The accuracy of Arctic lake chronologies has been assessed by measuring the 14C activities of modern carbon sources and applying these isotopic mass balances to dating fossil lake materials. Small (<1 km2) shallow (<25 m) Arctic lakes with watersheds <12 km2have soil and peat stratigraphic sections with14C activities ranging from 98 to 51% Modern. The14C activity of particulate organic carbon, dissolved organic carbon, and dissolved inorganic carbon from lake and stream waters ranges from 121 to 95% Modern. The sediment–water interface of the studied lakes shows consistent14C ages of ∼100014C yr, although the14C activity of living aquatic vegetation is 115% Modern. Radiocarbon measurements of components of the lacustrine carbon pool imply that the ∼100014C yr age of the sediment–water interface results from deposition of14C-depleted organic matter derived from the watershed.

Physical and chemical coupling of peatland vegetation, soils and landforms, and atmosphere creates feedbacks which may be important in ice-age initiation. A box diffusion CO2 exchange model shows that a transient forcing of 500 Gt C (the amount proposed to have accumulated in peatlands during the last interglacial–glacial transition) over 5000 yr results in a lowering of atmospheric CO2 by about 40 ppm. Proxy data indicate that a decrease in atmospheric CO2 may have occurred over the last 5000 yr up to preindustrial times, and the amount is similar to that calculated from Holocene peatland expansion (∼22 ppm). These results suggest that models should consider the role of peatlands in ice-age initiation.

Analysis of phytoliths and macroscopic charcoal from natural soils near Manaus, Brazil shows that the central Amazonian terra firme forest has not been stable during the Holocene. Phytolith taxonomic composition and radiocarbon ages indicate that a closed forest has continuously existed in the area since at least 4600 yr B.P., but that forest composition and species abundance changed significantly sometime during the Holocene Epoch. There is no evidence that the vegetation experienced swidden cultivation. The distribution and dates of the charcoal in the soils indicate that forest fires between 1795 and 550 yr B.P. burned an area of about 200 hectares. Vegetational changes and fires appear to be a result of climatic drying that may have affected large areas of the Amazon Basin over the past 5000 to 7000 years. Soil phytolith analysis is a promising tool for the study of environmental history in the humid tropics and may reduce the previous dependence on lake and swamp sediments, at least for the Holocene Epoch.

Macrofossil evidence indicates that the mid-Holocene hemlock [Tsuga canadensis L. (Carr.)] decline that occurred over a wide area in eastern North America was associated with phytophagous insect activity. In situ hemlock macrofossils and insect remains found in a paludified dunefield at the northern limit of hemlock testify that two defoliation events occurred at 4910 ± 90 and 4200 ± 100 yr B.P., respectively. The sharp coincidence of remains from hemlock needles with chewing damage typical of hemlock looper feeding, head capsules from the hemlock looper (Lambdina fiscellaria) and the spruce budworm (Choristoneura fumiferana), absence of hemlock fruiting remains, and tree-ring anomalies in fossil hemlocks that died prematurely (<165 yr) suggest that defoliation affected hemlock reproductive capacity and pollen productivity, or more likely caused mass mortality. Our findings indicate that defoliation can affect ecosystems for centuries, especially when long-lived tree species are involved.

Millennium-long tree-ring chronologies have recently been derived fromThuja occidentalis L. growing on the exposed cliffs of the Niagara Escarpment, Canada. Lengthening of these chronologies may be possible by incorporating tree-ring series from subaerially exposed and submerged subfossil wood, provided that its anatomy is not significantly influenced by time or treatment. In order to determine if Holocene-age samples would be suitable for incorporation into very long tree-ring chronologies, a study was undertaken comparing wood structure of living trees, dead trees exposed to the air during the late Holocene, and dead trees submerged since the early Holocene. The results show that the impact of time and environment on wood anatomy is considerably less than expected. The anatomy of annual rings in Holocene trees is similar to that found in modern specimens. High magnification scanning electron micrographs show that tracheid walls lose their rectangular appearance after 8000 yr, but few other signs of deterioration are present. The results mean that Holocene wood can be used to add to the long dendrochronological record that has already been produced from the Niagara Escarpment.

Pikas (Ochtona)—small gnawing mammals, related to rabbits—range today throughout parts of the Northern Hemisphere, but had a wider distribution during the Pleistocene. Nine caves from northeastern North America (a region not occupied by pikas today) have Pleistocene deposits containing remains of Ochotona. We examine 526 fossil specimens (ranging in age from approximately 850,000 to 8670 yr B.P.) from five of these caves. Two morphological forms of Ochotona lived in northeastern North America during the late Pleistocene—a large species (probably O. whartoni) and a small species (probably O. princeps).Ochotona of glacial age are not necessarily indicative of talus slopes and mesic communities. O. princeps-like of the Irvingtonian of West Virginia were living with an amphibian-reptilian assemblage found in the area today, implying winters not much, if at all, colder than at present. Late glacial and postglacial change in climate south of the ice sheets in effect would have isolated Ochotona in eastern North America, where they were unable to retreat to the west or north. Whereas western pika had the option of moving up in elevation, into boreal islands, eastern forms became restricted to ever-diminishing habitats, culminating in extinction and extirpation. Radiocarbon ages imply that Ochotona lived in eastern North America during the late Pleistocene (late Rancholabrean) and into the earliest Holocene. We describe the youngest remains of Ochotona in eastern North America and the youngest for the extinct large form, O. whartoni.

A radiocarbon-dated series of 75 beach ridges, formed at regular intervals averaging 72 yr over the past 5400 yr, provides further support for the existence of a 70-yr oscillation in Northern Hemisphere climate, postulated recently from instrument data representing less than two cycles of this climate oscillation. Results from this study lend support to the interpretation that internal variations in the ocean–atmosphere system are an important factor in climate fluctuations on a decadal–centennial time scale. A temperature oscillation with a period of about 70 yr has been a previously unrecognized but fundamental part of the global climate system since at least the middle Holocene.

A brief cold–warm–cold climate change during the middle Wisconsinan is described for the first time in North America, based on fossil beetle assemblages at Titusville, Pennsylvania. AMS dating of insect chitin and wood suggests the change occurred between 39,000 and 43,500 yr B.P. Basal peats in river terrace deposits contain arctic–subarctic beetle species representing a climate similar to that found at treeline in Canada, where mean July temperatures range from 10° to 13°C. These cold-adapted beetles were replaced by species representing a climate similar to the mixed coniferous–deciduous forests of southern Canada, where mean July temperatures range from 18° to 20°C. In turn, these warmer-adapted beetles were replaced by arctic–subarctic species in the upper part of the section, representing a climate with mean July temperatures in the range 10°–13°C. The brief warm interval is informally designated the Titusville interstade. The difference in temperature between the colder and warmer intervals is about 5°–7°C. The Titusville interstade is correlated with the Upton Warren Interstade of the British Isles which is of similar age and intensity. It is also correlated with the high sea-surface temperature interval between Heinrich events 4 and 5 represented in North Atlantic sediment cores.

Deformation structures interpreted as vertebrate hoof- and foot-tracks occur within upper-Pleistocene eolian sand-sheet deposits in the stabilized Kantishna sand sea of central Alaska and in the Nushagak lowland of southwestern Alaska. Exposures of tracks are generally limited to cross sections, which reveal concave-up deformation structures in which displacement of preexisting strata diminishes downward. Deposits in both areas contain tracks that are 6 to 16 cm in diameter and are divided by a central ridge, reflecting formation by artiodactyl (even-toed) ungulates. Larger (21–34 cm) tracks without a central ridge, observed in the Nushagak lowland, were formed by proboscideans, probably woolly mammoth. Large vertebrate tracks occur within irregularly stratified sand and silty sand that accumulated upon partially vegetated sand sheets, and within thin, even wind-ripple laminae of unvegetated sand sheets. The presence of tracks at multiple stratigraphic levels and preservation of roots and rhizocretions within the eolian deposits suggest that vegetated sand sheets may have formed a locally important grazing habitat for large herbivores during at least part of the last glaciation. Recognition that vertebrate tracks are preserved in eolian sand-sheet deposits, and in deposits of other environments as well, opens a new source of stratigraphic and paleoecological data to aid reconstruction of the vanished ecosystems of Beringia.

Fossil shells of the boring mussel Lithophaga lithophaga provide a means for dating changes in relative sea level. These bivalves, being among the first colonizers of bare calcareous substrates, can mark the earliest stages of marine submergence of caves. Here we report data concerning the deepest submerged speleothem presently sampled in a temperate area, at 48 m below present sea level off the west coast of Italy (Mediterranean Sea). A fossil mussel shell beneath encrusting layers from later marine colonists gave an AMS age of 9580 ± 35 14C yr B.P. (10,253 ± 72 cal yr B.P.).