We studied relations between natural seedling reproduction and above ground environment in a longleaf pine ecosystem. Forty-eight 0.05 ha circular plots were sampled under single-tree selection, grouptree selection and control stands in three main longleaf pine areas in south Alabama, USA. We measured six above-ground environment factors, viz. canopy closure, stand density, basal area, average tree height, understory cover and PAR under canopy. We employed forward, backward and stepwise selection regression to produce one model. Three main variables: canopy closure, stand density and basal area, were left in the model; light, PAR and understory cover were not incorporated into the model at the 0.10 significance level. Basal area was a positive parameter, while canopy closure and stand density were negative parameters. Canopy closure was the main parameter in the model. The model proved to be meaningful, and has potential to provide useful guidance for future work. We studied relations between natural seedling reproduction and above ground environments in a longleaf pine ecosystem. Forty-eight 0.05 ha circular plots were sampled under single-tree selection, grouptree selection and control stands in three main longleaf pine areas in south Alabama, USA. canopy closure, stand density, basal area, average tree height, understory cover and PAR under canopy. We employed forward, backward and stepwise selection regression to produce one model. Three main variables: canopy closure , stand density and basal area, were left in the model; light, PAR and understory cover were not incorporated into the model at the 0.10 significance level. Basal area was a positive parameter, while canopy closure and stand density were negative parameters. Canopy closure was the main parameter in the model. The model proved to be meaningful, and has potential to provide useful guidance for future work.