A STATISTICAL META-ANALYSIS OF LITHOLOGIC AND OTHER CONTROLS ON FOSSIL BONE CELLULAR AND SOFT TISSUE PRESERVATION

The organic phase of extant bone consists of four primary components: extracellular fibrous matrix (primarily collagen), blood vessels, intravascular contents, and osteocytes. Demineralization assays utilizing weak acids to isolate organics from biomineralized tissues have recently been applied by many researchers to explore preservation of these components in fossil specimens, revealing frequent retention of endogenous microstructures morphologically consistent with cells and other still-soft components within fossil bones. However, factors controlling the long-term preservation of such labile structures remain mysterious. To address this, we compiled a database of recovery of these materials from 29 recent studies, then conducted a statistical meta-analysis of these data to evaluate the importance of specimen age, taxonomy, entombing lithology, and bone tissue type on endogenous microstructure recovery. Our database encompasses results from 137 bones from 44 geologic formations as old as the Permian. Osteocytes, blood vessels, and fibrous matrix each exhibit bimodal recovery patterns in which most fossil bones either yield many or none of these microstructures. Although their relative abundance in any given fossil bone is extremely variable, statistically-significant Fisher’s Exact tests found that if a bone yields one of these components in abundance then the others are usually also abundant. None of the variables examined significantly influence osteocyte recovery, but Kruskal-Wallis tests identified burial lithology, geologic age, and taxonomic identity as each influencing the recovery of vessels and fibrous matrix. Subsequent pairwise Mann-Whitney tests revealed that bones collected from unconsolidated sediments, of Paleocene age, and/or of birds, amphibians, marine reptiles, or crocodilians to often yield few or no vessels and fibrous matrix. Although these findings hint at possible controls on cellular and soft tissue preservation in fossil bones, they should be viewed cautiously as they are demonstrably biased by uneven sampling. For example, many of these trends stem partially from over-representation of data from nonavian dinosaur specimens from Cretaceous fluvio-lacustrine deposits (comprising nearly half the dataset). Future demineralization assays should therefore focus on (1) non-mammalian specimens from the Cenozoic and (2) Jurassic and older non-dinosaurian specimens, especially those preserved in less-common depositional environments (e.g., eolian and marine settings).