In situ-produced cosmogenic helium (3Hec) provides a new tool for constraining histories of Quaternary geomorphic surfaces. Before general application of the technique, however, the systematics and production rates of 3Hec must be well understood. In a companion study, 3He and 10Be data from sandstone and granite boulders in the Dry Valleys region of Antarctica have been used to constrain the ages of an important moraine sequence formed by the Taylor Glacier. Data from these deposits also provide information about the systematics of 3He in quartz that has important implications for geochronology based on 3Hec. In contrast to previous results from olivine and clinopyroxene, crushing quartz in vacuo releases helium with high 3He/4He ratios (up to 148 × Ra, where Ra is the atmospheric 3He/4He ratio), indicating that crushing cannot be used to determine the isotopic composition of trapped (i.e., noncosmogenic) helium in quartz. Analysis of 3He in different size fractions of the same samples indicates significant 3 He loss not predicted by existing 3He diffusion data for quartz. The origin of the discrepancy is not clear, but loss from these samples is not as significant as suggested by the limited data of previous studies.

In situ produced cosmogenic nuclides provide a new technique for constraining exposure ages of glacial deposits. In situ 3He and 10Be in quartz sandstone boulders from Arena Valley, southern Victoria Land, Antarctica, provide chronological constraints for a sequence of moraines ("Taylor II-IVb" moraines) related to expansions of Taylor Glacier and the East Antarctic Ice Sheet. Mean 3He ages are 113,000 ± 45,000 yr, 208,000 ± 67,000 yr, 335,000 ± 187,000 yr, and 1.2 ± 0.2 myr, for Taylor II,III,IVa, and IVb moraines, respectively (mean ± 1σ). Corresponding mean 10Be ages for Taylor II and IVb moraines are 117,000 ± 51,000 yr and 2.1 ± 0.1 myr. For the older deposits the 3He ages are probably lower limits due to diffusive loss. Although the exposure ages appear consistent with the few previous age estimates, particularly with an isotope stage 5 age for Taylor II, each moraine exhibits a broad age distribution. The distribution probably results from a variety of factors, which may include prior exposure to cosmic rays, 3He loss, erosion, postdepositional boulder movement, and radiogenic production of 3He. Nonetheless, the exposure ages provide direct chronological constraints for the moraine sequence, and suggest a maximum thickening of Taylor Glacier relative to the present ice surface of ∼500 m since the late Pliocene-early Pleistocene.

Moraines in the canyons of Mono Basin are separable into relative-age groups on the basis of clast-sound velocities in exposed boulders, moraine morphology, and weathering features on boulder surfaces. Tioga, Tenaya, Tahoe, and Mono Basin moraine deposits each have distinct weathering characteristics and therefore constitute different relative-age groups. The Tioga glacial episode at June Lake may postdate ∼25,200(?)-yr-old basaltic lavas, and the Tenaya episode may have occurred ∼30,700 yr ago. A comparison of the glacial and lacustrine records of Mono Basin over the past 40,000 yr, based on new interpretations of radiometric ages, is consistent with the hypothesis that maximum glacial and maximum pluvial periods were not necessarily synchronous.

Geomorphic and sedimentologic evidence in the Grand Valley, which drained the retreating Saginaw Lobe of the Laurentide Ice Sheet and later acted as a spillway between lakes in the Huron and Erie basins and in the Michigan basin, suggests that at least one drainage event from glacial Lake Saginaw to glacial Lake Chicago was a catastrophic outburst that deeply incised the valley. Analysis of shoreline and outlet geomorphology at the Chicago outlet supports J H Bretz's hypothesis of episodic incision and lake-level change. Shoreline features of each lake level converge to separate outlet sills that decrease in elevation from the oldest to youngest lake phases. This evidence, coupled with the presence of boulder lags and other features consistent with outburst origin, suggests that the outlets were deepened by catastrophic outbursts at least twice. The first incision event is correlated with a linked series of floods that progressed from Huron and Erie basin lakes to glacial Lake Saginaw to glacial Lake Chicago and then to the Mississippi. The second downcutting event occurred after the Two Rivers Advance of the Lake Michigan Lobe. Outbursts from the eastern outlets of glacial Lake Agassiz to glacial Lake Algonquin are a possible cause for this period of downcutting at the Chicago outlets.

The Baxie loess section, just east of the Tibetan Plateau, contains evidence showing that the Asian monsoon climate experienced an abrupt reversal near the end of the last glacial age. Rapid deposition of dust under cool, dry full-glacial conditions gave way to an interval of soil development and reduced dust influx attributed to a strengthening of the warm, moist summer monsoon. A subsequent abrupt increase in dust deposition, a response to a weakening of the summer monsoon, was later followed by renewed soil formation as summer monsoon circulation again intensified during the early Holocene. By one interpretation, the thin upper loess is a manifestation of the European Younger Dryas oscillation; however, in this case the available 14C ages require either that (1) onset of loess deposition lagged the beginning of the Younger Dryas event in Europe by as much as 2000 calibrated 14C years or (2) all the 14C ages are too young, possibly due to contamination. Alternatively, the late-glacial paleosol, the top of which is synchronous with the abrupt end of the late-glacial δ18O anomaly in the Dye 3 Greenland ice core, records the Younger Dryas event. Such an interpretation is consistent with general circulation model simulations of Younger Dryas climate that show strong seasonality and a strengthened summer monsoon, and with marine cores from the western Pacific Ocean that contain evidence of pronounced cooling of surface waters during Younger Dryas time.

The Farmdale Soil occurs below late Wisconsinan loess throughout the U.S. Midwest. At the La Sena site in the central Great Plains, humates in the Farmdale Interstadial Soil have a corrected age of 21,000 yr B.P. Humates in a buried Bt horizon and a bulk sample of overlying loess 2.5 m above the Farmdale Interstadial Soil have ages of 17,000 and 19,000 yr B.P., respectively. In the Republican River Valley Picea (spruce) charcoal is common in the lower meter of Peoria loess. Near Bloomington, Nebraska, humates from burned organic matter only 60 cm above the base of Peoria loess have a corrected age of ca. 19,000 yr B.P.

A 34.5 m borehole, which was drilled near the Dead Sea coast (altitude -394 m) in the southern part of the fan delta of Wadi Zeelim, reveals the geological history of that area from the latest Pleistocene to present. The depositional time frame is based on six 14C dates and two U-Th dates. An erosional (or nondepositional) period is implied by the hiatus between 21,100 yr B.P. (U-Th age, depth 33 m) and 11,315 yr B.P. (14C age, depth 32 m). A subsequent arid phase is recorded by a 6.5-m-thick layer of halite; based on 14C dates this phase relates to the abrupt Younger Dryas cold period reported in temperate to polar regions. The fragility of the environment in this region is indicated by the fact that the region experienced such a severe, short aridification phase (less than 1000 yr), evidence of which is found widely in the desert fringes of the Middle East and North Africa. The aragonite found in most of the Holocene section indicates that the well site was covered by the lake for most of the Holocene. Exceptions are the intervals at 0-3 and 10-14 m depths which represent low stands of the lake.

The Holocene geomorphic history of southeastern Louisiana's middle Amite River is recorded in the stratigraphy of three alloformations, identified in decreasing age as the Watson (WAT), Denham Springs (DS), and Magnolia Bridge (MAG). The WAT meander belt formed by at least 9000 yr B.P., when sea level was lower and the Amite River was tributary to a larger ancestral drainage basin. The DS became an active meander belt by at least 3000 yr B.P., in response to relative sea-level rise and eastward progradation of the Mississippi River delta plain. The MAG developed its meander belt, in part, during the European settlement of the drainage basin, and is now attempting to adjust to modern anthropogenic influences. Geomorphic influences on the middle Amite River floodplain have temporal and spatial components that induce regional- and local-scale effects. Regional extrinsic influences caused meander belt avulsion that produced alloformations. However, local influences produced intrinsic geomorphic thresholds that modified channel morphology within a meander belt but did not induce alloformation development. Base-level influences of the relative sea-level rise and the Mississippi River delta plain were so dominant that the effects of possible climate change were not recognized in the Holocene Amite River system.

Soil tongues, or soil fingers, have been identified along the walls of an active gravel pit in northeastern Indiana. Locally enhanced pedogenesis, primarily carbonate leaching and clay illuviation, has resulted in wedge- and cone-shaped extensions of the B horizon into the underlying calcareous glaciofluvial sediments. Although previously described in the literature, the features at this site are considered unusual both for the numerous well-defined tongues and for the regularity of tongue spacing, which gives the illuviation front the appearance of a high-amplitude sinusoidal wave train. Depth to the illuviation front was measured at 0.25-m intervals along a 40-m exposure. Statistical tests indicate that the spatial pattern is nonrandom. The data were transformed using Fourier methods, and the resulting variance spectrum show a peak intensity at wavelengths bracketing 1.5 m. The results of these analyses suggest a surficial or near-surface topographic control. Climatic conditions following local glacier recession favored periglacial activity, and there is some evidence to suggest development of thermal contraction cracks and nonsorted or poorly sorted polygons 1.5 m in diameter. As an alternative hypothesis, laboratory experiments suggest that regularly spaced fingers can develop in homogeneous materials when instability occurs at the wetting front. In this case, soil tongues cannot be used as relict indicators of periglacial activity.

We present direct evidence for the relict nature of lichenspruce woodlands thriving at treeline, using 14C-dated stands growing in the humid climate of eastern Hudson Bay in northern Québec. Black spruce, Picea mariana (Mill.) BSP, forms two groups of relict stands: a Neoglacial group of postfire origin dating between 2000 and 900 yr B.P. and a pre-Neoglacial group most likely established between 4500 and 3200 yr B.P. The latter group shows no evidence of fire (absence of charcoal under topsoil organic horizon). The stands are exceptional because they probably are the direct lineages of the primeval forests that colonized the land sometime after deglaciation. The structure of the relict lichen-spruce communities gives strong support to the current thesis that considers lichen-spruce woodland as a self-perpetuating open forest, typical of the subarctic environment.

Paleoenvironmental changes dating back to 30,000 yr B.P. documented in a pollen record from central Brazil (lat. 19°S) permit the reconstruction of climatic changes related to shifts of the Antarctic polar fronts. The paleoclimatic inferences were obtained by a study of modern vegetation and pollen distribution, taking into account present-day climatic parameters. At 30,000 yr B.P. the climate must have been warmer and moister than today judging from the high amount of tree pollen taxa characteristic of floodplain forest. From 17,000 to 14,000 yr B.P. the climate was drier although tree pollen percentages were relatively high. After 12,000 yr B.P. Araucaria forest elements increased, suggesting a moister and cooler climate. The Araucaria forest disappeared during a short interval between 11,000 and 10,000 yr B.P. This could be related to the Younger Dryas event. At the beginning of the Holocene the climate became cool and moist again, as indicated by the reexpansion of the Araucaria forest. The latter was progressively replaced by a mesophytic semideciduous forest indicating warmer and drier climate after 8500 yr B.P. At 5000 yr B.P. an arid interval was followed by the expansion of mesophytic semideciduous forest elements.

Pollen and plant macrofossil analyses and dating (both TAMS and conventional 14 C) of wetland sediments from the Nodoroc Site in the Piedmont region of northeastern Georgia identified two depositional episodes of interstadial (ca. 26,000-22,000 yr B.P.) and mid- to late Holocene (<3600 yr B.P.) age. Interstadial pollen assemblages were dominated by Pinus and Quercus, with Carya, Picea, and Abies also present in significant amounts. The interstadial assemblages had lower Pinus and higher Quercus pollen percentages than apparently contemporaneous assemblages from sites in the southern Appalachians and Coastal Plain. Interstadial macrofossils included two needle morphotypes of Pinus subgenus Pinus, possibly indicating local occurrence of two Pinus species. One of the needle morphotypes is consistent with Pinus banksiana, but other species cannot be ruled out. Comparison of conventional and TAMS 14C dates illustrates problems in obtaining accurate age models from wetlands and ponds characterized by shallow or fluctuating water levels, low sedimentation rates, and vegetation growing on the depositional surface.

Twenty-two plant species were identified from leaves, fruits, or flowers, and 41 taxa from pollen, present in a macrofossil (leaf) layer in a peat swamp formed on Pleistocene dunes on the Aupouri Peninsula in northern New Zealand. Eight genera of gymnosperms are represented. With the exception of Lagarostrobos colensoi, all tree species abundant as macrofossils are also common as pollen. Macrofossils enabled the on-site flora to be compared with the regional flora, represented by the pollen rain. Studies on leaf decomposition rates indicate bias toward sclerophyllous species in the macrofossils. Identification to species level and treering data from preserved kauri logs allow quantitative comparisons with similar extant communities. Current climatic conditions at those analogue sites are cooler (2° to 3°C), cloudier (11%), and much wetter (85%) than those currently prevailing on the Aupouri Peninsula. Dendrochronological results also suggest that the far north of New Zealand had a cooler, cloudier, and wetter climate at the time the fossil leaf assemblage was formed. Radiocarbon dates from possibly contaminated samples suggest that a diverse mixed gymnosperm/angiosperm forest, dominated by kauri (Agathis australis), was present about (or sometime before) 41,00034,000 yr B.P., when the leaf layer was formed. Similar temperature reductions have been postulated for this period in New Zealand by other authors.

New magnetostratigraphic results may be used to determine the age of two locally important fossil vertebrate faunas near Geelong (Victoria), Australia. Paleomagnetic samples collected from sections located at Limeburner's Point and Limeburner's Bay, together with previously suggested stratigraphic correlations, now constrain the age of the Duck Ponds Local Fauna to be younger than 1.66 myr (and probably greater than 0.98 myr) and the age of the Limeburner's Point Local Fauna to be younger than 0.98 myr. These age determinations enhance the biostratigraphic importance of both local faunas on a continent characterized by a chronic lack of chronologically well-constrained fossil vertebrate localities.

A new extended radiocarbon calibration curve has allowed a reexamination of relative sea-level data based on pre-Holocene dates. At sites located far from any Late Pleistocene ice sheet, the effect of employing this new calibration curve is that the calibrated data (with ages up to 17,100 14C yr B.P.) now agree with the corresponding relative sea-level curve predicted by the ICE-3G deglaciation model. An important implication of this new calibration curve is that the last glacial maximum (ca. 18,000 14C yr B.P.) is inferred to have occurred 22,000-21,000 cal yr B.P. This allows for additional ice to be incorporated in a revised deglaciation model than suggested by ICE-3G. The predicted relative sealevel curves of this new model match the relative sea-level data as well as those of ICE-3G. Further, the total sea-level rise of 124 m at Barbados since the last glacial maximum predicted by this new model agrees with the estimated value of 121 ± 5 m obtained from the depths of drowned reef-crest corals.