Conversion to agriculture and plantations, development, and fire suppression have reduced the extent of savannas in the southeastern United States, and there is a need to catalog and classify the remaining savannas for both restoration and resource management purposes. The Big Savannah was a wet savanna in North Carolina that was destroyed in the 1950s, and subsequent vegetation classifications have generally not accommodated well the unique natural plant community of the Big Savannah. Vegetation reminiscent of that described for the Big Savannah was discovered north of the original site and designated as Wells Savannah. To evaluate the uniqueness of the savanna vegetation at Wells Savannah, we compiled a data set from permanent quadrats with information on vegetation and environmental variables from other Outer Coastal Plain savannas to compare with similar data from the natural community at Wells Savannah. We also inventoried an additional 26 quadrats on a tract adjacent to Wells Savannah that had experienced fire suppression. Results from multivariate analyses demonstrated clear differences between the Wells Savannah quadrats and other regional wet savanna quadrats based on both vegetation and soils. A number of species and several soil characteristics (higher clay percentage, and higher available iron and boron) distinguish Wells Savannah from other wet savannas. Although the fire suppressed quadrats near Wells Savannah had lower species richness, typical savanna species such as Ctenium aromaticum and Calamovilfa brevipilis were still present. Further exploration of fire-suppressed tracts in the area may yield more wet savanna inclusions similar to the former Big Savannah.

Multivariate analysis of vegetation and environmental variables from green pitcher plant (Sarracenia oreophila) bogs in northeast Alabama revealed three communities with unique species compositions and soil characteristics. Discriminant analysis of environmental variables revealed that A-horizon percent N, A and B horizon pH, humus layer thickness, and B-horizon K (kg/ha) were significantly related to the communities. A Quercus rubra, Arundinaria appalachiana, Pinus echinata community was found on upland seepage bogs located exclusively on Lookout Mountain close to the rim of the Little River Canyon and scattered along perennial streams. A Quercus falcata, Diospyros virginiana, Rhododendron canescens community was found on both Lookout and Sand Mountain in flat broad swales bisected by ephemeral streams. A Rhexia virginica, Dichanthelium scoparium, Carex glaucescens community was found primarily on Sand Mountain in open flood prone areas.

Most typical rock outcrop plants of the eastern United States occur either on calcareous or on non-calcareous outcrops, but not both. Often this is because their growth is inhibited in soil from the non-native substrate, as shown in this study for the granite outcrop endemic sedge Cyperus granitophilus when grown on limestone soil. For those rock outcrop species that can exploit both calcareous and non-calcareous substrates, it might be expected that they would do so by substrate specialization, with each population growing better on its native substrate than the other substrate. However, in most previously tested species, populations from both substrates grow well on the same substrate and both are inhibited on the second substrate. Populations found on the second substrate, though inhibited, are nevertheless able to maintain good health and grow sufficiently to maintain themselves there. In this study we show that the widespread sedge Cyperus aristatus ( = C. inflexus), a regular component of limestone, sandstone, granite, chert, and serpentine rock outcrop communities in the eastern United States, exploits both calcareous and non-calcareous substrates by substrate specialization, something previously shown in only one other eastern United States rock outcrop species.

Forest vegetation near Tallahassee, Florida was inventoried on a former cultivated field, called NB66h, which had been abandoned 43 yr earlier. Subsequently, the site suffered no disturbance and no colonization by alien invasive species. The canopy consisted of pines (Pinus taeda, P. echinata) and hardwoods, primarily Liquidambar styraciflua, Quercus nigra. Most tree species were characteristic of plant communities that historically occurred at lower slope positions in the landscape and not of presettlement upland shortleaf pine-oak-hickory woodland. Offsite species colonized surrounding uplands following abandonment of cotton plantations towards the end of the 19th century. Presettlement vegetation was represented by few species and individuals, and offsite species were abundant in every size class. Plant succession was arrested with little sign of directional development towards a predictable seral endpoint. Hypothesized seral trends towards proposed “climax” forests of magnolia-beech and southern mixed hardwoods were not supported.

Dense stands of Arundinaria species, or canebrakes, once were a dominant landscape feature along floodplains of the southeastern United States. However, human activities have reduced canebrakes to fragmented remnants representing <2% of their extent prior to European settlement. Canebrake restoration thus is a top priority for preserving and improving wetland biodiversity in the United States. Successful restoration requires an understanding of factors influencing establishment of the two most common United States Arundinaria species; therefore, this greenhouse study examined effects of inundation on A. gigantea and A. tecta. Both Arundinaria species were subjected to 0, 2, 4, or 6 weeks of inundation under long-day, warm temperature growing season conditions. Plant growth, mean net photosynthesis (P_n), and stomatal conductance (G_s) were measured on a weekly basis, and at the conclusion of the experiment, above- and belowground biomass were measured. We found significant correlations of P_n and G_s with duration of flooding and duration of post-inundation recovery, as well as a significant interaction among time, species, and flood duration, with both species responding similarly to flooded conditions. Once flooding was arrested, both P_n and G_s rates were higher in A. tecta than A. gigantea; the same was true for growth rates across the duration of the study. In conclusion, A. tecta appeared to be more flood tolerant than A. gigantea, reflecting habitats in which these species are known to occur. Future canebrake restoration projects may benefit from this information on differential flooding tolerance in selecting restoration sites or restoration species, based on hydrologic conditions.

Previous research has found significant differences in flowering time between sympatric, non-hybridizing populations of congeners in the genus Gelsemium (Gelsemium sempervirens and G. rankinii). An experimental approach using a common garden, reciprocal transplant experiments, and observations of natural populations were used to test the hypothesis that soil environmental variables (soil moisture and soil temperature) related to habitat specificity of the two species influence differences in flowering phenology. Replicated pairs of both species were planted into a common garden and into wet and dry habitats that typify the two species (dry for earlier flowering G. sempervirens and wet for later flowering G. rankinii). Two sites with sympatric natural populations of both species were also studied for comparison. The number of open flowers was counted every two weeks in 2007 and 2008 in all sites and in 2009 for the reciprocal transplant experiment. Soil moisture and soil temperature varied by site, date, habitat, and species but was not consistently related to flowering phenology across sites or species. Compared to natural conditions, plants in the common garden and experimental transplant populations did not show significant changes in differences in flowering times, suggesting that flowering phenology is under genetic control and is not influenced by soil environmental conditions.