It is proposed that the two preferred modes of temperature and circulation of the atmosphere which occurred over the past 100,000 yr correspond to two modes of partitioning of the poleward energy flux between the atmosphere and ocean. At present the ocean carries an appreciable fraction of the transport, for example about three-eighths at 30°N. In the cold mode it is suggested that the ocean carries less, and the atmosphere more, than at present. During the formation of the ice, at 50,000 BP, for example, the overall flux is expected to be slightly lower than at present and during melting, at 16,000 BP, slightly higher. The transition between the modes is seen as a natural imbalance in the atmosphere-ocean energy budget with a gradual warming of the ocean during an Ice Age eventually cluminating in its termination. At the present the imbalance is thought to correspond to a natural cooling of the ocean, which will lead to the next Ice Age.

The magnitude of temperature changes in the polar regions differ between the hemispheres in the same way as present seasonal changes, being larger in the northern than in the southern hemisphere.

Overall the atmospheric energy cycle was more intense during the Ice Ages than now.

Observational tests are proposed by which predictions from the present arguments may be compared with deductions about the environment of the past.

Data used for the present state of the atmospheric general circulation are the latest global data available and contain no known major uncertainties. However, data for the oceanic circulation and energy budget are less well known for the present and almost unknown for the past. Hence the proposed imbalances must be treated as part of a speculative hypothesis, but one which eventually may be subject to observational test as no solar variability is invoked.

Departures of mean annual precipitation and temperature for the decade 1961–1970 from the 1931–1960 averages are illustrated on maps of the tropical and subtropical portions of the American continents. Certain features of midlatitude climatic anomalies appear to be associated with concurrent anomalies in the tropics. There is an apparent southward shift of circulation features in this longitudinal sector of both hemispheres. A review of selected literature on latitudinal climatic shifts and atmosphere-ocean interaction suggests some similarities between the patterns of climate in the 1960s and the climate of the Little Ice Age.

This paper evaluates the geological and paleoecological implications of a Wenner-Gren symposium Stratigraphy and Patterns of Cultural Change in the Middle Pleistocene. The deep-sea, glacial-eustatic, loess, alluvial, and palynological records suggest between six and eight cold-warm cycles since the Brunhes-Matuyama magnetic reversal of 700,000 BP. Till and outwash stratigraphies are inadequate to provide a valid nomenclature for the numerous glacials preceding the Würm. Since at least five of the glacials since 700,000 BP were sufficiently severe to produce permafrost in midlatitude Europe, the “glacial Pleistocene” begins with the Brunhes-Matuyama. Although earlier cold-warm cycles extend well back into the early Pleistocene, with extensive glaciation and repeated floral decimations in higher latitudes, the first record of permafrost in the Rhine Basin 700,000 BP argues that major climatic oscillations of the Brunhes epoch were of glacial-interglacial amplitude. It is therefore recommended that a Lower-Middle Pleistocene boundary be linked to the practicable and universally applicable chronometric horizon provided by the Brunhes-Matuyama reversal, while the Middle-Upper Pleistocene limit can continue to be drawn at the base of the last, Eemian Interglacial (130,000 BP). The tropical African record presently contributes little to general understanding of the Middle Pleistocene, while the climatic cycles of higher latitudes are of limited value in analyzing mid-Pleistocene records of the tropical continents. Problems of stratigraphic control and environmental contexts for archeological sites are discussed.

Since chronometric data comparable to that available from the late Cenozoic succession of East Africa have not, and probably cannot be obtained in southern Africa, faunal dating methods retain their traditional significance in the latter area. Five successive late Cenozoic mammal ages, the Namibian, Langebaanian, Makapanian, Cornelian, and Florisian have been proposed as a framework for discussions relating to the chronology of southern African mammalian faunas. The Namibian fauna is poorly known, but is evidently of Miocene age. It is not discussed in this paper. The Langebaanian fauna is well known only from the prolific occurrence at the type site of Langebaanweg and is Pliocene in age. Makapanian faunas are best represented at the Transvaal Australopithecine sites and probably overlap the Plio-Pleistocene boundary. The Cornelian fauna is not as well known as others, the largest assemblage having been recorded from Elandsfontein, although this assemblage is one which is unfortunately temporally heterogeneous. Florisian faunas have been recovered from a relatively large number of localities, including several for which there are radiometric dates. The recorded southern African fossil Carnivora are listed and the local evolution of this group is discussed in terms of the mammal age chronology. It is concluded that secure faunal dating of individual fossil occurrences is enhanced by an appreciation of the nature of changes undergone in evolutionary lineages, while an uncritical knowledge of recoreded taxa is less useful.

Canonical correlation analysis, as described by Webb and Bryson, Quaternary Research 1972, provides a means of reconstructing past climates quantitatively from fossil pollen using a pollen-climate transfer function. This paper presents a method for analysis of variance of the transfer function model. This method is used to identify ecological relationships among the pollen and climate variables, to select climatically sensitive taxa, and to investigate the importance of site factors. Several criteria are presented, in addition to those used by Webb and Bryson, for choosing canonical variate pairs to include in the transfer function model, namely: the variate pair relationships should be ecologically meaningful; the transfer function model should yield stable paleoclimatic estimates; and, the variate pair relationships should be statistically meaningful. The application of these criteria to the set of variate pairs used in the transfer function model of Webb and Bryson is described and modifications are suggested.

A bog pond 4 km east of Yellowstone Lake has a pollen record starting with an Artemisia spruce assemblage, implying alpine vegetation. A layer of volcanic ash, dated as 14,360 ± 400 BP (probably Glacier Peak or Mt. St. Helens J), occurs within the zone, which terminates at 11,630 ± 180 BP. The rest of the pollen sequence is dominated by lodgepole pine, with reappearance of spruce pollen in modest quantities about 4500 BP, according to dating provided by a layer of Mt. Mazama ash (6600 BP). The present vegetation of the area is marked by forests of lodgepole pine with some stands of spruce and fir. The pollen sequence suggests that the upper treeline before 11,600 y. a. was perhaps 500 m lower than today. The climate then became warmer and/or drier than today (Altithermal interval). About 4500 y. a., a slight climatic reversal took place, roughly contemporaneous with the regrowth of glaciers in the western mountains (neoglaciation).

Emerged coral reef terraces on the Huon Peninsula in New Guinea were reported in a reconnaissance dating study by Veeh and Chappell 1970. Age definition achieved was not good for several important terraces, and we report here a series of new 230Th/234U dates, which further clarify the history of late Quaternary eustatic sea level fluctuations. More than 20 reef complexes are present, ranging well beyond 250,000 yr old: we are concerned with the seven lowest complexes. Major reef-building episodes dated by 30Th/234U are reef complex I at 5–9 ka (kilo anno = 1000 yr), r.c. IIIb at 41 ka (four dates), r.c. IV at 61 ka (four dates), r.c. V at 85 ka (two dates), r.c. VI at 107 ka (two dates), and r.c. VII at 118–142 ka. Complex II was previously dated by 14C at 29 ka: this age has not yet been confirmed, and may be only a lower limit. The reef crests were built during or immediately before intervals of sea level maxima, when rates of rising sea level and tectonic uplift briefly coincided. The culmination of each reef-building episode was only a few thousand years in duration, and multiple dates from the same reef complex generally group within the statistical errors of the individual dates.

Several methods can be used to estimate the altitude of each sea level maximum relative to present sea level. The least complicated is to calculate mean tectonic uplift rate for each profile of the terraces, and use the mean rate to calculate the tectonic displacement of each dated reef complex on that profile. The difference between the present altitude of a reef complex and its calculated tectonic uplift gives the paleosea level at the time the reef grew. We estimate uplift rates for six surveyed sections by calibrating against published paleosea level estimates from Barbados and elsewhere, viz 125 ka, paleosea at +6 m; 103 ka, −15 m; 82 ka, −13 m. For each section the individual uplift rates for reefs V, VI, and VIIb are within 5% of their section means. Using the mean rates. paleosea level estimates for reef crests II, IIIB, and IV are made for each section. Consistency of estimates between sections is good, giving −28 m for the 60 ka paleosea level, around −38 m for the 42 ka level and −41 m for the 28 ka level (if the age is older the paleosea level would be lower. Using the mean uplift rates, the 82 ka and 103 ka paleosea levels are also estimated for each section: all individual estimates are plotted graphically, and a sea level curve drawn. The reef stratigraphy indicates sea level lowerings between each dated reef crest: the crests probably represent the interstadials of the Wisconsin (Würm, Weichsel) Glaciation, and intervening lower levels correspond to stadials. Since the last time of eustatic sea level higher than the present (about 125 ka), five sea level maxima occurred at roughly 20-ka intervals, none being as high as the present.

Late Pleistocene radiolarian assemblages were examined in two cores (V24-58 and RC11-209) from the eastern equatorial Pacific Ocean, using Nigrini's 1970 method of recurrent group analysis. The radiolarian “climatic index” curves obtained by this analysis were in close agreement with fluctuations in carbonate levels established previously for the same cores. This correspondence suggests that equatorial radiolarian assemblages changed significantly in response to Pleistocene climatic oscillations. The radiolarian “climatic indices” of the two cores showed close agreement with each other. Our data indicate that equatorial radiolarian assemblages may prove useful for regional correlations between cores, and for paleoclimatic interpretations.

Examination of loess columns in coastal South Canterbury, New Zealand, and the recognition of paleosols with comparable morphology to surface soils allowed the division of the loess column into six members. The upper loess members 1–4 are grouped into a proposed Dashing Rocks Formation. This commonly overlies an erosion surface on Timaru Basalt, but in some situations is underlain by loess member 5 and mixed loess and weathered basalt member 6. The paleosol developed on loess member 5, on both morphological and chemical evidence, is indicative of a period of soil formation of longer duration or greater intensity than is indicated by the overlying paleosols or surface soils. Therefore members 5 and 6 are grouped into a separate formation.

A radiocarbon chronology for loess members 1 and 2 of the Dashing Rocks Formation, suggests loess accumulation phases from 9900 to 11,800 and prior to 31,000 radiocarbon yr BP, followed by periods of soil formation and contemporary peat development in surface depressions.

A hypothesis is presented suggesting that at least loess members 1 and 2 of the Dashing Rocks Formation accumulated during periods of glacial recession which in turn precipitated fluvial and eolian erosion, transport and redeposition of fluvioglacial deposits. The periods of soil development indicated by the paleosols were initiated during warm interstadial conditions and continued throughout the cooling of the following stadial. Such an interpretation has its parallels in the northern hemisphere but is in slight disagreement with previous glacial and fluvioglacial chronologies accepted in New Zealand.