Analyses of various methods used to quantify historical plant assemblages in east-central Mississippi

Five sets of data (faunal remains, charcoal, plant impressions in daub from house sites, pollen from soil samples, and General Land Office notes) from three archeological sites (Yarborough, Lyon's Bluff, and Josey's Farm) in the Black Belt region of Mississippi were used to test the adequacy of each database or their combinations to best estimate the past floral diversity.; The faunal remains indicated that the people ate a wide diversity of taxa and mostly small items gathered from all habitats. Some emphasis was indicated for forested terrestrial sites and backwater areas of aquatic sites.; Seeming differences among sites shown by the similarity indices actually resulted from differential sampling of each method. The canonical correspondence analysis resulted in an inadequate separation of floral assemblages between sites because of the large sampling variability in methodologies. The combination of GLO notes and charcoal gave the most similar floral assemblages. The only significant variable from the factorial ANOVA test was number of individuals across all sites and methods.; Within its limitations of selective sampling, daub has a potential for increasing the accuracy of historical floral reconstruction. These data from all three sites supported the hypothesis that the Black Belt physiographic region was generally an oak-hickory forest mosaic as opposed to a prairie-scrub successional forest.; Lack of complete sets of data from all sites and differential sampling, large variations, and biases of each method all contribute to the inability to present quantitative descriptions of the historical floral assemblage or distinguish between sites on a local basis.