Department of the Environment

About us | Contact us | Publications

Header imagesHeader imagesHeader images

Publications archive - Biodiversity


Key departmental publications, e.g. annual reports, budget papers and program guidelines are available in our online archive.

Much of the material listed on these archived web pages has been superseded, or served a particular purpose at a particular time. It may contain references to activities or policies that have no current application. Many archived documents may link to web pages that have moved or no longer exist, or may refer to other documents that are no longer available.

Threat Abatement Plan for Competition and Land Degradation by Feral Rabbits

Biodiversity Group Environment Australia, 1999
0 642 2546355

Measures to Address the Key threatening Process

Complete removal of rabbits from Australia is well beyond the capacity of available techniques and resources because the species is well established across a vast area. Most rabbit control programs now aim to achieve the long-term suppression of rabbit populations, to reduce the damage that rabbits cause to production and the environment in the most cost-efficient manner (Williams et. al. 1995).

Eradication of rabbits from small islands or of a localised, newly established population of rabbits has been accomplished and is feasible (McManus, 1979; Martin and Sobey, 1983) provided a sufficiently rapid, well-funded and persistent campaign could be mounted. For example, rabbits have been eradicated from several islands in Western Australia (Morris, 1985). For larger islands however, such as Macquarie Island (120 square kilometres), rabbit management has focused on control rather than eradication (Brothers et. al. 1982).

Williams et. al. (1995) undertook a comprehensive review of past and current management techniques for suppressing rabbit populations including predators, parasites, nutrition, trapping, biological control agents and warren ripping. Techniques currently available can be classed broadly as mechanical, chemical, and biological.

Mechanical control


Fences have often been used to exclude rabbits from an area. Simple electric fences are generally only effective for short periods, as rabbits soon negotiate them. Long-term exclusion requires the use of wire-netting fences (McKillop and Wilson, 1987). Electrified wire-netting fences can be very effective if properly maintained. Where the rabbit’s normal food source is some distance from the warren, temporary electric fencing can be used to prevent access to the food and to make baiting near warrens more effective.

Rabbit proof fences are expensive to construct ($3500 per kilometre) and to maintain (Korn and Hosie, 1988). They need regular maintenance and patrolling to be effective. Breaches of the fence resulting from burrowing animals, tree falls or rabbits climbing into the exclusion zone will quickly make it ineffective (Williams et. al. 1995).

Before erecting such fences the movements and dynamics of local fauna populations should be considered.

Harbour destruction

Rabbits living in severe climatic regions, such as the semi-arid and arid zones, depend on warrens for protection from climatic extremes (Hall and Myers, 1978; Parer and Libke, 1985). Destroying warrens by ripping can be a cost-effective and efficient way to suppress rabbit numbers and inhibit reinvasion of the treated area (Williams and Moore, 1995). The success of the technique depends on efficient destruction of the entire burrow system (Mutze, 1995), adequate follow-up control (Williams and Moore 1995) and using appropriate control measures on adjacent lands (Parer and Parker, 1986; Parer and Milkovits, 1994). Before ripping, dogs can be used to force rabbits to take refuge in the warrens, reducing the number of rabbits surviving to reopen warrens.

Warrens in relatively inaccessible places such as between trees and in narrow gullies can be ripped using a drag-arm ripper (Williams et. al. 1995). This is a ripper mounted on a hydraulic arm, allowing more flexibility in its placement.

Explosives can be used to destroy warrens in rocky areas, along rivers and in steep sandbanks where ripping is not possible. If used correctly, explosives will completely destroy the tunnel system (Barnes, 1983). The use of explosives is expensive and requires qualified personnel if the method is to be safe and effective.

Rabbits in some parts of Australia prefer surface refugia even when warrens are present (Wheeler et. al. 1981; Williams et. al. 1995). Habitat destruction in these situations may require the removal of the shrub layer, of exotic weeds such as blackberry and debris such as logs. This may be undesirable as such refugia also provide important habitat for amphibians, reptiles, small mammals, and ground dwelling or ground feeding birds. The relative benefits of this action for enhancing the recovery of endangered or vulnerable species would need to be carefully assessed against the risks to other native species if such action is proposed in areas being managed for conservation purposes.


Shooting can be a humane way of destroying rabbits but is rarely an effective means of reducing populations. For shooting to be effective in the long-term, a high rate of removal must be maintained to compensate for the rabbit’s high rate of increase. The commercial rabbit industry depends largely on the supply of field-shot rabbits to processing works and plays no significant role in reducing rabbit populations (Ramsay, 1994). This is due to a reduction in harvesting efficiency as populations decline, resulting in lower returns per unit effort for the shooter. To maximise returns, commercial hunters will move to areas where rabbits are more abundant rather than continuing to shoot rabbits in a lower density population.

Shooting may be used by land managers as a humane method for maintenance control once populations have been substantially reduced by other methods.


Leg-hold traps have been used to catch feral rabbits for many years. However, trapping is unlikely to result in any long-term decrease in rabbit densities. The use of the traditional steel-jawed leg-hold trap for the control of rabbits is still permitted in some jurisdictions. This threat abatement plan does not advocated the use of these traps as they subject animals to unnecessary pain and suffering. Where trapping of rabbits is considered appropriate, such as with small isolated populations, barrel traps or soft catch traps should be used (Korn and Hosie, 1988).


Chemical control

Lethal doses of poison can be administered through fumigating warrens, laying poison bait or using the ‘tarbaby’ technique.


Pressure fumigation or diffusion fumigation may be used to poison rabbits while they are in their warrens (Cooke, 1981; Williams and Moore, 1995). Toxins used for fumigation include chloropicrin, carbon monoxide, carbon dioxide, calcium cyanide, and phosphine (Hayward and Lisson, 1978).

Pressure fumigation requires the use of a pump to generate and force the fumigant throughout the warren. The fumigant usually comprises a mixture of thick smoke, carbon monoxide and chloropicrin (Parer and Milkovits, 1994). During fumigation all warren entrances are sealed. Entrances hidden in rocks or vegetation may be revealed by emerging smoke.

Pressure fumigation is a slow and cumbersome process and is suitable only for small areas (Williams et. al. 1995). In addition, the use of chloropicrin in pressure fumigation is both dangerous to operators and perceived to be inhumane: the identification and use of alternate fumigants is recommended.

Diffusion fumigation can use chloropicrin liquid, or aluminium-phosphide pellets, which give off phosphine gas (Korn and Hosie, 1988). Pellets are placed in absorbent paper towel or newspaper, then moistened with water and sealed as far into the burrow as possible. Unlike pressure fumigation, diffusion fumigation requires little equipment, and allows for the use of an alternate toxin such as phosphine which is possibly more humane (Williams and Moore 1995; Williams et. al. 1995). It may have an important role in follow-up maintenance after large-scale programs such as warren ripping or baiting (Korn and Hosie, 1988; Williams and Moore, 1995).


The poison 1080 (sodium monofluoroacetate) is the main toxin used in Australia for the control of a range of vertebrate pest species including, European foxes (Thompson and Fleming, 1994), dingoes/wild dogs (Thomson, 1986), feral pigs (McIlroy, 1983) and rabbits (McIlroy and Gifford, 1991). The popularity of 1080 is due largely to its effectiveness on the target species and its relative cheapness (Martin et. al. 1994). It also degrades rapidly under field conditions and is non-cumulative in animals which ingest a sub-lethal dose (Williams et. al. 1995). Very effective methods of use have been developed and these are generally target specific when properly applied. All of these factors combined have made 1080 the most widely accepted and suitable poison for rabbits.

Baits commonly used for rabbits include pieces of carrot, oat grains, or pellets manufactured from bran or pollard. These are either laid in furrows or broadcast from the ground or air (Korn and Hosie, 1988). Mortality rates following 1080 baiting of rabbits can be as high as 90 per cent (McIlroy and Gifford, 1991).

There are a number of disadvantages to 1080 poisoning and these include the absence of an antidote (Mead et. al. 1991) and high toxicity to livestock and domestic animals (Martin et. al. 1994). Solubility in water also leads to 1080 leaching from baits laid in damp environments or during wet weather (Martin et. al. 1994). In addition, not all individuals are susceptible to 1080 baiting because some may avoid baits (neophobia), while others which have previously consumed a sub-lethal dose may have become ‘bait shy’ (Oliver et. al. 1982). These disadvantages can be minimised if baiting programs are well managed. The ability of 1080 to leach from baits can be considered an advantage because it is rapidly broken down by soil micro-organisms (Wong et. al. 1992). This would reduce the time non-target species are at risk from any poisoned bait not collected following a poisoning campaign.

Concerns about the non-target impacts of 1080 have led to the study of alternative poisons for use in rabbit baiting campaigns. Poisons evaluated to date include the anticoagulants pindone (Oliver et. al. 1982; Eason and Jolly, 1993) and brodifacoum (Crosbie et. al. 1986), and the rodenticides 1,3-Difluoro-2-propanol (DFP) (Mead et. al. 1991) and cholecalciferol (Eason, 1993).

Anticoagulants are less toxic to livestock and domestic animals (Martin et. al. 1994), less likely to leach from the bait when used in damp environments (Oliver et. al. 1982) and there is a readily available antidote. In addition, they are slow-acting and cumulative, and hence individuals are more likely to ingest a lethal dose before developing an aversion to the bait (Crosbie et. al. 1986). Their disadvantages include being relatively expensive, persisting in the environment for over 12 months and bioaccumulating (Williams et. al. 1995). The use of brodifacoum for rat control in sugar cane is implicated in the decline of raptors in north Queensland (Young and de Lai, 1997). Animals may ingest several lethal doses of poison before they succumb, thereby increasing any non-target hazards. Broad scale baiting with second generation anticoagulants poses substantial secondary poisoning hazards to wildlife (Eason and Spurr, 1995) that generally exceed those of acute, non-cumulative poisons (Bird, 1995).

In some States, pindone is usually used near urban areas where the risks of accidental poisoning to humans and companion animals are greatest. In Western Australia the use of pindone is carefully controlled, and specific prescriptions require its careful use in the presence of certain native species, especially large macropods. Experience has shown that marsupials are very sensitive to this toxin. There are also considerable differences in the susceptibility of individual animals and species to pindone and this requires further investigation (Martin et. al. 1991; Martin et. al. 1994).

Rodenticides such as cholecalciferol are highly toxic to rabbits, but also very expensive compared to 1080 and pindone (Eason, 1993). There is insufficient information on the toxicity of cholecalciferol to non-target species, the availability of an antidote, and its acceptability by rabbits under field conditions (Eason, 1993). Further studies are required to determine if it has a role in the control of rabbits in Australia.

Frequent and ineffective poison baiting can lead to neophobia in rabbit populations that can reduce the effectiveness of baiting. Changing poisons does not overcome the problem (Williams et. al. 1985).

The ‘tarbaby’ technique is a possible alternative to baiting. This uses a toxin mixed with grease, which is placed down the entrance of the burrow. Rabbits which pass through the grease will attempt to groom themselves by licking the grease from their fur and will thereby ingest the toxin. The high concentration of 1080 required precludes its use with this technique, but alternative toxins such as anticoagulants may be suitable (Williams et. al. 1995). Advice has also been provided that it performs poorly in light sandy soils, such as those found in Western Australia.

Biological control

Biological agents, such as selected pathogens, may be introduced to control pests by increasing mortality and/or decreasing fertility in the host. Significant research has been undertaken in Australia over many years to identify potential biological control agents for rabbits.


The myxoma virus was imported into Australia in August 1936 to evaluate its potential as a biological control agent for feral rabbits (Bull and Mules, 1943). Since its release into the wild near Corowa in 1950 (Rendel, 1971), myxomatosis outbreaks have become common in Australian rabbit populations (Williams et. al. 1990). The effects of myxomatosis immediately following the 1950 release were not well documented (Williams et. al. 1995), but at Corowa 99 per cent of rabbits were thought to have died (Myers et. al. 1954).

Mortality and morbidity rates from myxomatosis vary considerably (Williams and Parer, 1972) and depend on genetic (Williams et. al. 1990) and non-genetic resistance (Sobey and Conolly, 1986; Williams and Moore, 1991; Parer et. al. 1995), age at infection (Cooke, 1983), and climatic conditions (Dunsmore and Price, 1972; Parer and Korn, 1989). Mortality rates also depend on the virulence of the infecting strain of the myxoma virus: strains are graded as I, II, III, IV and V from most virulent to least virulent (Fenner and Ratcliffe, 1965).

The strain released in 1950 at Corowa was considered to be a Grade I, having a 99 per cent mortality rate. Most field strains are now classed as Grade III, having mortality rates ranging from 40 to 90 per cent (Williams et. al. 1995). It has been observed that releases of highly virulent strains have resulted in relatively low mortality rates, due possibly to these strains killing rabbits too quickly for the virus to spread widely (Parer, 1991).

Myxomatosis caused extremely high mortality rates following its release in 1950. However, the virus quickly attenuated resulting in lower kill rates (Williams et. al. 1995). Rabbits surviving the myxoma virus develop immunity, making subsequent outbreaks less effective (Sobey and Conolly, 1986; Williams and Moore, 1991; Parer et. al. 1995). Immunity in rabbits previously exposed is also inherited (Fullager, 1977).

Despite a decrease in the effectiveness of the myxoma virus since its initial release in the 1950s, it continues to have an important role in rabbit control. Parer et. al. (1985) demonstrated that in the absence of myxomatosis, rabbit populations can increase rapidly.

Rabbit calicivirus disease

Rabbit calicivirus disease (RCD) is an acute and fatal infectious disease of rabbits first identified in China in 1984 (Liu et. al. 1984 cited in Xu and Chen, 1989). Clinical disease resulting from RCD has since been reported in many countries in Asia, Europe, Africa and in Mexico (Morisse et. al. 1991). In September 1991, RCD virus was imported from the Czech Republic to the CSIRO Australian Animal Health Laboratory in Geelong to study the effect of RCD on laboratory and wild rabbits as well as a range of non-target species. In March 1995 field investigations began on Wardang Island, South Australia, to investigate the behaviour of the virus in the Australian environment. In October 1995 the virus escaped onto the mainland, possibly as a result of windborne vectors (Cooke, 1996). In September 1996 RCD was accepted as a biological control agent under the Commonwealth Biological Control Act 1984 (covering the ACT). It was later accepted under relevant legislation in the States and Northern Territory.

Susceptible rabbits challenged with RCD develop clinical signs after an incubation period of 1 to 3 days (Marcato et. al. 1991). Death usually occurs as a result of acute respiratory and heart failure several hours to two days after the onset of clinical signs (Rodak et. al. 1991).

Little information has been published on morbidity and mortality rates in wild rabbits, although RCD outbreaks have been reported for wild populations in England, Germany, Italy, and Spain (Cancellotti and Renzi, 1991; Loliger and Eskens, 1991; Villafuerte et. al. 1994; Trout, 1996). In South Australia mortality rates exceeding 95 per cent were observed in populations not previously exposed to RCD (Cooke, 1996; Mutze et. al. 1998). Initial impressions on the impact of RCD in Australia indicate that it is more effective in arid than temperate areas (Anon, 1997). The factors influencing the impact of RCD are currently being studied.

Direct contact between infected and susceptible rabbits is believed to be one mode of transmission for RCD. The virus is also passed onto susceptible individuals who have contact with the secretions or excretions of infective rabbits, or items such as food and water that have been contaminated (Xu and Chen, 1989). Rabbits, which survive an RCD epizootic, may persist as carriers for up to a month (Gregg et. al. 1991). The virus may also be transmitted via people or implements that have had contact with carrier rabbits (Xu and Chen, 1989). An outbreak in Mexico was linked with the importation of frozen rabbit meat which was contaminated with RCD (Mason, 1989). The rapid spread of RCD in Australia, more than 400 kilometres per month (Kovaliski, 1998) suggests that wind-borne insect vectors may play an important role in its transmission in Australian rabbit populations (Cooke, 1996).

Mortality rates in excess of 90 per cent have been observed in some South Australian populations (Mutze et. al. 1998a) where naturally recurring outbreaks have kept the population at an average population level only 17 per cent of the long-term pre-RCD average (Mutze et. al. 1998b). More humid sites do not always fare as well with some sites observing little change in rabbit numbers following the arrival of the virus (Anon, 1997).

The long-term effectiveness of RCD as a means of rabbit control in Australia is yet to be fully evaluated.


Fertility control agents which use viral vectors have been advocated as a means of controlling vertebrate pest populations (Tyndale-Biscoe, 1994). The concept of virally vectored immunocontraception is based on encoding an antigen specific to elements of an animal’s reproductive system and inserting it into a virus (Tyndale-Biscoe, 1991). In theory, when an animal is challenged with the virus carrying the antigen, it develops antibodies, which reduce fertility. It appears that 60 to 80 per cent of female rabbits would need to be prevented from breeding to achieve a sustained reduction in rabbit numbers (Williams and Twigg, 1996).

Research into fertility control has been supported by animal welfare groups as a desirable and humane way of dealing with troublesome wildlife species (Russell and Pope, 1993). A central tenet of the development of immunocontraceptive techniques is that they are humane.

Fertility control is an attractive option but is still in an experimental stage of development. Substantial efforts are being made to develop immunocontraceptive vaccines for foxes, rabbits and mice. The Vertebrate Biocontrol Co-operative Research Centre is presently evaluating the feasibility of virally vectored immunocontraception for the control of Australia’s wild rabbit populations (Robinson and Holland, 1995). Research on immunocontraception of rabbits is presently examining the role of the myxoma virus for delivering the antigen (Holland and Jackson, 1994).

Before immunocontraception for the wild rabbit will become a reality many ecological hurdles will need to be identified and overcome (Williams, 1997).

Integrated control

It is clear from the information provided above that when any one technique of rabbit control is used in isolation it is less effective than when two or more techniques are carefully combined (Cooke, 1993). When reliance is placed on only one technique and follow-up control is not implemented, the initial gains made in controlling rabbits are soon lost as rabbits will readily recolonise an area in the absence of any further control (Parer and Milkovits, 1994; Williams and Moore, 1995).

Cooke (1993) demonstrated the value in following up an effective myxomatosis epidemic with ripping in South Australia. A property where warrens were ripped following a large decline in rabbit numbers remained almost rabbit free many years later, but a near-by property that was not ripped experienced high rabbit numbers again soon after, even though myxomatosis re-occurred at both sites in the following year. Similar results were obtained from a study in central Australia that assessed the impacts of ripping warrens on rabbit persistence (Dobbie, 1998).

Unfortunately, many land managers have tended to rely on the myxoma virus as the only form of rabbit control (Williams et. al. 1995). Similarly since the escape of RCD, land managers have had unrealistic expectations about its effectiveness and have relaxed their rabbit control efforts. They neglect an opportunity when they fail to initiate control programs using conventional techniques after disease outbreaks have reduced populations to more manageable levels (Cooke, 1993). Ideally these diseases should be viewed as one element in an integrated rabbit control program. Outbreaks should be followed up with conventional control techniques to achieve greater kill rates and, therefore, longer lasting control. State and Territory Government departments have instigated programs to encourage land managers to integrate RCD into traditional rabbit control programs (Hansard, 1997).

Sustained control is clearly the most cost-effective and efficient way to control rabbits under most circumstances (Williams et. al. 1995). Ideally, the relationship between pest density and impacts on environmental values would be known, but such data are not readily available. It is therefore recommended that the pest population be reduced to a manageable level using intensive control techniques and then maintained at that level with regular follow-up control. Ongoing monitoring of the effectiveness of the control in ameliorating the impacts of rabbits on threatened species or primary production is necessary to determine the level of control necessary for ongoing sustained management.

Integrated rabbit control in Australia normally involves a significant reduction of the population, followed by harbour destruction and subsequent follow-up control. The initial population reduction may be brought about by an effective outbreak of myxomatosis or RCD, poisoning, or even a drought. Techniques such as harbour destruction, poisoning, warren ripping, and fumigation may be effective in keeping rabbit populations low (Williams et. al. 1995). Using dogs to chase rabbits into their warrens prior to ripping or fumigation can increase the effectiveness of both techniques.

Should RCD prove to effectively reduce rabbit populations in the long-term, then its role in rabbit control will be similar to that of myxomatosis. RCD outbreaks will result in smaller and, therefore, more manageable rabbit populations. Outbreaks should be followed up with conventional control to achieve more lasting reductions in rabbit numbers.

Williams et. al. (1995) reviewed past and present management of rabbits at the State or Territory level and concluded that effective rabbit management requires an organisational strategy that includes:

Integrating control involves coordinated use of various control techniques, and integrating control with other activities. For instance, rabbit control may need to be integrated with the control of other pest species such as foxes and feral cats. It may also need to be integrated with weed management programs. Caughley (1989) states that the management of a plant-herbivore system that leads to a stable ecosystem is preferable to that which leads to an unstable ecosystem. It is important that action taken to control a rabbit population has a clearly identified outcome. That outcome may be the recovery of a specific endangered species or it could involve a whole ecological community that is under threat. Essentially integrated control involves planning control programs in such a way as to take advantage of environmental conditions and other activities in order to maximise control effectiveness and minimise costs.

Published June 1999 by Environment Australia under the Natural Heritage Trust.

Commonwealth of Australia