BOVINE TUBERCULOSIS IN THE BRUSHTAIL POSSUM (TRICHOSURUS VULPECULA)

Transcription

1 BOVINE TUBERCULOSIS IN THE BRUSHTAIL POSSUM (TRICHOSURUS VULPECULA) BEHAVIOUR AND DEVELPOMENT OF AN AEROSOL VACCINATOR SOLIS NORTON April 2001

2 ii

3 iii ABSTRACT The Australian brushtail possum (Trihosurus vulpeula) is a wildlife reservoir of tuberulosis (Myobaterium bovis) in New Zealand. The disease is endemi over one third of the ountry. Possum ontrol operations have redued the prevalene of disease in livestok but have not fully ontrolled infetion in wildlife or geographi spread of the disease. The disease is transmitted to livestok when they investigate the unusual behaviour of terminally ill possums. Redution of disease inidene in possums through vaination with baille Calmette-Guerin (BCG) has shown promise both in pen trials and field studies. Integration of vaination into existing ontrol programmes may redue transmission of tuberulosis among possums, and from possums to livestok. There are two parts to this thesis. Part one is a longitudinal, behavioural study of tuberulous and nontuberulous wild possums. Part two is a desription of an aerosol delivery devie (aerosol vainator) designed to administer aerosolised BCG vaine to possums in the wild, and a reord of its progressive development. The aim of part one was to identify aspets of behaviour of tuberulous possums that may influene disease transmission to livestok. Twenty two tuberulous and eight healthy possums were observed. Possums were radio traked weekly and live trapped at bimonthly intervals on a 56 hetare site in the Wairarapa, New Zealand. Generally possums remained within their ativity range apart from infrequent long distane forays. Possums were weak, lethargi and unoordinated during the terminal stages of disease whih lasted for one to three weeks. Only three possums made long forays when terminally ill with tuberulosis. The arasses of 17 tuberulous possums were reovered of whih 15 were in dense srub or on long grass under srub and two were on pasture. Of these 17 arasses, 14 were within or near (<200m) to their ativity range. Most tuberulous possums died in their ativity range and in srub. These possums represent little risk of infetion to livestok, but a risk to other wildlife. However, the small number of tuberulous possums that died on pasture present an important risk to livestok. Interations between diseased and healthy possums during long distane forays may ause onsiderable geographial spread of tuberulosis. Part two, the development of an aerosol vainator, onsisted of pen and field trials. The aim of pen trials was to evaluate the willingness of possums to investigate novel objets and the influene of soial hierarhy on this investigative behaviour. It also allowed refinements to aerosol vainator design.

4 iv Four aptive olonies were used. In eah olony most possums (80%) showed minimal neophobia and would atively investigate novel objets. A small proportion (20%) would not approah a novel objet. A loose soial hierarhy existed with one dominant animal and a hangeable middle order. In two of the four olonies, there was one possum learly at the bottom of the soial order. Soial hierarhy did not affet the proportion of a olony whih ould investigate a novel objet or vainator. However it did effet the order in whih individual possums would investigate. The aim of the field trials was to evaluate the effiay of an aerosol vainator with possums in the wild. During five field trials, the proportion of the possum population marked with dye from the devie steadily inreased and ranged from 0% 34%. Trials were onduted over eight months and during this period a total of 56% of the study population was marked with dye. Some possums would repeatedly use a vainator. These results justify further researh into aerosol vaination of wild possums with BCG. Three key avenues for future researh inlude determining the proportion of a possum population whih will use the devie, developing an aerosol ontainer suitable for dispensing BCG vaine and determining whether the ombination of vainator and aerosol vaine eliits a protetive immune response. The aerosol vainator may be use to deliver aerosolised materials other than BCG to possums. It may also be altered to suit use by other speies.

5 v ACKNOWLEDGEMENTS I have been very fortunate in being able to aept the Masterate projet as offered by the EpiCentre, headed by Professor Roger Morris. Over the last three years, this diverse and wonderful group has enlightened my understanding of epidemiology, the strength of statistial analysis, ultural flavours of the world and the phenomenal power of the omputer. To this group I owe my thanks. Thanks to Roger Morris, my hief supervisor, who has given me the opportunity to ontribute to the understanding of tuberulosis in New Zealand, and exposed me to the range of ongoing epidemiologial issues world wide. I am indebted to Leigh Corner for his everlasting, enthusiasti guidane and support, and for happily sharing his knowledge of tuberulosis in a way whih was easily understood. My thanks also to the tireless work of Donna Lewis and the late Ron Goile who have been an important part of the various Castlepoint researh projets over almost ten years. Also Marus Johnston and family for their hard work and provision of great ompany and a perfet haven from the halls of aademia, whih would on oasion beome overpowering to a ountry shepherd. I would like to reognise the efforts of Deb MCrae who most helpfully advised on, and expertly managed the administrative side of my projet. Fiona Dikinson is also thanked for her assistane in assembling various douments throughout the study. A projet suh as this plaes onsiderable stress on the personal life of the student at times. I wish to thank my friends for their ontinual support and enouragement. I also wish to apologise for the neglet of these speial people, in partiular my dear Sarah. Finally, I would like to thank Bill Maunsell, owner of Waio Station for the use of his land and the Animal Health Board for funding. Without the assistane of these people, this trying but rewarding hapter in my life would not have been possible. Thanks to you all.

6 vi

7 vii TABLE OF CONTENTS CHAPTER 1 LITERATURE REVIEW General Introdution Tuberulosis Introdution to myobateria inluding tuberulosis Introdution to the general disease and host range of tuberulosis Bovine tuberulosis History in New Zealand Pathogenesis Hosts in general Wild hosts in New Zealand Possums Deer Ferrets Feral pigs Hedgehogs The epidemiology of bovine tuberulosis in New Zealand Tuberulosis in the badger The Possum General introdution Possum biology Possum home ranges Preferred habitat and the bush/pasture margin Ranging harateristis Possum behaviour Neophobia Attratants Soial behaviour in the possum Possum ontrol The vauum effet Buffer zones and initial ontrol Bait stations Immunology Introdution to immunology... 23

8 viii Non speifi immune mehanisms Upper respiratory trat Muoiliary system Lower respiratory trat Speifi immune mehanisms Cell mediated immune response Humoral immune response Immune response to myobateria and in partiular tuberulosis Vaination Introdution Rabies: a positive ahievement using wildlife vaination Vaination with BCG Vaination of possums with BCG Oral vaination Vetor transmission Aerosol vaination Introdution to aerosol vaination The aerosol vaine Deposition Physial properties of air ways Deposition fores Hygrosopiity Studies of aerosol deposition Survival of aerosolized miroorganisms Aerosol vaination in pratie Poultry Pigs Cattle Other animals Aerosol vaination with BCG Aerosol vaination of possums with BCG Conlusion...43 CHAPTER 2 A BEHAVIOURAL STUDY OF BRUSHTAIL POSSUMS (TRICHOSURUS VULPECULA) WITH CLINICAL TUBERCULOSIS (MYCOBACTERIUM BOVIS) Introdution Materials and Methods...48

9 ix Study site Depopulation of the study site Radio ollars Naturally infeted possums Healthy possums Experimentally infeted possums Possum behaviour Den harateristis Radio traking proedure Radio equipment Radio traking method Trapping data Geographial position data Data analysis The kernel density estimator Definition of denning range Definition of a foraging range Definition of ativity range Definition of a total range Definition of a long distane foray Definition of a hotspot Statistial analysis Results Radio traking results Survival duration of radio traked possums Cause of death of radio traked possums Behaviour of radio traked possums Loation of tuberulous possum arasses Denning behaviour of possums Denning range for possum groups Denning behaviour of individual possums Comparison of possum denning range and foraging range Ativity range and total range of possums Correlation between number of observations and range size Long distane forays Length of forays Tuberulosis hot spots... 76

10 x Sequential use of the same den by tuberulous and healthy possums Den harateristis Disussion...81 Spatial aspets of tuberulous possum behaviour Temporal aspets of tuberulous possum behaviour Possum denning, ativity and total ranges Long distane forays Hotspots Physial qualities of the dens Limitations of the study Conlusion...90 CHAPTER 3 DESIGN AND TESTING OF A SELF-SETTING AEROSOL VACCINATOR Introdution...93 Part one: Pen trials of an aerosol vainator Objetive Materials and methods Results Observations of possums exposed to novel objets General observations on the approah and investigatory behaviour of possums The influene of dominane hierarhy on the investigation of novel objets by possums Pen trials of an aerosol vainator Introdution Type 1 MkI Type 1 MkII Type 1 MkIII Type 2 MkI Type 2 MkII Type 3 Mk I Type 3 MkII Type 3 MkIII Type 3, MkIV Type 3 MkV Type 3 MkVI Type 3, MkVII (Steel opies) Type 4 MkI (Wooden dupliates)

11 xi 3.5 Disussion of pen trials Neophobia Design path of an aerosol vainator Soial hierarhy The effet of aptivity on possum behaviour Part two: Field trials of an aerosol vainator Introdution Materials and methods Site desription Trial desription Results Field trial one trapping for dye only Field trial two August 99 bimonthly trapping Field trial three trapping for dye only Field trial four Otober 99 bimonthly trapping Field trial five depopulation of the study site Summary Disussion of field trials Preliminary estimates of vainator effiieny The influene of attratant on vainator effiieny Effet of onstrution materials and appearane on vainator performane Fators effeting density of vainator distribution Is a vainator attrative to animals other than possums? What is the next step? Conlusion CHAPTER 4 GENERAL DISCUSSION APPENDICES APPENDIX I Trap sites on the Castlepoint study site Den sites on the Castlepoint study site APPENDIX II The ativity range, total range and summary statistis for radio traked possums

12 xii APPENDIX III Radio traking form BIBLIOGRAPHY...207

13 xiii LIST OF FIGURES Figure 1. Aerial photograph of approximately two thirds of the Castlepoint study site...49 Figure 2. Radio traking equipment Figure 3. Survival of healthy, naturally infeted and experimentally infeted tuberulous possums Figure 4. Distribution of survival duration of healthy, naturally infeted and experimentally infeted tuberulous possums Figure 5. Distribution of deaths for radio traked possums...59 Figure 6. Density of den site use for all radio traked possums Figure 7. Density of den sites used by naturally tuberulous possums Figure 8. The denning range of Possums 5760: An example of single luster of dens Figure 9. The denning range of Possum 5700: An example of a bimodal denning range Figure 10. The denning range of Possums 5625: An example of a dispersed denning range Figure 11. The foraging range of all radio traked possums on the study site Figure 12. Examples of dispersed (Possum 0026) and lustered (Possum 0146) patterns of denning and foraging by individual possums Figure 13. The denning range, ativity range and total range of 28 possums Figure 14. The size of ativity range, total range and number of observations of 28 possums Figure 15. Monthly distribution of 22 long distane forays made by 14 possums Figure 16. Length of 22 long distane forays made by 14 possums Figure 17. An example of a long distane foray Figure 18. Hotspots identified on the study site between Marh 1998 and January Figure 19. Perentage of denning materials used by tuberulous possums and healthy possums Figure 20. Type 1, Mk II vainator Figure 21. Type two, Mk I vainator Figure 22. Type two, MkII vainator showing taper above the pressure plate Figure 23. Trigger mehanism and spring (arrowed) on Type three MkII vainator Figure 24. Can holding mehanism showing V shaped steel plate and lamp (arrowed) Figure 25. The final stage of development for the prototype vainator, Type 3, Mk VI Figure 26. The an holding and triggering mehanism used in the twelve opies of the final vainator prototype Figure 27. One of twelve steel opies (type 3, Mk VII) of the final vainator design after the hanges listed in setion

14 xiv Figure 28. One of two wooden vainators after the hanges listed in setion Figure 29. Vainator distribution on the Castlepoint study site during field trials Figure 30. Vainator distribution on the study site during the fifth field trial

15 xv LIST OF TABLES Table 1. Deposition of 1.5? m hygrosopi partiles within the human respiratory trat Table 2. Survival duration of naturally infeted, experimentally infeted and healthy possums 56 Table 3. Cause of death of radio traked possums Table 4. Reovery of arasses of tuberulous possums: individual possum details, loation of arass, and details of habitat where the arass was loated Table 5. Denning range (hetares) for the three groups of possums: naturally infeted, experimentally infeted and healthy Table 6. Distribution of denning range types for the three possum groups: naturally infeted, experimentally infeted and healthy possums Table 7. Ativity ranges and total ranges (hetares) for three possum groups: naturally infeted, experimentally infeted and healthy possums Table 8. Correlation between range size and number of observations Table 9. The proportion of healthy and tuberulous possums whih made at least one long distane foray and the median distane from their ativity range Table 10. The frequeny of long distane forays made by individual possums Table 11. Frequeny of use of individal dens by possums Table 12. Charateristis of dens used by possums: aspet of dens used by healthy, naturally tuberulous and experimentally infeted possums Table 13. Charateristis of dens used by possums: floor ondition of dens used by healthy, naturally tuberulous and experimentally infeted possums Table 14. Summary of vainator field trials

16 xvi

17 1 Chapter 1 LITERATURE REVIEW

18 2

19 GENERAL INTRODUCTION For over one hundred years New Zealand has struggled to ontrol the problem of tuberulosis in domesti livestok. This situation has been further ompliated by the introdution of the Australian brushtail possum (Trihosurus vulpeula). Tuberulosis aused by Myobaterium bovis is assumed to have been introdued to New Zealand during attle importation and was onsidered a serious publi health problem during the nineteenth and early twentieth entury. In reent years it has also been identified as a potential barrier to New Zealand s large export trade in meat and dairy produts (Animal Health Board, 2000). The organism M. bovis is a baterial pathogen whih infets attle primarily via the respiratory trat (Corner et al., 1990). Tuberulosis infetion is ontrolled primarily by the body s ell mediated immune system (Bloom, 1994). The possum was introdued to New Zealand in the 1850s to start a fur trade (Pray, 1974) and in the following years it spread rapidly throughout the ountry. Today it is onsidered a noxious pest with an estimated population at around 1990 of 70 million (Cowan and Bayliss, 1998). This problem was ompounded by the disovery that, in New Zealand, the possum ats as a wildlife reservoir for tuberulosis and is responsible for a major part of the spread of this disease amongst attle (Ekdahl et al., 1970). Extensive efforts are made to ontrol the possum population. These are based primarily on poisons and trapping, and onduted ontinuously at great expense. Aerosol vaination as a means of proteting animals against a range of diseases has been researhed in a range of speies, primarily poultry (Jadin, 1980) and pigs (Popa et al., 1982), (Newall and Hewinson, 1995). Aerosol vaination of possums with baille Calmette-Guerin (BCG) vaine has been shown to redue the effets of infetion with tuberulosis (Aldwell et al., 1995). There is potential to use BCG as an aerosol vaine to protet wild possum populations against tuberulosis. This may be used as an alternative to, or in onjuntion with urrent methods for the ontrol of tuberulosis. This review overs literature to date on tuberulosis as a disease and also as a problem in New Zealand. It overs the immunologial aspets of infetion with and protetion against the disease. It also overs the onept of aerosol vaination as a means of ontrolling disease in animals.

20 TUBERCULOSIS Introdution to myobateria inluding tuberulosis The genus Myobateria ontain rod shaped aid fast organisms whih are generally aerobi. They an basially be divided into slow and fast growing groups. Slow growers are divided into six distint omplexes, one of whih is M. tuberulosis. There are four distint speies in the M. tuberulosis omplex, M. bovis, M. tuberulosis, M. afrianum and M. miroti (Goodfellow and Wayne, 1982). Robert Koh identified the tuberle baillus in 1882 and the human and bovine types were differentiated by Smith in In the 1930s Wells emphasised the role of infetious droplet nulei in the transmission of this airborne, inhaled pathogen (Jakson et al., 1995). Human ases of tuberulosis numbered 3.4 million in 1997 and the number of new ases is inreasing annually by about 100,000 (WHO, 1999) Introdution to the general disease and host range of tuberulosis In general terms an infetion is the invasion and multipliation of miroorganisms in body tissues, espeially that ausing loal ellular injury due to ompetitive metabolism, toxins, intraellular repliation, or antigen-antibody response. (Blood and Studdert, 1988). In addition, organisms ausing respiratory disease suh as tuberulosis must be able to reah the respiratory muosal surfae before establishing infetion, usually by attahment to or penetration of the muosal surfae (Babiuk and Campos, 1993). The progression of tuberulosis from infetion to its various linial manifestations an be onsidered as a series of stages, as the organism gradually invades further into the host. Initially the invading baillus will multiply loally. This is followed by the formation of small aseous lesions whih may grow and multiply into larger lesions that an shed bailli into the blood and lymph systems. Caseous lesions may also liquefy and introdue bailli and their produts into the surrounding area, making arrest of the disease muh more diffiult (Dannenberg and Rook, 1994). Bovine tuberulosis has a very wide host range inluding humans, primates, arnivores, ungulates, marsupials, rodents and lagomorphs (Thorns and Morris, 1983).

21 Bovine tuberulosis History in New Zealand Bovine tuberulosis (tuberulosis) was probably introdued to New Zealand through imported infeted attle in the early nineteenth entury at whih time it was a serious publi health problem, espeially in hildren through ingestion of tuberulous milk (Pray and Kean, 1969). In addition exports of beef and dairy produts are majors soures of inome for New Zealand, making it essential that international market requirements are met. The disease was ontrolled by implementation of test and slaughter shemes, initially for beef and dairy attle in 1945 and then expanded to inlude deer in 1985 (Tweddle and Livingstone, 1994). These monitoring and ontrol methods are now standard animal health and disease ontrol proedures. Initially, the attle sheme produed enouraging results but it was observed that despite apparently suessful ontrol, herds in some areas would revert bak to infeted status. In 1971 epidemiologial evidene indiated that the Australian brushtail possum (Trihosurus vulpeula) was the ause of this reinfetion (Hikling, 1991). The West Coast of the South Island was the first of these areas to be identified. Trapping of wild possums in this region revealed twelve perent of the population was infeted with tuberulosis. The persistene of infetion in dairy and attle herds in the entral and south eastern North Island was identified soon after for the same reason (ONeil and Pharo, 1995). Possum ontrol programs were subsequently implemented, resulting in a substantial deline in tuberulosis levels by On the grounds of this suess, funding for the ontrol programs was redued and this aused a resurgene of the disease as well as its expansion into previously unontrolled areas (Livingstone, 1991). A review and subsequent inrease in funding for disease ontrol meant that by the end of 1991, 2.7% of attle herds and 3.9% of deer herds were under movement ontrol restritions (Tweddle and Livingstone, 1994). By the year 2000 these figures had been further redued to 0.8% of attle herds and 1.8% of deer herds infeted (Animal Health Board Annual Report, 2000). However, the area over whih the disease may be found (vetor risk area) has slowly inreased over time to urrently over 33% of New Zealand, with a further 20% labelled a fringe testing zone. Annual expenditure on vetor ontrol, researh and ompensation to farmers inreased slightly in 2000 to $52 million (Animal Health Board Annual Report, 2000).

22 Pathogenesis Myobaterium bovis is a faultative intraellular pathogen whih multiplies within ells of the host s immune system, primarily marophages (Aldwell et al., 1996). There are five stages assoiated with infetion with tuberulosis. Sneezing and oughing are exellent methods of aerosol generation (Stark, 1998). The aerosols generated by tuberulous individuals may be ontaminated with infetious bailli whih have an infetious distane of up to two metres (O Hara, 1976). The first stage of infetion is when ontaminated aerosols are inhaled and reah the upper respiratory trat, in partiular the tonsil and other lymphoid tissues, or the lower respiratory trat where they may adhere to the muous membrane or lung surfae. One the viable tuberle baillus has adhered to the surfae it is ingested and possibly destroyed by a marophage. In the ase of an M. bovis ell whih reahes the lung alveolus, ingestion is by an alveolar marophage. These are the body s primary mehanism for learing inhaled foreign material. The level of bateriidal ativation of the marophage and the virulene of the ingested baillus determine whether the baillus is destroyed. When the marophage does not have suffiient ativity to destroy the baillus, the baillus will begin to repliate within the marophage. The seond stage of infetion is termed symbiosis and begins when the baillus has repliated within the marophage. This auses redued viability of alveolar marophages whih may interfere with the ability of the immune system to reognise the invading bailli (Aldwell et al., 1996). It also auses the marophage to spread and tightly adhere to the alveolar surfae. Eventually the infeted ell bursts and the now free bailli will repeat the exerise in other ativated and non ativated marophages, inluding those whih migrate to the site from the blood stream. In time, marophages from the irulation beome ompletely responsible for the fate of the early lesion. Symbiosis begins here as the new marophages ingest bailli, but annot inhibit or destroy them as they are not ativated. The bailli repliate logarithmially but do not harm marophages beause the host has not developed tuberulin-type hypersensitivity. With time the number of ells in this situation inrease within the lesion (Dannenberg and Rook, 1994). Stage three is the beginning of aseous nerosis. This starts when the rate of repliation of bailli in the ore of the lesion delines sharply, usually 2 3 weeks after infetion, and the host beomes tuberulin positive. At this stage, partially ativated marophages destroy the non ativated ones in whih bailli have been repliating, ausing aseous nerosis whih is the formation of visible lesions. The host destroys its own tissue during this type of immune response. Bailli annot

23 7 repliate in the aseous material due to environmental onditions and there is muh baillary material, or antigen present in the area. At this point the T ell mediated immune response first develops. In attle lesions first beome visible 5 6 weeks after bailli are inhaled and exretion begins after approximately 14 weeks (Neill et al., 1991). Stage four involves the interplay of tissue damaging and marophage ativating immune responses in both suseptible and resistant hosts. In suseptible hosts, the bailli released from the edge of the aseous entre are ingested by inompetent (non ativated and poorly ativated) marophages. This means the host must ontinue its tissue damaging response to stop the intraellular baillary multipliation. As a result the aseous entre grows in size and loal lung tissue is destroyed. Bailli draining into the lymph system are not destroyed and may initiate infetion here. They may also drain from the lymph system into the blood stream resulting in the transport of infetious material throughout the body. In resistant hosts bailli are ingested by large numbers of ativated marophages whih ontrol baillary repliation. In this way bailli are onfined to the aseous entre and no additional self damage ours. The disease is effetively arrested, frequently for the rest of the host s life although bailli may survive for years in a dormant state in an apparently healed aseous fous. The final stage of infetion is liquefation and avity formation of existing lesions. This may our even where there is a strong ell mediated response. The liquefied material is an exellent growth medium for bailli and for the first time they may repliate extraellulary. In addition, it is toxi to marophages. The liquefied material may be disharged into the airways, reahing other parts of the lung and ating as a soure of infetion to other hosts through aerosol transmission. It may also lead to pneumonia (Dannenberg and Rook, 1994) Hosts in general Many speies may beome infeted with M. bovis (Thorns and Morris, 1983). Ungulates, partiularly attle and deer represent the greatest problem. The perentage of infeted domesti attle herds is highest in parts of Latin Ameria (Argentina 37%, Chile 28.5% and Bolivia 18%) (de Kantor and Ritao, 1994). There are also problem areas in Europe (Spain 10.7%, Ireland 8.8%) (Caffrey, 1994) with the badger (Meles meles) identified as an important host in Great Britain and Ireland (Cheeseman et al., 1989). Tuberulosis is present in domesti attle in both the United States and Canada. Elk, deer and bison from both wild and domesti populations have been reorded with tuberulosis in Canada (Munroe et al., 1999). It has also been found in the aptive deer industry in the United States (Essey and Koller, 1994). The spread of tuberulosis in a number of Afrian wildlife game parks has been identified as a serious problem and threat to

24 8 biodiversity. The key host is the Afrian buffalo (Synerus affer) and disease has also been deteted in several spill over hosts inluding the lion, heetah, leopard and baboon (Buddle et al., 2000). Tuberulosis may our in a range of rodents inluding the guinea pig, rat and mouse. Carnivores suh as the ats, dogs and ferrets may also ontrat the disease (Thorns and Morris, 1983) Wild hosts in New Zealand Tuberulosis is present in several speies of both wild and domesti animals in New Zealand. It is generally aepted that there are two reservoir host speies, possums and deer. In these speies the disease an be maintained by yling within the population. There are also several spill over or dead end hosts (Hikling, 1995). The most important speies is the possum Possums The first reorded ase of myobaterial disease in a possum was in India in 1895 (Moore (1903) in Buddle and Young, (2000)). Tuberulosis was first identified in possums in New Zealand on the West Coast of the South Island in 1971 (ONeil and Pharo, 1995). In 2000 tuberulous possums were found over 24% of the South Island and 9% of the North Island, or one third of New Zealand s total area (Animal Health Board Annual Report, 2000). Prevalene of tuberulosis varies within possum populations but usually lies between 1 15% (Pfeiffer, 1994), (Jakson, 1995). However there are oasional outbreaks where prevalene is muh higher, for example 60% in a low density possum population at Flagstaff Flat was reorded in 1992 (Coleman et al., 1994). Two subsequent surveys indiated disease levels had dereased, with prevalene estimates of 17% and 9% (Coleman and Cooke unpublished 1995). Tuberulous possums may be lustered in spae and in time in areas olloquially known as hot spots (Coleman et al., 1994), (Caley et al., 1999). These hot spots may be set amongst large areas of possums with negligible levels of infetion. Hot spots may be further divided into persistent sites, where tuberulous possums may be found over an extended time period, or sporadi sites, whih are more likely the temporary effet of a single tuberulous possum whih fails to transmit the disease to other possums at this site before dying (MKenzie and Morris, 1995). Studies on the experimental infetion of possums with tuberulosis were first arried out in New Zealand at the Wallaeville Animal Researh Centre (O Hara (1976) ited in Jakson and Morris, (1993)). Various studies have demonstrated diret and indiret transmission between possums

25 9 while showing disease due to experimental infetion to progress more rapidly than natural infetion in the wild (Buddle and Young, 2000). Buddle et al., (1994) experimented with infetion with very low doses (20 olony forming units) of M. bovis by intratraheal instillation. This resulted in loss of appetite in 3 5 weeks, followed by weight loss and linial signs of disease. Severe tuberulous pneumonia was evident by eight weeks and some possums had died within 5 9 weeks. At post mortem, M. bovis was ultured from the lungs and lymph nodes of all possums studied with lesions, and also found in the liver, spleen and kidney. Early hypotheses of major transmission pathways were entred around pasture ontamination, however researh has shown that this is not a signifiant fator (Jakson et al., 1995). Transmission between possums in the wild is now thought to be primarily by aerosols (Lugton et al., 1995). The minimum infetive dose by aerosol is between one and three infetive M. bovis bailli (Lurie (1964) ited in Dannenberg and Rook, (1994)). Aerosols have been shown to be infetious over distanes of up to two metres (O'Hara et al., 1976). There are three main pathways for transmission. The first path is diret horizontal aerosol transmission whih ours during agonisti or mating behaviour, partiularly around den sites. The seond route is indiret horizontal transmission whih ours when healthy possums beome infeted through using a den previously used by a tuberulous possum. The third pathway is pseudo-vertial transmission, where an infeted dam transmits the disease to her joey through normal maternal behaviour (Jakson et al., 1995). It is thought the majority of transmission between adult possums ours during high levels of soial interation, for example mating behaviour or onfrontations at or near den sites (Paterson, 1993), (MKenzie, 1999), (Pfeiffer, 1994). In infeted wild possums, disease progresses at a variable rate dependent on a range of stresses inluding environmental (notably adverse weather onditions), nutritional and the stress assoiated with latation (Pfeiffer, 1994). Examination by (Lugton et al., 1995) of 486 tuberulous possums showed that dissemination ours during the early stages of disease and the infetion of multiple sites is typial. Inguinal and axillary lymphoentres were most frequently affeted and also showed evidene of infetion in the very early in the disease proess. This is probably due to their entral role in lymphati learane of organisms. The lung was also ommonly affeted. The disease proess in possums was divided into four stages by (Lugton et al., 1995). During the first stage there were no detetable gross lesions, however it is possible exretion via the respiratory trat was ourring. In the seond stage some gross lesions were visible, partiularly in

26 10 the lung, and extensive dissemination of the disease was evident. The third stage was haraterised by generalised disease with exretion primarily by the respiratory route, sinuses and mammary gland. However, there was no evidene of loss of body ondition. In the short (1 8 weeks) final stage linial effets of the disease develop rapidly and physial ondition delines. Behavioural hanges were evident, and inluded lethargy, weakness and poor oordination. Terminally ill possums may also be found wandering on pasture in daylight. Exretion via urine and faees was noted, in addition to the routes already identified. In a study by (Pfeiffer, 1994) 50% of tuberulous possums survived for two months after detetion of palpable lesions, and by five months 81% of possums with lesions were dead. Geographi spread of the disease is thought to be primarily due to dispersal of tuberulous juvenile possums to uninfeted areas. (Cowan et al., 1996) monitored the dispersal of young possums and found 20% of radio ollared individuals dispersed greater than two kilometres with some moving as far as 12 kilometres. The reprodutive rate of the disease in possums is about 1.8 whih means that eah tuberulous possum is likely to infet two other possums in its life time (Caley and Ramsey, 1999). The type of immunologial reation whih the possum mounts in response to infetion with tuberulosis appears to be ineffetive, resulting in the rapid progression of disease. Possums an be grouped with ferrets and badgers in having an inability to wall off established lesions by formation of a granuloma. In the possum, this is a result of alveolar marophages having an immunosuppressive effet on lymphoyte ativity, regardless of whether the animal is infeted with M. bovis or not. Cells whih normally deal with bateria are poorly organised to deal effetively with M. bovis (Buddle and Young, 2000). This is despite the fat that possum marophages respond vigorously to infetion (Jakson and Morris, 1993). Normally, the granuloma leads to ontainment and eventual ontrol of the myobateria but in the possum, lak of granuloma formation leads to bateria spreading to form satellite lesions, hene produing generalised and fatal disease (Buddle and Young, 2000). The disease in possums is typified by the presene of aid-fast organisms in large lesions with overt nerosis and mirosopi granulomata with little or no nerosis. Infetion is usually deteted first in the lungs and assoiated lymph nodes, the bronhial lymph nodes and then the liver, spleen and kidneys as the disease invades the abdominal avity (Buddle et al., 1994), (Cooke et al., 1995) Deer Deer (Cervus spp) were first introdued to New Zealand in 1851 (Challies, 1990) with the wild population now approximately 250,000 and a domesti population of 600,000 (Animal Health

27 11 Board Annual Report, 2000). Tuberulosis is present in both wild and domesti populations and enlarged lymph nodes about the head and nek are typial of linial infetion (Lugton et al., 1995). Prevalene in domesti deer was measured at 1.8% in 1999/2000 (Animal Health Board Annual Report, 2000). Infetion of deer with tuberulosis has been reported in other ountries inluding North Ameria, Great Britain, Ireland and Continental Europe (Lugton, 1997), (Munroe et al., 1999), (Essey and Koller, 1994) Ferrets Ferrets (Mustela putorius) were first introdued to New Zealand in 1879 to help ontrol the rabbit population and also in aptivity to start a fur industry. The wild population is now widespread throughout New Zealand though densities are highest where rabbit numbers are high. Ferrets are regarded as pests, mainly as a result of damage to native animals, partiularly flightless birds. The diet of the ferret is mainly small mammals, supplemented with arrion. It is probable that wild ferrets beame infeted with tuberulosis through onsumption of diseased arrion, for example possums and domesti animals disarded in offal pits (de Lisle et al 1993). Tuberulosis was first reported in ferrets in the 1970s (Stokdale, 1975) and their importane and relationship with possums in regard to the disease has been the subjet of muh debate (Ragg et al., 1995). However, it is now beoming aepted that ferrets are not true maintenane hosts for the disease (Caley et al., 1996). Prevalene of disease ranges from 15% (Ragg and Walker, 1996) up to 90% (Lugton et al., 1995) and signs of infetion are typially lesions in the mesenteri lymph nodes and pulmonary region (Dunkin (1929) ited in Lugton, (1997)) Feral pigs Feral pigs (Sus srofa) are widespread throughout New Zealand. Wild pig hunting is a popular rereation and populations are maintained in some areas by the release of young pigs. The omnivorous, opportunisti feeding habits of feral pigs are well suited to onsuming potentially tuberulous arasses found in the wild. Pigs infeted with tuberulosis an be found in areas where the disease is onsidered endemi in other speies. Prevalene of disease was measured at 31% in Central Otago (Wakelin and Churhman, 1991). Disease is haraterised by lesions in the lymph nodes, whih regress as the pig ages (Ray et al., 1972) Hedgehogs In New Zealand, hedgehogs (Erinaeus europaeus) are abundant in developed lowland distrits but less ommon in forest environments and mountainous areas (Brokie, 1990). Reorded ases

28 12 of disease are unommon with lesions ourring primarily in the lungs and lymph nodes (Lugton, 1997) The epidemiology of bovine tuberulosis in New Zealand The epidemiology of tuberulosis in New Zealand is disussed in three parts. Aspets of the disease are overed first, followed by suseptible hosts and finally, important environmental fators. Tuberulosis is a myobaterial disease spread mainly by aerosol and aused by M. bovis. It is predominantly a disease of the respiratory trat, though spreads throughout the body in advaned stages. M. bovis is an obligate pathogen but an survive for variable periods in the environment. The duration of survival is influened mainly by temperature, moisture, ph and exposure to sunlight (Morris et al., 1994). The three most important environments for survival in New Zealand are possum den sites, pasture and the arasses of animals whih have died of tuberulosis. In possum dens, survival of the organism ranged between seven and 28 days. On the forest floor survival was less than four days in summer and between 14 and 28 days in winter. M. bovis survived less than four days on pasture regardless of season (Jakson et al., 1995). Survival of the disease in unsavenged arasses of dead hosts was at least four weeks, although most aessible arasses were savenged ompletely within two to three days (Pfeiffer and Morris, 1991). There are a number of different strains of M. bovis that an be identified by restrition endonulease (REA) typing, though there is no evidene of any differene in virulene between these strains (Morris et al., 1994). The key hosts in New Zealand are domesti attle and deer, wild deer, possums, ferrets and pigs. Transmission between hosts, other than the savenger speies, is primarily by aerosol. The majority of spread is between possums and this is hypothesised to our mainly around den sites (Pfeiffer, 1994). Long distane spread of the disease is largely due to dispersal of young possums seeking new home ranges and the ommerial trading and transport of infeted livestok (Pfeiffer, 1994). Transmission from possums to attle and deer ours when the possum is terminally ill. The disease auses the possum to beome weak and lethargi and to move around during daylight. This errati behaviour stimulates ative investigation by deer and attle. The normal response of a possum to suh threats is a loud hissing sreeh whih is ideal for generating infetious aerosols, whih may be inhaled by investigating animals (Paterson and Morris, 1995). There is a smaller amount of spread within attle and deer populations. Ferrets beome infeted primarily through the alimentary trat as a result of onsumption of tuberulous arasses (Caley et al., 1996). Feral

29 13 pigs an also beome infeted through onsuming tuberulous arrion though the disease proess is often arrested by the pigs immune system (Ray et al., 1972). Hosts may be divided in to two ategories. The first ategory, maintenane hosts, represents speies in whih the disease is self-sustaining. Possums and possibly deer are examples of wild maintenane hosts (Hikling, 1995). The seond ategory, spill over hosts, represent speies whih may beome infeted with tuberulosis, but annot sustain infetion long-term within the speies. In the absene of an ongoing external soure of infetion (for example the possum), disease will die out in these speies. Examples of spill over hosts in New Zealand inlude ferrets, wild pigs and wild deer (Morris et al., 1994). There are several environmental fators whih are important in the persistene of tuberulosis in New Zealand. Farmland whih ontains or borders on suitable possum habitat is most likely to be assoiated with a tuberulosis problem. The area where bush or srub and pasture meet is a prime example of high risk habitat (Caley and Coleman, 1998). In addition infetion is frequently lustered both in spae and time in possum populations (Coleman, 1988). Peaks in the level of tuberulosis reorded in possums oinide with peaks in adjaent attle, suggesting disease in one speies triggers that in another (Coleman et al., 1999). This suggestion is supported by DNA fingerprinting studies whih show that often the same strain of M. bovis is infeting both speies (Collins et al., 1988). Tuberulosis an only be eradiated from New Zealand by ontrolling infetion in both wildlife and domesti stok (Buddle et al., 2000). Although levels of disease in livestok an be maintained at very low levels, the inability to ontrol disease in wildlife will result in ontinued pathy outbreaks of disease (Morris et al., 1994). Currently New Zealand s test and slaughter strategy has redued infetion to 0.8% of attle herds and 1.8% of deer herds. Approximately $30 million is spent annually on vetor ontrol (primarily possums) in the form of poison baits and trapping. This is part of the $50 million spent in total on the tuberulosis problem annually (Animal Health Board Annual Report, 2000) Tuberulosis in the badger In Great Britain and Ireland, the badger (Meles meles) is assoiated with the persistent infetion of livestok with tuberulosis, in muh the same way as the possum in New Zealand. In Great Britain the first tuberulous badger was identified in 1971, after investigation of unexplained outbreaks of the disease in attle. As a result, badger ulling operations were implemented to ontrol the disease, in addition to existing ontrol methods. Culling operations redued rates of infetion in

30 14 attle in loalised areas, but failed to eradiate the disease. It is now known that tuberulosis is endemi in badgers in some areas, and that they funtion as a wildlife reservoir for disease (Cheeseman et al., 1989), (Little et al., 1982). Badgers are omnivorous, feeding mainly on earthworms, and live in groups of up to 15 animals in underground setts. The dark, onfined onditions within a sett provide exellent onditions for the transmission of tuberulosis (Roper, 1992), (Clark, 1988). Prevalene of disease over six years in Glouestershire varied from 0% to 8% (Cheeseman et al., 1989). Charateristis of infetion in badgers inlude lesions in the respiratory and urinary systems, and assoiated lymph nodes. Suppurating absesses may result from infetion via bite wounds, sustained during agonisti interations (Smith et al., 1995). Healthy badgers avoid ontat with attle and rapidly move away if approahed (Benham and Broom, 1989). However, the behaviour of badgers in the terminal stages of disease is similar to that of terminally ill possums, for example physial debilitation, lethargy, being seen abroad during daylight and not showing normal avoidane behaviour (Jakson, 1995). The sputum, urine and faees of tuberulous badgers have been shown to ontaminate pasture (Smith et al., 1995) but the most important route of transmission to livestok is still unlear. Inreasing publi disapproval of badger ulling indiates that alternative methods for the ontrol of tuberulosis in badgers and attle are required (Clifton-Hadley, 1996). One possibility is the vaination of badgers with BCG. Vaination of badgers enhaned ell mediated immunity, delayed exretion of the organism and prolonged survival of infeted badgers (Stuart et al., 1988) THE POSSUM General introdution The Australian brushtail possum (Trihosurus vulpeula) is a member of the Order Marsupalia; Family Phalangeridae. Adults weigh between two and four kilograms and are approximately the size of a domesti at. The possum is noturnal, prinipally arboreal and usually solitary. Possum populations are widely distributed within Australia and Tasmania where it typially inhabits open eualypt woodlands, ompeting with a number of other arboreal marsupials (Clout and Eriksen, 2000). The possum was first suessfully introdued into New Zealand in Riverton on the west oast of the South Island in 1858 (Pray, 1974). Numerous releases of possums were made during the 1890s primarily to start a fur trade. Government protetion and ative distribution of the possum was used to failitate olonisation up until By this date publi onern over environmental damage was suffiient to put a ban on all further releases and by 1951 a bounty sheme was

31 15 instigated to enourage the destrution of the possum (Jakson and Morris, 1993). Today, possums number approximately 70 million (Cowan and Bayliss, 1998) and oupy about 95% of New Zealand. They are found throughout a wide range of habitats (Pray, 1974) and are onsidered a serious environmental pest (Morris et al., 1994). The possum is regarded as a serious pest for various reasons. First and foremost it ats as a reservoir and spreading agent for tuberulosis. This is disussed in depth elsewhere (setion ). Seondly, the sustained browsing of some native tree speies by possums leads to tree death and in some ases omplete ollapse of the native anopy. It may also ause a hange in floristi omposition of the plant population, whih an our in as little as years. Rata and Kamahi forest areas in the South Island are prime examples of possum damage (Payton, 2000). Thirdly they are predators of native animals. Possums have been shown to eat the eggs, hiks and even adult birds of a range of native speies inluding the kiwi. They have also been found to prey on native snails and a range of insets (Sadlier, 2000). The final reason why the possum is regarded as a pest is their eonomi impat on New Zealand s primary prodution. Feeding damage has been reorded in at least 46 varieties of fruit and vegetable, although damage may be seasonal and pathy. In addition, the growth of young stands of Pinus radiata is redued by possum browsing (Buther, 2000) and willows and poplars planted for erosion ontrol and athment protetion are also frequently damaged (Thomas et al., 1984). The possibility of possum farming has been onsidered in the past with the primary produts being fur, pelts and meat. This idea was not suessful due to hanges in possum behaviour resulting from aptivity, in partiular the failure to breed (Leeh, 1979) Possum biology The possum is highly adaptable. So variable is the possums behaviour that observations and onlusions from one loality may be irrelevant when onsidering populations elsewhere (Cook, 1981). This adaptability is evident in a wide range of biologial fators suh as breeding, feeding and denning habits and population density whih is usually between 1 10 possums/hetare (ha) but has been reorded as high as 25 possums/ha (Coleman et al., 1980). The main breeding season is from Marh to June but in some areas there is also a smaller spring breeding season (Flether and Selwood, 2000). Possums eat a wide range of food whih varies with habitat and season. Major onstituents of the diet inlude foliage, pasture and seasonal foods suh as flowers and fruit (Nugent et al., 2000).

32 Possum home ranges Studies in a variety of habitats have determined the size of possum home ranges and degree of overlap of individual ranges. A summary of the findings is presented by (Cowan and Clout, 2000). In Australia home ranges were found to be on average 3 ha for males and 1.5 ha for females. Female ranges would frequently overlap, while male ranges were exlusive (Dunnet, 1963). In New Zealand extensive overlap of possum home ranges is ommon both between and within sexes (Crawley, 1973). In a dense native bush environment suh as the Orongorongo Valley (lower North Island, New Zealand), possum density was found to be 10.6 possums/ha and their home range to be 0.8 ha for males and 0.5 ha for females. In a pine plantation the home range was estimated at 0.7 ha for both males and females (Warburton, 1977). In mixed bush, srub and pasture habitat suh as in the Wairarapa, home ranges were estimated at 1.4 ha for males and 0.9 ha for females (Paterson et al., 1995). Within the home range an be found one or possibly several smaller foi whih onstitute a denning range (Ward, 1978). Where land is very steep and espeially at high altitudes possum home ranges extend further vertially than horizontally beause a greater range of plant types and hene food types are overed by hanging elevation (Green and Coleman, 1986) Preferred habitat and the bush/pasture margin Provided there are sheltered den sites and adequate food soures, possums live almost anywhere in New Zealand. They are found in native and exoti forests, grasslands, sand dune and swamp ountry, even urban and ity areas. Faeal signs have been reorded from sea level up to 2,400 m (Pray, 1974). Possums may be found in unusually high densities along the area where bush or srub and pasture meet (bush/pasture margin). This is beause bush and srub habitat provide many suitable den sites while pasture provides a favoured soure of food (Green and Coleman, 1981). Possums have been reorded moving up to 1,000 m through bush to pasture. These movements were found to oinide with peaks in pasture growth, indiating it is an important soure of food. One possums had moved from the bush to pasture, they seldom moved further than 300 m away from the edge of the forest (Green and Coleman, 1986). The signifiane of pasture as a onstituent of the diet of the possum is disputed by Harvie (1973) and Jolly (1976) who reorded possums feeding on pasture seven times, fruit trees 250 times and willow trees 80 times. Studies have found an unusually high prevalene of tuberulosis in possum populations on the bush/pasture margin (Coleman, 1988). This habitat was previously thought to be an important

33 17 fator in the transmission of disease from possums to attle and between possums. It was believed that diseased possums would ontaminate pasture, whih was subsequently grazed by attle and healthy possums. Transmission was also thought to result from frequent interations between diseased and healthy possums on pasture (Ekdahl et al., 1970). However, more reent studies have shown that pasture ontamination is an unlikely route for disease transmission. The proportion of attle infeted via the alimentary route was suggested as very small by Morrison et al. (2000) with most tuberulous attle having a lesion distribution indiative of infetion via the respiratory trat. In addition, the survival of M. bovis in a pasture habitat is unlikely to be more than four days (Jakson et al., 1995). Observations of possums on pasture by Paterson et al. (1995) and Brokie (1987) showed very few interations lose enough to failitate aerosol transmission Ranging harateristis The young adults of many wild mammalian speies disperse away from the territory of their parents in order to establish their own. This natural tendeny failitates geographi spread of possums into new areas. Young possums disperse between approximately ten and 24 months of age (Jakson and Morris, 1993). A greater proportion of males disperse, though on average females disperse further and movements of greater than 20 km have oasionally been measured (Brokie et al., 1987). Where the dispersing possum is infeted with tuberulosis, these movements are likely to result in geographi spread of the disease. One a home range is established, possums usually remain there for life (Crawley, 1973). Possums may make long distane forays, measured at up to 1,600 metres, to seasonal food soures suh as apples, walnuts, pine atkins or the spring growth on a range of speies inluding poplar and willow trees (Jolly, 1976) Possum behaviour Patterns of intraspeifi soial behaviour influene the fundamental ations of a speies, suh as mating, feeding and ompetition for shelter, to ensure the persistene of that speies (Grier, 1984). The aspets of possum behaviour were overed in broad detail by Montague (2000). There are three partiular aspets of possum behaviour whih will be addressed in this review. These are neophobia, soial hierarhy and interations between possums around traps or bait stations Neophobia Neophobia is the fear of novel objets, ausing avoidane of new, potentially dangerous stimuli within a familiar environment, thereby ensuring the safety of the animal. It is an important

34 18 omponent of behavioural eology. The degree to whih this aversion is present varies due to a range of fators inluding number of previous presentations of the novel stimulus, sex and soial rank of the subjet, and even individual identity (Sunnuks, 1998). Minimal neophobia is evident in wild possums. In ontrast to avoidane, (Short, pers. om., 1998) laims investigatory behaviour toward novel objets by wild possums to be very ative and aggressive, in omparison to aptive possums. Formal studies have also shown possums to be naturally urious, and to atively investigate novel objets (Carey et al., 1997). In omparison with other marsupials, possums show very little wariness or avoidane behaviour (Russell and Peare, 1971). Neophobia is present in some wild possums in the form of bait and poison avoidane. The major ause of this is previous sub-lethal doses of poison. Avoidane of poison baits was shown by 68% of mature possums trapped from a previously poisoned area. Approximately 20 25% of possums also rejeted poison baits during their first exposure, and some were found to avoid novel nontoxi baits. It is unlear whether this behaviour is learned or innate. Some of these animals ontinued to demonstrate avoidane after two and a half years. In omparison, bait avoidane was not demonstrated by young possums trapped in a poisoned area (Matthews and O'Connor, 1996). (Morgan, 1990) showed that although up to 34% of possums rejeted toxi pellet baits, non-toxi baits were rejeted by only 5 7% of possums Attratants Possums are solitary and noturnal. They have a large olfatory bulb, well developed vomeronasal region and many sent glands indiating sent is an important omponent of possum eology (Todd, 1995). Of partiular importane are those sents used for individual reognition, allowing subordinate animals to avoid areas oupied by dominants (Day et al., 1998). Effetive lures have been found for a number of mammals. They usually ontain ompounds reognised by the target animal, for example favoured seeds of deer mie and rie strains preferred by rats (Morgan et al., 1995). Olfatory lures are the traditional equipment for attrating possums and virtually every imaginable material has been tried at some time with highly variable results (Cowan, 1987), (Pray, 1974), (Morgan et al., 1995). The failure of the trial by Todd (1995) to identify a favourite amongst twenty attratants an be regarded as a typial result. To ompliate matters further there are many hunters who are onvined of their own partiular formula, while other experiened hunters note the yli trends in popularity of the more ommon types of lure. Although anedotal, the views of trappers have value. Their beliefs have a less sientifi basis than experimental trials but they depend on the effetiveness of their ompounds

35 19 for their livelihood. The results of sientifi trials are only slightly more informative. White light and an eletroni beep have been trailed as alternatives to olfatory lures. Results showed that a proportion of possums would not approah either of the stimuli. For those possums whih did investigate, there was no signifiant differene in approah or investigation between treatment and ontrol boxes (Carey et al., 1997). White material, for example flour, is reported to funtion as a visual lure possibly by imitating the white preipitate left after possum urine has dried. However, live trapping studies have found that use of a lure inreases the total number of possum athes, but neither onstant use of a single lure, nor daily hanges of lure inreased the proportion of the population aptured eah month (Cowan, 1987) Soial behaviour in the possum A soial hierarhy is present in a wide range of animals inluding the possum. This hierarhy or peking order determines how resoures suh as food, shelter and mates are distributed amongst a population. When resoures are abundant, they may be used by all levels of the soial order, however when resoures are sare, those at the bottom of the order go without. The ranking of individuals in a soial order is determined primarily by agonisti interations (Grier, 1984). The soial order of possum populations has been the subjet of several studies. In both New Zealand and Australia soial order is based on a system of mutual avoidane by o-dominants of both sexes. Dominane is probably established by initial enounters between individuals and is thereafter maintained mainly by memory (Winter, 1976). Soial postures and agonisti interations while defending den sites also maintain this hierarhy (Biggins and Overstreet, 1978). Winter (1976) studied wild possums and onluded that there was a linear hierarhy for eah sex and that the order of individuals was influened primarily by age and body weight. The idea that a hierarhy is determined by age and body weight is supported by both (Jolly, 1976) and (Oldham, 1986). Oldham also reports that interations between males were more frequent than between females and that in a mixed sex olony, females were usually dominant over males. In ontrast, an earlier study reported soial order in the possum to be of little if any value. It also laimed that the lak of onventionalised dominane and submission prevented group formation, sine the response to aggression in possums is fight or esape (Kean, 1967). The pattern of older, heavier individuals oupying the dominant positions and females being dominant over males has been found in the red kangaroo (Megaleia rufa) (Russell, 1970). In the grey kangaroo (Maropus giganteus) males were found to be dominant in aptivity and in the wild, and a soial order was present in both males and females (Grant, 1973).

36 20 Wild possums at bait stations show mutual avoidane of ospeifis. Hikling and Thomas (1990) report that feeding possums were alert to the presene of other possums nearby, but that bait stations were never used by two adults at the same time. If a possum approahed a bait station that was in use, it would either stop 5 10 metres away and wait for it to be vaated, or the new arrival would ontinue to approah until the possum at the station stopped feeding and moved away. There was no diret ontat between males, though lumps of fur were oasionally found around heavily used stations Possum ontrol The first efforts to ontrol the possum population New Zealand began in 1951 with a bounty sheme (Jakson and Morris, 1993). Control operations have been arried out regularly sine then for the protetion of livestok against tuberulosis and the protetion of native forests. The primary method of ontrol is the aerial distribution of died arrot or grain based pellets to whih sodium monofluoroaetate (ompound 1080) has been applied (Morgan, 1990). Maintenane ontrol operations, based on manual poisoning and trapping tehniques, must also be performed every one to two years, (Department of Conservation, 1999). In the year to June 2000 Animal Health Board expenditure on possum ontrol was $27 million. An additional $2.5 million is invested in researh for more effetive methods of possum ontrol (Animal Health Board Annual Report, 2000). While the existing ontrol methods are effetive at reduing possum numbers in the short term, areas an be reolonised as a result of the vauum effet (see setion 3.5.1). Overall they have had little suess at reduing the area of New Zealand in whih tuberulosis is endemi. There are inreasing publi onerns over the ontinued use of large quantities of poisons and the on going finanial ost of ontrol shemes (Buddle et al., 2000). In addition, possum traps and poisons are a hazard to other wildlife inluding New Zealand s native flightless birds (Warburton et al., 1997). Three aspets of possum ontrol are relevant to this review. These are the use of bait stations and the onepts of the vauum effet and buffer zones The vauum effet The vauum effet is a term used to desribe the permanent movement of possums from their original home ranges in unontrolled areas to adjaent regions whih have been subjet to ontrol operations. Following ontrol, possums in the surrounding area are attrated into the depopulated area by available den sites and inreasing food resoures as plant life reovers from possum browsing. This movement is distint from juvenile possums immigrating in searh of their own

37 21 home range. The vauum effet has been desribed in the past as a key part of possum reolonisation, partiularly in the two years following poison operations and in areas where maintenane ontrol is not arried out (Cowan and Clout, 2000). However, a number of studies indiate that the vauum effet may not be as important as ommonly believed. Green and Coleman (1984) showed very few possums moved into a 100 ha area up to three years after intensive ontrol operations. (Cowan, 1993) studied possums on Kapiti Island and also found no evidene of movement following intensive trapping operations. (Efford et al., 2000) found some hanges in home range, though adjustments to individual ranges were usually only about 50 metres into the ontrolled area Buffer zones and initial ontrol A buffer zone is a perimeter of partiular width around a previously ontrolled area, or area of importane, farms for example, whih is maintained at a very low possum density. The effet of this is to stop the possum population outside the buffer zone reolonising the ontrolled area within, essentially reating a possum free island. Where possum and attle populations are infeted with tuberulosis, farms plaed within these islands have a muh greater hane of eradiating the disease from livestok. Buffer zones are initially reated by ontrol operations designed to redue the possum population by 70 90% as determined by the residual trap ath of 5% or less, whih is the standard Animal Health Board target for suess of a ontrol operation. This makes the assessment of suess dependent on initial possum density (Livingstone, pers. om., 1999). Following initial ontrol, maintenane ontrol is used to keep the buffer zone below 40% of the pre-ontrol population as reommended by (Nugent et al., 1997). In a study on the effetiveness of a buffer zone, the perentage of reators from farms adjaent to forest buffer widths of 1, 3 and 7 km were ompared. The effet of these ontrolled buffer zones was to lower the inidene of tuberulosis from 1.22% 2.06% down to 0.35%. This redution was not signifiantly different aross the three buffer widths. After two years, the possum population in the 1 km buffer zone was reovering rapidly, partiularly around the forest edge of the buffer, while in the 3 and 7 km areas it remained low. The study onluded that wider buffer zones were more effetive at maintaining possum populations at low levels for longer periods, though the ost relative to width of buffer must also be taken into aount (Fraser et al., 1998). Repliation of this trial would be benefiial to onfirm Fraser s results, as it is the only substantial researh into the effets of buffer zones on possum populations and reator rates in New Zealand.

38 Bait stations Bait stations are small plasti weatherproof ontainers whih are used for dispensing pre-feed and poisons for possum ontrol. They are beoming inreasingly important in the ground based ontrol of possums. They an also be used to provide useful information about the loal possum population. This information an be divided into three types. Firstly, the size of the possum population in the immediate area an be determined by the volume of bait taken as eah possum onsumes on average 100 g of bait (Hikling and Thomas, 1990). This information an be used to alulate the volume of poison required to effiiently ontrol a range of possum densities. Seondly, observation of bait stations used by wild possums has improved the understanding of behaviour around fixed feeding stations. It has also shown the effet of bait station density on proportion of the possum population using the devies. Investigation in four habitats showed the number of possums attending inreased rapidly over the first two nights then levelled off (Hikling and Thomas, 1990). Between % of possums within 50 m would visit a bait station and 60 70% of possums within 500 m (Morgan et al., 1995). In another trial bait stations spaed at 100 m intervals were used by more than 90% of the possum population. The proportions of possums using bait stations at 100 m and 150 m were not signifiantly different from this, though there was a signifiant redution in use at 200 m spaing (Thomas and Fitzgerald, 1995). Thomas et al., (1996) found that bait stations spaed at 150 m along forest-pasture margins were used by 79% of possums aptured within 300 m of bait stations. In the forest this figure was over 80%. On the strength of these results Hikling and Thomas (1990) reommend that feeder stations should be spaed no more than 100 apart. Thirdly, when used as pre-feeders they an be used to train resident possums to frequent a partiular area. It is likely that repeated food rewards from a bait station would redue fear of investigating these novel objets. Pre-feeding with bait stations has been shown to inrease yanide kills (Morgan et al., 1995). It has also been shown to inrease the amount of 1080 taken from bait stations on the first night and overall by 70% (Thomas, 1998). Bait aeptane from stations has been shown to vary little (85 100%) between seasons, thought to be slightly lower at the bush pasture margin (84%) than deeper in the forest (95%). The suggested reason for this is that pasture speies are more nutritious than forest speies and the result of this is redued interest in additional food soures (Morgan et al., 1999). In ontrast to this data, Keber (1987) ited in (Hikling and Thomas, 1990) found that bait stations had little effet on the movement of possums, exept for those animals whih would have enountered the station either in the ourse of normal movements about their ranges, or as a result

39 23 of oasional long distane movements. Bait station use was unaffeted by the use of innamon oil as an attratant (Morgan et al., 1995) IMMUNOLOGY Introdution to immunology The mammalian immune system protets the body against foreign material (Blood and Studdert, 1988). This inludes pathogeni miroorganisms suh as bateria, myoplasma, virus and fungi (Villee and others, 1984). There are two key divisions to the immune system, non speifi and speifi immunity. Speifi immunity is further divided into ell mediated and humoral immunity (Kuby, 1997). A partiular immune response may involve a single immune proess, or a ombination of different immune proesses. This review fouses on those aspets of immunology relevant to infetion with airborne pathogens and begins with a disussion of respiratory trat and muosal immunity, followed by ell mediated immunity and loalised immune responses. The immune response partiular to tuberulosis onludes this setion Non speifi immune mehanisms Non speifi immunity is the basi resistane to disease whih an individual is born with. It utilises four types of defensive barrier, anatomi (skin and muous membranes), physiologial, (low ph of the stomah, temperature and hemial mediators) phagoyti (marophages) and inflammatory. It is the first immune defene against foreign material and destroys most miroorganisms within a few days (Klein and Horesji, 1997) Upper respiratory trat The upper respiratory system is the point where an inhaled pathogen first enters the body. It has several funtions inluding ating as a onduit for air during respiration and a non speifi immune mehanism by using muous membranes to filter and ondition the air. It is also responsible for olfation and phonation (Beeh, 1991). Inspired air is ontaminated with a wide range and varying amount of airborne partiles inluding miroorganisms, dust, smoke and soot. The average human inhales about 10,000 organisms daily (Klein and Horesji, 1997). The respiratory trat is omposed of many sharp twists and turns (bifurations) ause turbulent air flow during inhalation and exhalation. This turbulene auses the larger partiles to be thrown against the muous membrane or muus overed hairs, to whih they

40 24 adhere. The nasal surfae reeives the largest amount of foreign material, although the physiologial effet of material deposited here is minor ompared with the small amount penetrating further into the airways (Knight, 1973). The mouth and throat are proteted by a onstant flow of saliva whih washes deposited partiles to the stomah (Klein and Horesji, 1997) Muoiliary system Muous sereting membranes are present in the upper respiratory trat, onjuntiva, vagina and intestinal trat (Mesteky and MGhee, 1989). Muus is derived from goblet ells and defined as the free slime of the muous membrane. It onsists of fluid, salts, leukoytes (lymphoytes), plasma ells, glyoproteins, desquamated ells, various salts, and polysaharides. It is the long nature of this final onstituent s moleular struture whih makes muus visous (Villee and others, 1984), (Blood and Studdert, 1988). Muous membranes are overed by a ontinuous blanket of muus. This blanket has a thin sol top layer and thiker gel layer underneath. The blanket ats as a dynami learing mehanism, with the muus propelled by iliated ells in metahronal waves. Partiles whih land on the ilia stimulate prodution of a raft of muus (Hath and Gross, 1964) whih then arries the partile at a rate of approximately 18 mm per minute (Antweiler, (1958) ited in Hath and Gross, (1964)) toward either the pharynx or larynx (Tyrell, 1972;Beeh, 1991). From here the partile is swallowed and destroyed in the aidi environment of the stomah (Hensel, 1996). The physial properties of muus and the effiieny of traheal learane depend on maintenane of the balaned interation among several ell types (Basbaum, 1984). It is possible for the inhaled partiles to ross this muosal barrier and this is by way of M ells, whih transport antigens from the surfae into the body to lymphoid tissue for destrution (Klein and Horesji, 1997). Very high levels of air ontaminants redue the learing apaity of the respiratory trat and this an inrease the risk of lung infetion (Cox and Wathes, 1995). Overall, muous membranes protet a very large surfae area of the body against foreign material. However they are also the most frequent routes of entry for baterial, viral and parasiti infetion (Mesteky and MGhee, 1989). The muoiliary system is present throughout the upper airways but does not protet the lower respiratory trat Lower respiratory trat Protetion against airborne pathogens in the lower respiratory trat is the responsibility of alveolar marophages (West, 1987). Marophages are large phagoyti ells derived from white blood ells. They are found in many of the body s tissues inluding the alveoli of the lung (Villee and

41 25 others, 1984). Alveolar marophages at as savengers, roaming around the surfae of the alveoli and engulfing partiles whih have landed on the lung surfae from inspired air. Most of these partiles are dead but some are pathogeni. The majority of these partiles are digested in situ (Klein and Horesji, 1997) although in some ases marophages, having ingested the partile, leave the lung via either the lymphati system or blood stream. They may also migrate to the muoiliary system where they are transported to the stomah and digested or alternatively, expelled from the body by oughing (West, 1987). Marophages transport antigeni material and present it to lymphoytes. This is an important link from the non speifi immune system to the speifi immune response (Beeh, 1991) Speifi immune mehanisms In some situations a miroorganism is not ontained by the non speifi barriers. When this ours the speifi immune system works together with the non speifi proess to ontrol the invading miroorganism. The basi building blok of a speifi immune response is the lymphoyte whih is derived from white blood ells. There are two main groups of lymphoytes, the B group and the T group (Kuby, 1997). Lymphoytes atively identify and adapt to ontrol an enormous range of foreign invaders (antigens). The reation to eah antigen is highly speifi. Lymphoytes also maintain a long lasting memory of eah antigen to whih they are exposed. This memory allows muh faster, more potent and longer lasting ontrol of any antigen previously enountered. Lymphoytes are able to differentiate between self and non self (Kuby, 1997). The speifi immune mehanism is divided into two ategories, the ell mediated response and the humoral response. These areas are distint, although there is muh interation between the two. This is beause a loalised immune response an expand into a systemi response and usually involves ommuniation between different parts of the body (Klein and Horesji, 1997) Cell mediated immune response A ell mediated immune response uses T lymphoytes whih develop in the thymus. The T group omprises two subpopulations helper T ells and ytotoxi T ells. Helper T ells are the major antigen reognising ells in almost all speifi immune reations. Their funtion is to ativate the body s phagoyti ells when the appropriate antigeni material is enountered. They do this by sereting ytokines whih lead to the exitement of marophages, immunoglobulins and ytotoxi T ells (Kuby, 1997). Ativation of marophages is the major part of a ell mediated immune

42 26 response and onsists initially of their proliferation. The inreasing number of marophages also undergo hanges in their ellular properties inluding size, speed, energy onsumption, enzyme prodution and phagoytosis. This group of ativated phagoyti ells are muh more apable of antigen destrution. They remain loal to the site of the infetion while other marophages may also migrate to the site of infetion from the blood stream as a result of inflammation (Dannenberg, 1968). The magnitude of a ell mediated immune response is dependent on a number of fators, for example regional differenes within the organ system, age, speies and the virulene mehanisms of partiular pathogens (Wilkie, 1982). Cell mediated immunity is used to ontrol infetion with tuberulosis. The size of the ell mediated immune response of animals exposed to M. bovis an be measured by the lymphoyte stimulation assay (Buddle et al., 1994). The disease proess is overed in detail in the setion Humoral immune response A humoral immune response ats on antigen found in the blood stream, in ontrast to a ell mediated response whih ats at a speifi loation. A humoral immune response involves the B group of lymphoytes (antibodies). B lymphoytes develop in the bone marrow and further differentiate into immunoglobulins (Ig) mainly on the intestinal muosa and also in the bone marrow, spleen and lymph nodes (Brandtzaeg, 1989). There are five main lasses of immunoglobulin of whih IgG is the major type and found in the blood (Nugent and Lugton, 1995). IgA is found in muus of the respiratory trat, intestinal trat, tears, sweat, saliva and milk. The funtion of these immunoglobulins is the highly speifi reognition of antigeni material. On reognition of the appropriate antigen, the antibody binds to it, triggering events whih lead to destrution of the antigen (Villee and others, 1984), (Roitt, 1998) Immune response to myobateria and in partiular tuberulosis The respiratory trat is the most important route of initial infetion for tuberulosis, however as disease progresses lesions may also be found throughout most of the internal organs (Buddle et al., 1994;Cooke et al., 1995;MCool, 1979). Inhaled aerosols ontaining infetive bailli vary in size. Large aerosols (8? m or more) beome attahed to the muous membrane of the upper respiratory trat. Movement of muus will arry the partile to either the pharynx or larynx where it will be swallowed and destroyed in the aidi environment of the stomah (Hensel, 1996). Smaller aerosols generally do not beome aught on the muous membrane and penetrate further into the respiratory trat to the lung. The size best suited for maximum penetration is 4 5? m

43 27 (Walker and Stephen, 1977) and partiles of this size ontain a maximum of three bailli (Lurie (1964) ited in Dannenberg and Rook, (1994)). A proportion of these partiles will remain suspended in the air and be exhaled, but others may adhere to the lung surfae. Those adhering to the lung surfae will then be engulfed by alveolar marophages. While the ation of engulfing usually results in destrution of the invading organisms, in some ases the enzymes within the marophage fail to do this. When the marophage fails to destroy an infetive M. bovis, the baillus may then proliferate within the ell. This proliferation auses the ell to swell and eventually burst at whih stage many bailli are released to infet more marophages. The number of marophages infeted with bailli and tuberulin like produts (antigen) inrease logarithmially (see setion 2.3.2) for 2 3 weeks to the point where they beome an irritant to the body. Helper T ells reat to the irritant by ativating marophages, immunoglobulins and ytotoxi T ells (Kuby, 1997). The ativated marophages attempt to ontrol the disease by mounting a granulamatous response whih is the destrution of non ativated marophages whih ontain bailli. This leads to an aumulation of dead material known as aseous nerosis. If the ativated marophages are inapable of ontaining the disease within the resulting lesion, it spreads further in the loal tissues, lymph nodes and possibly on throughout the body. At eah new site of infetion the proess repeats itself (Dannenberg and Rook, 1994). The response to infetion with tuberulosis is loalised initially beause only marophages around the infeted site beome ativated to a high degree. Overall, the humoral immune system shows little reation to tuberulosis infetion, however, marophages in the humoral system an beome slightly more ative. This is thought to be a result of inflammation and small amounts of produt released into the irulation from loalised lesions (Makaness, 1964). The level of marophage ativation may lead to an immune state known as hypersensitivity. Hypersensitivity is the ondition ourring when the baillus is re-enountered and where lymphoytes and marophages are highly sensitive to the baillus as a result of the initial enounter. The ondition an be either detrimental or benefiial to a host, depending on the number of bailli presented in the seond enounter. If there are a high number, the vigorous response of hypersensitive ells an be seriously damaging to the host. The intensive destrution of infeted marophages extends to inlude destrution of healthy surrounding tissues. Inflammation due to this ativity ompounds the problem through redued loal irulation and, possibly, onentration of toxi produts released from dead and dying ells (Dannenberg, 1968). If the number of bailli is small, hypersensitivity auses the aumulation and multipliation of sensitised marophages. The effet of this is a stronger defene against bailli in the area loal to their presentation. Generally however, the number of inhaled bailli whih reah the alveolar

44 28 surfae is small, for example between one and three, and there is insuffiient tuberulin released to be toxi to marophages, but enough to eliit a protetive response (Dannenberg, 1968) VACCINATION Introdution Vaination is defined as the introdution of a prepared vaine into the body whih is strong enough to stimulate the reipient to make antibodies but not strong enough to result in the harmful effets of the disease. The aim of vaination is to produe a higher than natural state of immunity to a speifi disease (Villee et al., 1984). This immunity is referred to as either ative or passive. Ative immunity results from the ative prodution of antibodies by the subjet in response to a vaine. This type of immunity may be inreased by periodi boosters or exposure to the organism. Passive immunity is where the speifi antibodies are introdued diretly into the subjet, requiring no effort on the part of the reipient s body. This method produes immunity more quikly, though it is omparatively short lived, usually lasting only a ouple of weeks (Brand et al., 1971). This review will fous on the vaine used for protetion against tuberulosis, whih indues ative immunity. Vaines may be administered by a variety of methods, inluding orally or by inhalation of an aerosol, by subutaneous or intradermal injetion or infusion into the mammary gland (Blood and Studdert, 1988). The vaine usually ontains a similar but less virulent organism than the one against whih protetion is offered or ontains the disease organism in a harmless state, or ontains modified toxins of the organism. The purpose of vaination is to establish a state of memory immunity. This onsists of a long lived population of T lymphoytes that, on reognition of speifi antigens an mediate an aelerated reall of previously aquired resistane (Orme, 1999). The problem of how to eliit an effetive ell mediated response, as opposed to an antibody response, is still unsolved and is a major fous of urrent vaine researh (Salyers and Whitt, 1994). Some vaines an indue either ell mediated or antibody responses depending upon the dose of antigen administered, with low doses favouring the indution of ell mediated immunity. Resistane takes time to establish and it is therefore important that an immunising agent is slow growing and does not reah high levels before a protetive response has been established (Bretsher, 1995). To effetively protet a large population by vaination it is not neessary to vainate every individual. Instead, as the proteted proportion of the population inreases, a threshold is reahed whereby the number of suseptible hosts is too low for the disease to persist within the population

45 29 (Buddle et al., 1997). This onept is known as herd immunity and refers to a level of resistane in a population whih is suffiient to prevent the entry or persistene of a partiular disease (Blood and Studdert, 1988) Rabies: a positive ahievement using wildlife vaination The epidemiology of rabies (viral disease) in foxes (Vulpes vulpes) in Europe is similar to that of tuberulosis in possums in New Zealand. It generally moves in yles, drops to very low levels when host numbers are low and has a prevalene of 3 7% in areas where rabies is endemi. Vaination when repeated twie a year in spring and autumn for at least two years suessively was proven more effetive at ontrolling rabies than destrution of fox populations by shooting or gassing. Currently 13 ountries are involved in the vaination of foxes against rabies and from 1989 to 1995 rabies inidene has dereased in Frane by 99%. Complete elimination of rabies has been ahieved over large areas resulting in the vaination program being no longer required, for example Switzerland is now free from rabies as a result of vaination of foxes (Aubert, 1996), (erson et al., 1981) Vaination with BCG The vaine against tuberulosis was developed by serial passage of virulent M. bovis over ten years by Albert Calmette and Camille Guerin. In 1919, after 230 passages, the vaine was shown to be avirulent in guinea pigs, attle and horses. Today, the baille Calmette-Guerin (BCG) vaine is in widespread use throughout the world. However the degree of protetion afforded by this vaine is highly variable and has been the subjet of many studies. (Bloom, 1994) reports a olletion of these with a protetive effet ranging from 0 80%. The ause of this variation is the soure of muh debate and probably inludes methodologial differenes in studies, differenes between vaines, differenes in virulene between infetious strains of tuberulosis, BCG proteting against endogenous but not exogenous infetion and interferene with or masking of protetion by environmental myobaterial infetions. Two additional theories are given by Behr, firstly, that variability of protetive effet is due to the loss of a substantial number of genes during manipulation of the vaine over time. The seond possibility is that, by beoming austomed to laboratory ulture, BCG strains may have lost the ability to maintain a suitable infetion required to develop a strong and long lasting host immune response (Behr and Young, 1999). The aim of vaination with BCG is to protet the reipient against ontat with fully virulent tuberulosis. The proedure onsists of administering live, attenuated bailli whih invade host

46 30 marophages and repliate briefly before being killed (Salyers and Whitt, 1994). This repliation stimulates the seretion of IgA and IgG immunoglobulins (Wilkie, 1982) whih in turn funtion as part of an effetive ell mediated immune response and also memory ells whih will allow an aelerated response to the invader, should it ever be enountered again. Bailli must be alive and atively invade marophages to ensure suessful vaination. This is demonstrated by vaination of possums with killed M. vaae induing no protetion against a hallenge with virulent M. bovis when ompared with unvainated ontrols. In ontrast, vaination with live BCG resulted in fewer animals developing lesions and in a redution in number of lesions in diseased animals (Buddle et al., 1995). As attle have been the ornerstone of M. bovis persistene over time, it seems natural that a vaine strategy would primarily target this reservoir of infetion. Early trials evaluating the effiay of BCG began in the 1920s in attle and used omparatively high doses (1 x 10 7 to 1 x 10 9 olony forming units (fu)). Results were inonlusive though vaination did seem to redue the severity of disease (Franis, 1958). It was suggested that lower doses of BCG may preferentially stimulate the appropriate immune response for protetion against myobaterial infetions (Bretsher, 1992). This suggestion is supported by a study whih used subutaneous vaination with 1 x 10 4 to 1 x 10 6 fu of BCG in alves. Subjets were signifiantly proteted from the development of tuberulous lesions resulting from experimental hallenge (Buddle et al., 1995). However the situation in New Zealand onerning M. bovis and attle has additional onsiderations. Immunisation of attle with BCG results in some animals displaying positive skin test responses to bovine PPD (tuberulin). The PPD method is urrently used for identifiation of tuberulous attle in New Zealand, and the onfusion of results due to vaination has serious onsequenes. Primarily it limits the use of tuberulosis vaines in attle to those animals whih are not destined for the export market. The large sale vaination of attle in New Zealand would require international approval and be run in addition to ontinued wildlife ontrol programs. It would also require a guarantee that there were no risks to human health from vainated animals and aeptane by the various international regulatory bodies (Buddle et al., 1997). Regardless of the aeptane and suess of vaination of domesti livestok, it still leaves the problem of tuberulosis in New Zealand s wild life whih ensures domesti stok will remain permanently at risk and vaination ontinues indefinitely. Alternative options are required, in addition to urrent ontrol strategies, whih address the problem of a wildlife reservoir of tuberulosis, in partiular the possum. One of these options is vaination of possums with BCG.

47 Vaination of possums with BCG Several studies have observed the effets of vainating possums with BCG by a variety of routes. In general, results have shown the vaine to have a protetive effet (Buddle et al., 2000). An effetive vaine is the best option for the ontrol of the wildlife reservoir of tuberulosis represented by possums in New Zealand (Buddle et al., 2000). Vaination of possums with 4 x 10 6 fu of BCG markedly redues the severity of disease resulting from hallenge with virulent M. bovis. However the route of administration has a large effet on the level of protetion. The greatest redution in disease severity was observed when vaine was administered intratraheally or subutaneously in omparison with intragastrially vainated and non vainated animals (Aldwell et al., 1995). Effetiveness of protetion was measured by lung weight, lesion presene and type and also by lymphoyte blastogeni responses. A seond study ompared the intraduodenal (1 x 10 8 fu), intragastri (1 x 10 8 fu) and subutaneous (1 x 10 6 fu) routes of vaination in possums. Intraduodenal vaination provided the best protetion and all vainated animals showed better resistane to disease than non vainated ontrols. Protetion was measured by body weight hange after hallenge, hange in lung weight, numbers of aid fast bailli and lymphoyte blastogeni response (Buddle et al., 1997). A third study ompared vaination of possums with BCG by aerosol (4 x 10 6 fu), orally (3 x 10 8 fu) and subutaneously(1 x 10 6 fu). Results showed the aerosol and subutaneous routes to provide the highest levels of protetion as measured by minimal hange in body and lung weight following hallenge with virulent M. bovis. However lymphoyte blastogeni responses were low in aerosol and oral vainates ompared with subutaneous vainates. All routes of vaination redued the spread of disease resulting from hallenge (Aldwell et al., 1995). Several omputer models have investigated the possibility of eradiating tuberulosis from wild possum populations by vaination using the herd immunity onept (setion 5.1). One model indiates that protetion of 54% of the possum population is the threshold at whih the disease will be eradiated (Barlow, pers. om., 2000). A seond model laims maintenane of 40% of the population in an immune state will eradiate tuberulosis. It also suggests that the prevalene of tuberulosis in possums ould be redued to 10% of its pre-ontrol level within five years by vainating possums at a rate of 13% per year (Roberts, 1996). It is lear that vaination of possums by a number of routes an redue the level of disease resulting from infetion with M. bovis and that some methods appear more effetive than others.

48 32 However many additional fators must be onsidered when determining the method best suited to the large sale vaination of wild possum populations Oral vaination At present, the easiest method of administration would be orally; by baits beause parts of this system are already in plae as a result of urrent possum ontrol operations. These ontrol operations urrently manufature and distribute baits over large areas. However, this is an expensive and labour intensive exerise whih would probably have to be repeated at least annually (Tyndale-Bisoe, 1991). In addition, beause dominant possums may preferentially onsume attrative items of food, the appliation density of oral baits may need to be very high to ahieve adequate population overage. There are also potentially adverse effets of bait onsumption by non target speies, for example attle, deer and hildren. In addition, there is at present one serious biologial flaw in the onept of oral administration of BCG to possums. The oral route is ineffetive in possums beause of the stomah environment whih has a ph of 3 4 (Tyndale-Bisoe, 1973). This auses deativation of the bailli before reahing the small intestine, whih is their target for eliiting an immune response (Shwarting, 1948). Very high rates of ingestion would be required for suffiient vaine to remain viable after passing through the stomah. However, intraduodenal vaination of possums, whereby the stomah is bypassed, has been shown to provide a higher level of protetion than any other route (Buddle et al., 1997). This finding is enouragement to persevere with an oral bait. A potential solution to this problem would be to enapsulate the miroorganism in some way. For example, antigens have been administered with a solution of sodium biarbonate, or pakaged in gelatin apsules oated with substanes that are insoluble in aidi onditions (Mesteky and MGhee, 1989), releasing their ontents in the alkaline environment of the small intestine. The key to releasing the vaine from the apsule would be the hange to an alkaline ph. Unfortunately, the mouth is also a strongly alkaline environment, and would ause premature release of the vaine from its enapsulation. The release mehanism must be something unique to the small intestine. Prodution osts of a bait must also be low, beause of the volume required for distribution over very large areas and due to the strong possibility that bait distribution will need to be repeated every one to three years. Oral vaination has proven effetive in the past, in partiular against rabies in foxes in Europe and North Ameria (Artois et al., 1997), (Sanderson et al., 1981). However, failures have also been reorded, for example oral vaination of horses against the bateria ausing strangles failed to indue antibody prodution (Wallae et al., 1995).

49 Vetor transmission A self transmitting or vetor transmitted vaine is the seond potential method of vaination. The vetor would probably be some form of virus or parasite in whih genes from M. bovis have been inserted by geneti engineering. These genes would enode for antigens whih would be reognised by the host. There are two problems with this alternative, firstly, publi aeptane of this method would be diffiult to obtain on the grounds of the geneti engineering requirements (Bloom, 1994). The seond problem would be identifiation of a possum speifi vetor and ensuring that it ould not ross to other speies, partiularly humans. A third possibility is the use of an aerosolized BCG vaine targeting the respiratory trat of the possum. This avenue is disussed in more depth AEROSOL VACCINATION Introdution to aerosol vaination An aerosol is a solid or liquid partile, suspended in air or a gaseous environment (Blood and Studdert, 1988). Aerosols may ontain a wide range of small partiles inluding pathogeni miroorganisms. Experimental trials have demonstrated that mirobes an be sprayed from, or freeze dried as suspensions in solutions of all main lasses of water-soluble or water-ompatible materials inluding sugars, proteins, vitamins, dyes, faees and saliva (Cox and Wathes, 1995). As the understanding of airborne disease transmission grew, so did interest in the possibility of administering vaines by the same route (Middlebrook, 1961). The first reorded studies of aerosol vaination were for Newastle disease in hikens in 1952 (Hithner and Reising, 1952), followed by vaination against distemper virus in mink in 1954 (Gorham et al., 1954). Sine then many other speies and vaines have been investigated with respet to aerosol prophylaxis The aerosol vaine The onstituents of an aerosolized vaine and system by whih it is delivered are extremely important. They determine the volume and viability of vaine reahing the desired site and also the pattern of deposition. This in turn determines the degree of protetion afforded by the vaine (Buddle et al., 1997), (Jadin, 1980). There are three general onsiderations whih must be taken into aount when vaines are dispensed as aerosols. Firstly, the fundamental qualities of the ative agent within the vaine must be retained in an aerosolized form. Seondly, those vaines known to be allergeni by injetion should not be administered by aerosol. Thirdly agents ausing

50 34 hypersensitivity when injeted need to be thoroughly tested to ensure the effet is not magnified when administered by aerosol (Walker and Stephen, 1977). Organisms used in live vaines administered as aerosols are subjet to a redution in viability. This is due to the stresses of aerosolisation and exposure to the natural environment. The magnitude of this loss is a ritial fator when determining the volume of viable vaine required to generate a protetive response. In the ase of bateria, the situation is ompliated by the fat that standard ulture methods may not indiate true viability after aerosolisation. This is beause some environments into whih aerosolised bateria are introdued ause the bateria to enter a viable, but non-ulturable state (VBMC) (Roszak and Colwell, 1987;Heidelberg et al., 1997). Aspets of the delivery system must also be taken into aount, in partiular the volume of vaine to be administered. Spray volume is determined by the onentration of the drug in the spray suspension, air flow rates and the harateristis of the nebulizer. The fat that atomization inreases the onentration of a solution must also be onsidered (Walker and Stephen, 1977). The lung size and frequeny of breathing yle (respiratory minute volume) of the subjet is also an important issue when determining the amount of vaine whih must be inhaled to eliit a suitable immune response (Walker and Stephen, 1977). In any animal population there is a spetrum of immune responses indued by any form of vaination. Seletion of the most appropriate vaine type and method of administration will minimise the proportion of the population whih show little or no response Deposition Aerosol vaination with a live baterial vaine requires that the living miroorganism ontats a muosal surfae or the lung after passing into the respiratory trat. The proess of attahment involves interations between the miroorganism and host whih are primarily influened by loal immunity and the muous lining of the respiratory trat (Babiuk and Campos, 1993). While these immunmologial interations have already been desribed (setion 4.0), tehnially there are a number of fators that influene the way in whih aerosols are deposited within the respiratory trat. These are the physial qualities of the air ways, deposition fores and hygrosopiity Physial properties of air ways The anatomial arrangement and physial dimensions of the subjet s airways in addition to the breathing pattern, affet the rate of deposition by influening veloity of the air, times of transit of the air from plae to plae within the system and from moment to moment throughout the

51 35 breathing yle (Hath and Gross, 1964), (Padfield, 1987). From the trahea inwards, the respiratory airways are omposed of progressively branhing tubes of dereasing size and ross setional area. At first the airways are simply duts for air passage, but as their size dereases they also take on a respiratory role (Altiere and Thompson, 1996). There are a gradually inreasing number of tubes relative to depth within the respiratory trat resulting in an inreasing total ross setional area as depth into the lung inreases. The onsequene of inreasing ross setional area is a marked derease in the veloity of the air as it penetrates further into the lung. Even during gasping breaths, air enters the pulmonary air spaes with a maximum veloity of only a few entimetres per seond, resulting in essentially laminar air flow (Reid, 1973). The last quarter to one third of air volume inhaled during a single breathing yle (end tidal air) is unlikely to reah the pulmonary air spaes. As a result, fewer suspended partiles, espeially those 1 6? m in diameter, are deposited within the respiratory trat from this air ompared with air inspired during the beginning of inhalation (Hath and Gross, 1964) Deposition fores The shape of the airways influene the fores whih at on inhaled partiles. Bateria and viruses have been shown to exhibit the same aerodynami behaviour as other organi or inorgani partiles when airborne (Cox and Wathes, 1995). The prinipal fores involved are inertia resulting from the respiratory effort and sedimentation due to gravity and diffusion. Inertia and sedimentation are key fators when partiles are greater than 0.5? m in diameter, while diffusion beomes the dominant fore when aerosols are smaller than this (Knight, 1973;Martonen and Yadong, 1996). Fators affeting the inertia of an aerosol within the respiratory system inlude the size, density, onentration and shape of the aerosol partile. Deposition is also affeted by the solubility and hygrosopiity of a partile (Padfield, 1987). Where partiles are in the form of a spray, the onentration and veloity of the spray, in addition to possible propellant evaporation, will also influene the relative importane of the fores involved in partile deposition (Padfield, 1987) Hygrosopiity Hygrosopiity is the enlargement of small partiles (<6? m) due to their ontat with very humid (saturated) air. It beomes important very deep in the respiratory trat where air in the tertiary bronhioles and alveolar duts is saturated. The effet of hygrosopiity is to inrease the size and mass of small partiles. The fores of inertia and sedimentation inrease due to the inreased mass of the partile, thereby inreasing the rate of deposition (Knight, 1973), (Martonen and Katz,

52 ). The effet of hygrosopiity is relevant to an aerosol ontaining M. bovis as the baillus is approximately 5? m in length. Table 1. Deposition of 1.5? m hygrosopi partiles within the human respiratory trat (Knight, 1973) Area % of total air % deposition % deposition (hygrosopi) Nose * 25 Pharynx to seondary bronhi 10 1** 0 Tertiary bronhi to respiratory bronhioles 21 25** 10 Alveolar duts 63 21** 13 Total retained (nonhygrosopi) *24 % of 2?m partiles retained in inhalation: 12% of total inhaled partiles 4?m in diameter retained in exhalation due to aretion of water. **Retention as 4?m partiles Studies of aerosol deposition The patterns of deposition of aerosolized partiles have been examined in a number of studies and aross a range of speies. In general, the results of these studies have found that partiles with a diameter of 8? m or larger typially penetrate no further than the nasal passages and oropharynx. Optimum pulmonary deposition was ahieved when aerosol partiles were between 1.5 and 4? m (Walker and Stephen, 1977), (Wilkie, 1982). A partile size of 0.5? m has the lowest rate of deposition. Below this size, deposition inreases due to diffusion, while above it deposition inreases due to inertia and other fores mentioned above (see setion 6.3.2). During quiet breathing, the rate of deposition due to diffusion and gravity settlement are high, while during heavy breathing, a higher perentage of oarse partiles are deposited due to inreased inertia as a result of inreasing air veloity in the upper respiratory trat (Hath and Gross, 1964). Studies of aerosol deposition in hikens showed that partiles of 3 7? m diameter are aptured in the head and anterior trahea. As partile size dereased, there was an inrease in depth of partile penetration into the respiratory trat. Due to their tendeny to follow streamlines around obstales, rather than impating on them, partiles ? m were deposited primarily in the lungs (Hayter and Besh, 1974).

53 37 The deposition of aerosols in the human respiratory trat has also been thoroughly examined. Partiles greater than 6? m are typially trapped in the nose while those smaller than 2? m reah the lower respiratory trat and alveoli (Knight, 1973). Although larger partiles (10? m) are onsistently filtered in the nose aross a range of air-flow rates, inreasing flow results in higher levels of small partiles (1 2? m) also being filtered out (Hath and Gross, 1964). Most partiles smaller than 0.5? m remain in the tidal volume and are exhaled (Beeh, 1991). Exhaled end tidal air has been shown to be almost devoid of 6? m partiles and ontain less than half the inspired level of 1? m partiles (Hath and Gross, 1964), indiating high rates of deposition. A high rate of partile deposition has also been shown in guinea pigs, where one in every three or four inhaled infetive droplet nulei bearing a single baillus (approximately 5? m in size) would reah a suseptible lous (Middlebrook, 1961). In rats, the response to deposited partiles has been shown to vary between nasal, traheal, bronhial and bronhoalveolar ompartments of the respiratory trat (Hensel, 1996). In terms of an aerosol vaine, if a partiular site within the respiratory trat is shown to provide an inreased immune response, harateristis of the aerosol delivery ould be manipulated to ensure the maximum volume of vaine is direted at the appropriate site (Wilkie, 1982) Survival of aerosolised miro-organisms. The survival rate of any miroorganism, inluding those used as a live vaine, is a prerequisite for infetivity (Hensel, 1994). Survival rate in the natural environment is adversely affeted by a range of environmental fators. The magnitude of the effet varies depending on the organism and beomes greater with dereasing aerosol size. These fators are olletively known as the open air fator (OAF) and defined as the loss of viability due to exposure to open air (Druett, 1973). For most miroorganisms the effets inlude desiation, exposure to radiation, pollutants and even oxygen, whih has been shown to inativate some gram negative bateria (Cox and Wathes, 1995). A number of studies have measured the open air fator. (Druett, 1973) found that 1? m partiles of E. oli suffered a 10% loss of viability per minute under normal humidity onditions (60 100%) at night. Survival of Pasteurella multoida was found to be 80% after one minute through a range of humidity levels (Thomson et al., 1992). Under optimum onditions studied (55% relative humidity and 4 degrees elsius), the infetivity of pseudorabies virus in an aerosol was shown to be redued by 50% in less than one hour (Shoenbaum et al., 1990). Rotavirus demonstrated a similar pattern (Sattar (1984) ited in Shoenbaum et al., (1990)).

54 38 In ontrast to the above findings, other studies have found that the OAF has minimal impat on viability. For example, Beard and Sanderson (1967) found that two viable strains of myoplasma (M. galliseptium and M. meleagridis) ould be reovered from artifiially generated aerosols up to 24 hours after suspension. Additionally, in an investigation of the viability and distane traveled by aerosols from slurry spreading equipment, no onsistent assoiation was found between the infetivity of viable Serratia rubidaea (bateria) and relative humidity, air temperature or sunlight. (Hahesy et al., 1995). Survival and retention of antigeni properties after aerosolisation has also been shown in Streptoous suis (bateria) (Brown et al., 1997). Diret ounts and diret viable ounts of three other types of bateria suspended as aerosols (Serratia maresens, Klebsiella plantiola and Cytophaga allerginae) indiate all three remained viable for at least four hours after aerosolisation (Heidelberg et al., 1997). These findings indiate that there are a number of organisms apable of surviving the stresses of nebulisation, suspension in airborne partiles and impingement on a liquid filled air sampler. It is possible to manipulate the survival harateristis of aerosolized miroorganisms to some extent. Addition of serum or gluose, for example, inreased the survival rate of Atinobaillus pleuropneumoniae. Due to the inreased visosity resulting from these additives, it also enlarged aerosol size from approximately 1? m to 4? m. Manipulation of temperature and relative humidity has been found to influene the survival of pseudorabies virus (Heidelberg et al., 1997) Aerosol vaination in pratie Aerosol vaination has undergone onsiderable development sine the earliest trials in Today, it is being used ommerially in several fields, primarily against viral disease in ommerial poultry prodution and in the pork industry Poultry Aerosol therapy is well suited to intensive poultry systems with high densities of animals in a onfined, regulated environment. There are many studies investigating the effet of aerosol vaination of poultry against Newastle disease aused by a viral infetion. Full protetion was ahieved by (Hungerford, 1969), ompared with mortality of 10 30% in hikens vainated by spray or drinking water (Latif et al., 1981;Ibrahim et al., 1981). Ek (1990) found aerosol exposure twie weekly for two or three weeks gave the best protetive effet.

55 39 Aerosol protetion against the viruses ausing fowlpox (Deuter et al., 1991), fowl paralysis (Pridybailo et al., 1986), and infetious laryngotraheitis (Redmann et al., 1983), (Hilbink et al., 187) have also been investigated with enouraging results. Aerosol vaination of duklings against the hepatitis virus produed a protetive response (Balla and Veress, 1984). Immunity to Myoplasma galliseptium (myoplasma) infetion after aerosol vaination was signifiantly better than vaination by intratraheal injetion or intranasal instillation, and lasted for at least three months (Hayatsu et al., 1974). (Lin and Kleven, 1984) found aerosol vaination of hikens to be more effetive than eye drop administration for M. galliseptium although the differene was not signifiant. Three appliations of aerosol vaine against Pasteurella multoida (bateria) proteted 70% of turkeys against hallenge with virulent holera, ompared with 7% of ontrols. One aerosol vaination followed by one injetion provided similar levels of protetion (Mihael et al., 1986). Aerosol vaination against the bateria ausing listeriosis was found to be effetive when the immunising dose exeeded 5 x 10 9 bateria and was repeated after ten days (Eliseeva, 1976). The stress effet on the respiratory trat in hikens was not influened by the size of the aerosol aross a mean partile size range of ? m (Allan and Borl, 1980) Pigs Pigs were proteted against the virus ausing lassial swine fever (Pestivirus) by aerosol vaination with either lapinised C vaine or Celviva vaine (Popa et al., 1982). In ontrast, aerosol and intramusular vaination against Myoplasma hyopneumoniae failed to protet pigs against intratraheal hallenge (Murphy et al., 1993). Petzoldt onduted a study in whih pigs were vainated against the bateria ausing Erysipelas (Erysipelothrix rhusiopathiae) using a baterial onentration of 1 x 10 9 /ml. Subjets were exposed for 15 minutes to aerosols of ? m diameter, resulting in the inhalation of 0.13 g of vaine. Testing immunity two weeks after vaination resulted in the death of ontrols, while vainated animals survived the hallenge of 100LD 50 (Petzoldt et al., 1980). Six aerosol vainations proteted pigs against the bateria Atinobaillus pleuropneumoniae when administered to the respiratory or gut muosae. In omparison, six oral vainations offered poor protetion. The protetive effet of five oral vainations was improved when ombined with a single dose by aerosol (Nielsen et al., 1990), (Loftager et al., 1993).

56 40 A metered dose, propellant driven appliator was designed for delivering Streptoous suis into the respiratory trat of pigs. The devie was designed to disharge during the inhalation part of the breathing yle. It would generate respirable size aerosols (5? m) at a onentration of 40 mg/ml and baterial densities of 0.2, 0.4 and 1.0 x /ml. The inrease in both total released bateria and respirable bateria were less than proportional to inreases in the baterial densities within the propellant. Testing showed the respirable perentage of bateria to range from 47% to 67%. This indiates that a large fration of the baterial aerosols were lumped into sizes greater than 12? m. Those whih did not lump were predominantly in the 1 5? m range. Additional testing showed that aerosolized bateria retained their antigeni properties. When using a longer nose one, whih ated as an aerosol expansion hamber, better penetration was ahieved. The use of a nose one has also been shown to improve aerosol penetration in humans (Brown et al., 1997) Cattle (Mann et al., 1983) found that aerosol vaination with an attenuated strain of virus ausing IBR/IPV (infetious bovine rhinotraheitis and infetious pustular vulvovaginitis) proteted alves experimental infetion. In ontrast, aerosol vaination of alves with Pasteurella haemolytia in (Jeriho and Langford, 1982) study failed to provide protetion against experimental respiratory disease Other animals Foals (Equs equs) vainated by aerosol against viral respiratory infetions showed higher levels of protetion ompared with a subutaneous injetion of the same vaine(zabegina et al., 1999). Vaination by both aerosol delivery and intra musular injetion signifiantly redued mortality in mink (Mustela vison) due to the virus ausing distemper (Asztalos et al., 1983). (Sott and Glauberg, 1975) found that domesti ats ould be proteted against feline panleukopenia virus (FPL) by aerosol vaination. Aerosol immunisation of rats before hallenge redued the rate of retention of erysipelas in the respiratory trat. It also aused aeleration of the learane mehanisms in the upper respiratory trat (Hensel, 1996). Although aerosol vaination of mie against Aujeszky s virus provided less protetion than intra musular vaination, it still redued disease when ompared to ontrols (Neukirh and Bauer, 1977). Small aerosols (2.3? m) were used to vainate the upper respiratory trat of hamsters against infetion with Myoplasma hyopneumoniae. As well as providing resistane to subsequent

57 41 hallenge, aerosol vaination was found to more effetively ontain development of the disease when ompared to ontrols (Jemski et al., 1977) Aerosol vaination with BCG Aerosol vaination with BCG has been shown to indue strong protetive responses in a range of speies inluding monkeys (Ribi et al., 1971), mie (Orme and Collins, 1986), guinea pigs and humans (Lagranderie et al., 1996). Protetion with BCG by aerosol vaination is thought to be due to indution of ell mediated immunity in the lungs, whih results in the early ativation of alveolar marophages. The importane of improved loal ativation of marophages in tuberulosis infetion was earlier emphasised by Dannenberg (1968). This ativation is a response to the produts of bailli and produts of sensitised lymphoytes being at a higher onentration in the environment loal to infetion ompared to that found systemially (Klein and Horesji, 1997). (Middlebrook, 1961) found that inhalation of very small numbers of viable BCG organisms by guinea pigs, either as single ells or as lumps an result in a protetive response. In this study guinea pigs were vainated by aerosol with dose rates of 2.5 x 10 6, 2.5 x 10 5 and 2.5 x 10 4 organisms. After 40 days, all animals in the treatment groups whih reeived either high (2.5 x 10 6 ) or medium (2.5 x 10 5 ) doses showed a positive response to tuberulin, while only four of the six subjets in the low dose (2.5 x 10 4 ) group were found to be tuberulin positive. All ontrols returned negative responses. Seventy days after hallenge with 100 infetive units, high dose rate subjets averaged baterial ounts of 0.5 x 10 2 in the lung and 3 x 10 1 in the spleen; medium dose subjets returned ounts of 1.7 x 10 2 (lung) and 6.0 x 10 1 (spleen). The ontrol group, by omparison, had baterial ounts of 1.2 x 10 4 (lung) and 1.7 x 10 4 (spleen). In the same study, aerosol and sub utaneous vaination was ompared. The aerosol vaine ontained only 20 units of BCG. Twenty eight days after hallenge with 200 viable ells, baterial ounts in the lung and spleen of guinea pigs proteted by aerosol were 5 x 10 2 and 0.2 x 10 2 respetively. These were signifiantly lower ounts than those obtained from the subutaneously vainated subjets, whih had baterial ounts of 30 x 10 2 (lung) and 1.5 x 10 2 (spleen). Immunity resulting from aerosol vaination was found to persist for at least two years (Middlebrook, 1961). In ontrast, two experiments using aerosolized BCG to vainate rabbits and three to vainate mie failed to show a protetive response (Middlebrook, 1961). Results from an aerosol vaination trial in guinea pigs by (Lagranderie et al., 1993) suggest that loal immunization may prove superior to systemi immunization. It also showed that aerosol

58 42 vaination with BCG ould restrit the growth of virulent bailli at the sites of their implantation in the lungs. Similar results were ahieved by (Gheorghiu, 1994) who demonstrated higher ativation of bronho-alveolar marophages in animals vainated by aerosol in omparison with those vainated intradermally. The possible adverse side effets of high doses of BCG when administered by aerosol have been investigated in guinea pigs. Large doses of BCG (5 x 10 6 organisms) did not result in the suppuration of the traheobronhial nodes, or in enlargement of the ervial and hilar lymph nodes. The onlusion was that there are no adverse side effets to the lung or its funtion (Lagranderie et al., 1993). A suitable dispenser and method of storage for aerosolised BCG vaine is a major issue whih would need to be addressed before any large sale aerosol vaination program ould be onsidered. Vaine dispensers used in the ited studies were ompliated, expensive mahines, for example the Airborne infetion apparatus (Middlebrook, 1961), Turbair Vaanair apparatus (Latif et al., 1981) and Ultrasoni nebulizers (Lagranderie et al., 1993). These are all impratial for any form of ommerial use Aerosol vaination of possums with BCG There is limited information available on the effets of aerosol vaination with BCG in possums. The most informative study to date is by (Aldwell et al., 1995) who ompares the intranasal (aerosol), oral and subutaneous routes of vaination. Aerosol vaination used a dose of 4 x 10 6 fu, administered with a simple hand held atomiser nasal pump. The oral dose measured 3 x 10 8 fu and the subutaneous dose rate was 1 x 10 6 fu. All routes of administration markedly redued the severity of disease resulting from hallenge with 400 fu of virulent M. bovis. However the protetive effet varied between these routes. Protetion was measured by weight loss of subjets between hallenge and neropsy. Controls, subutaneous and oral vainates all lost an average of g. In ontrast the aerosol group inreased in weight by an average of 60 g. Protetion was also measured by granulomas in the liver. When measured in this way, aerosol vaination provided the best protetion. When measured by frequeny of granuloma in the spleen, subutaneous vaination was most effetive, followed by the oral and aerosol route and finally non vainates. Overall, the protetive effet of vaine when administered by aerosol was similar to that when administered by the subutaneous route. In addition, Aldwell notes that aerosol vaination with BCG may provide more effetive protetion against a field exposure to M. bovis by inhalation, than against an experimentally indued intratraheal hallenge. The results of the above study are supported by work in guinea pigs (Lagranderie et al., 1993;Middlebrook, 1961).

59 43 The dose of BCG administered by aerosol is diffiult to aurately ontrol. However, this may not be ritial in the ase of the possum as low doses of BCG administered by an appropriate route (sub utaneous or respiratory trat) have indued protetive responses in other animal speies (Buddle et al., 1995) CONCLUSION Tuberulosis aused by the baterium Myobaterium bovis was probably introdued into New Zealand during importation of attle in the nineteenth entury. Over the following thirty years it beame a serious publi health problem, with infetion in humans ourring mainly through the onsumption of milk ontaminated with M. bovis. Today tuberulosis is found in several speies of wildlife and a small perentage of domesti attle and deer herds. However, numerous vetor ontrol operations, herd testing and movement restritions on infeted herds have failed to stop the geographi spread of this disease. The brushtail possum (Trihosurus vulpeula) was introdued into New Zealand in the late 1800s to initiate a fur trade. The possum is a very adaptable animal and quikly spread throughout 90% of the ountry, despite various ontrol efforts. In 2000 the possum was regarded as the nation's major pest speies with a population estimate of 70 million. Feeding habits of the possum ause serious damage to native flora and fauna. The first possum infeted with tuberulosis was found in 1969 and today infeted possums are found over one third of New Zealand. Sine this first disovery, it has beome lear that the possum funtions as a wildlife reservoir for tuberulosis, ating as a soure of re-infetion into domesti attle and deer herds and greatly frustrating disease eradiation efforts. Tuberulous possums are also a soure of infetion for several other wildlife speies. Traditional methods of possum ontrol are based on the large sale aerial distribution of oral baits ontaining poison (ompound 1080), followed up with ground based poison and trapping operations. These methods have failed to effetively ontrol either the spread of the possum or the spread of tuberulosis. There is also inreasing publi onern over the detrimental effets of large sale poisoning operations on the environment and non target speies. An alternative method for ontrol whih may greatly assist in the ontrol and eradiation of tuberulosis is vaination of the wild possum population with BCG vaine. The vaination of wild foxes against rabies using oral baits has been suessful in Europe. Possums respond well to vaination with BCG by developing an improved immune response to virulent M. bovis. If

60 44 suessful, suh a method may be adapted for use in Great Britain and Ireland where a similar problem exists with the badger (Meles meles) funtioning as a wildlife reservoir for tuberulosis. Currently, there are three major hallenges faing the development of an effetive vaination program. Firstly the viability of a live vaine must be further researhed. Seondly, a ommerial produt to ontain BCG and dispense it as an aerosol must be developed. Thirdly, an effiient and ost effetive method of vaine delivery must be devised. The third hallenge is addressed in hapter three.

61 45 Chapter 2 A BEHAVIOURAL STUDY OF BRUSHTAIL POSSUMS (TRICHOSURUS VULPECULA) WITH CLINICAL TUBERCULOSIS (MYCOBACTERIUM BOVIS) A naturally formed hollow in a lay bank, sheltered by gorse and flax, forms a typial possum den. Den number, possum identifiation number and dates of use are reorded on the yellow tag.

62 46

63 INTRODUCTION The brushtail possum (Trihosurus vulpeula) was introdued to New Zealand to start a fur trade during the late 1800s. This highly adaptable animal rapidly spread throughout the ountry and by the 1960s the population had reahed natural peaks in most areas. An estimated 70 million possums now inhabit virtually the length and breadth of New Zealand (Montague, 2000). Bovine tuberulosis, aused by Myobaterium bovis, was probably introdued to New Zealand in about 1840 with imported attle. In 1999/2000 about two perent of attle and deer herds were infeted with the disease (Animal Health Board Annual Report, 2000). Tuberulosis was first disovered in the wild possum population in 1967 (ONeil and Pharo, 1995) and it is now onsidered endemi in this speies over approximately one quarter of New Zealand. In New Zealand the test and slaughter sheme has failed to eradiate the disease from livestok in areas with tuberulous possums. Possum ontrol operations have greatly redued the level of disease in livestok in the same area. It is now widely aepted that the wild possum population is a reservoir for the disease, ating as a soure of infetion to livestok and other wildlife speies. Wildlife reservoirs of tuberulosis have also been identified in other ountries inluding the badger (Meles meles) in Ireland and Great Britain, the Swamp buffalo (Bubalus bubalis) in Australia and the white tailed deer (Odooileus virginianus) in the USA. The route by whih disease is transmitted from an infeted possum to other possums and livestok is not well understood. Probable routes of transmission inlude by aerosol or ingestion, though in these speies the pattern of disease indiates aerosol transmission is predominant (Jakson et al., 1995, Costello et al., 1998). The tuberulous possum is most infetious during the terminal stages of disease (Jakson et al., 1995). Possums that are terminally ill with tuberulosis beome ative in daylight, are disoriented, unoordinated and do not show normal avoidane reations (Paterson and Morris, 1995). Both attle and deer have been shown to atively investigate possums behaving in this way (Sauter and Morris, 1995). The habitat where transmission most frequently ours has been the subjet of some debate. Paterson et al., (1995) found that tuberulous possums were more likely to transmit the disease to other possums and also domesti stok when in the viinity of their denning range in srub. Most possums den in small areas and tuberulous possums remained loser to these areas than healthy possums. Denning areas of tuberulous possums are high risk loations for disease transmission.

64 48 Pasture is another possible loation for transmission of disease from possums to livestok. This habitat is an important food soure for possums where adjaent to bush areas (Green and Coleman, 1986). Interations between possums feeding on pasture are infrequent and of low intensity (Paterson et al., 1995). Pasture is distint from other grass areas found within bush areas. It is a speifi mix of nutritious speies of high palatability to livestok. It is ommon for pasture to reeive artifiial fertilisation and be managed to maintain feed palatability and a high growth rate. Management of pasture is distint from that of grass areas found under srub and in learings within the bush. This observational study investigated the denning and ranging behaviour of possums from the development of linial signs of tuberulosis until death. The aim of the study was to identify aspets of behaviour of tuberulous possums that may influene transmission of the disease to livestok or healthy possums. The study site was first used to study possum populations with endemi tuberulosis in 1989 as reported by Pfeiffer, (1994). The longitudinal study initiated by Pfeiffer ontinued until MATERIALS AND METHODS Study site The study was arried out between Marh 1998 and February The study site was loated near Castlepoint (40? 51 S, 176? 14 E) in the Wairarapa on the east oast of the North Island, New Zealand. It onsisted of 56 ha of mainly dense manuka (Leptospermum soparium) and gorse (Ulex europeus) with pokets of flax (Phormium tenax), broadleaf forest remnants and groups of ponga (Cyathea dealbata). There were also areas of open savannah woodland and pasture. Elevation varied from 60 to 270 metres above sea level. The site was drained by four water ways whih were dry during the summer.

65 49 Figure 1. Aerial photograph of approximately two thirds of the Castlepoint study site On the site were 450 age traps set in fixed loations and divided into three trap lines (appendix 1.1). They were used in a longitudinal study of the resident possum population based on a apture-tag-reapture program. Traps were set at bimonthly intervals for three onseutive nights and baited with slies of apple dusted with innamon. At eah trapping session all trapped possums were sedated with 100 mg of ketamine (Parnell Laboratories New Zealand Ltd., New Zealand) and weighed. Age was estimated by the degree of wear of the upper first molar on a sale of 1 (unused, harateristi of a dependent joey) to 7 (worn ompletely flat, harateristi of

66 50 old age). Condition sore was reorded on a sale of 1 (emaiated) to 5 (fat) and for mature females their reprodutive status determined from an examination of their pouh and mammary glands. The trap number where the possum was trapped was also reorded Depopulation of the study site The longitudinal study at Castlepoint began in All possums on the site were depopulated in Marh 2000 over four weeks of intensive trapping. During the depopulation, all possums, inluding the radio ollared possums, were aptured in live traps or leg hold traps. Possums were euthanased by intraperitoneal injetion of 1.5 ml sodium pentobarbitone (Pentobarb 300, 300 mg/ml, South Island Chemials Ltd., Christhurh, New Zealand) and were subjet to an examination for marosopi lesions post mortem, bateriologial examination for M. bovis and histopathology Radio ollars Radio ollars were fitted to thirty possums during the study. These possums onstituted three groups: naturally infeted with tuberulosis (14 possums), experimentally infeted (eight possums) with tuberulosis, and healthy possums (eight possums) Naturally infeted possums Tuberulosis in naturally infeted possums was diagnosed by palpation for swelling of the mandibular, deep axillary, superfiial axillary and inguinal lymph nodes during the bimonthly trapping. Isolation of M. bovis from ulture of swabs taken from draining sinuses or from aspirates of swollen nodes was used to onfirm the diagnosis. Radio ollars were fitted when linial disease was deteted Healthy possums Radio ollars were applied to healthy possums to obtain ontrol data. Healthy possums were mathed on three harateristis with naturally tuberulous possums where possible. These harateristis were being resident on the same part of the site and mathed for age and sex where possible Experimentally infeted possums During the study experimentally infeted possums were released on three oasions. On eah of these oasions one mature male was seleted from eah trap line. They were anaesthetised with

67 51 12 mg of Saffan (Pet Elite Ltd, Lower Hutt, New Zealand) and infeted with virulent M. bovis via an intra-traheal annula into the lungs (Pfeffer et al 1995). A radio ollar was applied to eah possum before they were released bak on to the study site. Experimentally infeted possums were released in August 1998 (3 possums), February 1999 (3 possums) and August 1999 (2 possums) Possum behaviour The behaviour of eah possum was subjetively assessed as it left the den site during eah radio traking event. Aspets of the behaviour of interest inluded the ease and speed with whih the possum moved and the oordination of movements. Behaviour was used to judge whether a sik possum should be euthanased. Moribund animals were aught and euthanased by intraperitoneal injetion of 1.5 ml sodium pentobarbitone. Possums found dead or euthanased were reovered for post mortem examination Den harateristis Fourteen speifi attributes of eah den and den site were reorded on a speifially developed form (appendix III). These inluded height above ground level, degree of exposure of the den to wind, roof material, floor material, the number and size of entranes, and the predominant vegetation type. The ondition of the floor was divided into four ategories ranging from dry to saturated. The slope and aspet of the immediate area were also ategorised. Eah den was permanently identified using a plasti attle ear tag. Den number, date and the identity of the possum were reorded on the tag. The weather onditions, based on three ategories of temperature and preipitation, were reorded for the day, previous night and previous day. The study site was divided into three areas representing areas of high, medium and low density of den sites (appendix 1.2). These areas were used to evaluate the effet of den site density on the denning range of individual possums Radio traking proedure Radio equipment Radio ollars were two stage units that transmitted 60 pulses per minute, powered by volt battery, had an external aerial and weighed 25 grams (SirTrak, Havelok North, New Zealand). The reeption range of the transmitters was approximately 10km within a line of sight although in the rugged terrain of the study site this range was approximately 1km. Battery life span was

68 52 approximately 15 months. Traking equipment onsisted of a Merlin 12 reeiver and yagi aerial as detailed by Paterson (1993). Figure 2. Radio traking equipment Radio traking method Possums were traked to their den site weekly. Eah den site was examined visually and the harateristis listed above (setion 2.2.5) reorded. The den site was related to the nearest trap by distane (metres) and ompass bearing. Den sites outside the boundaries of the study site were inluded Trapping data The data on where possums were trapped at the bimonthly trapping events was used to supplement the denning data for eah radio traked possum. The data onsisted of the trap number and date. Data was olleted between November 1994 and February 2000.

69 Geographial position data The oordinates (latitude and longitude) of all trap positions, the perimeter of the study site, the major water ourse and fene lines were determined using Trimble Global Positioning System (GPS) equipment (Trimble Navigation Limited, Mapping and GIS Sytems, Sunnyvale, CA, USA). The equipment onsists of a Base Station (satellite detetor, GPS omputer, radio and antenna) and a roving Asset Surveyor (satellite detetor, GPS omputer, radio and antenna). The oordinates for the base station were determined by averaging 100 readings for latitude, longitude and altitude. The oordinates of eah point were determined using the rover unit. Point oordinates were generated by ombining the position of the rover unit with the known fixed position of the base station and satellite information through differential proessing. By using this method, points ould be defined with more speed and auray than with the rover unit alone. Where overhead vegetation prevented a trap position being determined with the GPS equipment, the distane and ompass bearing from the nearest trap with a known position was reorded, and the oordinates plotted using a omputer program. The software pakage Pathfinder Offie GeII (Trimble Navigation Limited) was used to down load data from the rover unit and also provided viewing and basi analytial tools. With the traps visible in Pathfinder and units of deimal degrees, the ompass bearing and measuring tool were used to manually generate a table of den site oordinates Data analysis Mirosoft Aess (Mirosoft, Aess97) was used for data storage and manipulation. ArView, version 3.1 ( 1996 Environmental Systems Researh Institute, In.) was used to visualise the data and alulate possum ranges using the extension spatialtools. This extension was obtained from the Alaska Biologial Siene Centre ( Animation of the movement between den sites was made possible with the extension AnimalMovement2 also obtained from this site, allowing the pattern of den use over time to be visualised. The sript Coordinate Preision was used to identify point oordinates with maximum auray (eight deimal plaes) in ArView The kernel density estimator The denning range, foraging range and ativity range for eah possum and tuberulosis hotspots (setion ) were defined using a kernel density estimator in ArView. This tehnique

70 54 provided a visual representation of the range for both individuals and groups of possums. It also took into aount the frequeny of den or trap use. Contours were used to divide the range of a possum into four areas, based on intensity of use. Range size was represented by the area within the outermost ontour (area ontaining 80% of possum loations). The innermost ontour enloses the area used with highest intensity. This area ontains the entral most 20% of possum loations. Eah suessive outer ontour (40%, 60% and 80%) represents a progressively wider geographi spread and lower density of loations. The size of the range was alulated using the onvex polygon funtion in ArView Definition of denning range The denning range was defined as the area that ontained 80% of all denning events for an individual possum. The denning range of some possums onsisted of two or more distint areas and in these ases the overall size was alulated by summing the individual areas Definition of a foraging range The foraging range was defined as the area that ontained 80% of the trapping events for an individual possum Definition of ativity range The ativity range was defined as the area that ontained 80% of ombined denning events and trapping events for an individual possum. The onept of ativity was derived from the ore range estimate as used by Brokie et al. (1987) Definition of a total range The total range was the area that ontained all denning events and trapping events for an individual possum and was derived from the estimate used to identify long forays by Brokie et al. (1987). The perimeter of this range was drawn to pass halfway between the outermost trap where a possum was aught and the next trap beyond it, away from the entre of the area. Where there was no trap beyond the outermost trap, the perimeter of the area was arbitrarily set at 50 metres beyond the outermost used trap Definition of a long distane foray Long distane forays were identified by both denning and trap data. A long distane foray was defined as a movement, of a distane at least three times the radius of the ativity range, from

71 55 whih the possum returns to its established ativity range. The minimum distane of a long foray equates to 200 m. These forays do not inlude dispersing movements as only adult possums were studied Definition of a hotspot There were two types of hotspots. Major hotspots onsisted of a luster of dens used by one or more tuberulous possums and were defined as the area in whih the likelihood of finding a tuberulous possum in a den was?60%. Minor hotspots were smaller lusters of dens and defined as areas where the likelihood of finding a tuberulous possum in a den was?20% and?60% Statistial analysis All statistial analyses were done using SPSS 8 (opyright, SPSS In., 1999). Distributions were heked for normality using the onfidene interval tehnique ited in Cramer (1998). The Mann- Whitney U test was used to ompare the size of possum ranges. Pearson s orrelation was used to examine orrelation between number of observations and range sizes. The duration of survival of eah possum group was ompared using the Kaplan-Maier survival urve. Graphs were produed in Exel (Mirosoft, Redmond, USA) RESULTS Denning and survival data was olleted for 30 possums: 14 naturally infeted, eight experimentally infeted and eight healthy possums. Radio ollars were applied to 16 possums in 1998 and 14 in Radio traking results During the study there were a total of 315 traking events of whih 204 were of tuberulous possums and 111 of healthy possums. A total of 233 dens were identified and on 63 oasions possums were using previously identified den sites. There was onsiderable variation in the denning behaviour of individual possums. On average eah possum was traked to a den 11 times, with a range from one to 56. On one oasion a possum moved before the den site ould be identified. Of the 22 tuberulous possums traked, the arasses of 17 were found.

72 56 Table 2. Survival duration of naturally infeted, experimentally infeted and healthy possums (weeks) Possum group n 1 Median Range Naturally tuberulous Experimentally infeted with tuberulosis Healthy n = number of possums in eah group A Kaplan-Maier survival plot (Figure 3) shows the variability in life span after radio traking began. The median survival time for healthy possums was 26 weeks, for naturally infeted possums 11 weeks and for experimentally infeted possums 10 weeks Cumulative survival Duration (weeks) Figure 3. Survival of healthy, naturally infeted and experimentally infeted tuberulous possums, solid line = healthy possums, oarse dashed line = possums naturally infeted with tuberulosis, fine dashed line = possums experimentally infeted with tuberulosis

73 Survival duration of radio traked possums The measured duration of survival within the study period for individual possums ranged from one week to 84 weeks. The distribution of survival periods for all possums is shown in Figure 4. The distribution of survival periods was skewed to the left. The median survival period for naturally tuberulous possums after first being diagnosed with linial tuberulosis was 11 weeks (Table 2), one week longer than for experimentally infeted possums. The survival of naturally tuberulous possums was signifiantly shorter than healthy possums (26 weeks) (Mann-Whitney U 2.35, z=-2.223, P=0.024). The survival of experimentally infeted possums was also shorter than healthy possums (Mann-Whitney U=8.0, Z=-2.534, P=0.010). The shortest survival time was 1 week (Possum 5644, first identified in the terminal stages of disease) and the longest survival was 84 weeks (Possum 5787). Possum 5787 was found with tuberulous lesions in August 1997, and was fitted with a radio ollar in Marh 1998, at the beginning of the study and euthanased during the depopulation. Infetion was onfirmed when M. bovis was isolated from aspirates on two separate oasions but at post mortem examination no lesions were deteted and M. bovis was not isolated from tissue samples olleted Survival duration Possum ID Figure 4. Distribution of survival duration of healthy, naturally infeted and experimentally infeted tuberulous possums. Blue bars = possums naturally infeted with tuberulosis, blak bars = possums experimentally infeted with tuberulosis, light bars = healthy possums

74 Cause of death of radio traked possums Possums died from a variety of auses (Table 3). Most died of tuberulosis, either naturally or were euthanased in extremis. A neighbouring farmer trapped four possums within 200 m of the northern boundary of the study site. One tuberulous possum was lost due to failure of the radio transmitter in the ollar. The arasses of 24 possums were examined post mortem inluding 15 tuberulous possums (Table 3). The arasses of three possums were very deomposed and were not reovered. One tuberulous female found dead had a small joey, approximately six months old. Gross lesions of tuberulosis were found in the joey at post mortem and were indiative of pseudo-vertial transmission. The radio ollar on one healthy possum aused an absess to form and was removed after five months. This possum was lassified as lost to follow up. Due to the failure of a transmitter on one radio ollar, possum 5885 was also lost to follow up. Table 3. Cause of death of radio traked possums Cause of death N 1 Tuberulous possums Healthy possums Post mortem examination Died due to tuberulosis Euthanased in extremis Euthanased during depopulation Trapped by neighbour Killed by bulldozer Died during heart bleeding Lost to follow up N/A Total n = number of possums

75 59 The distribution of deaths by month for 24 radio traked possums are shown in Figure 5. The death of experimentally infeted possums aused the peaks in Otober and November 1998 and April and Otober Apr-98 May-98 Jun-98 Jul-98 Aug-98 Sep-98 Ot-98 Nov-98 De-98 Jan-99 Feb-99 Mar-99 Apr-99 May-99 Jun-99 Jul-99 Aug-99 Sep-99 Ot-99 Nov-99 De-99 Jan-00 Count of possum deaths Date Figure 5. Distribution of deaths for radio traked possums. Blue bars = healthy possums, orange bars = possums naturally infeted with tuberulosis, blak bars = possums experimentally infeted with tuberulosis. This does not inlude the one tuberulous and four healthy possums euthanased during the depopulation Behaviour of radio traked possums Tuberulous possums showed a range of behavioural states. These were from behaviour assoiated with healthy possums through inreasing debilitation to the final stage where they were moribund and unable to move. For most of the time, the behaviour of diseased possums was indistinguishable from that of healthy possums. The level of physial debilitation inreased markedly during the terminal stages of disease, whih lasted between one and three weeks. The possums at this stage were dull, lethargi, and showed poor oordination and balane. They were muh less responsive to disturbane aused by the traker. They were unable to limb steep inlines and would frequently fall over. These possums were emaiated and often had visibly enlarged superfiial lymph nodes, some of whih had burst to form draining sinuses.

76 Loation of tuberulous possum arasses The arasses of 17 tuberulous possums were reovered. Thirteen had died and four were euthanased in extremis (Table 4). No possum arasses were reovered from within a den. Two were found on pasture. Three were found in long grass amongst sparse srub. The remaining 12 possums were found in dense srub. Two possums were found dead in pools of water in reeks, having apparently drowned, and a third was found next to a reek, having been drowned when the reek flooded. All three were found where the waterways ran through srub. The denning range was known for 12 of the 17 tuberulous possums for whih the arass was reovered (Table 4). Six possums were found within their denning range and six were found outside. The ativity range was known for all 17 possums. Eight died within their normal ativity range. Nine died outside this area. Of these nine, three possums had moved more than 200 m from their ativity range, and six had moved less than this distane. Ten of the 17 reovered arasses were found at the lowest reorded elevation point on their ativity area.

77 61 Table 4. Reovery of arasses of tuberulous possums: individual possum details, loation of arass, and details of habitat where the arass was loated Possum Soure Sex Within Within Low Distane Desription of number of Tb den range ativity range point (1) from entre of ativity range (m) (2) 0001 Natural F yes no 40 Srub site of death 0025 Exp. inf M yes no no 190 Long grass under srub 0084 Natural F yes no yes 230 Creek in srub 0111 Exp. inf M yes yes 105 Creek in srub 0151 Exp. inf M yes no yes 220 Srub 5625 Natural F yes yes yes 15 Srub 5644 Natural M no no 150 Long grass 5700 Natural M yes yes yes 0 Srub 5711 Natural M yes no 25 Srub under srub 5758 Exp. inf M no no yes 150 Pasture 5760 Exp. inf M no no yes 160 Srub 5768 Natural F no yes yes 90 Srub 5796 Exp. inf M no no yes 170 Pasture 5816 Natural F yes yes no 20 Srub 5831 Natural F no no yes 120 Creek in srub 5838 Exp. inf M no no no 260 Long grass 5871 Natural F yes no 0 Srub Exp. inf = possums experimentally infeted with tuberulosis 1 Low point = possum arass found at lowest reorded elevation 2 Movements shown in bold were defined as long distane forays (>200 m) Denning behaviour of possums Radio traking identified a total of 233 individual den sites and the distribution of these is presented in Figure 6. The preferred denning area for most possums overed 15 ha, or

78 62 approximately 25% of the area of the site. This area ontained the steepest slopes on the site. It was overed with srub and native forest pokets and was known as Ponga Gully. C line B line 100m A line Figure 6. Density of den site use for all radio traked possums. Crosses represent trap sites, and lines divide the study site into three trap lines, A, B and C. Contours represent den site density, from high (entral most 20% of den sites, darkest shade), through 40%, 60%, to low (80% of den sites, lightest shade) The distribution of den sites used by naturally tuberulous possums (Figure 7) was similar to that for all possums. However there were three distint pokets of dens used by this group. These small foi are indiated as areas 1, 2 and 3 in Figure 7. The den sites of healthy possums were grouped into two foi, one in Ponga Gully and a seond at the southern end of the C trap line (see appendix 1.8). Den sites of experimentally infeted possums were spread more evenly aross the study site (see appendix 1.11), indiating that the seletion of those possums to provide site-wide overage was suessful.

79 63 Area 3 C line B line 100m A line Area 1 Area 2 Figure 7. Density of den sites used by naturally tuberulous possums. Lines represent areas overed by the three trap lines (lines A, B and C). Crosses represent trap sites. The three distint pokets of dens used speifially by tuberulous possums are indiated as areas 1, 2 and 3. Contours represent den site density, from high (entral most 20% of den sites, darkest shade), through 40%, 60%, to low (80% of den sites, lightest shade) Denning range for possum groups There was suffiient data to alulate denning ranges for 22 of 30 radio traked possums (Table 5). The denning range overed approximately 20% of a possum s total range. There were no signifiant differenes between denning ranges of the three groups of possums.

80 64 Table 5. Denning range (hetares) for the three groups of possums: naturally infeted, experimentally infeted and healthy. Group n 1 Minimum Maximum Mean SD Healthy Naturally tuberulous possums Experimentally infeted n = number of possums in eah group The size of possum denning range was ompared aross three den site densities (appendix 1.2). The entral portion of the site had the highest density of dens and there the average denning range was 1.2 ha. The eastern part of the site had a medium density of den sites with an average denning range of 1.4 ha. The lowest density of den sites was on the western part of the study site and had an average denning range of 1.5 ha Denning behaviour of individual possums Denning ranges were generated for the 22 possums that were traked at least four times. Denning ranges for possums formed four broad types: a single group of dens (unimodal), two distint and adjoining groups of dens (bimodal), three distint and adjoining groups of dens (trimodal) and dens not showing any grouping. The distribution of different denning range types aross the three groups of possum is shown in Table 6. Most possums used den sites lustered in a single group (10/22) or two distint groups (8/22). Seventeen of 22 possums had one or more dens that lay outside the perimeter of the denning range as desribed above. Table 6. Distribution of denning range types for the three possum groups: naturally infeted, experimentally infeted and healthy possums Possum group n 1 Unimodal Bimodal Trimodal Dispersed Healthy (n=6) Naturally tuberulous (n=8) Experimentally infeted (n=8) Total (n=22) n = number of possums in eah group

81 65 The denning range of Possum 5760, whih overed 1.3 ha, is an example of a unimodal range (Figure 8). 100m Figure 8. The denning range of Possums 5760: An example of single luster of dens. Contours represent den site density, from high (entral most 20% of den sites, darkest shade), through 40%, 60%, to low (80% of den sites, lightest shade). Den loations are marked by blue dots and the dotted line represents an aess trak. Note the dens that lie outside the denning range

82 66 An example of a bimodal range is illustrated in Figure 9 by Possum 5700, the ombined area of the two foi is 0.9 ha. 100m Figure 9. The denning range of Possum 5700 as an example of a bimodal denning range. Den loations are marked by green dots. Contours represent den site density, from high (entral most 20% of den sites, darkest shade), through 40%, 60%, to low (80% of den sites, lightest shade). The dotted lines represent aess traks and fene lines present on the study site. Note the dens whih lie outside the denning range

83 67 The trimodal form of denning range was shown by two possums. The fourth type, that onsisted of several groups of dens dispersed aross a omparatively large area is illustrated by Possum 5625 in Figure 10. The denning range of Possums 5625 as an example of a dispersed denning range The total of the four denning areas was 1.2 ha although the dens were spread over 8 ha. 100m Figure 10. The denning range of Possums 5625 as an example of a dispersed denning range. Den loations are marked by red dots. Contours represent den site density, from high (entral most 20% of den sites, darkest shade), through 40%, 60%, to low (80% of den sites, lightest shade). The dotted line represents an aess trak through part of the study site. The two den sites with a single and double ring around them were used twie and three times respetively The areas on the study site with high, medium and low den densities eah have a similar proportion of possums showing the different types of denning range.

84 Comparison of possum denning range and foraging range There were 985 trapping reords used to alulate trap ath rates. The frequeny with whih individual traps aught a possum ranged from Some trapping reords for tuberulous possums were olleted before the diagnosis of infetion was made. For experimentally infeted possums most of the trapping data was olleted before the subjet was infeted. The possum population did not forage evenly over the study site. (Figure 11). There was one area learly favoured for foraging and it was haraterised by dense srub with a high proportion of flax, on steep slopes. Areas where possum foraging ativity was low were overed by a variety of vegetation types and inluded pasture on the valley floor. There were three small foi of low use areas. Two were on the top of the western side of Ponga Gully overed by manuka and gorse. The third was on the eastern side of the site, whih had a mixture of both open and dense srub, with small pokets of native bush. The terrain in these areas was of variable slope.

85 69 The area of highest foraging ativity for all possums as a group was different from the area of highest denning ativity (Figure 11). The distane between the two density entres was 200 m and they were separated by a small ridge. Area 2 C line Area 3 Area 1 B line 100m A line Figure 11. The foraging range of all radio traked possums on the study site. Crosses represent trap sites, lines define areas overed by the three trap lines. Contours represent density of trapping events, from high (entral most 20% of trapping events, darkest shade), through 40%, 60%, to low (80% of trapping events, lightest shade). Areas 1, 2 and 3 represent the small foi of low foraging ativity

86 70 The denning and foraging areas for individual possums were ompared. The entre of the denning area and the entre of the foraging area were less than 100 m apart for all possums, regardless of sex or disease status. There was onsiderable variation between individual possums in the total area overed by den sites and trap sites. Extreme examples are the patterns shown by Possum 0026, whih had a widely dispersed set of den and trap sites, and Possum 0142, whih had a tightly lustered pattern (Figure 12) m 0026 Figure 12. Examples of dispersed (Possum 0026) and lustered (Possum 0146) patterns of denning and foraging by individual possums. Dotted lines = fene lines and referene points, large dots = den sites, flags = trap sites where the possum was aught. Cirles enompass the data points for eah possum

87 Ativity range and total range of possums Ativity ranges were alulated using a kernel density estimate of ombined trapping and denning data. Total ranges were alulated using every data point from trapping and denning data. There was an average of 47 (range 5-89) data points for 28 possums. Two possums were omitted from this analysis beause the ombined number of data points for eah was less than five. The distribution of female ativity ranges and female total ranges is signifiantly skewed. The size of ativity range and total range was similar for eah possum group ( Table 7). Table 7. Ativity ranges and total ranges (hetares) for three possum groups: naturally infeted, experimentally infeted and healthy possums Ativity range Total range Group n * Median Max Min Median Max Min Healthy Naturally tuberulous Experimentally infeted Overall * number of possums in eah group The ativity range of 18 possums was unimodal. Seven ativity ranges were bimodal and the remaining three onsisted of three distint areas. A summary and diagram of the ativity and total range for eah possum is presented in Appendix II, where they are listed in order of possum identifiation number.

88 72 The relative sizes of denning, ativity range and total range for eah possum are ompared in Figure 13. Size (ha) Possum ID Figure 13. The denning range (green bars), ativity range (red line) and total range (blue bars) of 28 possums Correlation between number of observations and range size As the number of observations inreased, the range size also inreased (Figure 14). This orrelation was signifiant for all three types of range ( Table 8) Range size (ha) Possum ID Figure 14. The size of ativity range (blue bars), total range (red bars) and number of observations (green line) of 28 possums Count of observations

89 73 Table 8. Correlation between range size and number of observations Range type Pearson s orrelation statisti n 1 P value Denning range Ativity range Total range n = number of possums for whih the range type was generated Long distane forays Possums infrequently undertook long distane forays from their established ativity areas. All radio ollared possums were mature. For 28 possums there was suffiient trapping and denning data to examine long distane forays. Half of all radio traked possums made at least one long distane foray (Table 9). There was no onsisteny in the diretion of these forays. The median distane of forays by tuberulous possums was 290 m (Table 9) whih was not signifiantly different from healthy possums (256m) (Mann-Whitney U=29.5, Z=-1.021, P=0.319). The proportion of eah sex making a long distane foray was similar. The median distane travelled by eah sex was also similar. Table 9. The proportion of healthy and tuberulous possums whih made at least one long distane foray and the median distane from their ativity range Possum group n 1 Perent of group 2 Median distane (m) Healthy Tuberulous Males Females The number of possums in eah group 2 The perentage of eah possum group whih made at least one long distane foray Most possums made a single foray and no possums made more than three forays. The frequeny of movements of male and female and also tuberulous and healthy possums were similar (Table 10).

90 74 Table 10. The frequeny of long distane forays made by individual possums Number of movements by individual possums Possum group n 1 One Twie Three times Healthy Tuberulous Males Females The number of possums in eah group Most forays ourred during February, August, September and Otober (Figure 15). 5 4 Count of movements Jan Feb Marh Apr May Jun Jul Aug Sept Ot Nov De Month Figure 15. Monthly distribution of 22 long distane forays made by 14 possums. Forays were alulated using trapping data olleted over five years and den site data olleted weekly over 23 months. Dark segments = males, light segments = females Length of forays The frequeny and distane of long distane forays from the ativity range was examined. The length of 22 suh forays and a omparison between male and female possums is presented in Figure 16. The distribution of these movements is skewed to the right. The median distane for all forays is 300 m (range m). Males moved more frequently than females though the differene is not signifiant (P=0.164, Z=-1.413, Mann-Whitney U=-24.5).

91 Distane (m) m m f m f m m f f m m m m m m f m m m m m m Sex Figure 16. Length of 22 long distane forays made by 14 possums. Long distane forays were alulated using trapping data olleted over five years and denning data olleted over 23 months. Dark bars = males, light bars = females

92 76 A typial example of a long distane foray is shown by Possum 0055 in Figure 17. This possum moved 500 metres to the north west. It was trapped one at this loation and subsequently trapped bak within its ativity range. 100m Figure 17. An example of a long distane foray: Traps (red rosses) where Possum 0055 was aught and den sites (blue dots) it was known to use. Contours represent density of ombined trap and den loations for Possum 0055, from high (entral most 20% of loations, darkest shade), through 40%, 60%, to low (80% of loations, lightest shade). Flags = traps sites, broken lines = fene lines and aess traks Tuberulosis hot spots The loation of five hotspots on the site identified between Marh 1998 and January 2000 are shown in Figure 18. Hotspot loation was derived from 189 denning events for tuberulous possums. There were two areas of high density, a medium density area and two low density areas. The most prominent hotspot (hotspot 1) was on the steep sides of Ponga Gully. Five of the 22 tuberulous possums favoured this area. The large, entral area of medium density (hotspot 3)

93 77 was used by seven possums, with the denning ativity spread over a large area. There were three small hotspots in the north (hotspot 2), east (hotspot 4) and south (hotspot 5) of the study site. The northern fous was used by three possums. Two of these three frequently reused their own tightly lustered group of dens. The tightly lustered group of dens in the east and south were eah used by a single possum. The remaining five tuberulous possums used dens within the hotspots, as well as dens where tuberulous possums were seldom found. hotspot 1 C line hotspot 5 B line hotspot 2 100m hotspot 3 A line hotspot 4 Figure 18. Hotspots identified on the study site between Marh 1998 and January Crosses represent trap sites, lines define areas overed by the three trap lines. Contours represent den site density for tuberulous possums, from high (entral most 20% of den sites, darkest shade), through 40%, 60%, to low (80% of den sites, lightest shade) Sequential use of the same den by tuberulous and healthy possums At no time were two or more possums found in a den at the same time. There were 12 ases of two possums using the same den at different times and one ase of four possums using the same den at different times. There were five oasions when a healthy possum oupied a den that had

94 78 previously been used by a tuberulous possum. On four of these five oasions the time interval between diseased and healthy oupations was 1, 2, 23 and 62 weeks. In the fifth ase, a tuberulous possum used a den that was then used by three non tuberulous possums at 6, 8 and 18 weeks after the diseased possum had used it. No healthy possums under observation beame tuberulous. Dens were used only one by most possums, but in three ases a den was used by the same possum six times (Table 11). Table 11. Frequeny of use of individal dens by possums Count of times a den was used Frequeny Den harateristis Ninety-two perent of dens had a roof that overed?75% of the den floor, these were defined as fully enlosed. Two perent of 233 den sites had a roof that overed less than 50% of the den floor, these were defined as open den sites. The roof of the remaining 6% of den sites overed between 50% and 75% of the den floor. The most popular denning material was flax (Phormium tenax) and was found in 30% of dens. Gorse (Ulex europaeus) was the next most popular at 18%, followed by heaped dead vegetation (13%) and underground sites or reesses in banks (11%). The remaining seven types of den material were only reorded oasionally and are shown Figure 19.

95 79 50 perent of radio traking events root rakings pampas fern manuka nothing dead veg. utty grass vegetation type (roof) ponga soil flax gorse Figure 19. Perentage of denning materials used by tuberulous possums (dark bars) and healthy possums (light bars), the ategory labelled nothing desribes den sites where less than 50% of the sides and roof were enlosed The diretion faed by the slope on whih eah den was found was reorded by ompass bearing, then ategorised as having either a predominantly northern or southern aspet (Table 12). Most dens had a northerly aspet. The perentage of north and south faing dens used by naturally infeted, experimentally infeted and healthy possums were ompared. There was no signifiant preferene for the aspet shown by the three lasses of possums studied. Table 12. Charateristis of dens used by possums: aspet of dens used by healthy, naturally tuberulous and experimentally infeted possums Group North faing (%) South faing (%) Healthy possums Naturally tuberulous possums Experimentally infeted possums Of the dens identified, 99.5% were at ground level. The floor in most dens (65%) was dry with a few mildly damp (14%) or very damp (15%) (Table 13). Very few dens were lassified as wet or saturated.

96 80 Table 13. Charateristis of dens used by possums: floor ondition of dens used by healthy, naturally tuberulous and experimentally infeted possums Group Dry (%) Slightly Very Wet (%) Saturated (%) damp (%) damp (%) Healthy possums Naturally tuberulous Experimentally infeted

97 DISCUSSION Spatial aspets of tuberulous possum behaviour This study showed that the majority of possums dying from tuberulosis remained within their established ativity ranges, in srub. The arasses of seventeen tuberulous possums were reovered. Twelve were reovered from srub and three were found on grass areas within sparse srub. The arasses of only two possums were found on pasture, in both ases the possum had moved downhill, and was a short distane (less than that defined by a long distane foray, see setion 2.5.6) outside its established ativity range. Three possums made long distane forays during the terminal stages of disease but all remained and died in srub. It had been believed, based on anedotal aounts, that tuberulous possums migrated from their established ativity ranges to areas with easily aessible food and water, pasture in partiular. The data from this study show this speulation was untrue. Tuberulous possums showed a tendeny to move to lower elevation as the disease reahed its terminal stages. Ten of 17 tuberulous possum arasses were reovered from the lowest reorded elevation in their range. In three of these ases the arass was found in a water ourse and the possums had apparently drowned. The ranging behaviour desribed in this study is supported by the results of Paterson et al. (1995) who found most tuberulous possums remained within, or lose to, their denning range until death. Paterson also found that a small proportion of tuberulous possums made extended forays shortly before death. Inreasing physial debilitation was evident as possums progressed through the terminal stages of disease. Debility was haraterised by lethargy, weakness and poor oordination. For possums whose ativity range was on steep terrain, their gradual deent to lower elevations during the terminal stages of disease was probably due to physial debilitation. It may also have lead to the ases of drowning. It is likely that diseased possums desending from high in the srub to pasture, as a result of debilitation, were responsible for anedotal reports of diseased possums moving to pasture to feed. The anedotal observations ould have been biased, as possums dying on pasture would be easily observed, unlike those in srub. The anedotal reports were also made during a time when the prevalene of tuberulosis was muh higher on the study site.

98 82 The proportion of terminally ill possums whih die on pasture represent a small but serious risk of infetion to attle and deer. During the terminal stages of disease possums are not only debilitated, but are also at the peak of infetiousness (Jakson et al., 1995). The delining physial state results in behavioural hanges, suh as redued avoidane reation to livestok, foraging during daylight and a redution in ativity area (Paterson, 1993). The first two of these hanges greatly inrease the hanes of a normally noturnal animal enountering and failing to avoid inquisitive attle or deer (Paterson and Morris, 1995, Sauter and Morris, 1995). In omparison, healthy possums atively avoid livestok. Cattle and deer, attrated by the diseased possum s abnormal behaviour, show a high degree of inquisitiveness and will lik, sniff and even grasp the possum with their mouth. This lose investigative behaviour by livestok, ombined with the loud hissing and sreehing whih is behaviour typial of a threatened possum, provide exellent onditions for aerosol transmission of tuberulosis. Transmission of disease between possums is most likely to our when the tuberulous possum is still apable of engaging in affiliative or agonisti soial interation, whih is prior to the terminal stages of disease (Jakson et al., 1995). The frequeny of interations between a terminally ill possum and other healthy possums is likely to be low as a result of delining physial ondition and redued ativity levels. Tuberulous possums that die in dense srub represent a lower but variable risk of infetion to attle. This risk is low where attle are not permitted to forage through srub, but higher where stok management or poor fening allows attle to explore srub areas. Paterson et al. (1995) reports that attle will investigate bush areas not thought aessible, or attrative to them, inreasing the likelihood of enountering a diseased possum. Possums dying in srub also represent a soure of infetion to wild animals, partiularly deer, or savenging ferrets and pigs (Wakelin and Churhman, 1991). Tuberulous possums that die in dense srub represent a low risk of infetion to domesti deer. This is beause deer are typially fened suh that they annot forage through srub areas. Temporal aspets of tuberulous possum behaviour Survival of naturally tuberulous possums ranged from one to 84 weeks (median 11 weeks). Survival of experimentally infeted possums ranged from nine to 21 weeks (median 12 weeks). Variability in the survival of naturally diseased possums is to some degree artifiial as it was influened to a small degree by radio ollars being applied at different stages of disease. Examination of the population for linially ill tuberulous possums was onduted bimonthly. The period between examinations was large, relative to the survival duration of linially ill

99 83 possums, so tuberulous possums ould not onsistently identified early in the ourse of the disease. The duration of survival of naturally infeted possums in the urrent study was onsiderably shorter than that reported by Jakson (1995). He reported the survival duration, following the detetion of palpable lesions, to be about six months for 50 perent of animals, with a maximum of 31 months. Experimental infetion of possums resulted in a more rapid disease proess than that following natural infetion. The survival time after infetion in experimentally infeted possums was one week longer than the survival time, after linial diagnosis, for naturally infeted possums. In naturally infeted possums a prelinial period estimated from studies of aptive possums to be eight to ten weeks (Corner, pers. om., 2001) must be allowed for, when omparing relative survival times. There was one extremely unusual ase in whih a possum apparently self-ured infetion with tuberulosis. Eight months before the beginning of the study Possum 5787 was diagnosed with tuberulosis. Infetion was onfirmed on two separate oasions by the isolation of M. bovis, first from aspirated pus from a lesion, and then 2 months later from a swab taken from an open sinus. This possum survived for the duration of the study and was euthanased during the depopulation. The post mortem examination, subsequent histopathology, bateriology and ulture of tissues, all failed to detet any evidene of tuberulosis. Lugton (1997) suggests reovery from infetion with M. bovis is possible and presented six ases whih he believed demonstrated this proess. However, in none of these ases was a possum with tuberulosis onfirmed by ulture, reexamined and found to be ulture negative. The survival period of tuberulous possums is a rude estimate of the period where they may transmit disease to other animals. Transmission to healthy possums probably ours most frequently during the early to late stages of infetion, but before the terminal stages. During this period the infeted possum is exreting M. bovis but still behaving normally, for example involved in soial interations while seeking a den or during mating. In ontrast, the time at whih a diseased possum is most infetious to attle or deer is during the terminal stages, when debilitation limits the ability of a diseased possum to avoid investigation by livestok and levels of exretion are at their peak. Possum denning, ativity and total ranges The possums in the study were found to den and feed on different parts of the study site. Possums, as a group, most frequently denned in dense srub on the steep sides of Ponga Gully. This area represents about one quarter of the study site. The area most frequently used for

100 84 foraging, as revealed by trapping, was 200 m to the east, also in dense srub. The majority of foraging ativity also ourred on about one quarter of the study site. There was a small amount of overlap between areas of low denning ativity and low foraging ativity. It is unlear why possums hose these partiular areas in whih to den and forage as there were few apparent physial differenes between the two. There seemed to be a similar number of potential den sites in both areas. Vegetation of the two areas was similar in omposition, degree of over and density. The denning area was extremely steep, ompared with the moderately steep foraging area. The size of denning range was similar aross all possum groups. The denning range of an individual possum overed approximately one quarter of its total range, and was loated within part of the ativity range, but seldom loated on its edge, as previously found by Ward (1984) and Paterson et al. (1995). There is muh data published on the denning habits and home range estimation of possums (see the review by Cowan and Clout (2000)), but little data published on denning ranges. Detailed information of den materials was presented by (Cowan, 1989) but no estimates of denning range were given. There was muh variation in the denning habits of individual possums. The dens used by an individual possum were generally grouped in one to three foi. The size of denning range was similar aross these groups. Most possums were found on oasion denning well outside their established denning range. The pattern of den lusters with oasional outliers for individual possums was similar to that observed by Paterson et al. (1995). In areas where den density was high the average size of denning ranges was twenty perent smaller than in areas of low den density, indiating that the size of denning range is influened to a small degree by possum density. The denning range of individual possums frequently overlapped, both between and within sexes, as did that of healthy and tuberulous possums. In one instane an area of one quarter of a hetare was used simultaneously by four possums in the study, one healthy and three diseased, and quite possibly other possums. The overlap of multiple possum ranges is likely to inrease the frequeny of interations between the resident possums. Where a proportion of these ranges are oupied by tuberulous possums, there is a high hane of disease transmission to healthy possums. The demonstration of multiple overlap of denning ranges supports the hypothesis of Pfeiffer (1994) that the transmission rate of tuberulosis between possums is probably highest in areas of favoured denning.

101 85 There was onsiderable range in the popularity of individual den sites. Some dens were regularly reused, up to six times. Some dens were used by as many as four different possums, at different times. Simultaneous den sharing was not observed during the study. Healthy possums were found in dens previously used by tuberulous possums and in two instanes the interval between use was less than 28 days. Jakson et al. (1995) reported that, under ideal onditions, M. bovis ould survive for this length of time within a den. When tuberulous and healthy possums frequently reuse the same dens the possibility of healthy possums beoming infeted, due to den ontamination, is inreased. There ould also be an inreased risk of transmission through aggressive interations undoubtedly ourring as a result of ompetition for these favoured dens. Simultaneous den sharing was not observed during this study and has only rarely been observed at the Castlepoint study site in the past (Pfeiffer, 1994). However, it is more ommon in other areas of New Zealand (Caley et al., 1998). Simultaneous den sharing between diseased and healthy possums would be an exellent environment for transmission of tuberulosis, beause of the onfined air spae and still air whih would failitate the maintenane of infetious aerosols. The proportion of dens used by two or more possums found in the urrent study was very similar to that found when possum density was muh higher in an earlier study at Castlepoint (Paterson, 1993). The total range and ativity range found in this study were substantially larger than those previously reported at this site. The median total range of 5.9 ha and median ativity range of 2.2 ha were both greater than the ativity range estimate of 1.4 ha for males and 0.9 ha for females by Paterson (1993). The differene in results are due in part to differenes in data olletion and analytial methodologies. Paterson used radio triangulation to map the nightly movements of possums at fifteen minute intervals. By omparison, the urreent study identified den sites at weekly intervals and olleted trapping data at bimonthly intervals. Substantially more data (average 128 points/possum) was generated than in the urrent study (average 47 points/possum). He found that the range size for some animals stabilised after 100 loations while for others it ontinued to inrease for the duration of traking (12 months, 458 loations). Range alulation was by the onvex polygon, or alternatively harmoni mean method. It is likely that olletion of data over long periods, as in the urrent study, provides a more aurate estimate of possum ranging behaviour than short, intensive observations. Total range values for the urrent study (whih are very similar to home ranges) lay at the high end of home range estimates from other studies onduted in New Zealand and Australia (see the review by Cowan and Clout, 2000).

102 86 Long distane forays Half of all possums made at least one long distane foray during the study. Forays were identified by a distint movement to a single loation, for a brief period and then a return to their ativity range. These were made by tuberulous and healthy possums. Males made forays more frequently than females, whih supports the findings in a study by Green and Coleman (1984). The median distane was between 250 and 300 metres and the maximum was 660 metres. Travelled distanes were longer than those found by Ward (1984), although very muh shorter than those reorded for dispersing juvenile possums (Brokie et al., 1987). There was no trend in the diretion of long distane forays. The majority of the long distane forays ourred in February, August, September and Otober. Only three of seventeen tuberulous possums made a long distane foray shortly before death, indiating suh behaviour is not indued by the terminal stages of disease. Most long distane forays were probably assoiated with breeding behaviour. Forays oinided frequently with the minor breeding season, from August to Otober, and oasionally with the main breeding season, from Marh to May. This is onsistent with the results of Paterson et al. (1995), in whih possum ranges were largest during the autumn breeding season. The habitat on the study site is a mixture of srub and pasture and there were no apparent foi of fruit or nut speies, whih were laimed to indue movement of possums by up to 1600 m (Jolly, 1976). Faeal evidene indiated that possums regularly visited pine trees to feed on atkins, though these appeared to be nightly visits rather than a reloation over longer periods. All possums were mature, therefore these movements are distint from the permanent long distane movements harateristi of juveniles dispersing in searh of a new home range (Cowan et al., 1997). Long distane forays are likely to have important impliations for the spread of tuberulosis. Travelling possums will pass through the home range of established possums, and agonisti interations are more likely during this first enounter than with regularly enountered possums in whih the soial standing of eah individual has already been established. If the resident or travelling possum is tuberulous, disease transmission during these enounters has the potential to spread the disease to previously uninfeted areas. There are onfliting views on whether long distane forays should be inluded in possum ranges. This debate raises the question of whether a range is most aurately represented by all loations for a possum, or alternatively, the area in whih the animal spends most of its time. Ward (1984) argues strongly for their inlusion in range estimates, on the grounds that movements to seasonal food soures or during the mating season are important omponents of possum behaviour. However, the addition of these infrequent, outlying points an have a disproportionate effet on

103 87 the estimated size of a range, ompared with the majority of more entralised points. In the urrent study, the inlusion of an ativity range and total range presents both sides of this argument and allows the reader to ome to their own onlusion. Hotspots Most den sites used by tuberulous possums were lustered on one quarter of the study site, with a five hetare area of partiularly high density. The major hot spot was entred on the steepest area of the study site. Four minor hotspots were also identified on moderately steep to very steep terrain. Vegetation on all hotspots onsisted of dense gorse and manuka ranging in height from one to three metres and interspersed with lumps of flax. All hotspots ontained a high number of fully enlosed dens. The urrent study is in agreement with two harateristis of the hot spot preditor developed by MKenzie (1999), the number of fully enlosed dens and the height of overing vegetation, but at odds with the requirement for gently sloping or flat land. Den sites used by experimentally infeted possums were distributed more evenly aross the site than those of naturally infeted, or healthy possums. It was the goal of the trials to distribute the experimentally infeted possums evenly aross the study site. Therefore the harateristis of den sites used by this group do not ontribute to our understanding of hotspots. The most popular explanation for the loation of hotspots is the assoiation with high numbers of enlosed dens in favourable denning areas. However, this explanation appears inomplete as many other areas on the study site appeared virtually idential in all physial aspets, but supported low or negligible den density for tuberulous possums. One possible explanation is that, in addition to the fators urrently believed important for hotspot formation, hotspots are also influened by the pathways formed and used by wildlife, inluding migrating possums. Interations between resident and immigrant possums along these pathways would more frequent than found elsewhere. Where possums are infeted with tuberulosis, these spatially onentrated interations may lead to high rates of disease transmission, and hene a hotspot. A study involving the behaviour of feral urban ats found their home ranges were haraterised by a number of points of interest onneted by frequently used pathways, with the remaining areas rarely used (Leyhausen, 1965). In the urrent study, tuberulous possums denned in some areas also used by tuberulous possums eight years previously. Of the four minor hotspots identified, three were also noted by Pfeiffer (1994). It is not possible to determine whether tuberulous possums had been present in these minor hotspots ontinuously for the entire period between studies. However it is quite possible that these are permanent hotspot loations, as have been identified in other studies

104 88 (Caley et al., 1999). The major hotspot identified in the urrent study was not mentioned by Pfeiffer. Failure to detet tuberulous possums in this area may have been due to Pfeiffer using fewer trap sites, whih unlike the urrent study, did not enlose the primary hotspot. It appears from this omparison that some hotspots are present for long periods, while others are temporary. To effiiently target tuberulous possums with ontrol operations it is important to periodially re-evaluate hotspot loation, to take into aount the possible hanges in the presene and loation of hotspots over time. This study has provided an alternative explanation for the existene of temporary hotspots. Temporary hotspots were defined as a site where a single diseased possum was found during a ross setional survey (MKenzie and Morris, 1995). This definition assumes that a temporary hotspot is the permanent residene of a tuberulous possum, but where its ativity range does not ontain the neessary harateristis to support transmission of disease, therefore the disease is not maintained within that area. The alternative explanation is that a tuberulous possum was trapped while making a long distane foray, during whih it did not transmit disease to other enountered possums. The aurate definition of a temporary hotspot is ompliated by some harateristis of a permanent hotspot. A permanent hotspot usually funtions as a soure of moderate to high levels of infetion, but may also exist at very low levels of infetion, similar to a temporary hotspot, for a variable time period (Coleman et al., 1999). Hotspots, as areas with high rates of tuberulosis transmission between possums, are most aurately predited using den site data. Hotspots are frequently defined by trapping data due to the omparative ease with whih it an be obtained. The entre of a hotspot was found to vary in loation by up to 200 metres depending on whether it was defined using den site data, or trapping data. Transmission of disease is probably highest during agonisti interations (Pfeiffer, 1994)) and it is believed these are most frequent about favoured denning areas, as a result of ompetition for den sites, or oestrous females during the breeding season. They may also be due to the establishment and maintenane of soial hierarhies. Unfortunately it is substantially more diffiult to gather den site data than to ollet trapping data. The use of trapping data is a reasonable ompromise as long as it is seen as approximating the real hotspot. Physial qualities of the dens Possums most frequently used dens in flax, gorse or heaps of dead vegetation, similar to that reported in previous studies at the site (Pfeiffer, (1994), Jakson, (1995), Paterson et al., (1995)). Almost all dens were at ground level and the majority had dry floors, even during winter and after several days of wet weather. Oasionally dens were damp, but very seldom wet. No possum was

105 89 observed with wet fur as a result of water oming into a den. Most dens had a north faing aspet and a similar proportion of diseased and healthy possums used north faing dens. Good quality dens were still used by tuberulous possums during the terminal stages of disease. The possum population showed a preferene for north faing dens. This preferene may be due to the extra warmth and dryness assoiated with a northerly aspet. In ontrast, Pfeiffer (1994) found most possums denned on slopes whih faed in an ar between south and north west. This disparity may be due to the methodology by whih aspet was determined. The urrent study reorded the aspet of the half hetare surrounding the den, in this way the highly hangeable small sale topography was reorded as lose to the den as possible. Aspet over a greater area was reorded by Pfeiffer, whih may have aused his results to reflet the general aspet of the site, whih was south to north-west. Limitations of the study The prinipal limiting fator in the desription of possum ranges was the limited number of subjets and observations on eah subjet. The effet of this was redued somewhat by ombining trapping reords with den site data to generate an ativity range. We showed that the greater the number of observations used (maximum 128), and the longer the period of observation, the larger was the range. Even so, the values we have reported for both denning and ativity ranges are probably an under estimate of the true values, due to the limited volume of data available for individual possums. (Paterson, 1993) found that, where loations were taken every fifteen minutes over several nights, the range size for some animals stabilised after 100 loations had been determined, while for others the range size ontinued to inrease for the duration of traking (maximum 458 loations). Although the method of data olletion is somewhat different, the impat of inreasing observations on expansion of range size is again learly evident. Estimates of ranging areas are influened by the method of data olletion, as we have shown and as elaborated by Ward (1984). Disparities our beause there may be a onsiderable differene between where a possum dens and where it forages (Efford et al., 1994). By investigating the different data types individually and in ombination these disparities were demonstrated and an overall view of the various aspets of possum ranging behaviour was shown. The method of alulation an strongly influene the estimate of range size (Paterson et al., 1995). In this study, point data representing den sites and trap loations was onverted into an area, whih represents the range, by using kernel density estimates. This method of range alulation has two signifiant limitations. The first is variation in the estimate due to the number

106 90 of data points (MCallum, 2000). This effet was minimised by omitting from the analyses possums with less than five reords and by manipulation of ontour sensitivity, using the H value when alulating a kernel density estimate. The seond problem was the exlusion of a small number of outlying points from the estimate of area. These outlying points are extremely important in aurately refleting all movements of a possum (Ward, 1984). These outlying points were exluded from our alulation of ativity range but inluded in the total range, allowing an illustration of possum ranging behaviour with and without long distane forays CONCLUSION This study used radio telemetry and live-trapping data to observe tuberulous possums and identify aspets of their behaviour that may inrease the risk of disease transmission to livestok and other possums. The total range, ativity range and den range of tuberulous and healthy possums was similar in size. The behaviour of tuberulous possums was indistinguishable from that of healthy possums for muh of the disease proess. However, behaviour of tuberulous possums beame inreasingly debilitated during the terminal stages of disease, whih lasted 1 to 3 weeks. Debilitation makes it diffiult for a tuberulous possum to avoid investigation by attle and deer, inreasing the risk of disease transmission. In addition, it often auses possums to die at low points of elevation within their normal ranges. Most tuberulous possums died within or lose to their established ativity range in dense srub. However, a small proportion died on pasture. Although the number of tuberulous possums dying on pasture was small, they onstitute an important soure of infetion to livestok. Those dying in srub are a potential soure of infetion to wild animals suh as deer, ferrets and pigs. A small proportion of diseased possums made long distane forays of 200 to 300 metres in the last few weeks of life, but still died in srub. There was no temporal pattern in possum deaths. Tuberulous possums used den sites that were lustered into one major and four minor hotspots. A previous study on the same site had found tuberulous possums in three of these minor hotspots. The use of den site data, or alternatively trapping data, to define the loation of a hotspot aused variation of the loation by 200 m. Den site data provides a more aurate definition of hotspot loation than trapping data due to disease transmission between possums being most frequent in denning areas.

107 91 Chapter 3 DESIGN AND TESTING OF A SELF-SETTING AEROSOL VACCINATOR The spaious, lean, fully equipped work spae for vainator onstrution. Leaving the garage door open provided enough light to work by, while Cyril the motorbike rouhes in the bakground.

108

109 INTRODUCTION Tuberulosis (Myobaterium bovis) was probably introdued to New Zealand with imported attle during the 19 th entury. Today the disease is present at low levels in domesti attle and deer herds and represents a potential threat to the ountry s large export industry. The test and slaughter method plus restritions on the movement of animals from diseased herds are used to ontrol the disease in livestok. The brushtail possum (Trihosurus vulpeula) was introdued to New Zealand during the 1850s. This highly adaptable animal quikly spread over muh of New Zealand and in 1985 had an estimated population of 70 million (Cowan and Bayliss, 1998). Possums infeted with tuberulosis were first disovered in the wild in 1969 (Ekdahl et al., 1970). The disease is now onsidered endemi in possums over approximately one third of New Zealand (Animal Health Board Annual Report, 2000). The possum is onsidered a wildlife reservoir of tuberulosis, ating as a soure of infetion to livestok and wildlife. The ontrol of tuberulosis is based on the large sale aerial spreading of poison baits for possums and the test and slaughter method for infeted livestok (Montague, 2000). Redution, or ideally removal, of the wild life reservoir of disease would greatly improve the effetiveness of test and slaughter methods, possibly leading to eradiation of tuberulosis from livestok in New Zealand. In Australia, eradiation of tuberulosis from wildlife in partiular Swamp buffalo (Bubalus bubalis), has failitated the eradiation of the disease from domesti livestok (Cousins et al., 1998). In Great Britain and Ireland, the badger (Meles meles) is assoiated with persistent infetion of livestok with tuberulosis, in muh the same way as the possum in New Zealand (Cheeseman et al., 1989). In its national pest management plan for the years the Animal Health Board proposes expenditure on possum ontrol to average 60 million dollars per annum. The total proposed annual expenditure for the ontrol of tuberulosis in New Zealand is 85.5 million dollars with possum ontrol being the major omponent. The plan is to use existing ontrol tehniques but at higher intensity (Animal Health Board, 2000). Existing ontrol tehniques have redued the prevalene of tuberulosis in livestok, but have failed to ontrol geographial spread of the disease in wildlife. Baille Calmette-Guerin (BCG) is a live vaine whih has been used to protet humans against tuberulosis for many years. It has also been shown to protet a range of animals, inluding

110 94 possums (Buddle et al., 1997), attle (Buddle et al., 1995). Vaination of attle or possums is a possible alternative to urrent disease ontrol tehniques. However, there are two major ompliations faing the vaination of domesti animals. Firstly, immunisation of attle with BCG results in some animals displaying positive skin test responses to bovine PPD (tuberulin). The tuberulin test is urrently used to identify diseased animals. If vaination of livestok were to be implemented, a new method for detetion of tuberulous animals would be required. In addition, the large-sale vaination of attle would require approval by national and international regulatory bodies on the grounds of food safety. Seondly, the issue of tuberulosis in wildlife speies would not be addressed and urrent ontrol operations would be required to ontinue indefinitely. To avoid the problems inherent in vaination of livestok, it may be preferable to vainate wild possum populations. Researh has shown that delivery of vaine as an aerosol to the respiratory trat and onjuntiva of the possum protets against disease (Corner, pers. om. 2001). Computer modelling, based on the onept of herd immunity, suggests that when the proteted proportion of a possum population reahes 50 70% the disease an no longer persist within the population (Barlow, 1997). Vaination would probably be used in onjuntion with existing ontrol tehniques. The vaination of wild fox (Vulpes vulpes) populations against rabies using oral baits has proven an effetive method of disease ontrol in several European ountries (Aubert, 1996). In addition, it has provided a model on whih to base possum vaination. There are two key omponents required for the suessful aerosol vaination of wildlife. One is a simple form of aerosol dispenser and the seond is an aerosol an ontaining the vaine. The following hapter desribes the development of a self-setting aerosol vaine dispenser (vainator) to suit delivery to wild possum populations.

111 95 Part one: Pen trials of an aerosol vainator

112 96

113 OBJECTIVE The objetive of this study was to design and refine a self-setting aerosol vainator. To ahieve this, the study had three interdependent aims. 1. To study the investigative behaviour of possums toward novel objets. 2. To onstrut a vainator. This devie had to meet five essential design riteria:?? It must be attrative to possums?? It must be easily used by possums?? It must apply an aerosol spray to the fae of the possum. In pen and field trials, the use of aerosol ans ontaining dye required that the spray be direted at the head and shoulders of the possum to onform with animal ethis requirements.?? The devie must be simple?? It must be robust 3. To evaluate the influene of soial hierarhy on vainator use by aptive possums MATERIALS AND METHODS Pen trials were onduted between June 1998 and February Observations were onduted weekly for a period of approximately three hours. An observation period began when possums first emerged from their dens and finished at the onlusion of their first ative period. The observation pen was 12 x 15 m and 2 m tall. The floor was a mix of soil and vegetation. There were several raised wooden runways onstruted throughout the pen. Possums denned in hessian saks whih hung in two weather proof shelters within the pen. Possums were fed daily with a mixture of fresh fruit, vegetables, bread and ereal-based food pellets and water was available ad lib. A small hut funtioned as part of the pen wall and had a large hath whih opened into the pen for weather proof observation. Observation equipment onsisted of two small floodlights mounted on the pen roof. These were overed in red ellophane and powered by a 12 volt battery. Light was

114 98 foused on approximately one quarter of the pen area where the ground was gently sloping and free from obstales. Novel objets and vainator designs were presented in this area. There was also suffiient light provided to view possum ativity in ninety perent of the pen. Observations were reorded by video amera and written notes. An identity ollar positioned a refletive number above the head of eah possum, enabling aurate identifiation of individual possums at low light levels. The ollar onsisted of a plasti zip-tie, 8 mm wide. Attahed to the end of the zip-tie mast was a thin plasti plate 45 mm x 45 mm. Attahed to both sides of this plate was a number or letter, ut from refletive adhesive. A small fishing sinker (15 g) was also attahed to the zip tie. This ated as a ounter weight, keeping the refletive tag above the head at all times. One an identity ollar was applied, the rathet tightening mehanism of the zip tie was wired in plae ensuring that the ollar ould not be tightened if it beame aught, for example in the wire netting of the pen. The soial interations of aptive possum olonies were observed during their normal nightly routine and also during the investigation of various novel objets and vainator designs. The spetrum of interations were ranked in a system based on the study by Oldham (1986) presented below. 1. Avoidane. Atively moving around the spae of another animal to avoid a onfrontation. The radius of this spae was 1 m. 2. Appraisal. A nose to nose, or sniffing appraisal of another animal with no other obvious physial interation. 3. Threat. A defensive position using one or both forepaws, this may be aompanied by aggressive sreehing and the oasional light uff, or swipe whih is not reiproated. 4. Attak. Inludes a pouning movement, often from the side or bak of another animal. Biting and grabbing, often aompanied by some form of hase as the subordinate animal attempts to esape. 5. Fight. A fully reiproated attak where both ombatants bite, srath and kik, often as they roll around on the ground. Voalisations vary, but are often grunts and sreehes. 6. Compliant. Where two or more possums oupy the same spae and/or objet without any aggression. This is distint from appraisal as eah party has been identified with no ongoing investigation. The attention of individuals is foused on something other than possums.

115 99 Four possum olonies were used. There were fourteen possums in three olonies and twelve possums in the fourth. Three olonies were entirely omposed of males and one was largely female with a single male. Eah olony was observed on average seven times over eight weeks. Choolate was used as an attratant during all other observations exept during the first three observation periods and presentation of the first vainator design, when no attratant was inluded. Cinnamon powder was used to dust the seond vainator design presented, but not used after this. Attratant was positioned to be inaessible to investigating possums and was seldom reovered. During the first three observation periods no objet was presented. Following this, novel objets of gradually inreasing omplexity were presented. This series began with a heavy plasti ube with sides of 200 mm and progressed to inlude a buket, a hair, a gumboot, one metre lengths of pipe, 15 m or 25 m diameter, used as tunnels, with both open and losed ends and finally a ball made from wire mesh, ontaining attratant, hung 90 m above the floor of the pen. During the seond series of observations thirteen variants of vainator design were tested. The degree to whih eah vainator met the design riteria and areas whih required modifiation were determined by observation of possums interating with the various designs. Vainator designs were onstruted from 6mm steel rod and 25 mm x 3 mm steel plate. Additional materials inluded transparent perspex, wire mesh, 2 mm galvanised iron, eletrial double plug boxes, light springs, 2mm aluminium plate, tie wire and assorted fasteners. An oxyaetylene gas set was used for brazing and welding RESULTS Observations of possums exposed to novel objets Observations over a total of 37 hours were used to reord the behaviour of aptive possums in response to novel objets, before the first vainator was presented. The first three observation periods were used to study possum behaviour in a familiar environment without a novel objet. Possums were divided into three broad ategories depending on overall behaviour. The first ategory ontained most possums (70%), and was haraterised by

116 100 ative investigation of surroundings, partiularly food bowls and sent marking areas. Possums in the seond ategory (20%) were less ative and would remain stationary for up to 1 hour, at favoured positions on runways. Visits to food bowls were brief with a small amount of exploratory behaviour followed by a return to the favoured sitting area. These possums were alert and observant of the ativity and voalisations of other possums within the pen. There were few possums in the third ategory (10%). They were ative, though showed very little investigatory behaviour. These individuals would spend muh of the observation period moving around the perimeter of the pen or paing up and down a runway. All possums were strongly attrated to large flying insets, iadas in partiular, whih were aught and eaten.?? Category one: possums showing ative investigation of surroundings and novel objets.?? Category two: possums showing some investigation of their surroundings and novel objets, but also spend long periods stationary.?? Category three: possums ative but displaying minimal investigative behaviour of novel or familiar objets, and repeated paing of runways, or irling of the pen perimeter. Presentation of a novel objet initiated a lear pattern of investigatory behaviour. Possums in the first ategory would be the first to investigate the objet. The possums would approah the objet in a largely repeatable order. When most of the first group had ompleted their inspetion and had moved away, possums in the seond ategory would approah. Possums in the third group would rarely investigate novel objets. A heavy ube was the first objet presented. In most ases the investigation lasted less than thirty seonds. Investigation was primarily by smell as the possum irled around the objet before moving away. On some oasions an individual would briefly hew a orner of the ube. An inverted plasti buket was the seond objet presented and this resulted in muh more intensive attention (maximum 10 minutes) shown by all investigating possums. Possums in ategory one quikly learned the most effetive way to reover the attratant, whih was plaed under the buket. A weight plaed on top of the buket ensured that possums ould no longer reover the attratant. The buket ontaining the attratant was plaed inverted on the seat. This arrangement demonstrated the willingness of possums to limb on novel objets while exploring them.

117 101 Investigation was thorough and inluded smell and taste while seeking the attratant. The hair was frequently tipped over but possums were not deterred by the oasional fall. Presentation of a gumboot and two piees of pipe showed that possums will investigate small openings. The more ative individuals in partiular forefully pushed their way into the onfined spaes. When pursuing the attratant in a gumboot, at times only the bak legs were visible, slowly propelling the gumboot and possum aross the ground. The 15 m diameter pipe was not quite large enough for the bigger possums to enter, however they would attempt to squeeze in as far as possible. All possums ould easily enter the larger pipe. Approximately 20% of studied possums walked through the pipe and would frequently pause in the middle while sniffing the inside of the pipe. These possums were all in ategory one. Possums in ategory two and three were muh more apprehensive when exploring small spaes ompared with open objets suh as the hair. A wire mesh ball ontaining attratant was suspended 90 m above the pen floor. It ould be seen and smelled by the possums, but was out of reah. Many possums would sit underneath the ball and make upward grasping movements with both paws. In one ase a possum repeatedly limbed a post 1.5 m away and made grasping movements from the post. During the observation period up to five possums would sit or irle beneath the ball. In the last fifteen minutes, one individual jumped straight up and grasped the ball whih detahed from the string and fell to the ground. During a seond observation period using the same objet no possums were seen to jump for the ball General observations on the approah and investigatory behaviour of possums In summary, the basi behaviour of individual possums was onsistent aross the four observed olonies. General patterns observed in response to novel objets are listed below. 1. There was a range of investigative ativity between individuals. Some will diretly approah and inspet a novel objet, while others atively avoid anything new and maintain a distane of at least one metre from the objet. 2. During an investigation a possum s ears were priked forward, with onstant and sometimes loud sniffing as it approahed the objet. 3. A possum would sniff and inspet the ground along the path of approah as it moved toward a novel objet.

118 The sent of other possums on a presented objet appeared to redue wariness of investigating individuals, even when the sent was from a different group of possums. 5. Some possums would approah a novel objet several times before apparently beoming apprehensive and moving away. 6. A possum may streth it s head and nek forward to smell a novel objet, making its body low to the ground. This allowed a rapid withdrawal from the objet if required. 7. Possums often made a preliminary inspetion by irling the objet from a distane of m, before ommening loser investigation. 8. A lear esape route appeared to redue the level of apprehension in investigating possums. Steep or rough ground and the presene of other obstales whih ould not be readily limbed seemed to inrease apprehension. 9. Possums appeared to use a ombination of senses during investigation, in estimated order of importane they were smell, taste, sound and sight. 10. The degree of apprehension shown by a possum when approahing or investigating a novel objet appeared to inrease when other possums were within approximately 1.5 m. 11. Possums seemed to hek loations where attratant had previously been found when presented with a familiar objet. These loations would be heked up to five or six times during total inspetion. 12. In eah olony of possums, approximately eighty perent would investigate a novel objet, however, approximately twenty perent would not approah the area used for presentation of objets. Avoidane behaviour did not hange when observations were onduted without artifiial light, or when no novel objet was presented The influene of dominane hierarhy on the investigation of novel objets by possums Four aptive possum olonies were used to observe soial interations over a total of 58 hours. Observations were reorded when there was no novel objet present, and when a range of objets inluding vainator designs were presented. Observations were ondensed to highlight a number of trends whih are listed below.

119 Interations between two or more possums appeared to begin when their individual spae overlapped. The size of this spae was approximately one metre and was determined by the maximum distane at whih one possum would reat to the presene of another. 2. Eah possum olony ontained one learly dominant animal. The degree of aggression shown toward subordinate possums seemed to vary between olonies. 3. The hierarhy below the dominant animal (middle order animals) was subjet to hange. The outome of most middle order agonisti interations was dependent on the strength of position of the individuals speifi to that enounter. For example, the suessful possum would have a tatial advantage, suh as attaking the hind quarters or from an elevated position. 4. In two olonies there was one partiularly small possum. In both ases it was at the very bottom of the soial order and displayed affiliative behaviour suh as food sharing and nose to nose sniffing. In the other two olonies no possum was learly at the bottom of the soial order. 5. During most agonisti interations a low level of aggression was suffiient to determine the dominant animal, for example a threatening posture, hiss, pouning motion or short hase. During 58 hours of observations only three interations were learly lassed as fights. 6. The soial ativity of male olonies and female olonies was different. Up to six males were observed investigating an objet simultaneously with muh overlap of individual spae but only a small amount of low level aggressive behaviour. Females demonstrated less overlap of individual spae, more frequent agonisti interations and a higher level of aggression during these interations. As a result, a maximum of two or three females would investigate an objet simultaneously. 7. The dominant possum was first to inspet anything new in three of the four groups. Subordinate animals were not permitted to approah until the dominant animal had ompleted a thorough inspetion. At the onlusion of investigation the dominant animal would sometimes remain near the objet, in aeptane of the presene of investigating subordinate possums. Other times it would move away at whih point subordinate possums would approah. In the fourth group one middle order possum would repeatedly be first to investigate anything new, but would quikly be hased off by the approahing dominant animal. 8. Over the ourse of an observation period all possums attrated to an objet would have the opportunity to inspet it. In some ases, groups of up to six possums would investigate

120 104 simultaneously. On the other hand, some subordinate animals would wait for several hours to allow a solitary investigation with minimal disturbane. 9. Possums seemed highly aware of other individuals when approahing an objet, and subordinate possums appeared more apprehensive when making an approah in lose proximity to a dominant animal. 10. Subordinate possums approahed a dominant animal toward the side or hind quarters where possible. They appeared to avoid approahing the head of the dominant animal diretly.

121 Pen trials of an aerosol vainator Introdution The initial onept was to design and build a self-setting aerosol delivery devie (vainator), triggered by a possum. This onept was then onverted into five design riteria based on the funtions the devie was expeted to perform (listed in setion 2.0). A literature review provided information on the methods used for aerosol vaination. Aerosol generators desribed were of two broad types. Those omprising the first type were simple, hand held atomisers (Valois Spray atomisers, Douglas Pharmaeutials Limited, Aukland, New Zealand) or devies for intranasal instillation (Fort Dodge Pinnale I.N., Paifivet Limited, Christhurh, New Zealand). These produts were designed for manual operation and used when handling sedated animals. Those omprising the seond type were large, ompliated and extremely expensive mahines. For example the Airborne Infetion apparatus (Middlebrook, 1961), Turbair Vaanair apparatus (Latif et al., 1981) and the ultrasoni nebulizer (Lagranderie et al., 1993). Neither of these two types provided a pratial starting point from whih to develop a mass produible, self-setting aerosol vainator suitable for use in the wild. An aerosol delivery system has been designed for vaination of pigs (Brown et al., 1997). Parts of this design were relevant to the urrent projet, for example it was propellant driven and designed to deliver aerosol droplets of 5? m ontaining bateria, to the respiratory trat. Unfortunately the devie was hand held, hand operated, mehanially omplex and deliate. Finally, it required batteries and restraint of the animal. These fators made the devie unsuitable for the self vaination of wild possums. The five design riteria were used to onstrut the initial prototype whih was based on a modified age trap. The developmental proess began from there and was direted by observation of aptive possums interating with presented designs during pen studies. All designs inorporated an attratant of hoolate whih was inaessible to possums in most ases. The end point of development during the pen trial phase was repliation of the final design of the vainator in preparation for the first field trial. The first design was onstruted in November 1998 and modifiations were tested weekly until February Type 1 MkI The initial design was developed in onsultation with Rod MDonald and Joanne Short of HortResearh, Ruakura, Hamilton, New Zealand and Leigh Corner, EpiCentre, Massey University, Palmerston North, New Zealand. It onsisted of a walk through passage with one way

122 106 entry and exit doors. Where the aerosol an was ativated by the weight of the losing entry door. Charateristis of the initial prototype are presented below.?? A live animal trap was onverted into a retangular passageway, with one way entry and exit doors.?? It was a 240 mm square passage, 500 mm long.?? The weight of the entry door falling shut triggered the aerosol an by way of a wire loop from the door to the an nozzle. A plasti guide pressed on to the top of the aerosol an, kept the wire trigger in ontat with the valve of the an.?? Within the passage way was an enlosed spae whih housed the trigger and aerosol an. It was aessed through a small door in the roof.?? The doors and roof were made of sheet metal and the sides and floor were wire mesh. Aerosol an seletion The primary riterion for an seletion was a design urrently mass produed and requiring a minimum amount of energy to ativate. There are many shapes of an available, but essentially only two types of valve. These are vertially depressed and tilt ation valves. The sideways movement of a tilt ation valve requires onsiderably less fore to ativate than a vertially depressed valve. The long stemmed nozzle used for vertial release of tilt ation valves provided more leverage than nozzles designed for the horizontal release of these valves. It was also easier to attah a trigger to the long stemmed nozzle. The aerosol an seleted was the Airozone brand, 175 grams, using hydroarbon as propellant and produed by Rekitt and Colman, Aukland, New Zealand. Observations The initial prototype was not attrative to possums. The most frequent response (10/14 possums) was to avoid the area ompletely, or pass in an ar around it. Four possums in ategory one ame within 1.5 m of the devie, stopped and sniffed in its diretion before abruptly hange in diretion to avoid it. Before the seond trial, the prototype was thoroughly leaned with soap and water followed by dusting with innamon. Possums were attrated to the traes of innamon, sniffing more strongly about these areas. However, the entry door whih had to be pushed open ated as a barrier. Possums moved around the devie, inluding sitting on top but made no effort to push against any

123 107 part. On oasion a possum would move as lose to the bait as possible and reah through the mesh in an attempt to rake it loser. Conlusion During the first trial the main problem appeared to be the smell of either the vainator or the aerosol an. The an was briefly disharged in a test prior to the beginning of the observation period. However, one possums were attrated to the devie, it beame lear that they would not press against a barrier to gain entry. The whiskers about the nose, eyes and fore-paws of the possum are a highly sensitive method of identifying barriers during noturnal ativity. The fore required to push the door open omes from these sensitive areas, whih identify the door as a barrier. Observations with novel objets have shown possums will push their way into various things, for example a gumboot or narrow pipe. It was onluded that the important differene was the tapered entry into the boot, ompared with the flat surfae represented by the vainator door. The first prototype met the first design riteria, being attrative to possums, but failed the seond of useability, due to the onept of entry by way of a door Type 1 MkII Figure 20. Type 1, Mk II vainator

124 108 Changes?? Steel doors were replaed with transparent, slightly lighter perspex opies.?? The innamon was removed.?? Choolate was used as an attratant. Observations Possums were not enouraged to push against the door by being able see through it. Possums sniff with similar intensity around the whole of the devie. Conlusion It was lear that possums were not prompted to push against the door by being able to see through it. The fous of attention was direted at the vainator, but required further diretion toward the door. With the urrent model the attratant diffused through all parts of the devie. It was possible that the solid perspex doors enouraged possums away from the entrane as the sent would have been stronger from the porous mesh sides. This model (Figure 20) did not satisfy any more of the design riteria than the initial prototype Type 1 MkIII Changes?? Perspex sides were added to diret the flow of the attratant toward the entrane of the devie.?? A trapezium shape was ut out from the base of the entry door large enough for muh of a possum s head inluding the nose and eyes, to fit through. Observations Possums would follow the sent along the bottom of the perspex plate to the end of the vainator. The gap in the door was sniffed with greater intensity than the sides of the devie, presumably beause this was where the sent was strongest. Although possums frequently sniffed at the door, it was not touhed during the observation period.

125 109 Conlusion The shaped ut outs were designed to ahieve two goals. Firstly, they would make entry into the vainator easier by enabling possums to push on the door with the less sensitive area of the heeks, rather than the nose. This goal was not ahieved. The seond goal was to diret possums to the entrane of the devie by fousing the sent of the attratant out through the entry door. This goal was ahieved. Reports from other researhers (Short, pers. omm., 1998) indiate that possums would repeatedly use a swinging door, hinged from the roof, as used in the first three vainator designs. In ontrast to this information, aptive possums in this study showed a strong aversion to pushing on any type of flat barrier. Consequently development of the Type One line and the deision was made to generate a ompletely new design Type 2 MkI The Type Two vainator was based on a ombination of the simplest possible omponents required to ahieve the design riteria. These are an aerosol an, a mehanism by whih the possum will ativate the an and a basi frame on whih to mount the omponents (Figure 21).

126 110 Figure 21. Type two, Mk I vainator Changes?? The frame onsisted of a slightly tapered retangle, 500 mm long with an open entry end 350 mm square and a losed end of 240 mm square. The floor, sides and losed end were overed in wire mesh.?? The hinge point of a pressure plate was loated 200 mm from the losed end of the devie. The pressure plate was onneted by wire to the trigger of the aerosol an.?? The aerosol an was mounted in the top left orner above the pressure plate in a horizontal position. The spray was direted through the wire mesh and out of the devie suh that it would not ontat the possum.?? Attratant was loated under the pressure plate at the bak of the devie.

127 111 Observations Investigation was thorough though limited to the most dominant possum. After irling the devie twie the possum approahed the entrane and moved inside. Within the first hour this possum had disharged the aerosol an four times. This possum did not permitted any others to approah the devie during the observation period. When the an was first disharged the possum triggering the devie showed a strong fright reation, inluding a sudden jump and rapid esape. Three nearby possums showed a similar reation. During the third and fourth disharges of the an the degree of fright shown by the triggering possum and nearby possums was greatly redued. On three oasions the possum did not walk far enough onto the pressure plate to disharge the an. Conlusion This devie was attrative to possums and easy to use while also being extremely simple. It essentially satisfies three of the design riteria. The riteria of spray appliation to the possum may now be addressed, after onfirming that the devie is attrative to possums and easily used. Inreasing the length of the pressure plate would allow the possum to move further on to it, thereby inreasing the fore exerted on the nozzle of the an. Disharge of the an aused a similar degree of fright in possums lose to the vainator and the possum inside the devie. This indiated that either the sound or smell of the disharge was frightening to possums, rather than movement of the pressure plate. There now appeared to be two areas on whih to fous. One was repositioning the an to diret the jet of spray toward the possum. The seond was to devise a way of keeping the possum in the devie for long enough for the spray to be inhaled Type 2 MkII Changes?? A mounting frame for the an was made and attahed to the bak wall of the devie. When the an was attahed to the frame, the nozzle aimed diretly at the head of a possum on the pressure plate.?? The most diffiult hange was to onvert the vertial pull of the pressure plate into a horizontal fore required to ativate the an when in a vertial position. A very light steel able (2.5 mm thik) was used to hange the fore from vertial to horizontal. This passed

128 112 vertially up from the pressure plate, through a 90 degree angle about a pivot point, then horizontally to the valve.?? A six way adjustable hinge point for the pressure plate was added to vary the position of the plate with respet to the nozzle of the an and vary the fore exerted on the an nozzle.?? The pressure plate was lengthened by 30 mm to 150 mm.?? Between 100 mm to 200 mm from the rear of the devie additional taper was added to more aurately guide the possum underneath the an (Figure 22). Figure 22. Type two, MkII vainator showing taper above the pressure plate Observations This model delivered a dose of spray to the bak of the head and shoulders of the possum. However, the angle through whih the triggering able passed generated exessive frition. The result of this was errati disharge of the an. In addition, after the possum had left the pressure

129 113 plate the an would ontinue to disharge as normal for up to five seonds, then disharge as a small leak for a further 30 seonds. Heavy and light possums would trigger the an at different positions on the pressure plate. In extreme ases, a heavy possum would not have moved under the nozzle before it disharged and light possums would fail to set the devie off at all. This problem is ompounded by the weight and flex inherent in the pressure plate. Two of fourteen possums were sprayed. Conlusion The basi arrangement of omponents in this model worked during testing by hand, but observations showed the need for further refinement. By this stage the important requirements of vainator design were beginning to emerge. Primarily, a design was needed where the possum would enounter and aept the various novel parts individually, as it moved into the devie. For example, the basi frame and entrane are enountered during initial investigation. Entry allows aeptane of the roof and walls before enountering the pressure plate, taper to diret the head, and finally the nozzle of the an. The gradual presentation of omponents redued the degree of fear during investigation. As a onsequene there was more hane of the possum investigating the devie and less hane of it baking out before ativating the an. A seond requirement was that the proportion of weight ontributed by the possum to set the devie off, relative to the weight of the pressure plate and triggering system, must be as great as possible. As this proportion inreases, the range aross whih the leverage ratio of the triggering mehanism an be manipulated also inreases. Essentially this allows more aurate adjustment to ensure the devie is triggered by the widest possible weight range of possums. A third requirement was to minimise flex in the pressure plate, triggering mehanism and an holder as muh as possible. Minimising flex was neessary to aurately identify the types and magnitudes of fores operating during ativation of the devie. On the basis of this, a new design for the trigger was required whih was not able operated. There was also a need for a lighter, more rigid pressure plate Type 3 Mk I The basi design of the Type Three was a tapered retangle, similar to Type Two, but slightly smaller. The interior of the Type Three was divided into a front and rear hamber. The front

130 114 hamber inluded the entry, pressure plate and taper to diret the head. The rear hamber provided an empty spae and position to mount an attratant ontainer. The purpose of the empty spae was to redue the degree of onfinement imposed on a possum as it moved further into the devie, while still guiding the head under the an nozzle. Confinement aused apprehension and subsequent avoidane in some possums during earlier observations. Attahment of a ontainer for attratant enouraged possums to move aross the pressure plate until the devie was triggered. Changes?? The aerosol an was positioned above the head of the possum to more aurately target the head and shoulders of the possum.?? Size of the entry was redued from 350 mm square to 300 mm.?? The hinge point of the pressure plate was moved slightly loser (15 mm) to the entrane.?? Spae between the pressure plate and rear wall was inreased.?? The tapering to diret the possum s head was moved slightly loser (15 mm) to the entrane.?? The opening between the taper for direting the head of the possum was enlarged slightly.?? Ativation mehanism:?? The new ativation mehanism was basially a lever and fulrum (see-saw) design. Fore was exerted on one end of the lever by the weight of the possum and pressure plate. The fore was transmitted aross the fulrum to the other side of the see-saw whih ated on the nozzle of the an. Lever length on the nozzle side of the fulrum was shorter than the pressure plate side, this redued the degree of movement but inreased the fore applied to the nozzle of the an.?? A light spring was soldered on to the framing between the pressure plate and the floor of the devie. This was to offset the weight of the pressure plate whih previously rested on the nozzle of the an via the trigger. Observations The redued entrane size had no effet on the ease with whih possums entered the devie. Possums moved up onto, and further aross, the pressure plate with less hesitation. The spring rate was orret and the lever and fulrum onept was effetive. Overall, the trigger mehanism was a suess.

131 115 Small nothes had been worn on the nozzle of the an ausing the trigger to tilt the nozzle in a series of short, sharp movements, ausing errati disharge of the an. There was a small amount of flex evident in the mounting frame for the aerosol an. This greatly ompliated setting the leverage fore required by the an. The vainator was triggered three times by one possum and one by a seond. A third possum triggered the devie without being inside, by treading on the edge of pressure plate whih extends out through the side of the age. Disharge of the an aused the individual inside the devie and all possums within 1.5 m to flee. However, on two oasions, the first possum to return and reenter the devie was the one whih had previously been sprayed. One possum was partiularly aggressive and on two oasions worked either a paw or his nose under the side of the devie and then lifted upwards tipping it over. Six of fourteen possums were sprayed. When left in the pen overnight, the vainator was moved approximately five metres down a gentle slope (probably rolled) by investigating possums. Conlusion Attrativeness to possums was not affeted by the redution in size of the devie. The reason possums moved more easily onto and further up the pressure plate appeared to be beause they were not approahing the bak wall so losely. The apparent possibility of moving into the seond hamber lets them move under the valve more easily. Possums were unable to enter the seond hamber beause they must first ross the pressure plate, ausing a disharge of spray whih frightened them out of the devie. Additional support of the an holding mehanism and trigger were required to redue the amount of flex during an ativation. To fulfil the riteria of robustness an anhoring system was required to keep the devie upright and stationary Type 3 MkII Changes?? The mounting frame for the aerosol an was strengthened.?? The pressure plate end of the trigger arm was lengthened slightly to improve leverage.?? The wire loop attahing the trigger to the nozzle of the an was removed.

132 116?? A small plate was added to the end of the lever system to inrease surfae ontat of the trigger with the nozzle.?? The spring was moved from under the pressure plate and re-attahed from the trigger lever to the vainator roof. In this position the spring would still offset the weight of the pressure plate and also allow the resting position of the trigger to be set to minimise the resting fore on the nozzle of the an (Figure 23).?? Lateral (undesirable) movement of the triggering lever was redued by onfining movement at the fulrum.?? Peg holes were added to the orners of the vainator enabling it to be anhored to the ground. Figure 23. Trigger mehanism and spring (arrowed) on Type three MkII vainator

133 117 Observations The unusually aggressive possum was first to investigate. Investigation entailed grasping and shaking the whole devie, whih was pegged down. Possums regularly gathered at the bak of the vainator. There was more spae than neessary for a possum to move in the first hamber. This group of possums had now been exposed to three models of vainator and two in partiular were beoming adept at stealing the attratant. Some possums were seen to hew, or grab and wrenh the trigger and nozzle. Six of fourteen possums were sprayed. Conlusion The sent of the attratant should be direted through the entrane of the devie. The size of the first hamber ould be redued slightly. It was good for orienting the possum toward the pressure plate, but this funtion was now effetively performed by the additional taper between the two hambers. Some form of attratant ontainer was required. It should be fixed lose to the roof to guide the possum s fae as lose to the an nozzle as possible. An easy method of aess to this ontainer must also be onsidered. Some form of protetive struture was required around the trigger and an nozzle to stop damage aused by investigating possums, partiularly to the soft, plasti, removable nozzle Type 3 MkIII Changes?? The overall length was redued from 500 mm to 360 mm, with the first hamber shortened by 120 mm.?? The height of the vainator was redued from 300 mm to 280 mm.?? A perforated attratant ontainer was added. It measured 80 x 100 mm and 45 mm deep and attahed to the bak wall, just below the roof. A small hath in the roof of the vainator provided easy aess to the ontainer.

134 118?? The pressure plate was lengthened by 10 mm. The end opposite the hinge was bent up slightly to disourage larger animals from stepping over the plate into the seond hamber.?? Six shaped perspex plates enlosed the rear hamber, replaing the wire mesh. Silion putty was used to seal the joins between plates. Attratant ould now esape only aross the pressure plate and through the entrane of the devie.?? The perspex would also funtion as protetion for the trigger, but did not protet the nozzle.?? This model was initially tested without an aerosol an in order to reord the behaviour of possums when no frightening stimulus resulted from stepping on the pressure plate. Observations Possums would atively explore all parts of the vainator interior in searh of the attratant. Movement of the pressure plate aused by the weight of the possum did not ause alarm. The attratant ontainer was still aessed by the aggressive possum at the end of the observation period. An investigating possum was fored to arh its bak sharply in order to reah the attratant ontainer, indiating that the ontainer required reloation in a more aessible position. Six of fourteen possums were sprayed. Conlusion The disharging an was learly responsible for the fright shown by an investigating possum when the vainator was triggered. Previous observations had shown the sound of the an disharging, rather than the stream of spray, aused the most vigorous reation in possums. The frightening effet of the an was important in ensuring a possum left the vainator before it ould explore and possibly damage the moving parts. Some modifiations were required to the attratant ontainer to make it more approahable and seure Type 3, MkIV Changes?? The trigger mehanism was hanged to operate along the midline of the devie. Previously it was approximately 45 degrees off the midline.

135 119?? A 40 mm square plate was used to mount a fulrum for the trigger. This allowed exellent adjustment of the shaft to find the position where the trigger worked most effetively.?? The point where the attratant ontainer attahed was shifted to the midline of the vainator and lowered slightly. Observations A new olony of possums were used in the first observation period and their investigation was entred on the general pen environment rather than the vainator. Twelve of the fourteen possums approahed the vainator entrane, although it the was ativated less frequently than in previous sessions. Possums would frequently approah and briefly step on to the pressure plate, but then bak out of the devie, without triggering the an. From this, it was assumed that the opening from the end of the first hamber into the seond was too small. The framing to whih the trigger attahed was also making the opening seem small. The attention of possums was attrated to the trigger mehanism and an nozzle, rather than direted toward the attratant further inside. Two of fourteen possums were sprayed. Conlusion The opening between the two hambers required enlargement, both horizontally by manipulating the framing used to mount the trigger, and vertially by raising the position of the an and trigger. Presented with a larger intrane, possums should move more freely on to the pressure plate to ativate the spray Type 3 MkV Changes?? The plate, on whih the trigger was mounted, was redued in size.?? The trigger mehanism, an and holder were all raised 8mm.?? The width of the entrane to the seond hamber was also inreased slightly (8mm).?? Aerozone aerosol ans were replaed with speifially made ans ontaining four dye olours, made by Aerosol Produts, Aukland, New Zealand. Cans ontaining gentian violet dye were used for the remaining pen trials.

136 120 Observations Possums were muh less apprehensive about moving toward the attratant aross the pressure plate. The opening between the two hambers would appear larger if the plate to whih the trigger attahed was removed and the shaft attahed permanently. Gentian violet dye is an exellent marker. However effetiveness of the other three olours, whih are paint based, rather than a dye, was poor. Conlusion The hanges were effetive at enlarging the opening to the seond hamber, but this area would ideally be enlarged further still. It was now relevant to begin thinking about the impliations of making a number of opies of a vainator for trial in the field. Some adjustment would be required to ompensate for the small differenes in repliates. Inorporating adjustment into the attahment point of the an would allow the fine tuning of dupliates without the need to manipulate the intriate triggering mehanism. The fulrum of the trigger was found to be most effetive when plaed as lose to the aerosol an as possible and 25 mm above the nozzle. Gentian violet was the most effetive dye for marking possums Type 3 MkVI Changes?? The new an holding devie was based on a 250 mm length of 20 mm x 3 mm plate whih was bent to a V shape. The aerosol an was lamped to one arm of the V, while the other arm was lamped to the roof of the vainator.?? A strong lamp was welded to the roof of the vainator.?? Some modifiation of the roof was required to aept the new an holding design.?? An eletrial double plug box replaed the hand made bait ontainer. It was an ideal size, easy to attah and oarsely perforated.?? The plate used to fine tune plaement of the trigger pivot shaft was removed.

137 121 Figure 24. Can holding mehanism showing V shaped steel plate and lamp (arrowed) Observations The enlarged opening to the seond hamber, aused by the new an holding mehanism, improved possum aess. Adjustability of the V shaped an mount allowed the ontat between nozzle and trigger to be optimised. The nozzle of the an ould be adjusted in any diretion to suit the trigger. The new attratant ontainer was possum proof.

138 122 Figure 25. The final stage of development for the prototype vainator, Type 3, Mk VI Conlusion The urrent vainator (Figure 25) fulfilled the five design riteria. It was attrative to possums and ould be easily used. Marking the possums with dye was ahieved by direting the spray toward the bak of the head and shoulders of the possum. The devie was very simple and used only a lever and spring to dispense the spray. It was also robust, ould be firmly attahed to the ground and was adjustable to allow fine tuning of operation. At this point, 13 opies of the Type Three, MkVI vainator (Figure 25. The final stage of development for the prototype vainator, Type 3, Mk VI) were onstruted Type 3, MkVII (Steel opies) Three steel opies onstruted by Art In Iron (Palmerston North, New Zealand) were of almost idential design to the prototype. Minimal modifiations were required before they were suitable for field trials.

139 123 The ten opies of the final vainator design presented by Iron Village (Palmerston North, New Zealand) inorporated a very different trigger mehanism and an holding mehanism (Figure 26). Initially these devies would work effetively, although there was doubt about robustness and the degree of flex during operation, whih was muh higher than the prototype. Robustness of the repliates was tested during a preliminary field trial. The amount of use by possums was not investigated during this trial. After four nights only four of thirteen vainators were working properly, one of whih was the prototype. All opies were returned for modifiation. Figure 26. The an holding and triggering mehanism used in the twelve opies of the final vainator prototype Changes?? The steel pressure plate was replaed with an aluminium opy.?? The an holding mehanism was replaed with the prototype V design.

140 124?? The trigger mehanism was replaed with the prototype version of the lever and fulrum design.?? Lighter springs were used to suit the redued weight of moving parts. Observations Observations using aptive possums showed that modifiations had greatly improved the rigidity of the devie. Can movement was minimal and adjustment improved. The weight of moving parts was also redued. The trigger and an nozzle were muh more effetively proteted. Eight of fourteen possums were sprayed. Figure 27. One of twelve steel opies (type 3, Mk VII) of the final vainator design after the hanges listed in setion

141 125 Conlusion The vainators were now virtually idential to the prototype. Operation of a sample of three opies was satisfatory and all opies were now prepared for the first genuine field trial Type 4 MkI (Wooden dupliates) Two dupliates of the final steel prototype were onstruted from a ombination of wood and plasti (Figure 28). The aim of the opies was to ompare behaviour of possums toward vainators made from different materials and of different appearane. As with the steel opies, some modifiations were required, to ahieve an operating standard equivalent to that of the prototype, before pen trials ould ommene. Changes?? The trigger fulrum replaed with a stronger steel model to redue flex.?? The trigger was extended at the pressure plate end to inrease leverage.?? The wooden an holder was replaed with a V shaped an holder to redue flex.?? A spring was attahed from the pressure plate end of the trigger to the roof as used on the steel prototype.?? The steel pressure plate was replaed with an aluminium opy.?? The trigger mehanism presented worked at 90 degrees to the midline (aross the vainator). It was modified to operated along the midline of the devie. Observations Eight possums were found to approah and enter, or attempt to enter the wooden model ompared with twelve for the steel one. The steel devie has less flex, improving the auray of fine tuning. Small additions and subtrations to vainator design, or adjustment of angles is more easily aomplished with steel than with wood. Disassembly of the entire unit is required to aess some omponents of the wooden model, for example to enlarge the opening into the seond hamber. Srew holes fixing wooden parts have inreased flex when the devie is reassembled. Wood will swell, ontrat and warp in response to environmental onditions. Three possums were observed hewing the wooden sides. No possums were sprayed with dye.

142 126 Figure 28. One of two wooden vainators after the hanges listed in setion Conlusion The main onlusion was that wooden models appeared less attrative to possums than their steel ounterparts. In addition wood, as a onstrution material, had several qualities whih were undesirable for the preise prodution, attahment and fine tuning of small moving parts required by a vainator DISCUSSION OF PEN TRIALS A total of 75 hours were spent observing the investigative behaviour and soial hierarhy of four aptive possum olonies. A series of novel objets were presented first, followed by fourteen variations of vainator. The majority of aptive possums were found to be naturally urious animals and to display minimal avoidane behaviour toward novel objets. It was found that the best period for wathing was from when possums first emerged in the evening until the end of