Wide gape in the Ordovician brachiopod Rafinesquina explains how unattached filter-feeding strophomenoids thrived on muddy substrates

Strophomenoid brachiopods had thin, concavo-convex shells, were ubiquitous colonisers of Paleozoic muddy seafloors, and are hypothesised to have filter-fed in a concave upward orientation. This orientation would elevate their line of commissure out of potentially lethal lophophore-clogging mud. The paradox is that epibiont distributions on strophomenoids support a convex-upward life position, as do studies of strophomenoid stability and trace fossils formed by strophomenoid sediment-clearing. A premise of the concave-upward orientation hypothesis is a narrow gape, which causes narrow, high velocity inhalant currents, leaving strophomenoids vulnerable to sediment entrainment. Herein we investigate the gape angle of Rafinesquina using serial thin sections and peels, silicified specimens, computer modelling, SEM analysis, X-ray microCT, and 3-D printing. Hinge line structure suggests that, conservatively, Rafinesquina could gape 40–45°. Such a gape occurred when diductor muscle contraction could not cause any further rotation, hinge teeth and crenulations were disengaged, and interareas interlocked. In contrast, when closed, hinge teeth were locked in hinge sockets. This wide gape eliminates constraints on feeding orientation. In either convex-up or concave-up orientation, Rafinesquina could feed with slow, diffuse inhalant currents incapable of disturbing sediment, and could snap valves shut to forcefully expel enough water to clear sediment from the mantle cavity, explaining moat-shaped trace fossils associated with shells. Our findings demonstrate that Rafinesquina gaped at an angle approximately equal to the angle between the two interareas when the valves are closed. Our analyses also hint that other strophomenoids with similar interarea angles lived with their shells widely agape.

Methods
We employed a range of techniques for the analysis of the hinge morphology and function of Rafinesquina. We began with examination under a reflected light microscope, also documenting specimens by macrophotography, microphotography, and SEM imaging of the interiors of disarticulated valves with special attention to the articulating features of the hingeline and the distribution of muscle scars in the interior of the shells. We studied silicified Rafinesquina valves in the same way. Specimens from the large USGS collections of isolated valves were useful for comparing the features on opposing valves of similar size, as well as testing for fit. However, isolated valves do not always fit well even if similar in size because they did not come from the same individual. These specimens are also delicate. High resolution illustrations of all figured specimens in different orientations, as well as some of the exploratory scans and cross-sections used in this study, are available in this Dryad data set.

Small silicified specimens are often articulated, and we examined a suite of such articulated specimens visually and scanned them using a SkyScan 1172 X-ray microCT. 3D models of these specimens were examined to gain insights into hinge articulation and those data are available in MorphoSource. Because such specimens had between-valve silicification that could not always be distinguished from shell material, we mainly used these analyses to guide our approaches for serial sectioning and interpreting unsilicified specimens.

The most useful results were obtained through serial sectioning and serial grinding of articulated specimens, whereby we generated a series of sagittal sections through each brachiopod. We employed a slice spacing of ~3 mm in initial work, and based on those results decreased our slice spacing to ~0.5 mm. One 3D model based on serial sectioning work is presented herein, illustrated by different articulations of an oversize printed model. Slices from this work also permitted analysis of the function of each section in two dimensions and calculation of the centre of mass at different gapes. For mass calculations, when density is uniform, the centre of mass is equal to the centre of volume. Volumetric centres were computed before and after remeshing the model to a voxel size of 0.4 mm to smoothly interpolate between slices. Table S1 shows that remeshing did not substantially alter the computed volume. The centre of volume of the valves was calculated with gape angles of 0° (closed), 20°, and 45°, to compute how gape alters the distribution of mass for the highest density component of the living brachiopods - the shell. Additional methodological information is in Appendix S1.