The impact of rabbit haemorrhagic disease (RHD) on wild rabbit populations was assessed by comparing population parameters measured before the introduction of RHD into Australia in 1995 with population parameters after RHD. We used data from an arid inland area and a moist coastal area in South Australia to examine the timing and extent of RHD outbreaks, their interaction with myxomatosis and their effect on breeding, recruitment and seasonal abundance of rabbits. From this we propose a generalised conceptual model of how RHD affects rabbit populations in southern Australia. RHD decreased long-term average numbers of rabbits by 85% in the arid area. In the coastal area, RHD decreased numbers of rabbits by 73% in the first year but numbers gradually recovered and were only 12% below pre-RHD numbers in the third year. Disease activity generally begins a month or two after the commencement of breeding in autumn or winter, peaks in early spring and ceases to be apparent in summer. Where the disease is most active, the pattern of population change is almost the inverse of the former pattern. During the breeding season, RHD severely suppresses rabbit numbers. Compensatory recruitment of late-born young, protected by maternal antibodies until the disease becomes inactive at the end of spring (also the end of breeding), allows the observed rabbit abundance to increase during summer, albeit to lower levels than before RHD. Maternal antibodies are lost during summer and the population becomes susceptible to RHD. The seasonal peak in myxomatosis activity is pushed back from late spring to early summer or autumn. Survivors of myxomatosis breed after opening rains in autumn but many succumb to RHD before raising their litters. The reduced abundance of rabbits and changed pattern of seasonal abundance have potential consequences for vegetation recovery.