The Imbrie-Kipp method of paleotemperature estimation is rigorously tested by comparing Atlantic temperature equations independently derived from the microfossils of three biotic groups: the Foraminifera, Coccolithophorida, and Radiolaria. This method consists of two steps: factor analysis of the modern sea-bed data of the individual groups which resolves discrete biogeographic assemblages and regression analysis of the modern assemblage data with observed sea-surface temperature data to obtain paleotemperature equations. Assemblage biogeography shows a simple subdivision into warm (low latitude) and cold (high latitude) for all biotic groups. Between biotic groups there is greater similarity among high-latitude assemblages than low-latitude ones. Correlating the assemblage data with observed sea-surface temperatures to produce temperature distribution patterns shows differences of less than 2°C in their optimum and critical temperatures. Regression analysis produced accurate temperature equations for each biotic group, all with standard errors of estimate of less than or equal to 2°C. Multiple correlation coefficients were all greater than 0.970. Applying these equations to two multiple biotic data sets (the modern and ice-age sea-bed data) and comparing their temperature estimates using the standard error pooled, shows over 87% concordancy for both data sets. Unlike the modern data, the discordancy among temperature estimates of the ice-age data shows a distinct geographic distribution; its cause is believed to be oceanographic, a difference in the water-mass structure between the modern and ice-age ocean.

The presence of the Brunhes-Matuyama magnetic reversal in deep-sea core sediments makes possible an alternative to the usual K/Ar radioistope method of dating the reversal as found in rocks. The alternative method uses correlations of Northern Hemisphere summer insolation with oxygen-isotope ratios from tropical cores. The latitude-dependent insolation variations are calculated from planetary mechanics and thus provide a highly accurate astronomical time scale. The insolation variations strongly influence glacial-ice volume fluctuations that dominate the oxygenisotope ratio changes recorded in core sediments. The summer half-year insolation variations are identified with corresponding isotope-ratio changes in cores from the present through glacial Stage 20. Misleading effects of discontinuities or major nonuniformities of sediment deposition are avoided by an analysis of the uniformity of V28-238 and V28-239, the principal cores studied, and by comparisons with other cores. The top section (Stages 1 to 10) of V28-238 is uniformly deposited, and for this section an isotope-ratio time scale is chosen that agrees with the thorium-uranium date for the high sea stand of the last interglacial extreme. Over this interval, major glacial extremes (strong isotope-ratio minima) coincide consistently with major insolation minima at times of low orbital eccentricity. In addition interstadials are directly associated with precessional insolation peaks, and the envelope of isotope-ratio peaks resembles the envelope of precessionally dominated insolation peaks. The assumption that the glacial extremes depended similarly on insolation minima during Stages 10 to 20 permits minor age shifts of strong isotope-ratio minima in the two cores (relative to ages based on uniform overal deposition) to match the ages of low-eccentricity insolation minima. The age shifts reflect residual nonuniformities of deposition. The validity of this matching procedure is supported by a resulting consistent identification of principal isotope-ratio peaks with high- and low-latitude coincident insolation maxima. The Brunhes-Matuyama reversal is found intinterglacial Stage 19, and is dated on the astronomical time scale at 790,000 ± 5000 yr B.P.

The analysis of microparticles in a 101-m core from Amundsen-Scott South Pole Station, Antarctica has revealed a substantial increase in total particle concentration between approximately 1450 and 1850 A.D., a period encompassing the latest neoglacial interval or Little Ice Age. It is likely that this reflects a simultaneous increase in the concentration of particulate material in the Antarctic atmosphere. This is important climatologically, for the Antarctic atmosphere may represent the closest approximation to the natural background aerosol. Thus cores from East Antarctica may contain long and detailed records of the natural global background aerosol. Such records are unavailable from any other medium. Additionally, a cyclical variation which appears to be annual has been detected in the South Pole particle record. These features allow construction of a relative time scale for ice cores older than 100 yr from regions of low accumulation (<10 g a−1) where many traditional techniques are not applicable. This is especially significant, as the comparison of climatic data extracted from ice cores with other records of proxy data depends upon the ability to assign an accurate time scale to the ice core. An estimated nine-century record of net annual accumulation at the South Pole has been compiled and the calculated error in the time scale is ±90 yr.

The paleoceanographic–climatic record represented by deep-sea microfossils reflects conditions for only certain times of the year. Also, the relative abundances of microfossil species in deep-sea sediments do not usually reflect only one paleoceanographic variable, such as temperature. Rather, species distributions represent the integration of many factors that control biological production in the oceans. This influences the information on past climates that can be extracted from fossil material. The seasonal limitation is due to the cyclic nature of biological production in the open ocean. Case studies of the sediment record in the Atlantic for two species of planktonic Foraminifera, left-coiling Neogloboquadrina pachyderma (Ehrenberg) and Globigerinoides ruber (d'Orbigny), illustrate seasonal bias in environmental data reported by the relative abundances of species in deep-sea sediments. In addition, the study of G. ruber illustrates the operation of two oceanographic parameters in controlling a species distribution. These examples demonstrate that the environmental signal in the sediments is the result of the interplay of the ecological tolerance of the plankton species with seasonally variable biological and physical properties of the upper ocean.

An analysis of recent (1898–1972) seasonal rainfall data from California (winter) and Mexico (summer) provides no evidence of an inverse relationship such as has been reported by Winstanley for climatically analogous areas of Eurasia and Africa. However, California and Old World winter rain trends have been remarkably similar during this time period. Also, summer rain in northwestern Mexico (Mazatlan) correlates highly with winter rainfall in southern California (San Diego) when allowance is made for a 7-year lag. Winstanley suggested that the expansion and contraction of the circumpolar vortex is a primary control of rainfall trends on the margins of the subtropical deserts. The alternative proposed here is that, at least on this time scale, variations in the longitudinal position of troughs and ridges in the upper air westerlies are more important.

L. V. Worthington (1968, Meteorological Monographs 8, 63–67) hypothesized that a low-salinity lid covered the entire world ocean. By deconvolving isotopic curves from the western equatorial Pacific and equatorial Atlantic, W. H. Berger, R. F. Johnson, and J. S. Killingley (1977), Nature (London) 269, 661–663) and W. H. Berger (1978, Deep-Sea Research 25, 473–480) reconstructed “meltwater spikes” similar to those actually observed in the Gulf of Mexico and thus apparently confirmed the Worthington hypothesis. It is shown that this conclusion is unwarranted. The primary flaw in the reconstructed meltwater spikes is that the mixing intensity used in the deconvolution operation is overestimated. As a result, structure recorded in the mixed isotopic record becomes exaggerated in the attempt to restore the original unmixed record. This structure can be attributed to variable ice-volume decay during deglaciation, effects of differential solution on planktonic foraminifera, temporal changes in abundance of the foraminifera carrying the isotopic signal, and analytical error. An alternative geographic view to the global low-salinity lid is offered: a map showing portions of the ocean potentially affected by increased deglacial meltwater at middle and high latitudes and by increased precipitation-induced runoff at low and middle latitudes.

In the modern South Atlantic the transition between deep water and bottom water is marked by a clear change in the associated benthonic foraminiferal fauna. Uvigerina and Globocassidulina characterize oxygen-poor Circumpolar Deep Water which has long been isolated from the surface. Planulina and miliolids are found associated with the more newly formed, oxygen-rich North Atlantic Deep Water. Antarctic Bottom Water is characterized by “Epistominella” umbonifera. Analysis of the benthonic foraminiferal faunas in two sediment cores recovered from the Vema and Hunter Channels in the western South Atlantic shows that the level of the transition between deep and bottom waters shallowed sharply about 700,000 yr ago. This rise indicates a sharp, sustained increase in the volume of bottom water flowing through the South Atlantic after this time. Prior to about 700,000 yr ago, the amount of Antarctic Bottom Water entering the western South Atlantic was greatly reduced and Circumpolar Deep Water apparently accounted for the bulk of northward flow. Production of southward-flowing North Atlantic Deep Water seems not to have been affected. The timing of this change in circulation regime suggests a possible causal link to similar changes in records of terrestrial and sea-surface climate.

Benthonic foraminiferal faunal and isotopic data from Champlain Sea sediments (approximately 12,500 to 10,000 yr B.P. in age) in two piston cores from Lake Champlain provide a detailed, apparently continuous record of the evolution of the Champlain Sea. Cassidulina reniforme and Islandiella helenae are the dominant benthonic foraminifera during the initial phase of the Champlain Sea, and are replaced by Elphidium excavatum forma clavatum and Protelphidium orbiculare as the dominant species during the remainder of the sea's history. The oxygen-isotopic data show a gradual decrease in δ18O between approximately 12,500 and 10,900 yr B.P., followed by a more rapid decrease during the interval 10,900 to 10,000 yr B.P. The δ13C data have a similar trend as δ18O, with generally decreasing values up the section. The isotopic and faunal data suggest that nearly marine conditions were present in the initial plase of the Champlain Sea, followed by gradually decreasing salinities and increasing temperatures as the sea evolved. The beginning of the rapid isotopic decrease at approximately 10,900 yr B.P. marks the onset of the largest environmental change in the history of the Champlain Sea, probably reflecting a major pulse of meltwater from the Laurentide Ice Sheet.

Approximately 8000 km of continuous seismic reflection profiles throughout Lake Superior were examined for evidence of recessional moraines and other ice-margin deposits associated with the retreat of late Wisconsin ice. These features are correlated with the record of glacial-lake evolution in western Lake Superior. An offlapping sequence of glacial and glacial-lacustrine sediments overlying bedrock is recognized in west-central Lake Superior that is progressively younger to the northeast. The sequence underlies more recent glaical-lacustrine and postglacial sediments. Four facies are recognized on the basis of geomorphologic and acoustic properties and are interpreted to represent a southwest-to-northeast assemblage of: proglacial stratified drift (facies A), drift in major end moraines (facies B), till deposited as glacial retreat resumed, or possibly late-stage ablation till (facies C), and basal till (facies D). The prominent moraines of facies B are unusually thick and are believed to mark the ice-margin shorelines of successive major proglacial lakes that formerly occupied parts of western Lake Superior. The moraines are tentatively correlated with Glacial Lake Duluth (unit 1), Glacial Lake Washburn (unit 2), and Glacial Lake Beaver Bay (unit 3), the most prominent of lakes drained via the progressively lower outlets via the Moose Lake/ Brule-St. Croix Rivers, the Huron Mountains, and the Au Train-Whitefish regions, respectively.

The upper Enchantment Lakes basin in the North Cascade Range of Washington displays two moraine belts, each recording an episode of glacier advance after the end of the last glaciation. The inner belt, the Brynhild, 0.1 to 0.5 km beyond existing glaciers, postdates Mount St. Helens Wn tephra (∼450 yr old), which lies only beyond the moraines. The morainal surface is only slightly weathered, is almost barren of lichens, and is devoid of soil, evidence suggesting that the Brynhild moraines are no more than a century old. The outer moraine, the Brisingamen, 0.3 to 0.7 km beyond existing glaciers, is weathered and is covered with large lichens. On and behind the Brisingamen moraine the Mazama ash (6900 yr old) is present beneath the Mount St. Helens Yn and Wn tephras. Despite more than 7 millennia of weathering, the rock surface behind the Brisingamen moraine is measurably less weathered than the surface beyond, which was last glaciated during the Rat Creek advance about 13,000 yr ago. The age of the Brisingamen moraine therefore is probably early Holocene. The Brisingamen moraine evidently correlates with moraines near Glacier Peak, near Mount Rainier, in northeastern and central Oregon, in the southern Canadian Rockies, and in the northern U.S. Rocky Mountains. These regional effects suggest that a climatic episode of cooling or increased snowfall affected the entire region some time during the early Holocene.

Plant and animal remains found in packrat (Neotoma spp.) middens and cave fill from the eastern and southern Great Basin region reveal the presence of subalpine conifers and boreal mammals at relatively low elevations during the Late Wisconsin. Limber pine (Pinus flexilis) and bristlecone pine (P. longaeva) were important in the late Pleistocene plant communities throughout this region. Spruce (Picea cf. engelmannii) and common juniper (Juniperus communis) were present in some of the more northerly localities, and Douglas fir (Pseudotsuga menziesii) and white fir (Abies concolor) were present in southern and eastern localities. Single needle pinyon pine (Pinus monophylla), common across this region today, was apparently not present north of the Sheep Range of southern Nevada during the Late Wisconsin. Pikas (Ochotona cf. princeps), small boreal mammals present in only a few Great Basin mountain ranges today, were common throughout the region. Heather voles (Phenacomys cf. intermedius) have been found in two cave fill deposits in Nevada, though they are unknown in the Great Basin today. Limber and bristlecone pines are generally restricted to rocky substrates in modern subalpine habitats in the Great Basin, and this may also have been the case when these plants grew at lower elevations during the Late Wisconsin. Subalpine conifers were present on the rock outcrops sampled by the packrat middens, but shrub communities, perhaps dominated by sagebrush (Artemisia spp.), may have been present on alluvial valley-bottom substrates. Forested habitats would thus have been isolated habitat islands, as they are today. Boreal small mammals, including pikas and heather voles, were able to colonize the Great Basin mountain ranges during the late Pleistocene. We suggest that these mammals were able to survive in the intervening valley-bottoms under a cool-summer climatic regime, and that continuous forest or woodland corridors were not necessary for migration.

Pollen spectra from cores of organic spring deposits from the Transvaal provide evidence for the climatic evolution of the province during the last 35,000 yr B.P. or more. The past climatic phases are derived from palynological reconstructions of past vegetation types by comparison of fossil pollen data with modern surface pollen spectra from various localities. Evidence is provided for an early moist, cool phase with relatively mesic bushveld and expanded montane forest in the central Transvaal, followed by a drier period with drier bushveld which probably lasted until approximately 25,000 yr B.P. During the next phase, which at the latest ended about 11,000 yr B.P., the temperatures were probably 5°–6°C cooler than at present. At that time bushveld vegetation in the central Transvaal was replaced by open grassland with macchia elements. Climatic amelioration came and semiarid savanna returned to the plains, at first gradually and then developing into a warm Kalahari thornveld-type vegetation. After 6000 yr B.P. it apparently became slightly wetter and a more broad-leafed bushveld developed. About 4000 yr B.P. it again became cooler and slightly wetter and the bushveld vegetation on the central and northern plains was comparable to present open upland types. After 2000 yr B.P. conditions gradually became warner until about 1000 yr B.P., when the modern climate of the central Transvaal bushveld originated.

Voluminous and widespread tephras were produced frequently during the last 36,000 yr of volcanic activity at Mount St. Helens. Numerous tephra sets have been defined by D. R. Mullineaux, J. H. Hyde, and M. Rubin (1975, U.S. Geological Survey Journal of Research, 3, 329–335) on the basis of field relations, Fe–Mg phenocryst assemblage, and 14C chronology and are valuable marker beds for regional stratigraphic studies. In this study modal abundances and mineral compositions were determined (via petrographic and electron microprobe techniques) for numerous samples of individual layers within tephra sets W and Y to evaluate the degree of compositional variability within and between tephra layers and criteria by which to distinguish among Mount St. Helens and other Pacific Northwest tephras. Although individual layers within a set (e.g., We, Wn) cannot be distinguished from each other on the basis of mineralogic characteristics examined, mineral compositions allow distinction among layers W and Y and other Pacific Northwest tephras (e.g., Mazama, Glacier Peak). Fe–Ti oxide compositions and T–fO2 estimates derived using coexisting magnetite—ilmenite are especially useful due to the compositional homogeneity of these minerals both within and between samples of a given unit over a wide geographic area. The silicates show more compositional variability than the oxides, but SiO2/Al2O3 contents in hornblende and Fe/Mg ratios in hypersthene aid in distinguishing among Pacific Northwest tephras.

Ennadai Lake, in the forest-tundra ecotonal region of Keewatin, Northwest Territories, Canada, has been the subject of several paleoecological investigations (palynology, plant macrofossils, fossil soils). This study concerns Holocene insect fossils at Ennadai, a new approach in a region shown to be sensitive to climatic change. The Ennadai I site yielded 53 taxa, representing 13 families of Coleoptera and 7 families of other insects and arachnids, including abundant ants. These fossils range in age from about 6300 to 630 yr B.P. The Ennadai II site produced fossils of 58 taxa, including 13 beetle families and 15 families of other arthropods, ranging in age from 4700 to 870 yr B.P. The insect evidence suggests the presence of trees in the Ennadai region from 6000 to 2200 yr B.P. A conifer pollen decline from 4800 to 4500 yr B.P. at Ennadai has previously been interpreted as an opening up or retreat of forest in response to climatic cooling, but the insect fossils reveal the continued presence of trees during this interval. Both insect assemblages suggest trends of forest retreat and tundra expansion between about 2200 and 1500 yr B.P., presumably due to climatic cooling, with a return of woodland by about 1000 yr B.P.

A 1520-cm sediment core from Lake Patzcuaro, Michoacan, Mexico, is 44,000 yr old at the base. All parts of the core have abundant pollen of Pinus (pine), Alnus (alder), and Quercus (oak) with frequent Abies (fir). The interval dated from 44,000 to 11,000 yr ago has a homogeneous flora characterized by abundant Juniperus (juniper) pollen and frequent Artemisia (sagebrush). It is believed to represent an appreciably drier and colder climate than at present. The Holocene at Lake Patzcuaro is characterized by a moderate increase in Pinus pollen and the loss of Juniperus pollen, as the modern type of climate succeeded. Alnus was abundant until about 5000 yr ago; its abrupt decrease with the first appearance of herbaceous weed pollen may reflect the cutting of lake-shore and stream-course alder communities for agricultural purposes, or it may simply reflect a drying tendency in the climate. Pollen of Zea (corn) appears at Lake Patzcuaro along with low peaks of chenopod and grass pollen at 3500 yr B.P. apparently recording a human population large enough to modify the natural environment, as well as the beginning of agriculture. A rich aquatic flora in this phase suggests eutrophication of the lake by slope erosion. In the most recent period corn is absent from the sediments, perhaps reflecting a change in agricultural practices. The environment changes at Lake Patzcuaro are similar to and correlate with those in the Cuenca de Mexico, where diatom stratigraphy from the Chalco basin indicates fluctuations in lake levels and lake chemistry in response to variations in available moisture. Before 10,000 yr ago climates there were cool and dry, and the Chalco basin was occupied by a shallow freshwater marsh that drained north to Lake Texcoco, where saline water accumulated by evaporation. Increases in effective moisture and possible melting of glaciers during the Holocene caused lake levels to rise throughout the Cuenca de Mexico, and Lake Texcoco flooded the Chalco basin with brackish water. After 5000 yr ago such flooding decreased, and shallow freshwater ponds and marshes were restored in the Chalco basin. This environmental change coincides with the appearance of Zea pollen and suggests cultural control of lake levels and salinity.

Fossil pollen identified in the earliest sediments of three cushion bogs in the Cordillera Pelada (40°10′S, 73°30′W) dated 10,425 14C yr B.P. includes the subantarctic species Dacrydium fonckii, Tetroncium magellanicum, Astelia pumila, Gaimardia australis, Donatia fascicularis, and Drosera uniflora. All grow today in the Cordillera Pelada and range poleward to the southernmost Province of Magallanes; one species, Drapetes muscosa, included with the pollen of these plants in the earliest record, is no longer a constituent of the flora but is limited only to subantarctic Chile. Available evidence indicates that plants survived the last glaciation north of the glacial border with the course of postglacial migration southward following the wastage of the glacier complex. Holocene climatic and vegetational changes in the Cordillera Pelada are interpreted in the context of regional reconstructions which show maximum warmth about 9000 yr ago with a pronounced dry period lasting from 9000 to 6500 yr B.P. Maximum precipitation was later reached around 4000 yr ago but has decreased overall since then. The regional decline of the endemic gymnosperm Fitzroya cupressoides, which today is extensively destroyed in the Cordillera Pelada, follows this decrease in precipitation. These climatic data suggest a net south ward shift in the zone of westerly winds that bring rainfall to the region over the past 4000 yr.

Analyses of pollen and land snails from rocksheter sites in the Osage Hills of northeastern Oklahoma indicate that the period 2000-1000 yr B.P. was moister than today. During that time, colonies of the prairie vole Microtus ochrogaster were present in the Texas Panhandle. About 1000 yr B.P. the climate changed to dry conditions that have persisted to the present. Disjunct colonies of small mammals in Texas became extinct at the beginning of the dry episode, thereby establishing the composition of the modern fauna. The climatic model for the origin of the Panhandle Aspect (A.D. 1200–1500) is questioned on the grounds that the Southern Plains experienced a long period of dry climate commencing A.D. 950.

The bedrock valley of the Pecatonica River north of Freeport, Illinois, contains a thick valley-fill complex of alluvium and drift. Within the valley, loess-capped benches surround hills of silty Illinoian drift. Beneath these benches lie thick deposits of poorly sorted stony silt interbedded with thin lenses of silt, sand, and organic-rich loam. Channel deposits and peat cap the diamicton in places. We interpret the stony silts as solifluction debris shed from silty slopes within the valley-fill during the Early or Middle Wisconsinan (Altonian). Top and bottom radiocarbon dates from a 2.5-m section of peat overlying the diamicton are 26,820 ± 200 and 40,500 ± 1700 yr B.P., respectively. We informally refer to the stony silts, channel sediments, and peat as the “Martintown unit.” Geomorphic position, sediment input, and macrofossils suggest that the dated peat was deposited in a floodplain pond (oxbow?). The pollen record from the peat indicates that a boreal forest dominated this area during the Middle Wisconsinan (late Altonian and Farmdalian). Two pollen zones are recognized: a basal Zone I with Pinus slightly more abundant than Picea and with few herbs and shrubs, and an upper Zone II dominated by Picea and with a larger representation of herbaceous and shrub taxa. Little displacement of vegetation zones is indicated, even though ice advanced to within 100 km of the site during the time of peat accumulation. Because of the problems involved in clearly defining Middle Wisconsinan forest-tundra in mid-latitudes by using analogs of Holocene forest-tundra in high latitudes, caution is required in making geomorphic inferences solely from vegetation data. Together, though, pollen and sediment data indicate that during the Middle Wisconsinan, Pecatonica hillslopes progressed through a sequence of instability-stability-instability related to climatic fluctuations.

The Christensen Mastodon Site, located in central Indiana, contains a rich assemblage of vertebrates (including mastodon, caribou, and giant beaver), invertebrates, and plant macrofossils in situ in lake and bog sediments of late-glacial age. Studies of pollen and plant macrofossils suggest the existence of open, white spruce-dominated boreal forests from > 14,000 yr B.P. to ca. 13,000 yr B.P. The regional decline of spruce, local occurrence of black spruce, white spruce, and larch, immigration of many hardwood taxa (e.g., ash, oak, elm), and the initiation of bog development are recorded beginning about 13,000 yr B.P. Recent reconstructions of late-glacial and early postglacial vegetational changes provide a context for understanding the disappearance of mastodons. The dramatic and rapid restriction of boreal forests along the retreating ice margin from 11,000 to 9000 yr B.P. may have caused a substantial reduction of mastodon populations. A diminished population would be more susceptible to small-scale, stochastic events such as short-term extremes of weather, outbreaks of disease, or predation pressure from paleoindian hunters.