Abstract/Summary

<jats:p>I present a Bayesian phylogenetic predictive modelling (PPM) framework that allows the prediction of muscle parameters (physiological cross-sectional area, <jats:italic>A</jats:italic><jats:sub>Phys</jats:sub>) in extinct archosaurs from skull width (<jats:italic>W</jats:italic><jats:sub>Sk</jats:sub>) and phylogeny. This approach is robust to phylogenetic uncertainty and highly versatile given its ability to base predictions on simple, readily available predictor variables. The PPM presented here has high prediction accuracy (up to 95%), with downstream biomechanical modelling yielding bite force estimates that are in line with previous estimates based on muscle parameters from reconstructed muscles. This approach does not replace muscle reconstructions but one that provides a powerful means to predict <jats:italic>A</jats:italic><jats:sub>Phys</jats:sub> from skull geometry and phylogeny to the same level of accuracy as that measured from reconstructed muscles in species for which soft tissue data are unavailable or difficult to obtain.</jats:p>