Skip to main content
Open Access Publications from the University of California


PaleoBios bannerUC Berkeley

Epiplastral and geographic variation in Echmatemys, a geoemydid turtle from the Eocene of North America: A multi-tiered analysis of epiplastral shape complexity


Numerous geoemydid turtle fossils from the extinct genus Echmatemys have been recovered from the middle Eocene Uinta Formation, Uinta Basin, Utah over the past several decades. Here, we tested whether co-occurring Uintan species Echmatemys callopyge and E. uintensis can be reliably differentiated based on epiplastral morphology, and whether their geospatial distributions overlapped significantly. The geographic spatial and stratigraphic distributions of Uinta Basin E. callopyge and E. uintensis specimens were compared using ArcGIS and analysis of variance (ANOVA). The analysis revealed overlapping geographic distributions of these two species, and no significant differences in stratigraphic dispersal. This finding of extensive geospatial overlap between the two Uintan Echmatemys species highlights the need for accurate taxonomic identification, such as the gular scale morphology validated here. In addition, we sought to address a methodological question regarding the relative efficacy of data complexity in this context. Using epiplastra from three additional Eocene species of Echmatemys, we employed hierarchical analyses of increasing data complexity, from standard linear dimensions to 2D geometric morphometrics to 3D laser scans, to determine the degree to which data complexity contributes to taxonomic assessments within this genus. Uintan species E. callopyge and E. uintensis were found to differ significantly in epiplastral shape as captured by all three categories of data. These findings verify that these two co-occurring species can be differentiated consistently using the shape of the gular scale, and that the use of geometric morphometrics can improve identification of fragmentary specimens. Among the non-Uintan species, dorsal and ventral 2D landmark data reliably differentiated among species, but the linear dimensions were less useful.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View