Stanford University anthropologist Bridget Algee-Hewitt discusses craniometrics in the United States "melting pot" of ancestry, in two new papers in the American Journal of Physical Anthropology. (Photo: Kate Brimer, courtesy of Bridget Algee-Hewitt)

Craniometrics, the measurements of skull shapes, has been a critical anthropological tool. People with different ancestry look different, and the bones of the head are a good forensic lead, especially where skeletal remains leave little to no DNA.

But the increasingly complex racial mixtures in the United States have complicated what was once a simpler set of markers. So-called “admixture” of ancestry has been a major project of Stanford University anthropologist Bridget Algee-Hewitt.

Two recent papers by Algee-Hewitt in the American Journal of Physical Anthropology gauge the “subpopulations” created in the “melting pot” of America, both by geography and over time.

The latter study contends ancestral trends over generations can be calculated by looking at sets of minute measurements of skulls. The trends line up with much of what DNA has told about ancestry trends in America, Algee-Hewitt told Forensic Magazine.

“The admixture estimates are remarkably concordant between the two kind of data (cranial and DNA) and the different subsets of the U.S. population from which the participants or cases were drawn,” she said recently.

The 1,521 skulls from people with well-documented biographical data, including self-identified ancestral background, sex and birth years were assessed from the Forensic Anthropology Data Bank, housed at the University of Tennessee, Knoxville. Another 2,000 skulls helped determine a baseline—a “preexisting matrix of membership coefficients.”

A set of 20 measurements were rigorously scrutinized, first by ANOVA and regression, then by correlation analysis. The “unsupervised model-based clustering” eliminates the possibility of user bias and verifies the appropriate choice of samples, Algee-Hewitt said.

A few trends were identified, over the course of the 20th century (with people born between the years 1893 and 1991):

  • Those who identified as black in the samples had increasing admixture from white-European ancestry.
  • The “Hispanic” population had more and more Native American skull traits. The term “Hispanic” has proven to be a complex social term, Algee-Hewitt explained; for instance, those from the Dominican Republic or Puerto Rico or other parts of the Caribbean have vastly different lineage than people from Mexico or Colombia.
  • Less diversity was actually present in the white population over the century of births, she found.

“Population mixing is a fundamental process that shapes genomic variation and phenotypic expression, including the moderately to highly heritable cranial traits,” Algee-Hewitt told Forensic Magazine. “Admixture is obviously relevant to the study of U.S. populations, whose members derive their genome from mixtures of individuals that represent different ancestors, vary in their relative proportions of ancestries and display unique admixture signatures.”

For Algee-Hewitt, the most important direction for the future is to combine all biometrics, from “traditional” anthropological measurements, to the complications of DNA.

“To truly understand these patterns, it is critical that we start building databanks that contain multiple types of biological information for ancestry/admixture for the same individuals, like paired skeletal and DNA data, and complete information on geographic origins and birth years,” she said. “It is also critical that we continue to broaden the scope of our sample, to capture more and more of the finer-scale structure of the diversity within, and not just between, groups.”

The anthropologist’s wide-ranging work has been profiled over the last year in Forensic Magazine: her investigation into combining skull measurements and DNA, a new kind of “bone clock” for skeletal age estimations and her part in a major Stanford discovery of ancestral indicators and other information in the set of markers used in CODIS