A longleaf pine wiregrass ecosystem in the sandhills of north central Florida, upon which turkey oak gained dominance following a wild®re, was treated with applications of hexazinone (1.1 or 2.2 kg/ha) in May 1991. All applications successfully reduced competition from oaks in the overstory and understory (mortality >80%), resulting in progressive increases in the foliar cover of wiregrass, all graminoids and forbs through time. Broadcast application caused a decline in forb cover and species richness during the initial growing season, which recovered by the following year. The 2.2 kg/ha spot application resulted in an increase in species richness, while evenness declined with the continuing expansion of wiregrass. The entire site was then burned in June 1995 by prescribed ®re, which caused a widespread decrease in the cover of oaks, shrubs, wiregrass, all graminoids and forbs and plant species richness. In the following year, forb cover increased and oak cover remained signi®cantly lower on plots treated with the combination of hexazinone plus ®re than on ®re-only plots. The overall cover of forbs, graminoids, shrubs and longleaf pines continued to increase through time. Broadcast application initially exposed a greater number of understory plants to direct contact with herbicide, posing a higher mortality risk than may be acceptable in restoration efforts. Although recovery occurred in subsequent years, the lower selectivity of broadcast application makes it a less suitable restoration technique. Spot application of hexazinone was more selective in its effects upon the plant community. The 2.2 kg/ha spot application produced increases in the cover of wiregrass, all graminoids and forbs and the highest levels of species richness and diversity. The 2.2 kg/ha application rate was also most effective in controlling woody plant competition and is therefore recommended for restoring longleaf pine wiregrass ecosystems in sandhills and similar environments. Hexazinone application followed by prescribed ®re accelerates the rate of ecosystem restoration over that achievable by using ®re alone. The ecological bene®ts of controlling competition and rebalancing¯oristic composition rapidly achieved through this combination of treatments would likely require many cycles of prescribed ®re, if used as an individual treatment, over a period of several decades.