For years paleoclimatologists have held the general view that the last deglaciation began around 17,000 to 15,000 yr ago, that the shape of the globally integrated deglacial curve was smoothly sigmoidal with the fastest rate of change centered around 11,000 yr ago, and that the deglaciation ended around 7000 to 5000 yr ago. Recent studies have challenged several aspects of this consensus and have suggested that the mechanisms responsible for the deglaciation are significantly different from those previously proposed. As a result, an international workshop was held at Airlie House in Virginia during May 2–6 of 1983 to evaluate a wide range of evidence relevant to this controversy. The conference results suggested that (1) the decrease in global ice volume occurred in two steps, with the dating of the earlier step still in doubt, but the later step occurring at about 10,000–7000 yr ago and (2) the most likely feedback mechanisms for accelerating the initial forcing by orbital variations are delayed bedrock rebound, marine downdraw/calving, and CO2 heating.

Three time series based on precisely dated annual tree-ring widths have been used to reconstruct June plus July degree days for the central Alaska and northwestern Canada region. The time series are the longest recently developed chronologies for the area and represent 57 core samples from 27 trees. The degree-year-to-year variation and day reconstruction, extending back to A.D. 1524, exhibits much extended warming and cooling trends including a general warming trend from about 1840 to 1960. The reconstruction is in agreement with some subaretic glacial information and with data of percentage melting from arctic ice cores. This and similar reconstructions can provide quantitative information to compare with general circulation and energy budget models for longer time periods than are available in recorded meteorological data.

Values of the gradient of the former ice surface can be inferred at points along a flow line from deviations of esker paths or transitions in esker type and numerically integrated to give the profile. A profile calculated in this way shows that during formation of the Katahdin esker system about 12,700 yr ago the ice thickness at distances of 10, 20, 50, 100, and 140 km from the terminus, which is about two thirds of the distance to the ice divide, was approximately 200, 300, 600, 750, and 900 m. The terminal reach was computed by assuming an unknown uniform basal drag and matching the profile to its known elevation at the terminus and known gradient 10 km upglacier. Correction for isostatic rebound based on the elevations of contemporaneous deltas and of the marine limit proved unnecessary, because the tilt due to the difference in uplift at the two ends of the profile is only 0.1 m km−1. With other plausible assumptions as to sea levels in the past, elevations of the marine limit, or exact location of the terminus the profile could be as much as roughly 100 m higher. It hits Mount Katahdin about 500 m below its summit, which is at 1600 m, in agreement with the geological evidence farther west. The steepening of the upper part of the profile suggests that the mountain dammed and diverted the ice. Basal drag computed from the profile varies from about 20 kPa (0.2 bar) near the terminus to 30 kPa (0.3 bar) at 100 km to 70 kPa (0.7 bar) at 140 km. The relatively low values away from the influence of Mount Katahdin agree with independent evidence from deep-sea cores of substantial late Wisconsin ice-sheet thinning without comparable areal reduction. The method has potential for application over wide areas that were occupied by the Laurentide and Scandinavian ice sheets.

Near Summer Lake in southern Oregon, 54 tephra beds of late Quaternary age are exposed in pluvial lake sediments of Lake Chewaucan. Seven of the tephra beds near the top can be correlated with tephra deposits younger than 117,000 yr at Mount St. Helens, Washington, at Crater Lake, Oregon, and in northwestern Nevada in the deposits of pluvial Lake Lahontan. However, most of the section at Summer Lake lies below the correlated units, and contains 39 tephra beds older than 117,000 yr.

Major-element chemistry of tephra glasses was determined by electron microprobe analysis; petrography supports the correlations made from chemical evidence. Compositions correlated range from 70 to 76% SiO2; the least silicic Summer Lake glass contained 57%.

Extrapolation of depositional rate suggests that most of the sediments at Summer Lake are younger than about 335,000 yr, but older lake beds containing tephra layers occur at one place. The long lacustrine record suggests that Lake Chewaucan persisted through the last interpluvial stage, and that the lake may have dried up at the end of the Pleistocene due to diversion of the Chewaucan River by relict shore features.

Drill cores of Enewetak Atoll, Marshall Islands, reveal six stratigraphic intervals, numbered in downward sequence, which represent vertical coral growth during Quaternary interglaciations. Radiocarbon dates indicate that the Holocene sea transgressed the emergent reef platform by about 8000 yr B.P. The reef grew rapidly upward (about 5 to 10 mm/yr) until about 6500 yr B.P. Afterward vertical growth slowed to about 0.5 mm/yr, then lateral development became dominant during the last several thousand years. The second interval is dated at 131,000 ± 3000 yr B.P. by uranium series. This unit correlates with oxygen-isotope substage 5e and with terrace VIIa of Huon Peninsula, New Guinea, and of Main Reef-2 terrace at Atauro Island. The third interval is not dated because corals were recrystallized and it is tentatively correlated with either oxygen-isotope stages 7 or 9. The age of the fourth interval is estimated at 454,000 ± 100,000 yr B.P. from measured 234U238U activity ratios. This unit is correlated with either oxygen-isotope stage 9, 11, or 13.

Paleosols that formed primarily before the last glaciation cannot be adequately explained by a model of Holocene-like interglacial periods of about 10,000 yr in length which has been inferred from the marine 18O16O record. Geological and palynological evidence from terrestrial, midlatitude sites suggests that temperate conditions conducive to soil formation existed for periods greatly in excess of 10,000 yr during the middle and late Pleistocene. Available data on soils formed during this interval from Western Europe and midwestern United States are best explained by relatively long interglaciations or the development of composite soils over a number of temperate periods. Climatic contrasts between the interglaciations appear to have been overemphasized: widespread assumptions regarding greater warmth in interglaciations prior to the present one are generally unproven. In the absence of precise dating, progress in relating paleosols to fragmentary local records of Quaternary climatic changes will be achieved by detailed analytical techniques such as soil micromorphology.

The concepts of pollen source area and of production and dispersal biases in pollen representation are quantified by means of a simple theoretical model. Source areas and relative pollen representation are shown to depend on basin size according to functions that describe the amount of pollen remaining airborne at increasing distances from single pollen sources. The form of these functions is determined by physical processes. Standard formulas for elevated sources do not apply, but the integrated form of Sutton's equation for particle dispersal from a ground-level source gives useful approximations applicable to pollen transport over a forest canopy. Simulations using this equation yielded source areas that increased realistically with basin size, showed substantial differences between source areas for pollen grains with different deposition velocities, and predicted that lighter pollen grains should become better represented with increasing basin size. All of these predictions are qualitatively consistent with present knowledge of the characteristics of pollen assemblages in different depositional environments. The model further allows parameters that can be estimated by statistical calibration methods to be predicted from underlying physical quantities. This extension suggests procedures for testing the theory with quantitative data on surface pollen and forest composition. Preliminary results showed reasonable agreement between estimated and predicted values of dispersal indices for the most abundant taxa in pollen spectra from the northern midwestern United States.

Dissimilarity coefficients measure the difference between multivariate samples and provide a quantitative aid to the identification of modern analogs for fossil pollen samples. How eight coefficients responded to differences among modern pollen samples from eastern North America was tested. These coefficients represent three different classes: (1) unweighted coefficients that are most strongly influenced by large-valued pollen types, (2) equal-weight coefficients that weight all pollen types equally but can be too sensitive to variations among rare types, and (3) signal-to-noise coefficients that are intermediate in their weighting of pollen types. The studies with modern pollen allowed definition of critical values for each coefficient, which, when not exceeded, indicate that two pollen samples originate from the same vegetation region. Dissimilarity coefficients were used to compare modern and fossil pollen samples, and modern samples so similar to fossil samples were found that most of three late Quaternary pollen diagrams could be “reconstructed” by substituting modern samples for fossil samples. When the coefficients indicated that the fossil spectra had no modern analogs, then the reconstructed diagrams did not match all aspects of the originals. No modern analogs existed for samples from before 9300 yr B.P. at Kirchner Marsh, Minnesota, and from before 11,000 yr B.P. at Wintergreen Lake, Michigan, but modern analogs existed for almost all Holocene samples from these two sites and Brandreth Bog, New York.

A 33,000-yr pollen record from Carp Lake provides information on the vegetation history of the forest/steppe border in the southwestern Columbia Basin. The site is located in the Pinus ponderosa Zone but through much of late Quaternary time the area was probably treeless. Pollen assemblages in sediments dating from 33,000 to 23,500 yr B.P. suggest a period of temperate climate and steppe coinciding with the end of the Olympia Interglaciation. The Fraser Glaciation (ca. 25,000–10,000 yr B.P.) was a period of periglacial steppe or tundra vegetation and conditions too dry and cold to support forests at low altitudes. Aridity is also inferred from the low level of the lake between 21,000 and 8500 yr B.P., and especially after about 13,500 yr B.P. About 10,000 yr B.P. Chenopodiineae and other temperate taxa spread locally, providing palynological evidence for a shift from cold, dry to warm, dry conditions. Pine woodland developed at the site with the onset of humid conditions at 8500 yr B.P.; further cooling is suggested at 4000 yr B.P., when Pseudotsuga and Abies were established locally.

During the Quaternary glacial episodes, when sea level was considerably lower, Asia and North America were linked by large extensions of circumarctic land (Beringia), which remained unglaciated. This land mass served not only as a biogeographical bridge for plants, animals, and humans, but also supported a biome very different from present tundra or boreal coniferous forests, which was dominated by steppes and a rich mammalian megafauna. Carbon stable isotope ratios of Beringian late Pleistocene mummified remains of bison, equids, mammoth, caribou, musk-ox, moose, woolly rhino, and other undetermined species, found preserved in permafrost, indicate that these megaherbivores fed exclusively on C3 plants, and that C4 grasses were not differentially ingested by bison, as previously suggested. Paleoclimatic constraints probably prevented the formation of a warm-season (C4) guild during the later part of the growing season in the steppes of Beringia during the last glaciation.

This study tests the hypothesis that the number and distribution of some native American languages may be related to ice-margin changes of the Wisconsin glaciation. The analysis indicated that the number of languages per unit area is much greater in unglaciated areas of the last glacial maximum than in glaciated areas. The pattern of languge overlap between land areas sequentially exposed during deglaciation appears to indicate the direction of movement of populations from the periphery toward the core of the area once covered by the Wisconsin Ice Sheet. The data strongly indicate that North America was inhabited prior to the Wisconsin glacial maximum, because glacial maximum conditions apparently influenced linguistic distributions. Evidence suggests that ancestral Eskimo-Aleut and Na-Dene speakers occupied the northwestern edge of the continental ice mass, and that ancestral Algonquian speakers were south of the ice mass during the Wisconsin glacial maximum (approximately 18,000 yr ago). These three linguistic groups were the principal ones to spreas into areas exposed by the recession of the Wisconsin ice.