Postulations on the existence of forest refugia in parts of Beringia during the last glacial have been, in large part, based on ambiguous evidence. Existing data on radiocarbon-dated and identified fossil wood and macrofossils from Alaska and northwest Canada are synthesized here and are augmented by results of palynological studies in an effort to show the persistence of some, and total extinction of other, tree and large shrub species. Possible dispersal routes taken by species that reinvaded Beringia in postglacial times are also reconstructed from the fossil record. Macrofossil and pollen evidence, when combined with climatic factors, makes cottonwood a good candidate for survival during the last glacial. Larch and aspen are also candidates, though the evidence for them is less positive. Pollen and macrofossils of alder are very scarce in deposits of the last glacial age, and if it survived at all, it was probably in very isolated vegetatively reproducing clones. Shrub birch may have been present in Beringia, but tree birch probably was reintroduced during the Holocene. Spruce also appears to have been absent in Alaska from about 30,000 to 11,500 yr ago and probably reinvaded Beringia from a refugium south of the Laurentide ice sheet.

The various lobes and segments of the southern periphery of the Laurentide ice sheet reached their maximum extension at different times between 21,000 and 14,000 yr ago, but the CLIMAP date of 18,000 yr ago is taken as a reference level to review the distribution of major vegetational formations in central and eastern United States. Tundra was apparently confined to a narrow belt peripheral to the ice margin only in the Minnesota area and from northern Pennsylvania to New England, with extensions down the crest of the Appalachian Highlands at least as far as Maryland. Some areas south of the Great Lakes may later have been marked by treeless vegetation briefly as the ice retreated. The boreal forest to the south in the central United States was dominated by spruce; the jack pine that had prevailed during previous times was apparently eliminated by the time the ice reached its maximum. In the Appalachian Highlands and the Atlantic Coastal Plain, however, jack pine occurred along with spruce, which decreased in importance southward. The southern limit of the boreal forest in the Southeast was perhaps somewhere in southern Georgia and Alabama. Oak and other temperate deciduous trees were minor components of the boreal coniferous forests especially in the southern Appalacchians, but there is no evidence yet in the southeastern states for a relic mixed mesophytic forest 18,000 yr ago similar to the rich modern deciduous forests of the region, except possibly in the Lower Mississippi Valley. The climate in much of the Southeast was apparently dry as well as cool at that time; in Florida oak/pine scrub and prairie-like openings prevailed, and all but the deepest lakes dried up.

The causes of Antarctic glaciation are analyzed by means of numeral experiments based on the three-dimensional thermodynamic model of a large ice sheet. Refrigeration of the climate between the Eocene and the Oligocene was due to the opening of the passage south of Australia and to the formation of the South Ring Stream. Calculations have shown that this led to the development of the East Antarctic Ice Sheet which might have existed in spite of relatively high temperatures of the surrounding ocean air. A new cooling of the climate in the Middle Miocene is connected with the fact that the South Ring Stream found its way through the Drake Passage glaciers spreading on to the Western Antarctic. Between Miocene and Pliocene, glaciation of the South Polar regions was at its maximum due to the regression of the world ocean. In Quaternary time, sea level was lowering due to the glaciation of the Northern Hemisphere, which resulted in glacier growth in the Antarctic. The anticipated warming of the climate due to the activity of man is not likely to bring about any considerable change in the size of the East Antarctic Ice Sheet.

The marine Pliocene-Pleistocene sequence of Iceland contains four main assemblages of mollusks: (1) Venerupis rhomboides et al., (2) Serripes groenlandicus et al., (3) Portlandia arctica et al., and (4) recent species. Analysis of the stratigraphic distribution of 100 species shows that the percentage of extinct species in the assemblages decreases from 40% in the oldest to 4–8% in the youngest. There is also a change from south boreal affinities in the oldest assemblage to north boreal in the youngest. The molluskan assemblages of Iceland can be correlated with Pliocene-Pleistocene assemblages of England and North Europe: (1) Coralline Crag, (2) Red Crag—lower part of the Icenian strata, (3) upper part of the Icenian-Cromerian strata. The upper boundaries of the Pliocene and Eopleistocene are defined by paleomagnetic data and radiological dating. The upper boundary of the Pliocene presumably should be established in the basalts lying between the Tjörnes and the Breidavik deposits by changes between the second and third mollusk assemblages and by the Gilsa paleomagnetic event.

The scarcity of documented numerical relations between rock weathering and time has led to a common assumption that rates of weathering are linear. This assumption has been strengthened by studies that have calculated long-term average rates. However, little theoretical or empirical evidence exists to support linear rates for most chemical-weathering processes, with the exception of congruent dissolution processes. The few previous studies of rock-weathering rates that contain quantitative documentation of the relation between chemical weathering and time suggest that the rates of most weathering processes decrease with time. Recent studies of weathering rinds on basaltic and andesitic stones in glacial deposits in the western United States also clearly demonstrate that rock-weathering processes slow with time. Some weathering processes appear to conform to exponential functions of time, such as the square-root time function for hydration of volcanic glass, which conforms to the theoretical predictions of diffusion kinetics. However, weathering of mineralogically heterogeneous rocks involves complex physical and chemical processes that generally can be expressed only empirically, commonly by way of logarithmic time functions. Incongruent dissolution and other weathering processes produce residues, which are commonly used as measures of weathering. These residues appear to slow movement of water to unaltered material and impede chemical transport away from it. If weathering residues impede weathering processes then rates of weathering and rates of residue production are inversely proportional to some function of the residue thickness. This results in simple mathematical analogs for weathering that imply nonlinear time functions. The rate of weathering becomes constant only when an equilibrium thickness of the residue is reached. Because weathering residues are relatively stable chemically, and because physical removal of residues below the ground surface is slight, many weathering features require considerable time to reach constant rates of change. For weathering rinds on volcanic stones in the western United States, this time is at least 0.5 my.

A model for predicting the growth and decay of ice sheets based on the astronomical theory of climate change is presented. The purpose of the study in part is to isolate the role of the ice-sheet physics and earth response under varying ice load by simplifying to the extreme the role of the hydrosphere-atmosphere. Ice sheet physics and the response of the lithosphere-asthenosphere under the ice load are modeled explicitly. Insolation anomalies (taken at a fixed latitude) directly force latitudinal displacement of the snow line. Accumulation rate a, and ablation rate a′ evaluated at mean sea level are specificed as external constants; a,a′ decrease linearly with ice sheet elevation. Rough tuning of the model to the general shape of the ice-volume record of the last two major glacials determines the external constants. Model predictions of the ages of several events in the last major glaciation compare well with the radiological ages. The six glacial terminatios (I–VI) over the last 600,000 yr are identified and the predicted ages compare reasonably well with the δ18O record for two deep-sea cores. A direct comparison of model power spectra of ice volume as a function of period with spectra of the δ18O record shows apparent underprediction of power near 100,000 yr. When a quantitative but heuristic method for taking into account the “red noise” spectrum evident in the geological records is used, a much more favorable comparison is possible. The model prediction lends support to the hypothesis that the nonlinearity of the ice-sheet physics is responsible for the 100,000-yr periodicity in the geological record of the late Pleistocene.

Forest ecology suffers from a lack of long-term community records. Preserved pollen data are richer in such information than is generally realized. By applying suitable statistical techniques to pollen records, one can learn much about competition, succession, and population dynamics in past tree communities. In this study, preserved pollen records from Everitt Lake, Nova Scotia, are analyzed as time series. Time domain studies reveal the post-fire responses of individual tree taxa. Correlograms yield models of past forest succession patterns. The models explain some effects of changing fire frequency, thus suggesting mechanisms by which fire, competition, and climate combine to produce long-term forest composition changes. Frequency domain studies suggest relationships between disturbance cycles, stand composition, and forest mosaics. Fire frequencies are seen to be highest where fire-dependent species abound and most regular where tree stands have uniform, not mixed, composition.

Three unstable ephemeral-stream channels (arroyos), which drain source areas that have high sediment yields ranging from predominantly sand (Arroyo Calabasas) to a mixture of sand, silt, and clay (Sand Creek) to largely silt and clay (Sage Creek), were resurveyed to provide data on the rates and mechanics of erosion and sedimentation processes during periods ranging from 14 to 22 yr. Channel morphology changed significantly. Erosion occurred through nickpoint recession and bank collapse, but erosional reaches are separated by aggrading or stable-channel reaches. In general, sediment that is eroded, as the nickpoint recedes upstream, is trapped in the widened channel downstream. In this manner sediment is transported episodically out of these basins during a series of cut-and-fill cycles. The manner by which the channels aggrade and the morphology of the aggraded stable channels are controlled by the sediment type. The wide and shallow channel of Arroyo Calabasas is filled by vertical accretion of sand-size sediment. The narrow and deep channels of Sage Creek and Sand Creek are created by the lateral accretion of cohesive fine-grained sediment. The channel modification and the cut-and-fill episodes are dependent on high sediment yields, and therefore they are independent of subtle climatic shifts. Cut-and-fill deposits that have been created in this manner should not be equivalent in age from basin to basin, and therefore channel trenching and filling in the semiarid western United States during the Holocene need not be synchronous.

Temperatures for the past 2700 yr are estimated using well-dated pollen data from northwestern lower Michigan. The pollen data were sampled from sediment cores of four lakes along a 75-km transect, with fine-grained morainic soils around the two western lakes and sandy outwash soils around the lakes to the east. Climatic reconstructions based on the pollen data from the sandy sites show less temperature change than the reconstructions from the other sites, because variations in the composition of the vegetation at the sandy sites are edaphically restricted. One of the cores studied was dated by counting visible annual laminations (varves). The cores from the other lakes were dated based on three radiocarbon dates per core as well as the historically determined age of the settlement horizons. All the time scales were cross-checked using pollen-stratigraphic correlation between the four sites. A calibration function was developed using a network of modern pollen and climate data covering all of lower Michigan. Based on this calibration function, the 2700-yr reconstruction for Marion Lake indicates an estimated growing-season temperature range of 1.3°C between extreme 30-yr means. Mild conditions persisted prior to ca. A.D. 400, but a cold interval occurred between ca. A.D. 500 and 800. The well-marked warm period evident from ca. A.D. 1000 to 1200 was the last time when temperatures were about equal to the 1931–1960 mean. A prolonged longed cooling occurred after A.D. 1200 and reached 1°C below the 1931–1960 mean by the 1700s. A warming of 0.5°C is indicated from ca. A.D. 1750 to 1850. The estimated temperatures for the 1830s at Marion Lake agree with the instrumental data for that period and this provides some validation of the calibration-function results.

The sequence of vegetation phases in the late Glacial was studied in a sediment section from the bog Tourbière de Chirens by means of pollen analysis. 18O/16O ratios of samples of lacustrine marl, obtained from the same profile, reflect variations of 18O/16O in precipitation and thus provide an additional, independent paleoclimatic record. The observed 18O/16O variations agree well with the climatic history as deduced from pollen analysis. The climatic transition from the Oldest Dryas to the Bølling period sensu lato, as well as the beginning and end of the Younger Dryas cold phase, is marked by abrupt changes in the 18O/16O ratio which were observed also in other regions of the Alps. These drastic climatic changes probably took place simultaneously over large areas of Central Europe and occurred within short time spans.

Information from 240 km of high-resolution seismic reflection profiles has been analyzed to show the bathymetric and subsurface configuration of southern Lake George in the southeastern corner of the Adirondack Mountains, New York. Three units have been identified and sampled in 13 piston cores as long as 7 m and 4 grab samples; they are glacial drift, glaciolacustrine nonorganic clay, and Holocene lake mud rich in organic material. Three deep bedrock basins controlled glacial, glaciolacustrine, and postglacial deposition within the lake. Glaciolacustrine clay is more than 30 m thick in these basins but is generally absent in water depths less than 20 m. An unconformity separates glaciolacustrine clay from overlying Holocene mud in water depths less than 30 m, but the contact is conformable and transitional in deeper water. The unconformity may have originated from subaqueous or subaerial erosion during a low stage of lake level which probably occurred between 10,000 and 700 yr B.P. Holocene lake mud is thin to absent in the shallower waters separating the three basins, but reaches 15-m thickness near the entrance to The Narrows. A new radiocarbon date of 6950 ± 60 yr B.P. was obtained from a wood fragment which was found in the Holocene lake mud. We found no clear evidence of postglacial tectonic disturbances of the lake sediments although recent releveling profiles suggest that the Adirondack Mountains are undergoing contemporary uplift.

Examination of late Pleistocene packrat middens from the northern and central Chihuahuan Desert disclosed macrofossils of Colorado piñon (Pinus edulis) and Texas piñon (P. remota). Radiocarbon dating indicates that Texas piñon was widespread in Trans-Pecos Texas and northeastern Mexico between 30,000 and 11,000 yr B.P. Today it is found in small refugia east of its former range. In the late Pleistocene Colorado piñon occurred at lower elevations on the northern edge of the Chihuahuan Desert. Both species occurred in the Hueco Mountains, near El Paso, Texas. No clear evidence was found of the presence of Mexican piñon (P. cembroides), though today it is abundant in the Davis and Chisos Mountains. A paleoclimate is postulated that had the following characteristics: increased winter precipitation from Pacific frontal sources, reduced summer temperatures and precipitation, and milder winter temperatures due to a reduced frequency of Arctic airmass incursion. Winter precipitation appears to have decreased from north to south, while winter temperatures, and, possibly, summer precipitation, increased from north to south. During the late Pleistocene, the northern Chihuahuan Desert was dominated by woodlands of piñon pines, junipers, and oaks. The desert-scrub communities that characterize the area today are a Holocene phenomenon.

A number of ancient shorelines formed by late-Pleistocene proglacial lakes have been found in eastern upper Michigan. These shorelines delimit several water planes, the uppermost of which is correlated with the Main Lake Algonquin stage. This correlation is based on the continuity of the highest water plane with Main Algonquin shorelines in Wisconsin and Ontario, the strength of the shoreline features, its altitudinal relationship with lower water planes, and a reinterpretation of radiocarbon dates from the Sault Ste. Maria area. The isobases of this water plane have a bearing of S75°E. At the time of the maximum extent of Lake Algonquin, ca. 10,600 yr B.P., its northern, ice-limited border lay along the Munising moraine, the northernmost of the two main morainic systems of eastern upper Michigan. This interpretation lends support to the idea of a period of slow deglaciation from ca. 11,000 to 10,000 yr B.P. An ice lobe occupied the central Lake Superior basin until early Holocene time. Radiocarbon dates on wood found beneath till or outwash at several sites indicate a minor ice readvance from the central Lake Superior basin ca. 10,000 yr B.P. If true, this would have prevented the development of the post-Duluth series of glacial lakes in the western Lake Superior basin until ca. 9900 yr B.P., well after the end of the main Lake Algonquin stage.

Nonsorted polygons in the uppermost 2 to 3 m beneath Pleistocene surfaces indicate permafrost at 1340 m and higher elevations in the intermontane and piedmont plains of Wyoming during the Wisconsin, and perhaps earlier, glacial maxima. The polygons, as much as 10 m in diameter, are delineated by wedges that vary in depths, range from narrow to moderately flared forms, and deform host materials. The wedges have silty fine-to-medium sand matrices (largely eolian) with pebbles or clasts from hosts of gravel or bedrock. Some wedges may reflect seasonal cracking in a periglacial active zone, but most are either permafrost sand-wedge relics or, less commonly, ice-wedge casts. Alternative explanations are rejected largely because similar features are apparently lacking in the lower and warmer plains from eastern Colorado southward. The wedges suggest an arid, windy, periglacial environment whose mean-annual temperatures are conservatively estimated as some 10° to 13°C colder than those at present. Although late Wisconsin-early Holocene floral and faunal evidence indicates lowered montane biotic zones, the eolian and periglacial features indicate a lack of extensive forest cover on the basin floors. In conjunction with vertebrate-fossil associations of grazing and tundra animals, the wedges may provide a parallel line of evidence for a former periglacial steppe, or “steppe-tundra”, in the Wyoming basins.

New cross sections and dates from along the Pomme de Terre River clarify the complex local history of valley development and floodplain sedimentation. The observed history begins with a series of ancient bedrock strath terraces that record past bedrock valley positions at 15.5 to more than 58 m above the modern bedrock floor. Each strath is capped by 1–2 m of channel gravel and sand permeated by red clay. Sometime previous to ca. 140,000 yr B.P., a much lower bedrock valley only about 5–6 m above the modern level was excavated. By 140,000 yr B.P., accumulation of red and gray mottled silty clay had commenced, and had reached to 8.5 m above the modern floodplain before 48,900 ± 900 14C yr B.P. Sometime between ca. 49,000 and 45,000 14C yr B.P., erosion caused abandonment of an oxbow meander, and lowered the bedrock valley to about its present depth. Younger yellowish-red and gray mottled silty clay alluvium then began accumulating. This mid-Wisconsinan fill reached to 2.5 m above the modern floodplain sometime before 31,800 ± 1340 14C yr B.P., at which time another erosional phase was in progress. A late Wisconsinan olive clay accumulated between 27,480 ± 1950 and ca. 23,000 14C yr B.P., followed by approximate stability until 13,550 ± 400 14C yr B.P. After stability, an erosional episode began, but by 10,200 ± 330 14C yr B.P., deposition of a distinctive brown clayey silt was underway. This early Holocene fill reached to about the same level as the mid-Wisconsinan fill by 8100 ± 140 14C yr B.P. Erosion occurred between this date and 7490 ± 170 14C yr B.P., but the former floodplain level was rapidly reattained, and was apparently stable until ca. 5000 14C yr B.P. Finally, erosional unconformities and 17 dates from the brown clayey silt, and from younger grayish-brown silty sand underlying the modern floodplain, record subsequent episodes of floodplain erosion at ca. 5000, 2900, 1500 and 350 14C yr B.P. The timing of Pomme de Terre floodplain sedimentary regimes, characterized by net aggradation, erosion, or stability, may have been controlled by climate. In particular, both periods of stability appear to have been coeval to times of strongly zonal upper atmospheric circulation. Intensified zonal circulation would have resulted in less frequent large floods and an increased dominance by floods of small to moderate size. In contrast, there are no obvious parallels to be drawn between this local alluvial history and sea level or glacial outwash induced baselevel changes.

A vertically integrated ice-flow model suitable for use in climate studies is formulated. Large continental ice sheets may be characterized by two fundamental quantities: the height-to-width ratio, and the steepness of the edge. So it is natural to develop a model containing two parameters that can be chosen to give the right values of those characteristic quantities. The result is a model that is close to M. A. W. Mahaffy's (Journal of Geophysical Research, 81, 1059–1066 (1976)). The model is used to study glaciation in Europe. Dropping the level of zero mass balance creates small stable ice caps in the Alps and the Scandinavian mountains. If the drop exceeds 600 m (with respect to present-day conditions), the feedback between ice-sheet height and mass balance becomes dominating and the Fennoscandian Ice Sheet keeps growing. It does not reach an equilibrium state within 60,000 yr. An experiment simulating rapid onset of a glacial cycle shows that the growth of ice volume in Europe is smaller than that in northern America (J. T. Andrews and M. A. W. Mahaffy, Quaternary Research, 6, 167–183 (1976)). After 10,000 yr, the volume of the Fennoscandian Ice Sheet (2 × 1015 m3) is about half the volume of the Laurentide Ice Sheet. This leaves the “observed” sea-level lowering in the period 125,000–115,000 yr B.P. (estimates center around 50 m) unexplained.

The vertebrate fauna of the last 30,000 radiocarbon years in the Grand Canyon is reviewed. Faunas accompanied with 92 14C dates have been analyzed from nine cave sites (four systematically excavated) and 50 packrat middens. Reasonably precise chronological and environmental data of late Pleistocene and Holocene age were obtained through dung studies in Rampart, Muav, and Stanton's Caves; from the numerous packrat middens; and from a ringtail refuse deposit in Vulture Cave. The desert tortoise, 8 species of lizards, 12 species of snakes, 68 species of birds, and 33 species of mammals are identified. Extinct animals include the avian carrion feeder, Teratornis merriami, and the mammalian herbivores, Oreamnos harringtoni, Camelops cf. hesternus, Equus sp., and Nothrotheriops shastense. There is no apparent abrupt end to the late Pleistocene as observed in the Grand Canyon fossil faunal or floral record. Animal and plant taxa of the Grand Canyon responded individually to the changes in climate of the last 30,000 yr. Both animal and plant fossil assemblages indicate that a pre-full glacial, a full glacial, and a late glacial woodland community with many less dominant desert taxa were slowly replaced by a Holocene desert community. All woodland taxa were absent from the lower elevations of the Grand Canyon by 8500 yr B.P.

K-Ar ages (A. Abdel-Monem, P. D. Watkins, and P. W. Gast, 1971, American Journal of Science 271, 490–521; this paper) and revised paleontological determinations (J. Meco, 1977, “Los Strombus neogenos y cuatenarios del Atlantico euroafricano”, Las Palmas, Ediciones del Excmo. Cabildo Insular de Gran Canaria) show that “Quaternary” (R. Crofts, 1967, Quaternaria 9, 247–260; G. Lecointre, K. J. Tinkler, and G. Richards, 1967, Academy of Natural Science of Philadelphia Proceedings 119, 325–344) littoral deposits on Lanzarote and Fuerteventura are early Pliocene and late Pleistocene. Early and middle Pleistocene strand lines are not represented. Early Pliocene littoral and marine deposits contain a characteristic fossil assemblage: Strombus coronatus, Nerità emiliana, Gryphaea virleti, Patella cf. intermedia, and Rothpletzia rudista. Differences in elevation record differential post-Pliocene uplift of the coastal platforms on which they lie. Late Pleistocene beach deposits at low elevations belong to two groups, an older with Strombus bubonius and a younger without. Differences in elevation of early Pliocene littoral deposits are reflected by differences in elevation of late Pleistocene beach deposits nearby.

The paleogeographic and particularly palaeoclimatic evolution in the Congo at the end of the Quatemary has been investigated thoroughly with a study of grain surfaces using the scanning electron microscope. The succession of distinct physicochemical environment marks is characterized with the help of other methods (radiocarbon, micropaleontology). During a still-existing tropical climate, arid periods are contemporaneous with regressive phases (eolian, fluviatile, or pedogenetic reliefs of the grains) and humid periods correspond to transgressive phases (infratidal, Intertidal, deltaic, lagoonal marks).