Background/Question/Methods
Active management is being undertaken to restore native ecosystems in the Midwest Driftless Area ecoregion, but with little attention to historical vegetation and disturbance dynamics in the region.  An understanding of ecological history can assist by providing reference conditions from which to assess change and by providing an understanding of the potential outcomes of management activities.  Here we use the U.S. General Land Office’s original Public Land Survey (PLS) records to map pre-Euro-American (c. 1850) vegetation cover within the Driftless Area ecoregion.  Using cluster analysis, we develop a vegetation classification system based on surveyor designations of ecosystem types and bearing-tree data recorded within their notebooks.  We further assess the statistical robustness of our classification using cross-validation techniques. 

Results/Conclusions
We found a mosaic of prairie, savanna, and forest ecosystems.  Finer-scale, patchily distributed ecosystem types (e.g., wetlands, thickets) were not well represented as a result of the coarse resolution of our classification.  Dominant tree species in savanna and forest ecosystems included white oak (Quercus alba), bur oak (Q. macrocarpa), and black oak (Q. velutina).  Riparian forests, though confined to river valleys, were composed of a more diverse mixture of tree species (Ulmus americana, Acer saccharinum, Betula nigra, Celtis occidentalis, Fraxinus americana).  Mesic sugar maple-basswood (Acer saccharum-Tilia americana) forests, though common today, were generally confined to northern- and eastern-facing slopes.  Overall, the composition and structure of the historical vegetation suggests that both topography and fire played key rolls in the arrangement of vegetation patterns.  In addition to aiding in the establishment of restoration reference targets and management goals, our maps of pre-Euro-American vegetation can be used as baseline data in investigating vegetation-site relationships and in quantifying vegetation change over time.