The taxonomic implications of cranial shape variation in Homo erectus

The taxonomic status of Homo erectus sensu lato has been a source of debate since the early 1980s, when a series of publications suggested that the early African fossils may represent a separate species, H. ergaster. To gain further resolution regarding this debate, 3D geometric morphometric data were used to quantify overall shape variation in the cranial vault within H. erectus using a new metric, the sum of squared pairwise Procrustes distances (SSD). Bootstrapping methods were used to compare the H. erectus SSD to a broad range of human and nonhuman primate samples in order to ascertain whether variation in H. erectus most clearly resembles that seen in one or more species. The reference taxa included relevant phylogenetic, ecological, and temporal analogs including humans, apes, and both extant and extinct papionin monkeys. The mean cranial shapes of different temporogeographic subsets of H. erectus fossils were then tested for significance using exact randomization tests and compared to the distances between regional groups of modern humans and subspecies/species of the ape and papionin monkey taxa. To gauge the influence of sexual dimorphism on levels of variation, comparisons were also made between the mean cranial shapes of single-sex samples for the reference taxa. Results indicate that variation in H. erectus is most comparable to single species of papionin monkeys and the genus Pan, which included two species. However, H. erectus encompasses a limited range of variation given its extensive geographic and temporal range, leading to the conclusion that only one species should be recognized. In addition, there are significant differences between the African/Georgian and Asian H. erectus samples, but not between H. ergaster (Georgia + Africa, excluding OH 9 and Daka) and H. erectus sensu stricto. This finding is in line with expectations for intraspecific variation in a long-lived species with a wide, but probably discontinuous, geographic distribution.