Testing for taphonomic bias in deep time using trilobite sclerite ratios

Taphonomic sorting can be assessed directly in fossil assemblages by comparing expected and observed proportions of elements of multielement skeletons. Trilobites are model organisms for this approach because each individual possesses one cranidium (head) and one pygidium (tail). Departures from an expected 1:1 cranidia:pygidia (C/P) ratio reflect taphonomic processes such as size- or shape-sorting. We analyzed a dataset of >16,000 secondarily silicified cranidia and pygidia from subtidal, storm-influenced fades of a highstand systems tract in the House Limestone (Lower Ordovician) in Utah. Species fall into four distinct isotaphonomic groups, which we define as sets of morphologically similar species likely to have similar responses to taphonomic processes. All isotaphonomic groups have median C/P ratios that depart significantly from expected proportions; micropygous groups show strong enrichment of cranidia in all samples, whereas isopygous groups include some pygidia-rich samples. Despite this, rank orders of abundances and C/P ratios are not correlated for any isotaphonomic group, indicating that sorting bias is not controlling abundance patterns. Cluster analysis of genus abundance data defined two biofacies, each of which included unique dominant taxa, and which characterized early and late highstand strata. The same groupings of samples were readily recognizable using ordination (non-metric multidimensional scaling). Rank orders of C/P and positions of samples along ordination axes are not correlated, so that sorting bias does not influence biofacies groupings. Rank order of species richness of samples, both before and after rarefaction also shows no correlation with C/P. The results indicate that paleoecological analysis is possible despite clear evidence of taphonomic sorting. In this case, sorting has shuffled sclerite ratios without having a significant impact on taxon abundances and species richness. However, taphonomic bias may be problematic in more proximal marine environments where frequent winnowing produces extensive sorting and differential breakage of skeletal material. (C) 2016 Elsevier B.V. All rights reserved.