Background: Prescribed burning is an important management tool in jarrah (Eucalyptus marginata Sm.) forests of southwest Western Australia to reduce the risk of damaging bushfires. In 1986 to 1987, we established long-term study sites in dry (mean annual rainfall~700 mm) and moist (mean annual rainfall~1000 mm) jarrah forests to assess the effects of prescribed burning, as well as other fire treatments including fire exclusion, on the composition, richness, and relative abundance of forest understory vegetation. Results: Over almost 30 years, species assemblages within all fire treatments changed significantly through time, but the changes were independent of treatment. The pattern of change in composition of fire response types over elapsed time was different between sites. At the dry site, changes in species assemblages were initially relatively large in the first decade or so of the study, but slowed thereafter. At the moist site, compositional change driven by obligate seeding shrubs occurred faster and more uniformly with time across all treatments. Species richness was also independent of fire treatment at both sites. Species richness decreased with elapsed time on the moist site but increased with elapsed time on the dry site. The stronger elapsed-time effect rendered the time since fire effect on loss of species richness on the moist site to be insignificant, but there was an inverse relationship with time since fire at the dry site. Within each site, there were clear patterns of changing abundance based on life form and fire response groups, but this was not consistent between sites. Conclusions: At the scale of the current study, jarrah forest plant communities displayed resilience to imposed experimental fire regimes. Over time, either elapsed time or time since fire, species assemblages across all treatments changed regardless of fire treatment, but changes were mostly associated with specific plant life forms and fire response traits, suggesting that the process is both deterministic and stochastic. While many species changed in abundance over time, no species were lost as a result of the fire treatments. Within the fire frequency and intensity ranges investigated in this study, there was flexibility in the application of prescribed fire to achieve management objectives without loss of plant diversity.