In developing an improved understanding of the dynamics of understory plant composition and productivity in Coastal Plain forest ecosystems, we examined the influence of site preparation and phosphorus fertilization on the successional trends of shrubs and herbaceous plants growing on lands of widely ranging sub- soil texture in Arkansas, Louisiana, and Texas which are man- aged for southern pine production. Burn-inject, chop-burn, chop-burn-disk, double-chop, shear-burn, shear-windrow, and shear-windrow-disk site preparation methods were applied in a completely randomized split-plot design to sites with subsoil tex- tures consisting of loam, gravelly-clay, silt, silty-clay, and clay, both fertilized with 73.4 kg P/ha and unfertilized. Site prepara- tion method, subsoil texture, and fertilization influenced produc- tion of paspalums and other forbs the first growing season fol- lowing treatment, but no treatment combination affected plant groups in subsequent years. Total herbaceous production increased 24 to 35-fold over pretreatment levels the firt growing season after treatment. While site preparation methods had little influence on herbaceous biomass, subsoil texture affected herba- ceous production the first year after treatment, with loam sub- soils being most productive. Although annual composites were the most abundant herbaceous group the first year after treat- ment, they were largely replaced by perennial grasses by the third post-treatment growing season. By the seventh growing season folowing treatment, herbaceous production declined on all subsoil textures with composition and yield approximating pretreatment estimates. Subsoil texture influenced shrub density only in the first and third growing seasons after treatment. During the first few years after site preparation, herbaceous pro- duction appeared inversely related to shrub density. In the first and third post-treatment growing seasons, fertilization signifi- cantly increased total herbaceous production and biomass of composites and legumes. But 7 years after application, total herbaceous production and biomass of bluestems, other grasses, and sedges was greater on unfertilized areas. The absence of dif- ferences among treatments by the seventh post-treatment grow- ing season indicates an overall long-term similarity in the degree of disturbance caused by application of each method in this ecosystem