The 89th Annual Meeting of the American Association of Physical Anthropologists (2020)

The effect of Procrustes superimposition on the quantification of hominin cranial morphology


Department of Anthropology, University of Pittsburgh

April 17, 2020 , Platinum Ballroom Add to calendar

In a geometric morphometric analysis (GMA), when the Procrustes superimposition eliminates the effect of orientation in a group of superimposed landmark configurations, the sole criterion is to rotate them until the sum of squared differences between every pair of corresponding landmarks is minimized. On the other hand, descriptive morphology is based on orienting specimens by the Frankfurt plane.

Therefore, this study assesses whether shape differences captured by GMA can be incompatible with observed morphological differences, due to different ways of aligning specimens. For this purpose, four surface semilandmark sets were collected from 3D models of hominin crania for analyzing the gross morphology of the overall cranium, posterior cranium, temporal bone, and frontal bone. Principal component analysis (PCA) was performed to simplify the process of visualizing shape differences.

Results indicate that some aspects of shape differences revealed by higher-ranked principal components, such as the degree of frontal rise, the direction that the supraorbital region faces, the trajectory of the squamosal suture, and the inclination of the nuchal plane, cannot be directly associated with observed morphological differences. Evaluating the results of the Procrustes superimposition shows that the superimposed landmark configurations are all rotated to different angles. This inconsistent rotation contributes to the incompatibility between aspects of shape differences captured by PCA and observed morphological features. Thus, when using GMA, especially with the aid of PCA, to analyze the overall cranial morphology of hominins, extra attention should be paid to the effect of the Procrustes superimposition.

Slides/Poster (pdf)