Reconstructing an ancient fish: Three-dimensional skeletal restoration of the head of Mawsonia (Sarcopterygii, Actinistia) using CT scan, and an adjusted model for body size estimation in fossil coelacanths.

Mawsonia constitutes one of the most conspicuous fossil coelacanth taxa, due to its unique anatomy and possible maximum body size. It typifies Mesozoic coelacanth morphology, before the putative disappearance of the group in the fossil record. In this work, the three-dimensional cranial anatomy and body size estimations of this genus are re-evaluated from a recently described specimen from Upper Jurassic deposits of Uruguay. The 3D restoration was performed directly on the material based on anatomical information provided by the living coelacanth Latimeria and previous two-dimensional restorations of the head of Mawsonia. The montage was then scanned with computed tomography and virtually adjusted to generate an interactive online resource for future anatomical, taxonomic and biomechanical research. In general terms, the model constitutes a tool to improve both the anatomical knowledge of this genus and its comparison with other coelacanths. It also facilitates the evaluation of possible evolutionary trends and the discussion of particular features with potential palaeobiological implications, such as the anterior position of the eye and the development of the pseudomaxillary fold. Regarding the body size, a previous model for body size estimation based on the gular plate was submitted to OLS, RMA, segmented linear and PGLS regressions (including the evaluation of regression statistics, variance analysis, t-tests and residual analysis). The results point to a power relationship between gular and total lengths showing a better support than a simple linear relationship. The new resulting equations were applied to the studied individual and are provided for future estimates. Although an isometric evolutionary growth cannot be rejected with the available evidence, additional models developed with other bones will be necessary to evaluate possible hidden evolutionary allometric trends in this group of fishes, thus avoiding overestimates.