Spinosaurids as ‘subaqueous foragers’ undermined by selective sampling and problematic statistical inference

Abstract

Fabbri et al.[1] claim that the huge sail-backed dinosaur Spinosaurus aegyptiacus was a "subaqueous forager," diving underwater in pursuit of prey, based on their measure of bone "compactness." Using thin-sections and computed tomographic (CT) scans of thigh bone (femur) and trunk rib from various living and extinct vertebrates, they claim to be able to distinguish taxa with "aquatic habits" from others. Their conclusions are undermined by selective bone sampling, inaccuracies concerning spinosaurid bone structure, faulty statistical inferences, and novel redefinition of the term "aquatic."

Sinking a giant: quantitative macroevolutionary comparative methods debunk qualitative assumptions

“…In contrast, extant taxa that are anatomically less specialized for subaqueous foraging (potentially with lower bone densities) could be classified as such based on observation of their in-life behaviour. Therefore, unless Myhrvold et al 1 is suggesting that stem whales, such as remingtonocetids, basilosaurids, and protocetids, and sauropterygians, including plesiosaurs, were not subaqueous foragers, we don't see how this challenges our results. Furthermore, some of the extinct taxa that we include are crown members of recent clades of aquatic animals like phocids or otariids or closely related stem members of recent groups with similar ecologies (e.g.…”

“…(2) Concerns about our statistical procedure, including the scope of our comparative dataset and potentially high misclassification rates. (3) That, in their view, we presented a novel redefinition of the term 'aquatic'.We closely examine these concerns, supporting the validity of our original bone density measurements 2 even based on the new scan data presented by Myhrvold et al 1 Importantly, we show that…”

“…1) Accuracy of the quantification of skeletal density in spinosaurid dinosaurs, its relationship to mass reduction, and inference of subaqueous foraging in Spinosaurus and Suchomimus based on novel data Universal definition of density: as every scholar knows, the density of an object is defined as the ratio between its mass and volume. Therefore, mass cannot be assumed to have a linear relationship with density, contrary to the suggestion of Myhrvold et al 1 . A clear example of this basic physical concept applied to buoyancy is the difference between granite and pumice, which have an average density of 2.7-2.8 g/cm 3 and 0.5 g/cm 3 3 , respectively.…”

“…Before tackling these concerns in detail, we want to specify that the depiction of our methods by Myhrvold et al 1 is incorrect and misleading. Our ecological inference 2 is not based on the visual interpretation of the regression lines and scores in PGLS models or violin plot distribution, as they imply.…”

“…This bone density value was substituted to the one of the neotype in our original dataset to infer ecological behaviour in this new individual, following the methods of our study. Similarly, we also substituted our previous value of bone density for the femur of Suchomimus with a novel estimate obtained from a cross section extrapolated from the low-quality CT scan data presented by Myhrvold et al 1 : the section we used come from an area closer to the fourth trochanter (sections 4-3 in Supplementary Figure 1 in Myhrvold et al 1 ). The resulting bone density is only slightly higher (bd=0.756) than our original value for Suchomimus and might be related with the low resolution of the scan data.…”