[en] To date, we still lack evidence of the evolutionary forces driving the diversity of sagittae. To better understand the pattern of their shape diversification, we collected the largest dataset of teleost fish otoliths, including 697 species from 68 orders and 309 families. We used geometric morphometric methods to quantify the mesial and dorsal faces. Phylogenetic information was recovered from pruning the consensus time-tree from Betancur et al. (2017). First, we found little evidence of phylogenetic signal on size and shape variation across taxa. Then, we fitted various models of continuous trait evolution to decipher the main dynamic of otolith evolution. Among tested models (e.g. Brownian motion, accelerating and decelerating models), the Ornstein-Uhlenbeck model with a single optimal trait value for all taxa was the best supported. Globally, results from tests of phylogenetic signal and trait modeling reveal that species trait value are uncorrelated. Convergence would be a good candidate to explain such an evolutionary pattern. Next, we assessed morphological integration between sulcus and otolith outline and we surprisingly found high morphological covariation. Finally, we found the accumulation of otolith disparity is unrelated to orders’ age but we highlighted a tendency of higher shape (and not size) diversity in clades evolving at faster rates. Our preliminary analyses reveal that the evolution of sagittae is not random. Shape convergence at high taxonomic scales probably highlights morpho-functional constraints. We ask for further studies focusing on eco-morphological hypotheses to reveal more macroevolutionary drivers of in inner ear functional morphology.
