The recent extension of (1) the residual Δ14C curve back to 11,400 cal yr B.P. and (2) the lake-level reconstruction in the Jura back to ca. 13,500 cal yr B.P. offers the opportunity of testing by proxy data the relationships between climate, atmospheric 14C, the sun, and the ocean recently suggested from the atmospheric 14C record. The climatic significance of the Jura record is supported by correlations with climatic oscillations reconstructed in the Alps from glaciers and timberline movements. Correspondence between the 14C and paleoclimatic record from the Jura suggests a working hypothesis: two intervals within the Holocene can be distinguished in the middle latitudes of western and central Europe. An early Holocene period shows abrupt climatic oscillations linked to ocean forcing. Major colder climate phases developed between ca. 9000 and 8800, and between ca. 8000 and 7000 cal yr B.C. that coincide with higher Δ14C values. After 6000 cal yr B.C., a second period is characterized by smoother multicentury climatic oscillations linked to solar forcing.

Since the turn of the century, a Pliocene-Pleistocene connection between the Death Valley-Owens River pluvial system and the Colorado River drainage basin has been frequently postulated. The two most commonly proposed routes involve (1) a southward overflow from the Death Valley Lake system or (2) southward migration of the Mojave River between its present course and a more southerly route. Under the present topographic regime, a Death Valley Lake capable of overflowing the bedrock saddle at Ludlow, California and discharging southward into the Bristol, Cadiz, and Danby Lake basins (and eventually the Colorado River) would be over 12,000 km2 in size. Few surface and subsurface indicators exist to support either a fluvial or lacustrine connection. Evidence from deep cores and boreholes drilled in Soda, Bristol, Cadiz, and Danby dry lake basins indicate that a hydrologic connection has not occurred during the past 4 myr. No well-documented paleoshoreline features have been located at elevations corresponding to the above hydrologic systems in Death Valley, Silver-Soda, Bristol, or Danby Lake basins. In the Cadiz, Silurian, and Broadwell basins these features have not been found at all. Therefore, we conclude that a hydrologic connection between the Death Valley-Owens River system and the Colorado River has not occurred along either of these routes since the middle Pliocene.

δ18O measurements of benthic ostracodes are used to reconstruct the δ18O history of Lake Huron and Georgian Bay water between 10,600 and 7600 14C yr B.P. This δ18O record was calibrated using a comparison of the δ18O values of modern ostracodes and Lake Huron water, where a fractionation of 1.0358 was measured between the oxygen isotope ratios of the most isotopically positive ostracode Candona subtriangulata and lake water. The most positive shell δ18O value was used because it is precipitated in the cold (0° to 2°C) water common to both deep and shallow environments. The δ18O of Lake Huron water reflects a large glacial meltwater component for much of its history before 7600 14C yr B.P. Times of low lake level correlate with the highest ratio of glacial meltwater to local precipitation in the basin (most negative δ18O values). Georgian Bay water was more negative in δ18O than Lake Huron water of the same age; this reflects a higher proportion of glacial meltwater in Georgian Bay and its separation from Lake Huron during times of low lake level.

The floodplain along a 75-km segment of the Brazos River, traversing the Gulf Coastal Plain of Texas, has a complex late Quaternary history. From 18,000 to 8500 yr B.P., the Brazos River was a competent meandering stream that migrated from one side of the floodplain to the other, creating a thick layer of coarse-grained lateral accretion deposits. After 8500 yr B.P., the hydrologic regime of the Brazos River changed. The river became an underfit meandering stream that repeatedly became confined within narrow and unstable meander belts that would occasionally avulse. Avulsion occurred four times; first at 8100 yr B.P., then at 2500 yr B.P., again around 500 yr B.P., and finally around 300 yr B.P. The depositional regime on the floodplain also changed after 8500 yr B.P., with floodplain construction dominated by vertical accretion. Most vertical accretion occurred from 8100 to 4200 yr B.P. and from 2500 to 1250 yr B.P. Two major and three minor periods of soil formation are documented in the floodplain sequence. The two most developed soils formed from 4200 to 2500 yr B.P. and from around 1250 to 500 yr B.P. These changes on the floodplain appear to be the result not of a single factor, but of the complex interplay among changes in climate, sediment yield, and intrinsic floodplain variables over time.

High-resolution δ18O and δ13C analyses of a stalagmite from Lobatse II Cave reveal late Pleistocene environmental changes in Botswana. Large shifts in δ18O and δ13C are observed between two main periods of deposition. The first period, between 51,000 and 43,000 yr B.P., was warm and humid and may have been associated with some C3 vegetation. The second period, between 27,000 and 21,000 yr B.P., had temperatures approximately 2°C lower, and vegetation dominated by drought-adapted C4 plants. The intervening period, between 43,000 and 27,000 yr B.P., bounded by two major hiatuses in stalagmite growth, produced discontinuous speleothem formation, probably under dry conditions. New 230 Th/234U mass spectrometry age determinations for the stalagmite generally agree with previously measured 230 Th/234U ages.

River mouths on the steep, high-relief coast of the French Riviera exhibit thick sequences of Holocene marine, estuarine, deltaic, and river channel-floodplain sediments that overlie basal fluvial Pleistocene gravel. Gravel is uncommon in most of the early to middle Holocene aggradational-progradational marine, estuarine, deltaic sediments, despite an ample supply from rock units in the steep adjoining uplands. River-mouth gravel is common only in late Holocene river channels and in barrier beaches perched on finer-grained nearshore sediments. Neither downslope grain-size fining on alluvial fans nor sediment stacking patterns during sea-level (base-level) rise readily account for the lack of early to middle Holocene gravel in the river-mouth sediment wedges. Holocene sea-level rise led to the storage of fine-grained sediments in shallow marine, estuarine, and deltaic environments in the present coastal zone. We infer that humid temperate conditions, a dense forest cover, landscape stabilization, and a regular quiescent river flow regime associated with the Atlantic climatic optimum limited gravel supply in the adjoining catchments and gravel entrainment downstream during the early Holocene. Sea-level stabilization in the middle and late Holocene coincided with a marked change in bioclimatic conditions toward the present Mediterranean-type regime, which is characterized by a less dense forest cover, soil erosion, and episodic catastrophic floods. The late Holocene was thus a time of downstream bedload channel aggradation, fine-grained floodplain and paludal sedimentation, and seaward flushing of clasts leading to the formation and consolidation of the gravel barrier beaches that bound the rivermouths and embayments.

Between the western and eastern Andean cordilleras in Peru and Bolivia, there are three main lacustrine basins: Lake Titicaca, Lake Poopó, and the group of Coipasa-Uyuni. For the past few millennia, highly variable environmental conditions have been recorded in their sediments. Today a latitudinal meteorological gradient influences the lakes' status, leading to specific deposits and ostracod communities. Lake Titicaca in the north is oligohaline, whereas Lake Poopó further south is polyhaline. In the south, the Coipasa-Uyuni depression is characterized by a 12,000-km2 surficial salt crust. During the Late Pleistocene (ca. 40,000 to 25,000 yr B.P.), the water depth and salinity in paleolake Poopó fluctuated widely and paleolake Titicaca was slightly larger than at present. Sedimentation was mostly biocarbonate in the shallower areas and it was detrital-organic in the deepest zones. During the Holocene, a dry period transformed Lake Poopó into a "salar" with evaporite precipitation. Lake Titicaca registered a large decline in water level (8100-3600 yr B.P.) initially inducing gypsum precipitation followed by short influxes of water, with an ostracod faunal composition similar to that of the modern brines of Lake Poopó. Lake Titicacas' present condition only appeared between 2200 and 1500 yr B.P.

Arabian Sea sediments record changes in the upwelling system off Arabia, which is driven by the monsoon circulation system over the NW Indian Ocean. In accordance with climate models, and differing from other large upwelling areas of the tropical ocean, a 500,000-yr record of productivity at ODP Site 723 shows consistently stronger upwelling during interglaciations than during glaciations. Sea-surface temperatures (SSTs) reconstructed from the alkenone unsaturation index (UK′37) are high (up to 27°C) during interglaciations and low (22-24°C) during glaciations, indicating a glacial-interglacial temperature change of >3°C in spite of the dampening effect of enhanced or weakened upwelling. The increased productivity is attributed to stronger monsoon winds during interglacial times relative to glacial times, whereas the difference in SSTs must be unrelated to upwelling and to the summer monsoon intensity. The winter (NE) monsoon was more effective in cooling the Arabian Sea during glaciations then it is now.

Cultural shell middens, a common feature of coastal landscapes, preserve a record of past human use of coastal resources and often provide important paleoenvironmental information. Molluscan remains from two shell midden sites, located 0.7 km from the modern Pacific coast at Seaside, Oregon, suggest the former presence of a small sheltered bay at this locality. The modern sand beaches in the Seaside area, stretching for 30 km southward from the mouth of the Columbia River, are exposed to high-energy surf where razor clams (Siliqua patula) are today the only common bivalve. By contrast, over 90% of molluscan remains (by weight) recovered from the archaeological sites represent species that favor sheltered water. Geologic studies in the Seaside area have shown that the beach prograded some 2 km over the last 4 millennia. The midden evidence suggests that the progradation was accompanied by the infilling of a former bay.

Pollen data from two sites provide information on the postglacial vegetation and climate history of the Cascade Range. Indian Prairie in the western Cascade Range was colonized by subalpine forests of Pinus, Picea, and Tsuga and open meadows prior to ca. 12,400 14C yr B.P. The treeline lay 500 to 1000 m below its modern elevation and conditions were cooler than at present. From ca. 12,400 to ca. 9950 14C yr B.P. Abies became important and the forest resembled that presently found at middle elevations in the western Cascade Range. The pollen record implies a rise in treeline and warmer conditions than before. From ca. 10,000 to 4000-4500 14C yr B.P., conditions that were warmer and effectively drier than today led to the establishment of a closed forest composed of Pseudotsuga, Abies, and, at lower elevations, Quercus and Corylus. During this period, Gold Lake Bog in the High Cascades was surrounded by closed forest of Pinus and Abies. The early-Holocene pollen assemblages at both Indian Prairie and Gold Lake Bog lack modern analogues, and it is likely that greater-than-present summer radiation fostered unique climatic conditions and vegetation associations at middle and high elevations. In the late Holocene, beginning ca. 4000-4500 14C yr B.P., cooler and more humid conditions prevailed and the modern vegetation was established. A comparison of these sites with others in the Pacific Northwest suggests that major patterns of vegetational change at individual sites were a response to large-scale changes in the climate system that affected the entire region.

Pollen analysis of sediment cores from Grandfather and Ongivinuk Lakes reveals a record of postglacial vegetation and climate change in the northern Bristol Bay region. The chronology is based on six conventional 14C dates of bulk organic matter from the Grandfather core. A mesic herb tundra dominated the landscape 13,000-9800 yr B.P. Betula shrubs probably first appeared in the region 11,300 yr B.P. but were restricted to favorable microhabitats until 9800 yr B.P. The later establishment of Betula shrubs and relatively low Betula pollen abundance in these records compared to other areas of eastern Beringia suggest that postglacial warming in southwestern Alaska was dampened by regional climatic controls, possibly low sea-surface temperatures of the North Pacific Ocean. Between 10,800 and 9800 yr B.P., diminished Betula shrub cover, along with decreased aquatic productivity as recorded by Pediastrum cell nets and biogenic silica, suggest a brief reversion to colder and drier climatic conditions possibly associated with the Younger Dryas event. Around 9800 yr B.P., Betula shrub tundra and meadow communities expanded, probably in response to increased temperature and precipitation. Alnus arrived and formed extensive thickets within the region ca. 7400 yr B.P. The establishment of the modern boreal forest-tundra ecotone is marked by the arrival of Picea glauca at Grandfather Lake ca. 4000 yr B.P. and the subsequent increase to present population densities ca. 2000 yr B.P. The unique features of these pollen records emphasize the spatial complexity of late Quaternary vegetation and climate history in eastern Beringia.

Sediments from Rapid Lake document glacial and vegetation history in the Temple Lake valley of the Wind River Range, Wyoming over the past 11,000 to 12,000 yr. Radiocarbon age determinations on basal detrital organic matter from Rapid Lake (11,770 ± 710 yr B.P.) and Temple Lake (11,400 ± 630 yr B.P.) bracket the age of the Temple Lake moraine, suggesting that the moraine formed in the late Pleistocene. This terminal Pleistocene readvance may be represented at lower elevations by the expansion of forest into intermontane basins 12,000 to 10,000 yr B.P. Vegetation in the Wind River Range responded to changing environmental conditions at the end of the Pleistocene. Following deglaciation, alpine tundra in the Temple Lake valley was replaced by a Pinus albicaulis parkland by about 11,300 14C yr B.P. Picea and Abies, established by 10,600 14C yr B.P., grew with Pinus albicaulis in a mixed conifer forest at and up to 100 m above Rapid Lake for most of the Holocene. Middle Holocene summer temperatures were about 1.5°C warmer than today. By about 5400 14C yr B.P. Pinus albicaulis and Abies became less prominent at upper treeline because of decreased winter snowpack and higher maximum summer temperatures. The position of the modern treeline was established by 3000 14C yr B.P. when Picea retreated downslope in response to Neoglacial cooling.

Previously published pollen studies from Maine have not identified any late-glacial reversals despite evidence for them from adjacent areas. The distribution and abundance of midge flies are strongly correlated with summer surface-water temperatures. We analyzed fossils of larval midge flies from the sediments of two ponds (Trout and Tilley) from Maine. Each site reveals a major and a minor oscillation during which inferred summer surface-water temperatures fell by 7-13°C and 2.1-2.6°C, respectively. We tentatively correlate these events with the Younger Dryas cooling and Killarney Oscillation reported from adjacent Maritime Canada, where they have been AMS 14C-dated at 10,770-10,000 and 11,160-10,910 yr B.P., respectively. A third oscillation occurs at the northernmost site, Tilley Pond, and may represent the effects of a local ice advance in northern Maine.

Many sites in Fennoscandia contain pre-Late Weichselian beds of organic matter, located mostly in the flanks of eskers. It is a matter of debate whether these fragmentary beds were deposited in situ, or whether they were deposited elsewhere and then picked up and moved by glacial ice. The till-mantled esker of Harrinkangas includes a shallow depression filled with sand and silt containing, for example, several tightly packed laminar sheets of brown moss (Bryales) remains. It is argued that these thin peat sheets were transported at the base of the ice sheet, or englacially, and were deposited together with the silt and sand on the side of a subglacial meltwater tunnel. Subglacial meltout till subsequently covered the flanks of the esker near the receding ice margin. Information about the depositional and climatic environments was obtained from biostratigraphic analysis of the organic matter. Pollen spectra for the peat represent an open birch forest close to the tundra zone. A thin diamicton beneath the peat contains charred pine wood, recording the former presence of pine forests in western Finland. The unhumified, extremely well-preserved peat evidently originated during the final phase of an ice-free period, most probably the end of the Eemian Interglaciation. It was redeposited in the esker by the last ice sheet. Reconstructions of the Pleistocene chronology and stratigraphy of central Fennoscandia that rely on such redeposited organic matter should be viewed with caution.

Season of rainfall can be a major influence on the relative proportions of C3 and C4 grasses in a landscape, since their relative distribution is essentially constrained by temperatures during the growing season; cool growing seasons favor C3 grasses, whereas warm ones favor C4 grasses. 13C/12C ratios in tooth enamel carbonate of grazers from the site of Equus Cave, northern Cape Province, South Africa, were used to determine vegetation and rainfall seasonality shifts during the late Quaternary. The results show shifts in the relative proportion of C3 and C4 grasses during the latter stages of the Pleistocene, and domination by C4 grasses in the Holocene. The data show that the northwestern Cape did not fall within a winter rainfall zone during this period as has been postulated in recent climate models. Instead, a more complex situation is suggested in which at least two periods of enhanced winter rainfall occurred within a predominantly summer rainfall regime.

A 12,500-yr pollen record from Loon Lake, Wyoming provides information on the climate history of the southwestern margin of Yellowstone National Park. The environmental reconstruction was used to evaluate hypotheses that address spatial variations in the Holocene climate of mountainous regions. Loon Lake lies within the summer-dry/winter-wet climate regime. An increase in xerophytic pollen taxa suggests drier-than-present conditions between ca. 9500 and 5500 14C yr B.P. This response is consistent with the hypothesis that increased summer radiation and the expansion of the east Pacific subtropical high-pressure system in the early Holocene intensified summer drought at locations within the summer-dry/winter-wet regime. This climate history contrasts with that of nearby sites in the summer-wet/winter-dry region, which were under the influence of stronger summer monsoonal circulation in the early Holocene. The Loon Lake record implies that the location of contrasting climate regimes did not change in the Yellowstone region during the Holocene. The amplitude of the regimes, however, was determined by the intensity of circulation features and these varied with temporal changes in the seasonal distribution of solar radiation.