A body of moderately well sorted and well stratified ice-cemented sand, here informally called the Coral Ridge sand body, was deposited across eastern Taylor Valley before the deposition of a veneer of glaciogenic deposits related to late Pleistocene incursions of the Ross Sea ice sheet. The Coral Ridge sand body is more than 50 m thick where preserved in a north-south trending ridge that is transverse to the long axis of the valley. The ridge forms a drainage divide that stands 100 m above sea level and separates the basin of Lake Fryxell on the west from the seacoast to the east. An erosion surface having about 35 m of abrupt local relief was apparently developed on the sand body before the latest incursions of Ross Sea ice.

The Coral Ridge sand body accumulated in a fluviatile or possibly fluviomarine deltaic environment following deposition of coarse diamictons and interbedded layers of sand in a fjord that once occupied the site of Taylor Valley. From its sedimentary characteristics and topographic distribution, the sand was deposited across the valley following filling of the fjord. The 1) comparatively high degree of sorting, 2) general lack of clay and silt, and of very coarse detritus, and 3) common fluviatile cross-stratification, argue for stream transport and for accumulation principally above sea level. Coarse glacial detritus, which must have existed in the source area, was not transported to the area of sand deposition. Very fine glacial detritus that presumably was transported with the sand was deposited some place beyond the area of sand deposition. Evidence bearing on the direction of transport, and thus the source area, has not yet been developed. The source could have been a grounded ice sheet in the Ross Sea to the east, or alternatively, the source could have lay to the west in the area of the Lake Fryxell basin at a time when the basin was occupied by a more extensive Taylor Glacier.

Evidence bearing on the age of the Coral Ridge sand body also is not fully developed. West of the Coral Ridge divide, the upper 2-3 m of a 14-mthick section of Coral Ridge sand near the top of hole DVDP-11 is reversely polarized. All other beds of the ice-cemented sand have been found to be normally polarized. Two possibilities exist: 1) the Coral Ridge sand body was deposited during late Pliocene time, mainly during a time of normal polarity of the Gauss polarity epoch, or 2) the sand body is much younger and was deposited during the Bruhnes normal polarity epoch of Pleistocene time, less than 730,000 years ago (in which case, the reversely polarized strata are anomalous). Additional subsurface and surface geological and paleomagnetic study is required to resolve the problems of age and source, critical to deciphering the late Cenozoic glacial and structural history of Taylor Valley and environs.