Potential use of myxoma virus and rabbit haemorrhagic disease virus to control feral rabbits in the Kerguelen Archipelago

etal. CSIRO PUBLISHING www.publish.csiro.au/journals/wr Wildlife Research, 2004, 31, 415 420 Potential use of myxoma virus and rabbit haemorrhagic disease virus to control feral rabbits in the Kerguelen Archipelago B. D. Cooke A,D, J.-L. Chapuis B, V. Magnet B, A. Lucas B and J. Kovaliski C A Estación Científica Charles Darwin, Fundación Charles Darwin, Casilla 17-01-3891, Quito, Ecuador. B Muséum National d Histoire Naturelle, Département Ecologie et Gestion de la Biodiversité, USM 2699 Inventaire et Suivi de la Biodiversité et UMR 6553 Ecobio, 36 rue Geoffroy Saint-Hilaire, 75005 Paris, France. C Animal and Plant Control Commission, GPO Box 1047, Adelaide, SA 5001, Australia. D To whom correspondence should be addressed. Email: bcooke@fcdarwin.org.ec Abstract. Rabbits have caused enormous damage to the vegetation on seven islands in the sub-antarctic Kerguelen archipelago, including the main island, Grande Terre. Rabbit sera collected during 2001 02 were tested for antibodies against myxoma virus and rabbit haemorrhagic disease virus with a view to considering the wider use of these viruses to control rabbits. The results confirmed work done 15 20 years earlier that suggested that myxoma virus has not spread across all parts of Grande Terre and occurs at low prevalence among rabbits. By contrast, on Ile du Cimetière, where European rabbit fleas were introduced in 1987 88, the prevalence of myxoma antibodies is high and the rabbit population is relatively low, supporting the idea that the fleas are effective vectors of myxoma virus. Consequently, there should be benefits in releasing fleas on Grand Terre to enhance disease transmission. Reactivity of some rabbit sera in RHD-specific ELISAs suggested that a virus similar to RHDV may be present at low prevalence on Grande Terre but most rabbits are likely to be susceptible and this virus could be considered for use as a future biological control agent. WR03084 Myxomatosis B. D. Cooke and RHD in Kerguelen Introduction The Kerguelen Archipelago (48 25 50 S, 68 25 70 35 E) lies on the Antarctic convergence and, as a consequence, has a severe climate characterised by high winds and low temperatures (monthly means range from 1.9 C in July to 7.4 C in February). Despite this, European rabbits have persisted since their introduction in 1874. They are now distributed across much of the main island, Grande Terre, particularly in coastal areas below an altitude of 200 m and are also found on six smaller islands (Boussès 1991; Chapuis et al. 2002). The effect of rabbits on the original vegetation at Kerguelen has been rapid and severe (Werth 1911; Aubert de la Rüe 1964; Chapuis et al. 1994a). Plants such as the Kerguelen Cabbage (Pringlea antiscorbutica) and cushion plants (Azorella selago) have been eliminated over large areas and persist only in localities where rabbits cannot reach them, such as rocky cliffs and rocky plateaus at altitudes above 200 m. With the loss of mixed communities of cabbage and cushion plant from the islands shorelines there has been a major decline in habitat for burrowing seabirds (Weimerskirch et al. 1989). The loss of major plant communities, including deep-rooted Azorella, and their replacement with a herb field dominated by the relatively shallow-rooted Acaena magellanica (Chapuis et al. 1994a) also means that organic soil built up over time is not as well protected from wind and water erosion, especially in winter. The rate of degradation of soil has further increased because A. magellanica has declined as a consequence of more frequent summer drought during the last decade (Chapuis et al. 2004). Severe loss of soil is apparent in many areas and underlying clays and parent rocks support little vegetation. The rabbit population is also an important food resource for feral cats (Derenne 1976; Pascal 1980; Pontier et al. 2002). Rabbits and, secondarily, house mice (Mus musculus) help maintain the cat population between seabird nesting seasons (Chapuis et al. 1994a) and the removal of rabbits should reduce seabird predation by cats in the longer term. On this basis, it must not be forgotten that any attempt to limit rabbits should take into account the possibility that cats might kill more birds in the short term. There is clearly a case to be made for substantially reducing rabbit numbers at Kerguelen, but owing to the CSIRO 2004 10.1071/WR03084 1035-3712/04/040415

416 Wildlife Research B. D. Cooke et al. extreme climate and rough terrain, methods such as poisoning using anticoagulant (chlorophacinone) on grain baits have so far been limited to smaller islands where there is a high chance of eradication (Chapuis et al. 2001). Repeated poisoning on Grande Terre is simply not practical because of its size (6500 km 2 ) and, despite applying bait at carefully chosen times, this also adds to the risk of poisoning other species such as Eaton s pintail duck (Anas eatoni), the Dominican gull (Larus dominicanus), and the sheathbill (Chionis minor), by primary poisoning (Chapuis et al. 2001), and potentially the brown skua (Catharacta skua), by secondary poisoning. There has long been an interest in using biological control agents on Kerguelen and, indeed, myxoma virus was introduced to the archipelago in 1955 56 (Lésel 1967). Even though the usual arthropod vectors, rabbit fleas and mosquitoes, are absent the virus has nevertheless persisted in a highly attenuated form (Chapuis et al. 1994b). More recently, in 1987 88, the European rabbit flea (Spilopsyllus cuniculi) was introduced experimentally to Ile du Cimetière (3.1 km 2 ) and became widely established across that island within several years (Chekchak et al. 2000). The research described in this paper provides further information useful in considering how biological control agents might be used. Our main objective was to gather information on the seroprevalence of antibodies to myxoma virus as this would indicate how widespread the virus has become and how frequently it affects rabbits. Collecting data from selected sites such as Ile du Cimetière also provided a way of checking whether European rabbit fleas would be likely to increase the prevalence of myxomatosis if they were more widely released across the archipelago. Additionally, the recent spread of rabbit haemorrhagic disease (RHD) through rabbit populations across the world (Cooke and Fenner 2002) also showed that rabbit haemorrhagic disease virus (RHDV) was worth considering as a potential biological control agent for use in Kerguelen. Checking rabbits for specific anti-rhdv antibodies was a simple additional task that might indicate whether RHD might be used as a new biological control agent. Methods Small samples of rabbits were collected from November 2001 to March 2002 from Ile du Cimetière and from a range of widespread localities on the main island (Fig. 1), in particular Baie de l Observatoire, Fjord Bossière, Port-Christmas and Val du Retour where no myxoma antibodies had been observed in rabbits collected in 1983 88 (Chapuis et al. 1994b). Rabbits were obtained using a shotgun usually at night with the aid of a head torch, or in the late afternoon. Immediately each rabbit was confirmed dead, the chest cavity was opened and a sample of blood drawn from the heart using a needle and vacuum tube (Vacutainer). Each sample was individually labelled and the date and location were also recorded. Some rabbits were dissected to obtain reproductive data and weighed. On Ile du Cimetière, rabbits were combed using a small flea comb and fleas were collected in an aspirator and counted. 49 N 49 30' 0 10 20 30 40 50 km 69 8 Péninsule Loranchet 10 Péninsule Rallier du Baty Calotte Glaciaire 1. Anse du Pacha (PAF) 2. Baie de l'observatoire 3. Estacade-Ratmanoff 4. Fjord Bossière GRANDE TERRE 69 30' 1850 m 4 5. Ile du Cimetière 6. Pointe du Morne 7. Pointe Molloy 8. Port-Christmas 11 Port-aux-Français 6 1 7 The blood was allowed to clot and was then spun in a small bench-top centrifuge to obtain clean serum samples. Subsequently, the samples were analysed in the laboratory of the Animal and Plant Control Commission, Adelaide, South Australia, using ELISAs to detect antibodies to myxoma virus and RHDV. For antibodies to myxoma virus the methods described previously by Kerr (1997) were used. Results were expressed as either positive or negative, and no equivocal samples were found. For antibodies against RHDV, a competition ELISA and an IgG isotype ELISA were used as these are most sensitive and also have the potential to indicate the presence of non-pathogenic viruses closely related to RHDV (Cooke et al. 2002). The sera were also screened for anti-rhdv-specific IgA and IgM antibody isotypes following the methods of Capucci et al. (1995), as adapted by Cooke et al. (2000). Although competition ELISA results can be expressed as a reciprocal titre, for the purpose of screening large numbers of samples the results are usually expressed in terms of percentage inhibition of a 1:10 serum dilution and for the Kerguelen samples the cut-off between positive and negative sera was taken as 50% inhibition. Samples with higher percentage inhibition were considered further according to isotype ELISA results. Serum samples that were positive in IgG ELISA were titred and results expressed as a reciprocal titre. Results Prevalence of antibodies against myxoma virus Excluding data from Ile du Cimetière, 8 of 72 rabbits were seropositive to the myxoma virus. On the main island, seropositive rabbits were observed at four localities: Anse du Pacha, Pointe Molloy, Pointe du Morne and Port-Couvreux (Fig. 1; Table 1). By contrast, 7 of 9 rabbits shot on Ile du Cimetière were seropositive. Despite the small sample, the 9 ILES KERGUELEN 2 70 Péninsule Courbet Golfe du Morbihan 5 Rabbit distribution (2003) 70 30' 9. Port-Couvreux 10. Val du Retour 11. Val Studer Fig. 1. Location of studied sites and rabbit distribution in the Kerguelen Archipelago. 3

Myxomatosis and RHD in Kerguelen Wildlife Research 417 Table 1. Number of rabbits in which myxoma antibodies and RHDV antibodies (IgG reactivity) were detected in 2001 02, compared with data from 1983 88 The 1983 88 data, for myxoma antibodies only, are from Chapuis et al. (1994b) Location 1983 88 2001 02 Sample size rabbits Myxoma antibodies Sample size rabbits Myxoma antibodies RHDV IgG antibodies Grande Terre Anse du Pacha (PAF) 145 4 8 1 0 Baie de l Observatoire 58 0 6 0 0 Estacade-Ratmanoff 54 3 10 0 1 Fjord Bossière 69 0 5 0 0 Pointe du Morne 2100 80 8 1 0 Pointe Molloy 882 35 8 3 0 Port-Christmas 31 0 6 0 0 Port-Couvreux 96 26 8 3 0 Val du Retour 35 0 5 0 0 Val Studer 199 5 8 0 1 Ile du Cimetière 246 106 9 7 0 Total 3915 259 81 15 2 Table 2. Details of rabbits shot on Ile du Cimetière Note the generally higher numbers of fleas on the female rabbits Rabbit no. Sex Bodyweight (g) Reproductive state Estimated no. of fleas 43 F 2225 Lactating, not pregnant 20 44 F 2275 Lactating, recently parous 350 47 F 2375 Lactating, not pregnant 200 48 F 2300 Not lactating or pregnant 200 49 F 2175 Lactating, recently parous >200 50 F 2100 Lactating, 22 days pregnant >200 45 M 2125 Testes down, large 10 46 M 750 Testes abdominal 10 51 M 950 Testes abdominal <10 proportion of rabbits seropositive for myxoma antibodies on Ile du Cimetière clearly differs from that seen generally on the main island. If all data are analysed together, including data from Ile du Cimetière, a significant difference between samples is shown (G adj = 30.42, G crit = 18.31) whereas there are no significant differences when the sample from Ile du Cimetière is excluded (G adj = 13.67, G crit = 16.92) (Sokal and Rohlf 1995). The two seronegative rabbits from Ile du Cimetière were the only juveniles shot in the sample. On this island, all rabbits collected carried fleas: young and adult males carried 10 or fewer fleas and adult female usually carried 200 fleas or more, except for one female with 20 fleas (Table 2). The three females most heavily infested were recently parous or pregnant. Prevalence of antibodies detected by RHDV-specific ELISAs The sera of 2 of 81 rabbits reacted positively in RHD IgG ELISAs at titres of 1:40 and 1:80. Using competition ELISA, four samples showed percentage inhibition exceeding the 50% cut-off value (52, 52, 53 and 69%); but with no supporting evidence of antibodies from isotype ELISAs they were considered negative. No additional reactivity was detected in any rabbit when using the IgA and IgM isotype ELISAs. Rabbit abundance At most sites on Grande Terre a sample of rabbits was collected within a couple of hours despite high winds and at times driving sleet or snow. By contrast, it took over half a day to collect eight rabbits from Ile du Cimetière, where the population was noticeably lower than on Grand Terre. More importantly, plants such as Acaena magellanica were less closely cropped on Ile du Cimetière than on any part of Grand Terre where we collected rabbits; nonetheless, some grazing remained evident on Ile du Cimetière in comparison to nearby Ile Guillou, where rabbits had been removed by poisoning (Chapuis et al. 2001).

418 Wildlife Research B. D. Cooke et al. Discussion Myxoma virus About 50 years after the introduction of the myxoma virus to the Kerguelen archipelago near the Port-aux-Français station, data collected in 2001 02 have provided remarkably similar results to those obtained in an earlier survey between 1983 and 1988 (Chapuis et al. 1994b). They show that some areas of this archipelago the Péninsule Loranchet (Port-Chrismas, Val du Retour) and some localities of the Plateau Central (Baie de l Observatoire, Fjord Bossière) still appear to be unaffected by the virus. They also show that the prevalence of myxoma virus infection on the main island in 2001 02 (11%) was equivalent to that reported generally between 1983 and 1988 (mean ± s.e.: 9.2 ± 2.8%, n = 19 sites) (Chapuis et al. 1994b) although for the specific sites revisited in 2001 02 the prevalence was found to be generally higher than previously recorded. Although we did not specifically measure rabbit abundance it is clear that rabbits are now less abundant on Ile du Cimetière than was the case before rabbit fleas were introduced and certainly less abundant that they are in the rabbit-infested coastal areas and valleys of Grande Terre. During the initial period of colonisation by fleas of Ile du Cimetière in 1991 93 (Chekchak et al. 2000), hunters took samples from an average of 62 rabbits (67, 52, 67) for every two days of hunting. In 1994 96, as fleas were colonising the last sections of the island, the average was 52 (56, 50, 52) rabbits for each two days and by 1998 it took 2.5 days to shoot 34 rabbits. Our rate of sampling in 2003 (8 rabbits for a half day) suggests that rabbit numbers still remain relatively low. The poor state of rabbit-grazed vegetation on Grande Terre and its improved, but far from pristine, condition on Ile du Cimetière also supports the view that significant changes have occurred since the fleas were released. Together, the serological results and the observations on rabbit density confirm the earlier conclusions of Chapuis et al. (1994b) that, in absence of arthropod vectors, the transmission of myxoma virus by contact is very low under the conditions on the Kerguelen archipelago and that the myxomatosis virus generally plays a minor role in the regulation of the rabbit population. On Ile du Cimetière, the high prevalence of myxoma antibodies among adults (7 of 7) in 2001 02 confirms data collected in 1994 98 (after the island had been completely colonised by Spilopsyllus cuniculi), where 91.2% of adults (n = 170) had myxoma antibodies (Chekchak et al. 2000). These results can be taken as strong evidence that rabbit fleas are effective vectors of myxoma virus and have contributed to the low density of the rabbit population since the end of the 1990s (Chekchak et al. 2000). Consequently, there appears to be a good possibility of enhancing the rate of transmission of myxoma virus on Grande Terre simply by introducing European rabbit fleas. This could also pave the way for the use of myxoma viruses that are more virulent than those currently circulating. Bringing new fleas to Kerguelen presents its own difficulties because they must survive a long boat trip (Chekchak et al. 2000). Nevertheless, if the voyage were commenced with flea eggs instead of adult fleas, the larvae that hatch after two days could be maintained at room temperature on a diet of ground blood meal and yeast if humidity were carefully controlled at ~80% RH. These larvae would progress through to the pupal stage within ~14 days and the first adult fleas would emerge from cocoons at ~20 days, a period sufficient for the boat to reach Kerguelen. This would certainly be preferable to beginning the voyage with flea pupae. Physical disturbance of pupae during travel invariably causes them to emerge and mortality of unfed fleas is high. Few unfed fleas survive more than a week (Cooke and Skewes 1988). In view of the number of rabbit fleas carried by pregnant females on Ile du Cimetière in 2001 02, equivalent to observations made in the 1990s (mean ± s.e.: 182.1 ± 21.8, n = 98) by Chekchak et al. (2000), it seems that sufficient fleas could be obtained for making at least localised introductions to Grande Terre although it would take a long time for fleas to become distributed across the whole island. Rabbit haemorrhagic disease virus Rabbits that survive virulent RHDV normally have moderate to high titres of anti-rhdv antibodies detectable by both competition ELISA and IgG ELISA. These titres are usually in a ratio of 1:4 respectively. Additionally, it is usually possible to detect antibodies using the specific IgA ELISA while IgM antibody isotypes may also be present, especially if rabbits have recently recovered from infection. The detection of anti-rhdv IgG isotype alone, as was the case with two of the Kerguelen rabbits, is considered atypical and such antibody profiles are arguably the result of infection with another virus partially related to RHDV (Cooke et al. 2000, 2002). Such viruses are likely to have been the precursors of virulent RHDV (Fenner and Fantini 1999) and Capucci et al. (1996) have, in fact, isolated a non-pathogenic strain of rabbit calicivirus. Despite the detection of anti-rhdv antibodies in the two Kerguelen rabbits the anomalous antibody profiles make it improbable that RHDV is present in the archipelago. Nevertheless, the antibody profiles of the rabbits were sufficiently similar to those seen in Australian rabbits prior to the spread of virulent RHDV (Cooke et al. 2000) that they should not be discounted as being false positives. The indication of these antibodies is intriguing because, if they really indicate the presence of a virus closely related to RHDV then, due to the isolation of Kerguelen, it is possible that such viruses arrived in 1874 with the original rabbits. It raises the possibility that this group of viruses may have been associated with European rabbits more than 100 years before

Myxomatosis and RHD in Kerguelen Wildlife Research 419 RHD was first identified in China in 1984 (Liu et al. 1984; Xu 1991). Moreover, the rabbits released at Kerguelen were probably taken from Robben Island, South Africa (Eaton 1875; Boussès 1991), where a Dutch domestic rabbit population had been established in 1656 (Cooper and Brooke 1982). This information is consistent with recent arguments by Moss et al. (2002), who suggested that non-pathogenic forms of RHDV may have been present in European rabbits for many hundreds of years. It would be worth studying RHD antibodies at Kerguelen further to see whether more light could be shed on the history and evolution of lagomorph caliciviruses. This might resolve some of the conjecture that surrounds the sudden emergence of virulent RHD as a new disease among European rabbits (Fenner and Fantini 1999). The evidence that an RHDV-like virus may be present at Kerguelen nevertheless remains extremely limited and the low seroprevalence of antibodies and the low titres recorded make it unlikely that specific antibodies to the Kerguelen virus would interact with RHDV and reduce its effectiveness if the latter were used to control rabbits at Kerguelen. Research in Australia and New Zealand suggests that antibodies to a putative pre-existing virus related to RHD did not confer much protection when virulent RHDV first spread in 1995 except in areas where they were present in rabbits at high titres (Cooke and Fenner 2002). Nevertheless, it would be important to test both seronegative and seropositive rabbits captured at Kerguelen by direct challenge with virulent RHDV to assess their susceptibility to RHD before any future use of the virus was considered. Conclusions The introduction of rabbit fleas to Grande Terre is well worth considering as a serious option for better managing the rabbit population. Given the specificity of myxomatosis for a few species of lagomorphs and the widespread introduction of host-specific vectors such as the European and Spanish rabbit fleas (S. cuniculi and Xenopsyllus cunicularis respectively) into Australia, this can be done without direct risks to native animal species (Fenner and Fantini 1999). The information collected over 14 years following the release of S. cuniculi on Ile du Cimetière supports the idea that fleas would persist and enhance the spread of the myxoma virus in other parts of the Kerguelen archipelago to reduce rabbit numbers. The potential benefits of having fewer rabbits on Grande Terre are enormous because without effective rabbit control the future looks bleak for soils, plant communities and the fauna that depends on them for food and safe nesting sites. The main problem we foresee in enhancing myxomatosis as a biological control agent is the possibility that a reduction of rabbits could be harmful to native birds in the short term if cats turned to alternative prey or there were other unpredictable consequences (Courchamp et al. 2003). Similar concerns were expressed in Australia (Newsome et al. 1997) prior to the release of RHD, but analysis of ecological consequences since its establishment indicate that improved regeneration of vegetation has occurred while a shift in the diet of cats towards species other than rabbits has been off-set by a reduction in cat abundance (Cooke and Fenner 2002). Consideration of the possible use of RHDV in Kerguelen will require more time. Nevertheless, this paper provides sufficient background information to at least say that there appear to be no immunological obstacles to the introduction of this second form of biological control if it were required. Most rabbits appear to be fully susceptible to RHD on the basis of their lack of antibodies indicating prior exposure. Acknowledgments This research was supported by the Institut Polaire Français (IPEV, Programme 276) and the CNRS (Zone atelier antarctique et subantarctique). The logistics were organised by Alain Lamalle (IPEV) with assistance from the staff of the Port-aux-Français station. We especially thank the Animal and Plant Control Commission, South Australia, for supporting this work by providing laboratory resources for antibody testing. Dr Peter Kerr, CSIRO Sustainable Ecosystems, supplied advice and reagents for myxoma ELISAs, and Dr Lorenzo Capucci, Istituto Zooprofilattico Sperimentale della Lombardia e dell Emilia, arranged the supply of reagents used in RHD ELISAs. References Aubert de la Rüe, E (1964). Observations sur les caractères et la répartition de la végétation des îles Kerguelen. Comité National Français des Recherches Antarctiques 1, 1 60. Boussès, P. (1991). Biologie de population d un vertébré phytophage introduit, le lapin (Oryctolagus cuniculus), dans les îles subantarctiques de Kerguelen. Thesis, University of Rennes I. Capucci, L., Frigoli, G., Ronshold, L., Lavazza, A., Brochi, E., and Rossi, C. (1995). Antigenicity of the rabbit hemorrhagic disease virus studied by its reactivity with monoclonal antibodies. Virus Research 37, 221 238. doi:10.1016/0168-1702(95)00033-m Capucci, L., Fusi, P., Lavazza, A., Pacciarini, M. L., and Rossi, C. (1996). Detection and preliminary characterization of a new calicivirus related to rabbit haemorrhagic diease virus but non pathogenic. Journal of Virology 70, 8614 8623. Chapuis, J. L., Boussès, P., and Barnaud, G. (1994a). Alien mammals, impact and management in the French subantarctic islands. Biological Conservation 67, 97 104. doi:10.1016/0006-3207(94)90353-0 Chapuis, J. L., Chantal, J., and Bijlenga, G. (1994b). La myxomatose dans les îles subantarctiques de Kerguelen, en l absence de vecteurs, trente années après son introduction. Compte Rendu de l Académie des Sciences, Sciences de la Vie 317, 174 182. Chapuis, J.-L., Le Roux, V., Asseline, J., Lefèvre, L., and Kerleau, F. (2001). Eradication of the rabbit (Oryctolagus cuniculus) by poisoning, on three islands of the subantarctic archipelago of Kerguelen. Wildlife Research 28, 323 331. doi:10.1071/wr00042 Chapuis, J.-L., Frenot, Y., and Lebouvier, M. (2002). Une gamme d îles de référence, un atout majeur pour l évaluation de programmes de restauration dans l archipel de Kerguelen. Revue d Ecologie (Terre Vie) (Suppl.) 9, 121 130.

420 Wildlife Research B. D. Cooke et al. Chapuis, J.-L., Frenot, Y., and Lebouvier, M. (2004). Recovery of native plant communities after eradication of rabbits from the subantarctic Kerguelen Islands, and influence of climate change. Biological Conservation 117, 167 179. Chekchak, T., Chapuis, J.-L., Pisanu, B., and Boussès, P. (2000). Introduction of the rabbit flea, Spilopsyllus cuniculi (Dale), to a subantarctic island (Kerguelen archipelago) and its assessment as a vector of myxomatosis. Wildlife Research 27, 91 101. Cooke, B. D., and Fenner, F. (2002). Rabbit haemorrhagic disease and the biological control of wild rabbits Oryctolagus cuniculus in Australia and New Zealand. Wildlife Research 29, 689 706. Cooke, B. D., and Skewes, M. A. (1988). The effects of temperature and humidity on the survival and development of the European rabbit flea, Spilopsyllus cuniculi (Dale). Australian Journal of Zoology 36, 649 659. Cooke, B. D., Robinson, A. J., Merchant, J. C., Nardin, A., and Capucci, L. (2000). Use of ELISAs in field studies of rabbit haemorrhagic disease in Australia. Epidemiology and Infection 124, 563 576. doi:10.1017/s0950268899003994 Cooke, B. D., McPhee, S., Robinson, A. J., and Capucci, L. (2002). Rabbit haemorrhagic disease: does a pre-existing RHDV-like virus reduce the effectiveness of RHD as a biological control in Australia. Wildlife Research 29, 673 682. Cooper, J., and Brooke, R. K. (1982). Past and present distribution of the feral European rabbit Oryctolagus cuniculus on southern African offshore islands. South African Journal of Wildlife Research 12, 71 75. Courchamp, F., Chapuis, J. L., and Pascal, M. (2003). Mammal invaders on islands: impact, control and control impact. Biological Reviews 78, 347 383. doi:10.1017/s1464793102006061 Derenne, P. (1976). Notes sur la biologie du chat haret de Kerguelen. Mammalia 40, 530 595. Eaton, A. E. (1875). Natural history of Kerguelen s islands. Nature 12, 35 37. Fenner, F., and Fantini, B. (1999). Biological Control of Vertebrate Pests. The History of Myxomatosis an Experiment in Evolution. (CAB Publishing: Wallingford, UK.) Kerr, P. J. (1997). An ELISA for epidemiological studies of myxomatosis: persistence of antibodies to myxoma virus in European rabbits (Oryctolagus cuniculus). Wildlife Research 24, 53 65. doi:10.1071/wr96058 Lésel, R. (1967). Contribution à l étude écologique de quelques mammifères importés aux îles Kerguelen. Terres Australes et Antarctiques Françaises 38, 3 40. Liu, S. J., Xue, H. P., Pu, B. Q., and Quian, N. H. (1984). A new viral disease in rabbits. [In Chinese] Animal Husbandry and Veterinary Medicine 16, 352 255. [Veterinary Bulletin (1985) 55, 5600 [Abstract]]. Moss, S. R., Turner, S. L., Trout, R. C., White, P. J., Hudson, P. J., Desai, A., Armesto, M., Forrester, N. L., and Gould, E. A. (2002). Molecular epidemiology of rabbit haemorrhagic disease virus. Journal of General Virology 83, 2461 2467. Newsome, A., Pech, R., Smyth, R., Banks, P., and Dickman, C. (1997). Potential Impacts on Australian Native Fauna of Rabbit Calicivirus Disease. (Environment Australia.) Pascal, M. (1980). Structure et dynamique de la population de chats harets de l archipel de Kerguelen. Mammalia 44, 161 182. Pontier, D., Say, L., Debias, F., Bried, J., Thioulouse, J., Micol, T., and Natoli, E. (2002). The diet of feral cats (Felis catus L.) at five sites on the Grande Terre, Kerguelen archipelago. Polar Biology 25, 833 837. Sokal, R. R., and Rohlf, F. J. (1995). Biometry: the Principles and Practice of Statistics in Biological Research. (W.H. Freeman and Co.: New York.) Weimerskirch, H., Zotier, R., and Jouventin, P. (1989). The avifauna of the Kerguelen Islands. Emu 89, 15 29. Werth, E. (1911). Vegetation der subantarktischen Inseln Kerguelen, Possession und Heard-Eiland. Deutsche Sudpolar-Expedition 1901 1903, 8. Botanie 2, 223 371. Xu, W. (1991). Viral hemorrhagic disease of rabbits in the People s Republic of China: epidemiology and virus characterization. Revue Scientifique et Technique de l Office International des Épizooties 10, 393 408. Manuscript received 9 September 2003; accepted 27 February 2004 http://www.publish.csiro.au/journals/wr