Summary1. Demographic changes in response to surgically imposed female sterility were monitored in 12 free-ranging rabbit populations in south-western Australia over a 4-year period. This was part of a research programme aimed at examining the potential for virally vectored immunocontraception to limit the abundance of rabbits (e.g. using a recombinant myxoma virus) and other mammalian pests. Sterility levels were 0%, 40%, 60% and 80% of all females in year 1, with a similar proportion of female recruits sterilized surgically in subsequent years.2. There was a signi®cant decrease in rabbit productivity with increasing sterility level. This was overcome by increased survival of kittens and adults on the highsterility sites, such that the base-level numbers of rabbits were maintained, and mean annual rates of increase (r) were near zero for all treatments in all years. However, in the high-sterility populations this compensation was insucient to overcome the eects of sterility totally, and there was a marked decrease in the seasonal peaks in rabbit abundance for these treatments. 3. Survival and recruitment were dependent upon the level of sterility, and consequently the density of rabbits, with greatest survival of adult rabbits occurring on the 80% sites. Survival of sterile females was greater than that of other adults, probably because of their increased ability to maintain body condition during times of low pasture biomass (summer drought). Thus two density-dependent processes were identi®ed: the ®rst was operating through increased survival of juvenile rabbits, the second through increased adult survival, particularly sterilized females. 4. Because the proportional impact of immigration was greater (i.e. immigrants constituted a greater proportion of the population) and emigration was less, from the 80% sites, the eects of sterility may have been underestimated on these sites. 5. The abundance of European rabbit¯eas, a vector of myxomatosis, was signi®-cantly lower on the 80% sites, but this did not appear to aect the transmission of myxoma. Myxomatosis occurred as an annual epizootic in three of four years, with > 90% of rabbits on site after each epizootic testing positive for myxoma antibodies. 6. To achieve a sustained long-term reduction in rabbit abundance, 60±80% of female rabbits would need to be prevented from breeding. This could be achieved by a recombinant strain of myxoma provided the strain retained good transmissibility and all infected rabbits became sterile for life.

Demographic changes in three free-ranging rabbit (Oryctolagus cuniculus) populations were monitored over 4 years in southern Western Australia. Peak densities followed periods of high rainfall and pasture biomass. The breeding season was prolonged, often extending from at least April to November, with some pregnancies occurring outside this period. Fecundity, determined by the autopsy of pregnant offsite rabbits and the known length of each breeding season, appeared to be relatively high, with the potential for 34–39 kittens doe-1 year-1; however, because not all females are pregnant in all months, the overall productivity of these populations was estimated at 25–30 kittens adult female-1 year-1. Exponential rates of increase varied from 0.13 to 0.30 during the breeding periods and –0.05 to –0.14 during the nonbreeding season. Kitten survival was generally low whereas some adults lived for more than 5 years. Two patterns of myxomatosis were observed: annual epizootics of the disease (3 of 4 years) and an epidemic that slowly spread over many months. European rabbit fleas were most abundant during winter–spring and attained highest densities on adult female rabbits.

Regardless of their sex and age, the persistence of 76 rabbits (Oryctolagus cuniculus) translocated onto 5 different sites with relatively low rabbit densities was identical to that of resident rabbits. Emigration and exploratory movements by rabbits from 12 discrete populations were positively correlated with rabbit density and mainly undertaken by adult rabbits. Adult males moved significantly more often and further than adult females. Two peaks in immigration were observed; a large peak (usually in January) immediately following the breeding season, and a second but smaller peak in March which preceded the start of the next breeding season. Again, significantly more adult males than females immigrated. The proportion of rabbits seen in spotlight counts was positively correlated with rabbit density, which suggests that biases in population estimates could result in some situations. No clear patterns on the effects of a variety of weather variables on spotlight counts could be established, but increasing rainfall, wind speed and moonlight may have reduced rabbit activity.

The impact of rabbit haemorrhagic disease (RHD) on free-ranging rabbit populations, their immunological response, and the abundance of insect vectors, were monitored in detail in the southern agricultural region of Western Australia. Further, a broad-scale rabbit monitoring program was also established at nine locations across a rainfall gradient in the southern half of Western Australia to monitor the natural arrival, or controlled release, of rabbit haemorrhagic disease virus (RHDV). Changes in rabbit populations and the immune status of RHDV antibodies were monitored in these areas to enable further understanding of the epidemiology of RHD, and its impact on rabbit numbers.RHDV had the greatest impact on rabbit populations in the arid and semi-arid areas (<360 mm per annum), where rabbit numbers were reduced to, and maintained at, 10% of pre-RHD levels. Conversely, the effects of RHD on rabbit numbers in higher-rainfall areas (360–700 mm per annum) were highly variable and patchy, and in some instances RHD had little apparent impact. In higher-rainfall areas where RHD was effective, rabbit numbers were reduced by 50–78%. RHDV was first confirmed in the southern agricultural region of Western Australia in early September 1996, ~1 year after its escape from Wardang Island, South Australia. At the detailed monitoring site (485 mm rainfall per annum), rabbit numbers declined by 65% within 2 weeks of RHDV being detected. However, ~70% of the remaining rabbits had antibodies against RHDV, indicating that they had survived the disease. There was also a demographic shift towards young rabbits (<1 year old) at this time. Further, even though there was no clinical evidence of RHD recurring in this population during the 3-year study, the presence of IgM antibodies in some rabbits well after the initial epizootic suggests that low-level transient outbreaks of RHDV had occurred at this site. Although the impact of these low-level outbreaks on population dynamics were uncertain, rabbit numbers at this site had recovered to pre-RHD levels within two breeding seasons.The abundance of insect vectors on the detailed monitoring site was similar between years with and without RHD outbreaks. Thus, the failure of RHDV to develop clinical disease on this site after the initial epizootic was unlikely to have been caused by the lack of suitable transmission vectors. The apparent lack of disease recurrence at the detailed monitoring site may have been caused by the presence of a non-pathogenic form of RHDV, which seemed to impart at least some cross-immunity to RHDV in these rabbits. The presence of RHDV also caused a shift in the timing of natural epizootics of myxomatosis at this site.

The acceptability of four different bait station designs (drum, slab, tyre, corrugated iron) to rabbits was tested in the field using unpoisoned oat bait. The drum (200 L, cut longitudinally) and the raised concrete slab (60 × 60 cm) designs were the most acceptable to rabbits. The raised tyre design was unacceptable, and this was supported by later field efficacy trials that compared the drum and tyre designs using 1080 One-shot oats. The efficacy of three of these designs (drum, slab, tyre) against ‘urban’ rabbits was assessed more fully using pindone oat bait. The tyre stations were again found to have little impact on rabbit numbers. With the exception of one drum site where pindone bait stations were totally ineffective, the proportional reductions in rabbit numbers for the remaining sites were similar between the drum (69%, n = 3) and slab (70%, n = 5) designs. However, the slab design provided much easier access to bait by non-target species (particularly birds), and we therefore recommend that the drum design would be the best bait station for controlling rabbits. The overall proportional reduction in rabbit numbers achieved with pindone bait stations was 48% (range 0–80%, n = 13), which is less than that usually achieved during broadacre control programs with pindone (60–90+%). In addition, these kills took 30–60 days to achieve, and as rabbit damage still occurred over this period, the use of pindone bait stations did not always result in damage mitigation or, ultimately, an economic benefit. Some potential problems associated with the use of pindone bait stations, such as the possibility of the development of ‘resistance’ to pindone bait and the risk to non-target species, are also discussed. The combined use of track counts and a ‘digs’ index proved a reliable indicator of changes in rabbit abundance.