During the last glacial maximum in West Antarctica separate ice caps developed on Alexander Island and on Palmer Land, became confluent in George VI Sound, and discharged northward from latitude 72° S. Radiocarbon (>32,000 yr) and amino acid (approximately 120,000 yr) age determinations on shell fragments (Hiatella solida) found in basal till suggest a Wisconsin age for the glaciation that incorporated them. The pattern of ice flow differed from that deduced for this area in the CLIMAP reconstruction. Following the maximum stage, there was a stadial event when outlet valley glaciers flowed from smaller ice caps into George VI Sound. More widespread recession permitted the George VI ice shelf to deposit Palmer Land erratics on eastern Alexander Island before isostatic recovery raised them to final elevations of about 82 m. The ice shelf may have been absent at about 6500 yr B.P., when large barnacles (Bathylasma corolliforme) were living in the sound. Small glaciers readvanced to form at least two terminal moraines before the ice shelf re-formed and incorporated the barnacle shells into its moraine on Alexander Island. The shells gave a 14C age (corrected for Antarctic conditions) of about 6500 yr B.P. and an amino acid ratio consistent with a Holocene age. Valley glaciers readvanced over the ice-shelf moraine before oscillations of both valley glaciers and the ice shelf led to the formation of the present sequence of contiguous ice-cored moraines, probably during the Little Ice Age. Such oscillations may represent a climatic control not yet observed in the dry valleys of Victoria Land, the only other part of Antarctica studied in detail for glacier fluctuations.

Detailed measurements of the altitudes of the shorelines of former ice-dammed lakes in Glen Roy and vicinity in the Scottish Highlands prove erroneous the conventional view that former shorelines in areas affected by glacio-isostatic uplift are uniformly tilted and/or gently warped. The measurements demonstrate differential uplift of blocks of the earth's crust. The surfaces of some blocks have no detectable tilt whereas others have gradients up to at least 4.6 m/km and are tilted in different directions. In three areas 0.7–2.0 km long, shorelines are distorted by crustal movements; all these areas have landslides attributed to earthquakes that accompanied the release of stress. In the area of greatest local distortion three shorelines rose about 3 m above their average altitudes for the immediately surrounding area, and a fault scarp was produced. Possibly these local distortions accompanied catastrophic lake drainage by jökulhlaup, analogous to crustal movements associated with man-made lakes. A relationship between crustal movement and the limit of a glacial advance is demonstrated. These findings have implications for other parts of the world affected by glacioisostatic uplift.

Six methods for approximating late Pleistocene (Pinedale) equilibrium-line altitudes (ELAs) are compared for rapidity of data collection and error (RMSE) from first-order trend surfaces, using the Colorado Front Range. Trend surfaces computed from rapidly applied techniques, such as glaciation threshold, median altitude of small reconstructed glaciers, and altitude of lowest cirque floors have relatively high RMSEs (97–186 m) because they are subjectively derived and are based on small glaciers sensitive to microclimatic variability. Surfaces computed for accumulation-area ratios (AARs) and toe-to-headwall altitude ratios (THARs) of large reconstructed glaciers show that an AAR of 0.65 and a THAR of 0.40 have the lowest RMSEs (about 80 m) and provide the same mean ELA estimate (about 3160 m) as that of the more subjectively derived maximum altitudes of Pinedale lateral moraines (RMSE = 149 m). Second-order trend surfaces demonstrate low ELAs in the latitudinal center of the Front Range, perhaps due to higher winter accumulation there. The mountains do not presently reach the ELA for large glaciers, and small Front Range cirque glaciers are not comparable to small glaciers existing during Pinedale time. Therefore, Pleistocene ELA depression and consequent temperature depression cannot reliably be ascertained from the calculated ELA surfaces.

Amino acid enantiomeric (D/L) ratios in the mollusk Mercenaria are compared with recently published biostratigraphic and/or U-series solitary coral data from 22 Quaternary localities on the central and southern Atlantic Coastal Plain. In all cases, local relative aminostratigraphic sequences are consistent with relative ages inferred from U-series or biostratigraphic data, although occasionally more depositional events are recognized by aminostratigraphic than biostratigraphic methods. However, if the U-series data are used as age calibrations for the D/L values, latitudinal trends of “isochronous” D/L values are highly variable and conflict with trends expected from the present temperature gradient, which is smooth and nearly linear between 45° and 25° N. Age estimation can be performed independently of the U-series data using a kinetic model that relies on the assumption that Pleistocene temperature gradients have also been smooth functions of latitude, although significantly steeper than the present temperature gradient. Within the uncertainties of this assumption, kinetic model age estimates for localities in the coastal plain fall into the following groups: 70,000–130,000 yr, 200–250,000 yr, 300,000–400,000 yr, 500,000–600,000 yr, 700,000–800,000 yr, and > 1,000,000 yr. Major conflicts between these model age estimates are observed for localities near Charleston, South Carolina and in central Virginia. These conflicts could indicate that the basic temperature assumptions of aminostratigraphy are incorrect, and that apparent local aminostratigraphic sequences (clusters of different D/L values) could be due to factors other than age difference. Alternatively, some of the U-series dates may be only minimum ages for these localities.

The results of an integrated study comprising litho- and biostratigraphic investigations, uranium-series coral dating, amino acid racemization in molluscs, and paleomagnetic measurements are compared to ascertain relative and absolute ages of Pleistocene deposits of the Atlantic Coastal Plain in North and South Carolina. Four depositional events are inferred for South Carolina and two for North Carolina by all methods. The data suggest that there are four Pleistocene units containing corals that have been dated at about 100,000 yr, 200,000 yr, 450,000 yr, and over 1,000,000 yr. Some conflicts exist between the different methods regarding the correlation of the younger of these depositional events between Charleston and Myrtle Beach. Lack of good uranium-series dates for the younger material at Myrtle Beach makes the correlation with the deposits at Charleston more difficult.