217-12 A Blind Evaluation of Morphological Keying With Genetically Identified Southern African Rodents

Session: Paleontology, Paleoecology/Taphonomy

Presenting Author:

Charlotte Hohman

Authors:

Hohman, Charlotte Jeanne Heinz¹, Campbell, Timothy Lee², Bell, Christopher Jackson³, Thies, Monte LeRoy⁴, Lewis, Patrick J⁵

(1) Department of Biological Sciences, Sam Houston State University, Huntsville, TX, USA, (2) College of Health Sciences, Midwestern University, Downers Grove, Illinois, USA, (3) Jackson School of Geosciences, The University of Texas at Austin, Austin, TX, USA, (4) Department of Biological Sciences, Sam Houston State University, Huntsville, Texas, USA, (5) Department of Biological Sciences, Sam Houston State University, Huntsville, Texas, USA,

Abstract:

The ability to morphologically distinguish among differing taxa is fundamental to the field of paleontology. Rodents form the most speciose order of mammals today and their study has supported a wide range of paleontological research areas including paleoecology. Due to close associations found in some taxa with specific habitat types, rodents are frequently used in paleoenvironmental reconstructions. This is especially true in Africa, where rodent fossils have played a major role in understanding the habitats associated with early hominins, as well as broader paleoecological patterns across the continent. These applications rely on accurate identification of abundant, although often fragmentary, rodent skeletal material from caves and other fossil-bearing deposits. This task is complicated by high levels of morphological convergence, substantial intraspecific variation, and the prevalence of cryptic species. To address the frequency of misidentification among Quaternary rodents, we conducted a blind morphological identification study using 76 rodent specimens from South Africa, Namibia, and Botswana with published GenBank sequences or karyotype data. Specimens representing 13 genera and 3 families were photographed and keyed separately by cranium and dentary using the Linchamps et al. (2023) key to Quaternary rodents of South Africa. The identifying individual (the lead author) was blind to the genetic identity of the specimens. Specimens were photographed and digitized by coauthors and assigned unique coded specimen identifiers to ensure anonymity and reduce biases. Morphological identifications were then compared to published genetic identifications to calculate error rates. For crania, genus-level error was 31% and family-level error was 1%. Within Muridae (88% of the sample), no subfamily-level error occurred, and tribe-level error was 7%. Dentary identifications were more problematic, with genus-level error at 42% and family-level error at 5%; subfamily- and tribe-level errors within Muridae were 4% and 11%, respectively. Disagreement between cranium and dentary identifications for the same specimen occurred in 26% of cases at the genus level and 3% at the family level. Given that complete crania and dentaries were available in this study, unlike the fragmentary remains typically encountered in the fossil record, the observed error rates are cause for concern. These results suggest caution is warranted on the reliability of fossil rodent identifications and, by extension, on the paleoenvironmental reconstructions based upon them.

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

doi: 10.1130/abs/2025AM-9161

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