SEDIMENTOLOGY AND GEOCHEMISTRY AS A KEY TO THE TAPHONOMY OF THE UPPER TRIASSIC VIRGINIA SOLITE QUARRY LAGERSTATTE (DAN RIVER/DAN RIVER BASIN: NEWARK SUPERGROUP)

The Virginia Solite quarry, straddling the North Carolina/Virginia border yields phenomenally preserved invertebrates and vertebrates in a thin (34-mm-thick) layer. Previous hypotheses for the exquisite preservation of microtrichiae (insects) and skin impressions (of the protorosaur Tanytrachelos) relied primarily on grain size, color, and cyclostratigraphic interpretations to suggest that the Lagerstätte formed in anoxic conditions at great depths in paleo-Lake Danville. Our fine-scale sedimentological and chemostratigraphic analysis of a 332-mm-thick portion of the Van Houten cycle encasing the “insect layer” challenges this hypothesis by showing a transition from black siliciclastics to dark brown dolomitic laminates in the “insect layer,” separated by 1-3 mm of dolostones that is laterally traceable for >200 m. The presence of microspar, chemical variations between dolomite crystal cores and rims, and low permeability and porosity that preclude large-scale replacement all indicate that the dolomitic laminates of the “insect layer” are primary. Carbonaceous interlaminations within the dolomite suggest microbiallly mediated dolomite precipitation, and we interpret the “insect layer” as the most oxygenated in the succession, thereby inferring that it formed in shallow conditions. Our analysis indicates that paleo-Lake Danville was in fact shallow, saline, and alkaline enough to allow primary dolomite precipitation. The combination of high F levels in the “insect layer” with sufficient cations to precipitate dolomite is attributed to lake-margin groundwater seeps. We hypothesize that it was the toxicity of those seeps, as opposed to great water depth, that excluded bioturbation, predation, and/or scavenging of the fossils. Our interpretation of the Lagerstätte as a shallow, yet toxic, lacustrine environment therefore explains the fine preservation without resorting to cyclical 200-m fluctuations in lake level inherent in previous models.