Introduced European rabbits, Oryctolagus cuniculus, can severely damage Australian native vegetation but the problem is difficult to quantify because simple methods to estimate rabbit impacts are lacking. Management decision-making is often uncertain because of unknown relationships between rabbit density and damage. We tested simple quantitative sampling methods using belt transects to detect differences in critical characteristics of perennial vegetation communities affected by rabbit browsing: damage to individual juvenile shrubs and trees, and loss of recruitment cohorts. Rabbit density and relative abundance of larger herbivores were estimated from dung pellet density. The prevalence of identifiable rabbit browse on juvenile plants increased with increasing rabbit density and was higher for plant species considered by previous authors to be highly palatable than for moderately palatable or unpalatable species. At densities of =0.5rabbitsha-1, highly palatable plant species were severely damaged as juveniles and cohorts in 0.3-1.0m height classes and 5-20mm basal diameter classes were missing. Similar damage became apparent in moderately palatable species at 2rabbitsha-1 but was rarely seen in unpalatable species. Within species, size cohort evenness was inversely related to the proportion of surviving juveniles with identifiable rabbit damage. The effect of rabbits on native vegetation condition can be recorded in a simple manner suitable for identifying density-damage relationships and changes in vegetation condition over time. It is particularly useful in setting target densities below which rabbits must be managed to maintain natural plant recruitment and ecosystem function in conservation reserves and pastoral grazing properties of southern Australia.