Faces Across the World

Chapter 5

Faces Across the World

Clines, not types

Human faces and skulls vary across the globe — that part is not in dispute. Nose width, jaw projection, cheekbone height, and dozens of other cranial and facial measurements differ, on average, between people whose ancestors lived in different regions. What the evidence does not support is the older assumption, built into two centuries of craniometry (Chapters 2 and 3), that this variation sorts cleanly into a small number of discrete "racial" types with hard edges between them. Population geneticists and biological anthropologists now describe most human physical variation as clinal: it changes gradually across geography, tracking climate, ancestry, and the distance a trait has to travel through generations of migration and mixture, rather than jumping between fixed categories. A cline has no natural boundary — any line drawn across it, separating "type A" from "type B," is a human classification choice imposed on a continuous gradient, not a division the biology itself provides. This is the throughline of this chapter: variation is real and geographically patterned; discrete facial "races" are not.

The framing has a name and an origin. In 1962 the physical anthropologist Frank Livingstone distilled it to a slogan — "there are no races, only clines" — arguing that human variation grades continuously across geography with no natural cut points, which is not the same as claiming populations don't differ.¹ It was a position in an argument the field was actively having, not a settled consensus every anthropologist signed onto — but it is the paper that gave "clines, not types" its name.

Howells and the multivariate turn

One of the most influential attempts to measure global cranial variation rigorously came from the American physical anthropologist William W. Howells. Over decades of fieldwork, Howells assembled a dataset of standardized cranial measurements from thousands of skulls drawn from populations across every inhabited continent, then analyzed the set using multivariate statistics capable of handling many measurements at once rather than the single ratios earlier craniometrists had relied on. His 1973 monograph presented the results as a study of patterns of difference among population samples — differences of degree along multiple, overlapping measurement axes — rather than as evidence for a small number of discrete racial types.² Howells's own populations could be statistically distinguished from one another on average, which is unsurprising given real geographic variation, but the overlap between neighboring groups on any individual measurement was substantial, and the boundaries between his statistical clusters did not correspond to any small, universal set of racial categories. The dataset remains a standard reference collection in forensic and biological anthropology today, used less to sort people into races than to model how cranial form varies continuously across geography.

Climate and the nose: a real, partial adaptation

Some of the clearest geographic patterning in the face concerns the nose. In 1923, the anatomist Arthur Thomson and the anthropologist L. H. Dudley Buxton compared the nasal index — the ratio of nasal width to nasal height — across populations and climates, and reported a statistical association: narrower, relatively taller noses tended to occur in colder, drier climates, while broader, shorter noses tended to occur in warmer, more humid ones.³ The proposed mechanism was functional: a longer, narrower nasal passage gives incoming air more time and surface area to be warmed and humidified before it reaches the lungs, which is more valuable in cold, dry air than in air that is already warm and moist. A century of follow-up research has refined and complicated the original picture considerably — later studies point to humidity, not temperature alone, as the more consistent predictor, and the association, while real at the population level, is a statistical tendency with wide individual variation and exceptions, not a rule that predicts any one person's nose from their ancestry. It is a genuine example of climate-linked adaptation, and also a caution against overreading a real average difference between populations as a strict marker of type.

Later work did more than qualify that pattern; it re-measured it. A 2011 study measured three-dimensional nasal-cavity landmarks across ten world populations and found the climate signal split by structure — the bony nasal cavity tracking temperature, the nasopharynx tracking humidity — with overall shape falling along a genuine gradient from cold-dry to hot-humid climates.⁴ The 1923 ratio was refined, not overturned.

Forensic ancestry estimation: real practice, contested ground

Craniofacial variation is not only a topic for historical or evolutionary study — it is actively used today. Forensic anthropologists routinely estimate the population ancestry of unidentified skeletal remains from cranial and facial measurements, as one input alongside age, sex, and stature estimation, to help investigators narrow a missing-persons search. Statistical classification methods built on large modern reference samples can predict a skeleton's population ancestry at rates well above chance, and practitioners in the field argue this reflects real, clinally-patterned biological variation rather than a revival of the fixed racial typology this tour has spent several chapters dismantling.⁵ But the method carries real limits and real stakes. Accuracy varies substantially by sample, region, and the reference population a given skeleton is compared against; it degrades for individuals of mixed or underrepresented ancestry; and the social categories investigators need for a search — categories shaped by history and self-identification, not biology — do not map cleanly onto the biological clines these methods actually measure. Misclassification can send an investigation in the wrong direction entirely, and the practice sits uneasily alongside this chapter's central point: it works, when it works, precisely because population ancestry leaves a real statistical signature in the skeleton, not because that signature sorts humanity into the discrete "races" nineteenth-century craniometry invented and this tour has spent several chapters unwinding.

The most consequential case of that gap is Kennewick Man. In 1996 the forensic anthropologist James Chatters examined a roughly 9,000-year-old skeleton from the Columbia River and judged the skull "Caucasoid" — a craniometric classification a later study pushed toward Ainu and Polynesian affinities, fueling a two-decade legal fight over repatriation. A 2015 genome sequence overturned the reading: Kennewick Man was in fact more closely related to modern Native Americans, including one of the tribes that had claimed him, than to any other population.⁶

What the variation actually shows

Put together, Howells's global samples, the nasal-index climate pattern, and forensic ancestry estimation all point to the same underlying picture. Human facial and cranial variation is real, it is substantial, and much of it tracks geography, ancestry, and climate in statistically detectable ways. None of that requires — or supports — sorting that variation into a small number of fixed, bounded "racial" types. The face varies the way most biological traits vary across a widely dispersed species: continuously, along overlapping gradients, shaped by where people's ancestors lived and how populations moved and mixed over tens of thousands of years. That is a more complicated picture than the typological one it replaced, but it is the one the measurements actually support.

Further reading