The formation of upper Paleozoic (Viséan to Sakmarian-Artinskian) Euramerican cyclothems, which resulted from base-level fluctuations of up to 100 m, commonly are attributed to large-scale waxing and waning of Gondwanan glaciers. However, evaluation of the geographic and chronostratigraphic distribution of Gondwana deposits reveals that glaciation was not the primary cause of base-level changes of that magnitude.Gondwana strata contain three non-overlapping glacial successions. Glacial I (Frasnian to possibly Tournaisian) and Glacial II (Namurian to lowermost Westphalian) rocks were deposited by alpine glaciers. Water sequestered by these glaciers was insufficient to account for the base-level changes. In contrast, Upper Carboniferous (Stephanian) to Lower Permian (Sakmarian-Artinskian) Glacial III rocks were widespread and indicate deposition by ice sheets that may have covered a total area of between 17.9 and 22.6 × 10 6 km 2 . Complete ablation of a single ice sheet of this size would produce eustatic changes of ~100 m. However, multiple ice sheets were likely present, which would have resulted in considerably smaller fluctuations in sea level during Glacial III deposition.The argument that Glacial I and II deposits were originally comparable in extent to those of Glacial III, but were eroded during the advance of Glacial III ice-sheets, is untenable. Weathered granite profiles on the pre-Glacial III unconformity occur scattered over a 1200-km length of the central Transantarctic Mountains. The profiles indicate prolonged subaerial exposure and, thus, an absence of ice cover. These and non-glacial successions in Gondwana constrain the size of ice sheets before Glacial III deposition and imply that glaciation prior to Glacial Episode III was not the primary cause of base-level changes linked to upper Paleozoic Euramerican cyclothems.

Figure 1. Carboniferous and Permian paleogeographic map of Gondwana (after Powell and Li, 1994), showing several hypothetical ice sheets. ABSTRACTEvidence from Antarctica indicates that a 2000-km-long section of the Transantarctic Mountains-including Victoria Land, the Darwin Glacier region, and the central Transantarctic Mountains-was not located near the center of an enormous Carboniferous to Early Permian ice sheet, as depicted in many paleogeographic reconstructions. Weathering profiles and soft-sediment deformation immediately below the preglacial (pre-Permian) unconformity suggest an absence of ice cover during the Carboniferous; otherwise, multiple glacial cycles would have destroyed these features. The occurrence of glaciotectonite, massive and stratified diamictite, thrust sheets, sandstones containing dewatering structures, and lonestone-bearing shales in southern Victoria Land and the Darwin Glacier region indicate that Permian sedimentation occurred in ice-marginal, periglacial, and/or glaciomarine settings. No evidence was found that indicates the Transantarctic Mountains were near a glacial spreading center during the late Paleozoic. Although these findings do not negate Carboniferous glaciation in Antarctica, they do indicate that Gondwanan glaciation was less widespread, and, therefore, that glacially driven changes to other Earth systems (i.e., glacioeustatic fluctuations, climate) were much smaller than previously hypothesized.

No abstract