The influence of plane of section on the identification of bone tissue types in amniotes with implications for paleophysiological inferences.

The proportion of woven bone (WB) to parallel-fibered bone has been extensively used to infer bone growth rates and resting metabolic rates of extinct organisms. The aim of this study is to test in a variety of amniotes how reliably WB content can be measured using transverse sections. For this, we analyzed femoral transverse mid-diaphyseal thin sections of 14 extant and extinct taxa and the corresponding longitudinal sections for comparative purposes. We used the following characters to identify WB in transverse sections because they are known to be distinct from those observed in parallel-fibered bone: an isotropic bone matrix at tissue scale; an anisotropic microlamellar arrangement in former osteoblast secretory territories at cellular scale; no alignment between osteocytes; and canaliculi running radially from large irregular osteocyte lacunae. Our null hypothesis predicts no differences between the amount of WB quantified in the transverse and longitudinal sections of a given long bone. Qualitatively, when a stripe or a patch of WB was identified in a transverse section, the corresponding stripe or patch of WB was always found at the same location in the corresponding longitudinal section. Quantitatively, a Wilcoxon signed-rank nonparametric paired test did not detect a significant difference in the WB content of the two section planes. Thus, the null hypothesis is not rejected. Considering that paleohistology is a destructive method, we recommend a workflow to efficiently establishing the proportion of WB: quantifying it in transverse sections; preparing and analyzing longitudinal sections only in cases where an ambiguity remains; reanalyzing the corresponding transverse sections.