Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Biological Sciences

First Advisor

William J. Platt


The main objective of this dissertation is to describe and examine how altering fire regimes in southeastern longleaf pine savannas may affect the abundances and distributions of indigenous plant species. In chapter one, I investigate whether herbs, compared to shrubs, have greater utility in delineating communities along elevation gradients of several meters in longleaf pine savannas. Herbaceous and shrub abundance data were analyzed separately for areas in Louisiana and Florida using cluster analysis and ordination. In Louisiana, three plant communities were sharply delineated and strongly correlated with gradients in elevation, surface soil moisture, and other surface soil properties. Herbaceous species indicated the existence of three communities in Florida, but species distributions and edaphic factors were less discrete than in Louisiana. In both regions, classification of shrub species clustered data into uninterpretable groups suggesting broad distributions. Shrubs were less sensitive than herbs to changes in ground level elevation and surface soil properties. Expanded distributions of shrubs may be a result of past fire suppression and dormant season burning. In chapter two, I examine the experimental effects of fire regimes on shrubs. The effects of geographic region (Louisiana and Florida), habitat (upslope savannas and downslope seepages), fire regime (dormant versus growing season; single versus repeated fires), and species characteristics (root crown versus rhizomatous species; woody versus suffrutescent species) on resprouting of shrubs following fires were monitored over four years. Regardless of geographic region, shrubs collectively resprouted significantly more following dormant season fires than growing season fires, especially when conducted biennially in downslope seepages. In Florida downslope seepages, Hypericum microsepalum and H. brachyphyllum, both suffrutescent species with root crowns, resprouted more following dormant season fires than growing season fires. Of the two species, only abundances of H. microsepalum were inversely related to maximum fire temperatures. Growing season fires did not result in appreciable complete kill of shrubs; densities of shrubs were comparable to pre-bum numbers. Present abundances of shrubs may be a result of dormant season burning. Shifting fire regimes to the growing season alone will not likely result in a rapid return toward conditions once present in longleaf pine savannas.