Allometry of bony sound reception structures

The origin of sensory structures provides an excellent framework for studying how constraints and selective pressures affect the evolution of complex features. The evolution of the mammalian middle ear from the jaw hinge of non-mammalian synapsids offers a deep time perspective on sensory evolution but is limited by a poor understanding of early synapsid hearing. This work tests hypothesis that the reflected lamina of the angular in non-mammalian synapsids followed a strict, negative allometric trend that may be expected for a sound reception structure. Allometry is first investigated in the pterygoid of chameleons, which was co-opted for hearing in some species and represents a possible analog for the synapsid reflected lamina. Results indicate that chameleons with a pterygoid ear converge on a similar allometric slope, while other species have variable slopes, suggesting an optimum allometric pattern in sound receivers. In the reflected lamina, we find reduced variation around the allometric trend in therocephalians and non-bidentalian anomodonts, and evolutionary modeling suggests constraint in these groups. These results are consistent with a mandibular middle ear in non-mammalian synapsids, suggesting that selective pressures for hearing ability were present long before the evolution of the mammalian middle ear.

Methods
We photographed museum specimens of chameleons in ventral, lateral, and ventromedial views, and synapsids were photographed in lateral view with a Nikon D3100 camera. Additional photographs of inaccessible synapsid specimens were graciously provided by other researchers and museum staff. 59 synapsid specimens were also 3D surface scanned using an Artec Space Spider blue light scanner with a voxel size of 0.3 mm.

For chameleons, ImageJ V1.51j (Schneider et al., 2012) was used to measure basal skull length, pterygoid plate area, orbit diameter, and coronoid process height from photographs. In synapsids, we measured reflected lamina area in ImageJ and mandible length and orbit diameter using a Neiko digital caliper. Area was measured by outlining each structure with the “polygons selection” tool in ImageJ. When photographs and/or caliper measurements were unavailable, comparable measurements were taken using Meshmixer V3.5.474 (©Autodesk, San Rafael, CA, U.S.A) and the R package molaR (Pampush et al., 2016) on 3D surface scans.