Conifer wood, probably spruce (Picea sp.), of middle Wisconsinan age (29,200 ± 500 yr B.P.) was recovered from late-glacial lake sediments from Upper South Branch Pond, Maine. If the wood was derived from a local source, deglaciation of part of northern New England is suggested for this time. The occurrence also has implications for understanding the problem associated with radiocarbon dating of bulk lake sediment containing small amounts of organic matter.

Pollen influx and percentage diagrams were prepared from an 11.4 m core from Moulton Pond, Maine. The pond basin was deglaciated about 14,000 y. a., after which it was located on an island in a sea of subarctic character until about 12,400 y. a. when the surrounding area emerged from the sea. The terrestrial vegetation was tundra until about 10,000 y. a. A change in the tundra vegetation is synchronous with the emergence from the sea, but synchroneity with the Pineo Ridge glacial readvance, which reached its maximum 50 km to the east of the pond about 12,700 y. a., is also possible because of imprecision in the dating. Comparisons of the Moulton Pond results with late-glacial pollen sequences elsewhere in eastern United States and adjacent Canada reveal a lack of synchroneity in vegetational changes casting doubt on claims of major broad-scale climatic shifts over the entire area.

The tundra period at Moulton Pond ended with a transition of a few hundred years to partly open, relatively xeric forests of low diversity dominated by white pine, oak, and birch trees. There was no intervening boreal forest. In the postglacial period the vegetation was continually changing, including in the early portion a series of immigrations of temperate tree taxa which later became important in the forests. The transient nature of these assemblages is further indicated by their differences from the closest modern analogs. From about 7100 y. a. until settlement by Europeans 200 y. a., the forests were closed. A major decline of conifers centering about 4700 y. a. was followed by maxima of mesic hardwoods about a thousand years later. In the most recent 2000 yr, the pollen record suggests greater environmental severity, evidenced by increasing spruce. But for the entire postglacial period, the closest modern vegetational analogs are all in the conifer-hardwood region. Much of the postglacial pollen sequence is inexplicable in climatic terms, as evidenced by nonsynchronous behavior of hemlock and beech.

The pollen influx diagram is useful for distinguishing tundra from forest, but for the postglacial period it is difficult to interpret. Pollen influx data are strongly affected by shifts in the pattern of sedimentation in lakes. We propose that such shifts account for the major changes in influx in mid- and late-postglacial time at Moulton Pond and at Rogers Lake, Connecticut. This complicates the interpretation of influx data which otherwise are superior to percentage data.

By mapping and summarizing 478 pollen counts from surface samples at 406 locations in eastern North America, this study documents the relationships between the distributions of pollen and vegetation on a continental scale. The most common pollen types in this region are pine, birch, oak, and spruce. Maps showing isopercentage contours or isopolls for 13 important pollen types reflect the general N-S zonation of the vegetation. The maps and tabulations of average pollen spectra for the six major vegetational regions indicate high values for the following pollen types in each region: (1) tundra-nonarboreal birch, sedge, and alder; (2) forest/tundra-spruce, nonarboreal birch and alder; (3) boreal forest-spruce, jack pine (type), and arboreal birch with fir in the southeastern part; (4) conifer/hardwood forest-white pine, arboreal birch, and hemlock with beech, maple, and oak in the southern part; (5) deciduous forest-oak, pine, hickory, and elm, with beech and maple in the northern part, and highest values of oak and hickory west of the Appalachian crest; and (6) southeastern forest-pine, oak, hickory, tupelo, and Myricaceae. In some cases, less abundant pollen types are diagnostic for the region, e.g., bald cypress in the southeast. In the conifer-hardwood region and southward, pollen of weeds associated with deforestation and agriculture is abundant. The maps also show that much of southeastern U.S. and the area just to the east of Hudson Bay are in need of additional sampling. At 51 of the sites, absolute pollen frequencies (APF; grains/ml lake sediment) were obtained. These confirm the major conclusions from the percentage data, but differences are evident, e.g., the percentages of alder pollen peak in the tundra whereas alder APFs peak in the boreal forest, and spruce percentages peak in the forest-tundra whereas spruce APFs peak in the boreal forest. Because the APF data reflect the patterns of absolute abundance of individual taxa in the vegetation as well as the overall forest densities, future counts of modern pollen should include APF determinations. The effects of sedimentation processes on APF quantities indicate that APF samples should be obtained from moderate size lakes of similar morphology and hydrology and that, in each lake, several samples from the profundal zone should be pooled to create a sample representative of that lake.