RESIDUES VERSUS RADIOGRAMS: BIASES IN DIAGENESIS AND PREPARATION OF SILICIFIED FOSSILS IN LIMESTONE

Silicified fossils of calcareous organisms are common objects of study in invertebrate paleontology; they often retain excellent morphological detail due to fine-scale replacement, and are easily recovered from limestone matrix through acid maceration due to silica's high resistance to dissolution. However, such experimental methods can be problematic to paleoecological studies relying on silicified specimens. There are two potential types of bias introduced: differential likelihood that a given specimen will silicify during diagenesis, and differential likelihood that a given specimen will, once silicified, survive preparation and be recovered. By comparing the original makeup of partially silicified limestone to the residues recovered after acid maceration, we can assess and quantify these biases. Following a methodology modified after Pruss et al. (2015), abundance, species richness, evenness, and body size of invertebrate assemblages are compared across X-ray micro-computed tomography (μCT) 3D imagery and virtual thin sections and recovered residues from sample rocks.

Hand samples for this study were collected from the Liberty Hall member of the Middle Ordovician Edinburg Formation of Virginia, a deep-ramp/basinal unit of black shales and argillaceous marls well known for silicified fossils. Cores of approximately 2.5 cm diameter were cut from samples and subjected to μCT before being macerated in 10% acetic acid. μCT projections were stacked to create 3D images, and a single 2D projection was randomly selected from each core to serve as a virtual thin section. Fossils, silicified and calcareous alike, resolved in μCT 3D imagery were measured for body size and identified taxonomically, producing a dataset of all fossils present in the cores. Virtual thin sections were counted for all recognizable fossils and insoluble residues were picked for silicified specimens, producing subset assemblages of fossils identified through those methods.

For each core, the 3D imagery, virtual section, and recovered residue assemblages were compared in order to assess recovery bias. The effects of specimen size and taxonomic affinity on likelihood of recovery were tested using Pearson's correlation coefficient, aiding in correction for this bias in future research.