Abstract Environmental changes during deglaciation are among the most profound and rapid of any during the Quaternary. Palynological evidence reveals the changing patterns of vegetation on late Quaternary landscapes, and provides abundant evidence for large shifts during the late Wisconsin termination in particular. Changes in the oxygen-isotope ratios in marine microfossils show that the large, rapid termination of the most recent ice age is similar to those during previous cycles back to at least 400 ka (Broecker and van Donk, 1970). The abundant evidence for changes from 18 to 6 ka (thousands of radiocarbon years B.P.) therefore provides clues to the possible behavior of terrestrial and marine ecosystems during earlier times. Palynological evidence for major changes in the North American vegetation began with the contributions of pioneers like Deevey (1939, 1949, 1951), Sears (1942), Leopold (1956), and Fries (1962) and has grown dramatically since Wright and others (1963) and McAndrews (1966) demonstrated that the prairie-forest border in Minnesota had moved considerable distances in apparent response to Holocene climatic changes. More recently, increases in both the amount of data and in the use of computers to compile and evaluate these data has permitted mapping of population expansions (Davis, 1981a) and population densities (Bemabo and Webb, 1977; Webb, 1987). These maps reveal broad-scale patterns and complex behavior in the vegetational changes across eastern North America during the past 18,000 yr. Here we use three separate approaches to the reconstruction of vegetation during the most recent deglaciation. The first involves mapping the changing patterns of distribution and abundance of selected plant taxa across eastern North America. The second approach uses the combined evidence from several taxa to illustrate how the different taxa interacted through time and to reveal patterns in the changing composition of the vegetation. The third is a numerical approach designed to reveal the rate of change in the vegetation per unit time, and thus to show whether periods of relative constancy have existed between periods of especially rapid change, or whether changes have been gradual throughout the past 18,000 yr. Each of these approaches produces evidence of vegetational changes that are not only climatically controlled but also ecologically mediated by local events such as fire and other disturbance. By using the three approaches together, we attempt to differentiate the nature, timing, and causes of these changes. Fossil pollen stratigraphies provide the primary data for our study. These stratigraphies come from among the several hundred sites in eastern North America that have been studied over the past 25 years by modern palynological methods and radiometric dating. Most of these data are now in a computer-based repository at Brown University and thus available for systematic study.