22-12 Detangling the enigmatic cricetid Leidymys: a 2D geometric morphometric approach to understanding species variation

Session: Working Up an Apatite: Teeth as Paleo -Ecological and -Climatological Archives

Presenting Author:

Authors:

Abstract:

Fossil rodents are mostly known from isolated teeth, which are challenging to identify taxonomically. Members of Cricetidae are commonly identified by the morphology of the first molar. However, these identifications are difficult in worn teeth. Diagnoses are particularly subtle in the Oligocene cricetid genus Leidymys. The eight species of Leidymys share similar occlusal morphology and are mainly differentiated by size and geography. Three of these Leidymys species occur in the John Day Formation. No discrete dental features diagnose the species, and the ranges of dental dimensions overlap among the three species. The presence of three species was inferred from a trimodal size distribution in a large sample. Even the generic diagnosis of Leidymys is so subtle that members of other genera (e.g., Paciculus) have been misidentified as Leidymys, confounding efforts to accurately assess fossil rodent diversity. Preliminary work using geometric morphometrics (GM) to solve this difficulty compared dental variation in the fossil genera Copemys and Leidymys to that found in the modern cricetid genera Reithrodontomys and Perognathus. The species-level GM identification matched previous identification for roughly 70% of the Copemys specimens, ~30% of Leidymys specimens, and ~70% of modern cricetid specimens when plotted in separate morphospaces. We continue to study this variation using Elliptical Fourier Analysis (EFA) of molar outlines by combining both modern and fossil cricetid species into a single morphospace. Photographs of a sample of several extant and extinct species were digitized using 64 equally spaced semi-landmarks and six homologous landmarks. Coordinates were analyzed with EFA, which describes the shape as a sum of harmonics of decreasing wavelength, yielding four Fourier Coefficients (FCs) per harmonic. Principal component analysis of the resulting FCs yielded a rich dataset to describe tooth shape. The first five principal components (PCs) account for approximately 90% of the variation found within the outlines. Using the first five PCs, linear discriminant analysis showed that the majority of specimens can be matched to their previous species identification. These findings contribute to the understanding of the variation encompassed by current cricetid species and help define the morphological limits of species within Leidymys and Copemys.

Geological Society of America Abstracts with Program. Vol. 57, No. 6, 2025

doi: 10.1130/abs/2025AM-9184

© Copyright 2025 The Geological Society of America (GSA), all rights reserved.