Geometric morphometrics suggests different environmental pressures on small and large <em>Polygnathus</em> conodonts during the recovery after the Hangenberg crisis (latest Devonian-earliest Carboniferous)

The morphology of feeding structures is driven by the nature of the food source, itself influenced by environmental factors, resulting in an indirect relationship between the form of feeding structures and the surrounding environment. However, food sources may change along growth, leading to potentially complex interactions between ontogenetic trajectories and environmental variation. The present study investigates these interactions by analysing the platform shape of P1 elements in unornamented Polygnathus from a latest Devonian - early Carboniferous succession from the Montagne Noire, France, deposited after the Hangenberg Event mass extinction. Allometric trajectories were found to vary along the record, starting from different small-sized shapes but converging towards a common adult morphology. To assess variation of environmental forcing along ontogeny, the specimens from each stratigraphic level were split into three size classes. The shape of the small size class varied primarily with conodont biofacies, the medium size class with sedimentary microfacies, while the large size class followed a random walk unrelated to environmental variation. These results agree with a change of diet along ontogeny, and with an increase of functional constraints related to occlusion in large-sized elements, leading to their convergent adult morphology. Biofacies and microfacies trace different signatures of bathymetry and distance from the coast, suggesting that post-Hangenberg unornamented Polygnathus varied in their exploitation of the water column along ontogeny. Finally, the high disparity of small-sized elements suggests a relaxation of selective constraints related to occlusion, and/or different foraging strategies, making them more susceptible to environmental influences than the large-sized elements. Methods Description of methods used for the collection/generation of data: Polygnathus P1 elements were photographed in oral view using a Nikon DS U3 stereomicroscope camera, a digital camera, and the associated NIS software. Images of dextral specimens were mirrored in order to measure all conodonts with a comparable orientation. The two-dimensional outline of the platform was quantified using a set of three landmarks positioned at the rostral insertion point, the tip of the platfor,m and the caudal insertion point, complemented by 13 sliding semi-landmarks positioned at equal intervals along the margins between the tip of the platform and the respective insertion point. Landmarking was performed using TPSdig. Methods for processing the data: A Generalized Procrustes Analysis (GPA) was performed on the coordinates of the landmarks and semi-landmarks to standardize the size, orientation, and position of the specimens. During the GPA, the semi-landmarks are allowed to slide between the previous and next (semi-)landmark, according to the bending energy criterion. The logarithm of the centroid size (i.e., the square root of the sum of squared distances from the landmarks and semi-landmarks to the centroid of the configuration, referred to as (LogCS) was considered as a size proxy for the Polygnathus P1 elements. Aligned coordinates and LogCS are provided in the present dataset.