Anthropology, University of Arkansas
Saturday All day, Plaza Level
Femoral head size is commonly used to estimate fossil body mass. Unfortunately, many femoral heads are damaged, precluding an accurate estimate of size. 3-D computer models offer the potential to estimate femoral head size in fragmentary fossils by fitting a sphere to a digital 3-D model of the preserved surface, and measuring the sphere diameter. Here we evaluate the accuracy of this method using data on extant apes and humans.
3-D surface scans for 123 specimens of Gorilla, Pan, Pongo, Hylobates, and Homo were gathered using a Konica Minolta Vivid 9i laser scanner, and analyzed using Polyworks software. Femoral head diameters of the scans were measured mimicking standard caliper measurements. To mimic fragmentary remains, femoral heads were divided into a series of 16 patches defined by a series of planes. Spheres were fit using whole, half, quarter, eighth and sixteenth sections of the head, with patches selected at random for repeated estimates for each specimen. Spheres were also fit using randomly selected pairs of patches.
Error rate for the smallest patches was 1.21%, and declined to less than 1% for whole surfaces. Spheres fit using pairs of patches also yielded lower error rates. Estimated femoral head diameters closely matched linear distances, but showed some bias, meaning that estimates of femoral head from spheres either should be corrected for bias empirically, or only compared to estimates based on fitted spheres. The method was illustrated using a series of South African hominins, including estimates for 5 fragmentary femoral heads.
Supported by NSF, The LSB Leakey Foundation, and Wenner Gren Foundation