A systems approach is necessary for effective control of feral rabbit (Oryctolagus cuniculus L.) populations in the arid environments of Australia. Localized control procedures may result in local extinctions, but the persistence of the overall population will depend on the probability of recolonization, and hence, the degree of isolation of each local population unit. Genetic markers obtained using allozyme electrophoresis, temperature gradient gel electrophoresis (TGGE) and heteroduplex analysis (HA) were used to characterize the degree of structuring and extent of gene flow among rabbit populations in arid Queensland, Australia. Allozyme allele frequency data showed that there was no significant differentiation among sites (average FST = 0.005) and no isolation-by-distance or environmental discontinuity effects. TGGE/HA results also revealed no significant differentiation in mitochondrial DNA Control Region haplotype frequencies among sites and low interpopulation nucleotide divergence estimates (NST = 0.013). Therefore, rabbit populations exhibited a high degree of gene flow over large geographical areas (1600 km2) and were essentially a single panmictic unit. Unpredictable environmental conditions together with the spatial configuration of habitats which possess different probabilities of extinction may have resulted in repeated local extinctions followed by recolonization and homogenizing gene flow. These data suggest that current rabbit control strategies based on individual warren management may not achieve effective control in arid Queensland.