the community{oracceptance) measures could include keeping the animal indoors during the operation(or at least the duration ofthe operation around the town)i or holding the cats offtheisiand for the full duration of the elimination project (to be returned, desexed, oncesuccess has been determined).

The Southland Regional Pest Management Strategy (RPMS) requires all pets to beneutered and micro'chipped on Stewart Island. Ifenforced, this will greatly assist incontrolling cat numbers on the island during and afier anyelimination. There may bescope to seek additional changes to the RPMS to further rest ct the ownership of cats(and other animals) to maintain the biodiversity gains of any elimination project.

Discussions with Iocal DOC stafl on Stewart Island have suggested thar the distributionofpossums is limited, possibly driven by availability of preferred habitat. This theoryneeds investigation, however it may open up the possibility of treating sections of theisland for possums (rather than the whole island).

Based on experiences from other eradication operations (e.g. Kapiti Island, CodfishIsland), it seems unlikelythat poisoning can be relied on to aceount for all possums onthe island. As such, we need to determine how to conduct the follow up work to detectand kill the remaining animals. The current use ofdogs and hunters is logistically and

4 Harper, c.A. (2oos) Nrne ticdl and functianal rcsponse of ferol cars (felis catus) t o ranations in abundonce

technically infeasible on an island the size and scale of Stewart Island. We need new toolwith sufficiently high enough sensitivity to detect low numbers of animals across largeiandscapes.

While a number ofmainland sites have been treaied for possum and rat control withaerially applied ro8o with good results, these have been contrcl operations by nature,they are not seeking to remove every individual. Understanding the issues with inter-species competition in an eradication context (e.g. does one speeies have a competitiveadvantage that may result in the creation of gaps in the bait coverage such that someindividuals do not ever encounter any bait) wili enable robust pianning ol themethodology to reduce this risk.

Target mal<e-up and sequencing:There is a choice to be made as to what suite ol predators is targeted in what order. Ther<appears to be merit in investigating the costs and benefits of a range of targetingapproaches (e.g. single species at a time, multi species, etc). The choice made here willhave direct flow-on effects to the other areas of the project - namely, biosecurity (e.g.

possums are easier to keep off an island than rodents).

Critica] to the decision oftarget make up will be {ull understanding otthe ecologicalconsequences oftaklng one (or more) predator out of the system, and leaving othercpreseni. Will iaking cats out create a mesopredator release situation whereby the ratpopulations explode? If cats and rats are removed, will mice establish and erupt? Arethere ecological beneliis (or costs) in removing possums but leaving the rats and catspresent for a time? It willbe critical to understand the population dynamics andinterdependencies ofthe target predators (and their prey), in order to design the optimalprogramme to enhance biodiversity gains on the island.

One potential scenario raised in a number o{ discussions dunng the investigation stageof this project has been the removal of possums from the entire island in the immediatefuture. Given the constraints outiined above (e.g. not able to secure complete eradiationusing toxin alone, the number ol dog teams required to hunt ihe island, etc), removrngpossums from Stewart Island does not appear feasibie at this time. Advancements inpredator (possum) detection tools and understanding of reinvasion behaviour (r,rhichcould allow us to 'defend'areas) may enable a possum eradication to occur in the future,potentially targeting zones one at a time rather than the entire island.

Methodr

Aerictl application:Our current best practise of using helicopters to spread the bait would not work for anisland ofthis size. Currently, helicopters can spread around Bookg of bait per llying hour:Therefore, spreading 4240 tonnes of bait over Stewart Is]and would take one helicopter53oo hours or 663 days flying I hours a day (or alternatively, 67 helicopters flylng for B

hours a day lor ro days)l As sueh, we need to Iind improved efficiencies in delivering baitacross large landscapes that significanily speeds up and increases coverage ofthe targetarea (e.g. utilising fixed wing aircraft that can carry and distribute larger loads).

Presently, in eradications, we apply a considerable amount ofbait to the island to ensurethat every individual has access to the quantity needed to ki]1 them. However, we do not

s lsland Eradication Advisory Group neetins notes (15 16 December2olo): Stewan lsland R€view key srase l(do.dn 6876l)1).

1(

know if we are overloading the area with more toxin than is sufficient (to killallindividuals). In some operations, this is likely to be the case given some bait is stillevident on the ground some weeks after the aerial application has taken place. IIwecould deiect in real time (or close to it) ho1.r many individuals are still present after eachapplication, it could allow for different treatment regimes over time, potentially resultingin less toxin and labour being used.If we are able to reduce the amount of toxin used toonlythat which is necessary, it could result in significantly lower application times andrequirements, saving time and resources.

G r oun d - b a s e d c omp on e nt s :

Our current tools for detection (e.g. tracking cards/tunnels) and trapping (e.g. DOCseries traps in boxes) are too labour intensive for a project of this scale. The track networ(and subsequent maintenance) alone would be beyond our capacity (as well as beingundesirable across those ecosystems) not to mention the sheer number o{ devicesneeding regular checking, baiting, resetting, and maintaining. The need for new andimproved detection tools (with high sensitivity to detect low numbers of animals overlarge landscapes) is abundantly apparent.Ifwe knew where the animal(s) is, we would b<

in a posiiion to target our follow-up or response actions with greater precision.

Our current Iures Ior traps (e.g. egg, rabbit meag etc) do not have the optimum'pull'todraw in the target animals over a large area, nor to dlaw them consistently into the devicthey encounter. As such, we are required to have our devices rclativelyclose togetheracross the landscape which resuits in a high number of devices and a maintenanceregime that becomes prohibitive in a project this size. IIwe could develop a'super lure'(or a series ollures for each target species) it would enable us to set our device networkswider and reduce the number of devices within them pulling the animal to the device,rather than putting the device where the animal is.

One area that does not appear to have been considered in ary great detail before now isthe idea of deterrents or repellents. Couldwe develop some form of deterrent that repelsanimals away from eertain areas? This technology could then be used much Iikepredator-proof fencing, and could allow us to establish zones or bu{fers that the targetanimals would be very reluctant to enter (or cross) because of the power olthe detefient.

The township o{ Oban will Iikely need to be treated using ground based techniques (e.9.

toxin in bait stations, traps, etc). How this is integrated into the methodology and timingofthe rest ofthe operation will need serious consideration, to avoid issues with targetanimals moving between untreated and treated areas. The idea of a predator'proof fenceclose to the township has been raised for this project. The merits of this approach willneed to be investigated, taking account ofthe community's views and perceptions (e.g.

being excluded from the rest ofthe island, etc). Predator prool fencing is discussed agairlater in the document.

Zoning:The potentially limited distributions of possums and cats add weight to the possibility oldividing the island into zones to be targeted separatel, rather than attempting theelimination over the entire island. Some eradications on large islands (e.g. MacquarieIsland) have applied bait over multiple days, and have retreated small buffer zones (overpreviously ireated areas) to redllce the risk oI rats moving from untreated into previousbtreated zones and avoiding encountering bait. The curreni South Georgia Island Norwayrat eradication has been operating under a strategy oftreating the island as a sedes ofislands (or zones) due to the many large glaeiers that divide up the island (and are large

enough that mts can not cross them). However, no continuous island has been broken utr

into zones for treatment at different times,

We would need to better understand the invasion behaviour of the target animals, as we1

as our ability to eliminate predators from areas facing constant invasion pressures, andour ability to defend those areas from reinvasion. Critical to this will be our ability toestablish and maintain some kind of buffer area between the zones, be it with a predatorproof fence and/or a network ofdevices etc.

Tools

Brodifacoum is the most commonly used toxin for rodent eradications on islands. ro8o(sodium flouroacetate) has been used for a number ofyears to target possums and rats itcontrol operations in New Zealand. These are the onlyviable toxins that are currentlyregistered in New Zealand for aerial application under speeific conditions.

Currently, brodifacoum can not be broadcast aerially on Stewart Island due to the 1abelrestrictions (for aeriallyapplied use oniy on unstocked off shore islands or mainlandareas behind predator-proof {ences). Stewart Island would be considered part oi themainland lor the purpose of an elimination campaign ofthis scale (A. Fairweather, pers.

comm). Therefore, it would require a change to its registmtion - there is a precedent herwith the registration changed to allow brodifacoum to be aerially applied at mainlandsites behind predator'proof fences.

The recent development of new humane toxins has largelyproven unsuccessfu], withonly paraminopropiophenone @APP) showing promise thus far. PAPP has recently beerregistered in New Zealand for use in fresh meat bait in bait stations to target cats andstoats. The Australian authorities are trialling some methods for encapsulating PAPP

fand separately, ro8o] into a bait {or aerial application6 - this work is in its early stages. 11

is possible that aerial application of PAPP targeting cats would be ofbenefit for thisproject, therefore it will requirc additional trialling and registration here in New Zealand.

There is a considerable amount of research going on into new vertebrate toxic agents(VTAs) lor possums at present. These include zinc phosphide (nearly registered); sodiunnitrite; and C*C (where two toxins, cholecalciferol and coumatetralyl are combined). Foranyolthese VTAs to be registered in New Zealand by the Agricultural Compounds andVeterinary Medicines Group, a data efficacy package needs to be submitted that showsthe VTA is humane. PAPP and sodium nitrite could be considered the most humane ofthese + VTAs, but all ofthem will be suffieiently humane to pass registration.

There are field tdals underway to Iook at the eflieacy of aerially distributedcholecalciferol; zinc phosphide; and sodium nitrite. However, it will likelybe 5* yearsbefore any ofthese would be available for aerial distribution @ecause aerial applicationofthese VTAs will need to be approved bythe Environmental Protection Authority) (AFairweather, pers. comm.).

6Johnston,M., Gislionl f., O'Donoghue M., Holdswodh, M., RobiGor! S., HerodA. and Eklom, K (2op)field assessnenf of rne Cu/iositya bdt fat tudnogetuent af fetdl cdts in *e seni-atid zane €linders RdnsesNotioroiPorl). Arthur RyLah Ilstitute for lnvilonmental Research Technical Report Series No.234.

Traps and devices:The'Good Nature'series of selfresetting traps (for possums and rats, with a cat trapbeing t alled by US authorities in Hawaii) ofier a way to reduce the trapping burden(mentioned above in the ground based operations section). This technology is improvinall the time, with the number of resets per traps increasing from a single reset (atprototype) to 24 resets lorthe stoat/rat tlap; and 12 resets Ior the possum trap. TheDepartment is currently undertaking trials ofboth the rat/stoat; and the possum resettirtraps to investigate their reliabilitn effectiveness and their efficacy compared with theDOC series oftraps. As this technology develops and our understanding grows ofhowbest to use it, these resetting tBps may offer effective ways ofcarrying out the ground-based components of an eiimination strategy.

Coupled with the development of PAPP, outlined above, has been the development of anew toxin delivery device known as the'Spit[ire', being deveioped by Connovation Ltd.,in association with Lincoln University. This device squirts toxin onto the animaj's bodyas they move through the box. The animal ingests the toxin as they self-groom and thendies. The mechanism ofthe Spitfirc is the same for all target species, but the'housing'will be tailored for each species for example, the rat and stoat spitfire mechanism can Iinto a standard DOCzoo trap box. It is hoped the Spitlire device will have the ability todeliver approx. too doses oftoxin before it requires refilling. Successful pen trials havebeen conducted at Lincoln Universityusing PAPP for stoats & cats;1o8o for ship &Norway rats; and zinc phosphide for possums @. Murphy, pers. comm). Field trials arebeing pianned for all species.

Use of predator-proof fencesThe use oIa predator proo{ fence has been suggested by many people when theycontemplate this project and its scale. One area that could benefit fiom having such afence would be behind the township area. Predator proof fencing technoloov issufficiently advanced to enable an order ofmagnitude shift in the biodivercity gainsbeing aehieved around the town. It could potentially serve as a clear demarcation ofwhere aedal application methods end and ground based operations siart (depending onhow large ihe area to be treated is); and it could prove to be a key biosecu ty tool instopping invading animals from reaching the rest ofthe island once clear olpredators- C

course, before any fence can be constructed, the Stewart Island community need to agre(to it.

However, if fencing technology is going to be used in this project, we need to understan(it better. While a number of conservation projects in New Zealand are currently usingpredator proof fencing (e.g. Zealandia, Maungatautiri Ecological Island), we considerthere are some critical knowledge gaps which need to be answered, such as how animalsinteract with these fences, how different designs (height, hoods, mesh shape and size, etcaffect pe ormance, and how to manage 'leaks' or breaches of the fence, particularlyaround the ends (e.g. underctanding the use o{'wings'at the fence ends protruding dowreliffs or into the ocean).

Deel lepellents:With deer being excluded lrom this projec! it is important that we use cost-effectivemethods that limit the impact on the deer populations and ihe associated huntingactivities. Key to this will likely be the need to further develop effective deer repellents(both operationally and cost-wise) that prevents deer from ingesting toxin or interactingwith devices.

!L13, |:',: O.* .*9 a deer repellent rhat is cur.enr'y used :n aer;at roso op"rarrons,but.rl .s labour inror s;vp ro apply to the ba r and rnerelore is verv e\pensivF. Current.yad-dingdeer repellent to aerial logo possum operations using very low bait applicationrares aods approx. g€j.per ha ro the project costT as a crude indication, that would add amrnlm rm ol $2.02 million to the cost ol rhe bat (ar rhe volumes L,s"d aoove). \or tomention wherher it is even possible to manulacture rhe quantity of rep"fl""i fil"iy a U"required for an operation ofthis scale.

Deer repellent has never been used in New Zealand in any toxin other than lo8o (AFairweather, pe,". comm). Research with deer repelle,rt in roao hasl"* "fr.*"," *a.*red deer by-kill; however lurther research is n""d'"a to a.t"r.in" *l.r"i;;il;;;.equally (or more) effective in baits containing brodifacoum. Likewise,."*"J f"

"""a"atotesr the effectiveness of deer repellent on.whit" t"l"a a""r 6tr" _or" ."J.""ii.""ffyvaiued of the deer species on Stewarr Istand). L

"dditi.r, _;;; ;;i.."i#,r,"effect deer repellents have on the atrractivene"", pd"t^bilitn.nd .iii;;;;";;"" .,devices for the target animals (and other non-target species e.g. native birds allresearch/trials to date indicate that the attractiveness of baitsio nati ru norl

".o",speciFb is not in. reased by rne rs. of rne r"pel,ert).

Baitmatix:Stewart Island's climate offers a challenge when it comes to the right bait matrix. lncurrent eradication operations, brodifacoum is used in a cereal balea U"fi n,t"-,.*Howe""r. t\is narriy do.s ror I and le wot condirions well n"".* i"^ ._,,.r..,. . -,rpotentially losing palatabitity. Stewart rstand ""*"s* ,".,;;i;;;J" r"",'ia"r. _,,f.t m n o1 rarn or nor-) ac ording to N, WA.s N"t,or"t C, i."," o","i". j.;;. ii"'b:;","r,".(a Us company) produced brodifacor:m bait in a more_weather resistant matrix for theRat Island Norway rat eradication (in the Aleuti", r"t""d") ;;;;;. iii."i" #i" ,* *"aon Stewan Islano. test;ng wi ,b^ required to en"ure tnar rh" marrii s"l;;;;; -wrthstand condirjons on ihe island without adversely alfecting the attrectiveness orpalatabilityto the target species (and non target species).

The bait matrix used to house the toxin could be coupled with the potential advances inlure deve'opm.nr (outlin.d aoove) o claate a stronger a.tra.,ion [o. the tarqet animals. lt.outd a,low us to use less bait by ,pulling.ar imals to rhe oa,t, ra-her tr an pla'. irq rhe bairunare lh. target ani-nal is. Ol .ourse. any changas in ihF ba;j matrrx make up wiltn^eq robe thoror.rghly res-eo lo. alracrivaress and pal"."Uiti,V i, Uoti, ,1" .";i";;;;; ,"rn",

species.

Detection/MonitoringAs ou lrned abov". large scale oerection rools ar" one of m" key missinq oiec€s oItechnology lor this project. These detection.to.f. ,"rfa

""t ""ifi"iir"",i*" ,r,*"a.",individuals,tut ideally they would also be ab1" to taf "" -i.,"r" iiro".l.i*"i" ,* a U"n,.wir h. As sucn. it could change how we approac\ th. wnole op"rarion -;;hr;;; -.sFrecr.vely rarga. areas with particular tools and/or oevices at ." ecred densiLils

dependr g on ihe s:7e ol rhe poprlarions detected. These roo s cou.d ooen uD rhepossibility to 'spot rreat'areas or divide the rstu"a int" ,o.r"., ,"ii,"rlhJ,l ,",I.i,"ff,unnec"ssarily needing to cov"r th. ent:r- is,and with roxin or devi.es

T IEAG neetins nores (rs 16 Decenbe.2olo): Srewart Island Revi€w keystase 1(docdm 687601).

14

Non-target species

Deer:Although deer are out ofthe scope ofthis project, they do require serious considerationespecially iftoxic bait is used. Deer have the potential to consume large quantities of thebait (unless an effective repellent is used) and create gaps in the coverage, while alsokilling some ofthe deer. These gaps could result in not all target species having access trthe bait and therefore surviving that aspect of the operation.

Any eonsumption ofbait (ifbrodifacoum is used) may result in sub lethal levels of thetoxin being present in deer. Brodilacoum residues have been found in some deetsamples, in the Iivers and (to a lesser extent) in muscle tissue, following somebrodifacoum operations3. Due to the toxin s persistence in the food chain, a ban onconsumption ofdeer meat from Stewart Island will need to be investigated. The currentwithholding or caution period used by DOC is 36 months (for a bait station operation).With the presence of deei on Stevvart Island it is likely a withholding period ofthisIength would be needed (A. Fairweather, pers. comm.).

Obviously, anywithholding periods will have an impact on deer hunting on ihe island.We considerthat the deer hunters should be involved in investigating mitigationmeasllres (e.g. trials ofdeer repellent involving white tailed deer, closing zones of tireisland to hunting during treatment periods, etc) to reduce the impact on deer andhunting activities. This involvement could go some way to achieving greater buy-in frorrthat sectlon ofthe community, and reduce the perception of any research being capturedor biased by DOC.

Notive species:In all eradication operations, there is some Ievel of non-target by kill ol native species(namely birds via primary and secondary poisoning). However it is impossible toaccumteiy predict the levelof non-target losses.Itwill be important to understand whichnative species on Stewart Island are at most risk (via bait acceptance tdals etc), so thatappropriate mitigation measures (e.g. eaptive holding ortranslocation off the island forreturn post-operation) can be investigated and tested.

Seocl,There is a small number of sheep on Rakiura Maori Land Trust land near The Neek.These animals will need to be prevented from eating the bait (if used) this may best bedone by removing the animals ftom the island to be retumed after the operation iscompleted.

Ngai Tahu

Stewart lsland/Rakiura lies \,r'ithin the t bal area of Ngai Tahu. It is absolutely criticalthat Ngai Tahu is represented on any governance structure that is set up forthis project.Withoui their support for a project olthis magnitude, it is extremely unlikelythis projeetwill succeed.

The recent eradication work undertaken on the Maori owned Titi (muttonbird) Islands,off the south-west coast ol Stewart Island, has been hugely suecessful in demonstratingthe biodiveBity gains that can be made here. Not onlyhas this work illustrated the

0 Broome. KG.; Faitueathe. A.A.C-: Fisher. P. 2ot2: Brodifdcoutu Pesticide lnfomation ,Revi€r. Ve6ior

damage being done by invasive rats, but also our ability to halt that damage throughcurrent eradication techniques. A number ofland owners have changed their views ontoxins, eradications, etc, moving ftom 'anti'to 'pro' due to seeing the operation and theresults firsthand @. MeClelland, pers. eomm). These Iand owners could be strcng alliesand advocates in progressing the Siewart Island project.

Kiore is a valued taonga species by some Maori. Initial conversations with local iwimembers suggest that the removal of kiore from Stewait Islandwill not create significanconcern, so long as a population of kiore is maintained within the tribal area o{ Southlanr(S. Bull and G. Thompson, pe$. comm). Further discussions will be required with localiwi to determine an appropriate management approach for kiore to address the concernsthat a population of kiore is maintained.

Rakiura Maori Lend Trust

Rakiura Maori Land Trust manages approx. Sz of Stewart lsland, on behalf ofMaoriIandowners. This land is predominately on the east/south-east coast ofthe Island.It isabsolutely criticalthat they are pafi olany community engagement (and any govemanctstructure that is set up for this project). Without their support and the ability to accesstheir land, it is extremely unlikely this project will have any chance of success.

Biosecurity

There are a number ofpathways for potential reintroduction of predators onto StewartIsland Ierry, plane, fishing vessels, Ireight, etc. We need to understand these path\aayrisks and the subsequent invasion and establishment risks. This knowledge will enable uto develop appropriate interventions on those pathways that could greatly reduce the rislo{ an incursion. As noted above, the choice oftarget species has implications for thebiosecurity risks associated (as diflerent measures will be needed to keep possums oll thisland, than would be required to keep rats out).

Dete ction prcbobi litie s:There have been recent advances in our understanding of detection probabilities withregards to island biosecurity, particularlythe work on Barrow Island in Australia.lt isimportant that we take those learnings and develop robust New Zealand-cent cmethodology for understanding the probabilities of arrival, and (arguably moreimportandy) of detection of an invading individual. This knowledge will enable us todevelop appropriate methods ofintervention at Iikely depature and arrival/detectionpoints that could greatly reduce the risk of an incursion developing into an establishingpopulation. There are clear links between this knowledge and the innovaiion required onlarge scale detection iools mentioned above.

Invasion biology:Related to the wider suite of biosecurity issues, we do not know how invading individualsbehave at large scale sites. Experiments at smaller sites indicate that individuais movethroughout the area (potentially searching lor a mate), but we do not know ifthis type ofbehaviour holds true over large scale areas nor howthe behaviour changes as theestablishing populations grows and invades futher. This knowledge, coupled with thedevelopment of large scale detection tools (outlined above), will Iikely result in a bettermethods Ior responding to biosecurity incursion events (e.g- spot treatment of areas).

lmpacts on way of life:It has been suggested that the biosecudty standards required to maintain predator{reestatus on Stewart Island would need to be akin to that when entering New Zealand. It ishighly unlikely that the residents and visitors to the island would tolerate such anintensive process. Ifthe biosecurity requirements are considered too onerous, people areless likelyto conform to them - biosecurity needs to be seen as a'normal'part of life.Coupledwith the developments in detection theory and detection tools, we need todevelop a biosecurity system that is effective, but does not undulyimpact on individualsand their lifestyles on the island.

Maintenance af the system:Biosecurity is forever (or for as long as the island is deliberately maintained {ree fromthat species). But biosecuity measures cost - both in the prevention and the responseside. Any system that is developed will need to take into account the ongoingmaintenance involved in having prevention and response components. Likelvise, systemwill need to be established around who is responsible for deteeting and responding toincursion events.

WHERE TO FROM HERE?

As noted above, using current best practise techniques, methods, and technology, it is ncfeasible to eradicate rats, possums, and feral cats from Stewart Island at this time.However there are a number of challenges that if met would go a long way todetermining whetherwe could be successful with a project ofthis scale.

The Halfmoon Bay proiectAs outlined above, we recommend that the immediate operational goal for this projectshould be to significantly enhance the biodiversity gains around Halfmoon Bay. Ourproposal to achieve this is to target an area encompassing the township (somewherebetween roooha and soooha) for elimination of all the target species (ineludinghedgehogs). We consider that this seale ol project is feasible using our current suite oftools and techniques (so long as the community wants the projeet to happen).

In order to faeilitate community engagement and subsequent ownership ofthis project,we believe that the exact size ofthe area and methods used to eliminate these predatorsshould be for the community to decide. In essence, we recommend the communitybeallowed to make the biodiversity gains'on their terms'.

This smaller'scaie project would enable us to ensure a close relationship is in place withthe community; and their motivations, attitudes to eradication and the techniquesinvolved, and their willingness to engage with or own a project of scale are completelyLlnderstood. At the same time, the project will'bring more biodiversityto thecommunity's backdool building enthusiasm and understanding of the connectionbetween eliminating predators and restodng native biodivercity

A small scale project around the town will provide the testing ground to determine if wecan eliminate predators in the presence of a resident human population. Furthetmore, itwould help identifywhether we are able to delend those biodiversity gains, in thepresence of constant invasion pressure. This invasion pressure would allowvaluablelearning about reinvasion rates onto the island, identilying common pathways andorigins for invading species.

We consider that the initial work towards this aspect of the project - namely, thecommunity engagement and building that relationship could begin immediately. Thetiming ofthe subsequent components, depending on what those are (e.g. construction ofa predator proof fence, elimination ofthe target species, funding models, planningpermissions, etc), will be determined in conjunction with the community.

Critical technical innovationsWe have identilied several critical technical innovations that would'change the game' fothis project (and most large scale predator control work in New Zealand):

Latge scale detechon toors - we do not have any proven method that is technically andlogistically feasible to detect very low numbers ofpredators across large landscapes athigh sensitivity. Our current tools (tracking cards, etc) are not appropriate {or this scaleof project. This is our number one technical innovation challenge.To;geted, applic.ttion offoxin [this innovation would flow on from the detectioninnovationl if we are able to solve the detection problem, it could significantly changehow we use toxin. Rather than covering all areas at all times with toxin and bait as we donow (to ensure that every animal can access it), we could potentially selectively target

lTff::3:l:,:""r:f6ait spreadiae - our curent best pracrise of using heticopterswlrn ouckets to spread the bair becomes cost and capacity prohibitive at the scaie of thlsproject This is compounded bv our ,se of bait in q.,.ntities ttrat urrsr."" ,rr-irar"ia""r".r^11_::"..":,: l], 1,, Tnowing

wherher ihis is e,".ist,,e o. .,ot. We ,,";; ,;;";i",i:::: :l:: "1 -.:i.9"Ios,es hat r equirF ess bair. andlor . ild ways ro dFrrver baj r acrossrarge la'l0snapes lhdt e'gnilican lv so..ds :p "overage oI tho rarg"r ar"a te.g. lixed wingaircraft).Lure development - o rr current .ure. (for rraps ano bait marrix) do not have opti,n rm'pull -odraw,arge'arimalsinf.omlarg.ar"as.Thisholdsrruelora'fof,fr""io.,species of this project. We need to deve)op 1u*" ,f.,r, riff g"""*;#;"*; ;::;""landscape scale areas, thereby reduce the neea t" prt a""i*" .. i.rJ*'fr"." t" ."r_a ,"but rather draw the animat(s) to where the ded"";.;"';.i;;iliii"., _" rl""r*" ,r*"rnargreatty-ncrea"Frheirteract:or rar"olanimalsor cprhey.ncounterorrdjces rnotherwords, we need to get more animals goi"g i"a

"". "";".'-_"Deerrepellent - in order.for this project to gain the public support to proceed, we need touevarop mernoos rha. reduce he irnpact on deer. Key o tn:s wjll be he o"vclopmenr otelecrrve oeer repoIenl thdt prevFnrs deer lrom lngesting roxin or trreractit q withde\,l.es. Associated with .har is rhe need ro rnd"rsrand rhe eff^.t o.". ,"p"tti..,t.'l.or. onrhe use of roxins o' dereces ror the targer anir""'. 1"ra .if.,", "", .r;":'"";;J ";native birdsr

Xnon l"ao" o.o"In addirion ro the critical technical inno.have been uncovered d.,.ing thi"

"n"ry"iitio"s'

a number ofimportant knowledge gaps

Pledator i,tero,ctions.. we need to understand how these target species behave in thepresence of each other on Stewart Island, especiaily in relatio"n to i_rabfi" ""^* *acompe. iiion for resour.es. Tnero are a numbe, "f k"y q .",tion" tn"t;";;; ;:;".,"g

""they will-sreatly inform the methods, rechniqu*, *;";;"1;;;f '";"1,.,.","j]"a,r-,"timing of any operation.Mice: we need to know if mice are present on the island. This question has never beencaiegorrcany answered. We need to und-orstand the likelihood of mouse establishment,a-nd a,conseguential eruption. rf all (or some) oi,f," ,".g",

"f"";";." ,"_.".jir._ ,t "rsrand: and rhe ecological consequences of such an event. This knowredqe wir enableoenrstons to be made or critical matrers s rc\ as biosecur irv roquir"menis orc.

Predator-proof fencing relinerlena if fencing t""lnotogy i" g;i;;i.l" *"a ,, ,fr,"projecr (poranria.ly .o crea,p biodivers: y gains around H;lfm;on B-"r, ,r" ,""i i"uno,arsrand rr bFtfer. ( .ili.a,issues .o b. rocussed or include \ow animals int.ract wrththese.fences, how different designs affect perf.r-"".",;;;;;;;;;;;;" ilil::.breaches of the fence.New toxin crpplicatioa metiods.r we need to expand the development of toxinapplrcar,ior methods. b r ldingon rne registrarion ol pApp ano ri e in"*,., .t ,f."-Jprxre

rnro avenues and methods applicable to large scale projects (e.g. aerially appliedPAPP for cats er. \Erosecurity: we n'eed to undersLand the pathway and incursion risks for the isrand. weneeo to better understand rhe Drobahiliiics of arrival and of detection, of an invadingindiv.rdu€l This l.nowledge wiiJ enabJe us to d",r"top

"pprop.iri" *";;;;;i;;;;",,."

that could grearlv red r"e rh. risk ol an incu-s:or oevelopinq :nro an ".r"briJi_-population. There are ctear tinks between this k,"*t"dg"

""J;" ;,;;;;"" i"r'..,."a ."large scale detection tools.Invd.sio bioloy: rclated to the wider suite of biosecurity issues, we do not know howinvading individuals behave at large scale sites. This knowledge, corri"d.;ilil^

19

development of large scale detection tools, will likely result in a better methodology forresponding to biosecudty incursion events.

It is our opinion that to drive this innovation forward and addrcss the importantknowledge gaps will require a minimum investment ofgl million per annum Ior the nexl5 Years.

Governance and Technical GuidanceHowthis project is governed depends on how it proceeds into the future. There are twoclear streams ofwork outlined in this report:

r) OperationaL the Halfmoon Bay project planning, developing, and executing thlelimination ofpredato$ to achieve community driven goals

z) Technical solving the innovation challengesj while being informed and learnin{from the operational stream

Ifthe Halfmoon Bay project goes ahead, it would appear logieal to have a Stewart Islandcentric governance group guiding the operational stream. As a community drivenproject, the govemance group also needs to come largely from the community. Therefolmembership should span the community- capturing the cultural, economic, societal, an,enyironmental views ofthe Stewart Island residents (and community at large).

The technical innovation challenges in this report are not Stewart Island-specific, butrather challenges that we will face at most large scale projects throughout ihe country. Asuch, the technical group guiding this work could have a much wider focus. Itsmembership could include experts from the Crown Research Institures (CRI,s),univercities, central govemment, non Government environmental organisationq and th€technological community.

FEUlrlzUXoztrl

I

o-o

oo-

(o

5o-r3oo

3

(a

ord

€.

E RE ! g E T;1cl:rq

= o- iJ I ^! \o.: 5 +,

E 6 i i E_d< -a : < o< - +.rc!o:'1X D d Id a d g*

E 3! E 6

'. eE. {qo o- le#BI

,

p-F

5=

E&EI.r6E&

N

E ! o-- G- ! a'r o .l ! -;? 6;9;3'l .Pd 60 Et A.:H 3 i; g 8- f-J .; i:. *; S;ia-:G E 0:. 1d€ t': 3.3 0€!o o ;': ) '3 ) ';5 I o o : oj q.o ;o 5 Y 6 o- ,b 6 6 s- # 6 o di

a ts =;; 8 x +; E a ? 6I .1ii;A.x sH;Elae'*='q3itE.{P:.r} E+3E qE HE-iglda i?f i:;18:;r+3 ii+i.-i31+rit;EeE=B';_=&+ d:!Et; i+L?E6 ii,: e l€ ; ;s;-0n::5; B:e ,i, q

: lEgY-"-; g=5 d Fd&;iJsB 3ii': I= ;i!o-qg; q 3 ; I;4eE_3e= & i E fi:E.e :i € I E 3raSj a* E ! * elsid !; 3 + i "r i{: ;3 I ; 6 b:t, 1Ze-l6E :d : a i Aq 3?.E :3- : 3 e

=5 93r ; : a e iit.5i ; Y ; s !

+a c +o a ro5: x 5; ;,4f ; E- 3'c B 15C. aO rg

. a5- i,P 5,;id I g P d t

;i; d i c ,3^-9o

dii @

--o H t'! aH,otGl ,:rt X < r !r ;ai! 6i ab o-.q. ; 19

6-q i5 3d c_

<c !o o-Ca a a n',i5 DEJ a

,qa E- g

55 5 0

nl} c

oa 3 x'o- i1i@ + :l

9+ *

o

tt- tts- r!- ttsri id- t r-ds 6g6g6s6g6g

;;;;;;

i-. 66g6q6g

B

a

t

I

6

6'

I!g

HR

;=

529.

oo-9

z.<3Ev1P^

iE [ 9f.i?aqz 4 !ooha

;PB95'

o.Ii'r

F

.:r 5

;

Esn6

:

€f-

;

F

3.fi :,36; i E-

o-6 € 9

tq I

gr +e!{9E:a!:1 dF! u

t: f;g * tE i[tsu &i6! +io 5- ot'".i ol,b,v o *1 j.= : i { 5:5 91r'; - 5-H : D

= J +r o XE&;-:9 6s €q 1;:iiq rSigE ?:,;-?q:iE^';! ai s9,663-=iFf;[*I] fE lilso i I ! c': d d E. -3 ii

=E=n::ir 1q;: i.eii6=6_-i 6ax= ;j- o

=i o; o 3 4: q

sp;?io Elits ^-€r;i;'i rsII i

2:*i i 5i1: +o O^o - oin 5^ Xo- 5 Y-o ;, r. !- o

3€?r 1 .-q E ;+ noi o ^

1<7 9.i - &o 9 o: r ir ,! o-r, i! =E ra I j

=S. A Y

1 6-< 6 o E oi Eg 5! I *l a tr@.ci + d E. P

-t) \h a

*u p n 9 u p u'1 I QP \9,i9u g qf Ps [43E iE Et 3t€; i 6- 3; [iu q o o 5+(q -

o=.1!:; o:r35 3 i E ? FE3 E +i"e & I7:tb a R d F ',i --

'- ** 3a s6 F*"g-^ d ;o &r I :1.g E= q Iq€Er'iisRq aa;o-1.6-o:r9;99-cdo'a1;:e*5-pE3ALo?g

l9

!J.

6g6s@r- cDF td- @r- dt- d-

.r5 +6 11 6 i.5 i'5 f6:cs :Es rq ra 5a re

EdF tnF tI.- raFa6 aL BOx0

.r. ,<. t

5H@!

E3n,o

EO

3[EP

3E

o.

$

F

vHs aJ+Phd id

,1*

n

oo

Edts

oq6-

6P

8o

3

::9 6q

A@6 {

5ZTo !iq -.rid ooi I

EE .

9;

So,P-

6-3x9-a6-b"o

tx.

E3

€s

CP

a;

!ii

rtg. r'

qPo-t

nli;Ea

I

tE.

E.

'i' 6

;

d9-d9-

;;.i6.1 5

;; E

Bc

Ic