How subtle are the biases that shape the fidelity of the fossil record? A test using marine molluscs

•We test preservation biases at the species level via experimental taphonomy.•Chitons are morphologically homogeneous, yet have different preservation potentials.•Preserved diversity is correlated to decay rates (faster decay, less fossil taxa).•Preservation is non-uniform and determined by species' biology and individual size.

Biases in preservation shape the fossil record, and therefore impact on our reconstructions of past environments and biodiversity. Given the intensive recent research in the general fields of taphonomy and exceptional preservation, surprisingly, fundamental questions remain unanswered about species-level variation in skeletal preservation potential at low taxonomic levels (e.g. between genera from the same family, or between taxa from related families) across myriad groups with multi-element skeletons. Polyplacophoran molluscs (chitons sensu lato) are known from the late Cambrian to Recent, and possess a distinctive articulated scleritome consisting of eight overlapping calcareous valves. The apparent uniformity of living chitons presents an ideal model to test the potential for taphonomic biases at the alpha-taxon level. The vast majority of fossil chitons are preserved as single valves; few exhibit body preservation or even an articulated shell series. An experimental taphonomic programme was conducted using the Recent polyplacophorans Lepidochitona cinerea and Tonicella marmorea (suborder Chitonina) and Acanthochitona crinita (Acanthochitonina). Experiments in a rock tumbler on disarticulated valves found differential resistance to abrasion between taxa; in one experiment 53.8–61.5% of Lepidochitona valves were recovered but 92% of those from Tonicella and 100% of elements from Acanthochitona. Chiton valves and even partly decayed carcasses are more resistant to transportation than their limited fossil record implies. Different species of living chitons have distinctly different preservation potential. This, problematically, does not correlate with obvious differences in gross valve morphology; some, but not all, of the differences correlate with phylogeny. Decay alone is sufficient to exacerbate differences in preservation potential of multi-element skeletons; some, but not all, of the variation that results is due to specimen size and the fidelity of the fossil record will thus vary intra-specifically (e.g. between ontogenetic stages) as well as inter-specifically.