Why Humans Look So Different: Skin, Altitude, and Milk

In her office at Penn State in the early 2000s, the anthropologist Nina Jablonski was doing something deceptively simple. She took measurements of skin reflectance, recorded from indigenous populations all over the world, and plotted them against satellite maps of how much ultraviolet light actually strikes the ground at each location. What emerged was not a messy cloud of dots but a clean, sloping line. The intensity of sunlight at a place predicts the skin color of the people who have lived there for many generations with a tightness that is rare in human biology. As Jablonski herself has put it, skin color is one of the strongest examples we have of natural selection acting on the visible human body.

That single scatter plot does a quiet, radical thing. It takes one of the traits most loaded with social meaning, the trait that nineteenth-century race science treated as the master key to human difference, and explains it as a thermostat responding to a beam of light. This article asks the obvious follow-up question. If race is not a meaningful biological category, and a century of population genetics says firmly that it is not, then what does real human variation actually look like? The answer comes through three small, concrete case studies: the color of skin, the lungs of mountain dwellers, and the ability to drink a glass of milk.

Variation that comes in gradients, not boxes

The starting point is a distinction that sounds abstract but turns out to be the whole game. Human biological variation is real, abundant, and well documented. What it is not is categorical. The variation is clinal, meaning it changes gradually across geography rather than jumping in discrete steps at continental borders. Skin color does not flip from one value to another when you cross a coastline; it shades almost imperceptibly as you walk from the equator toward the poles. The same is true of nearly every trait once you look closely.

Crucially, the variation is also trait by trait. Skin color follows one gradient, body proportions follow another, blood-group frequencies follow a third, and these gradients do not line up. A line drawn to separate populations by skin pigmentation will cut through completely different groups than a line drawn by, say, resistance to malaria or the shape of the nasal passage. This is why knowing where someone falls on one cline tells you almost nothing about where they fall on another. The boxes that older race science tried to draw around humanity assumed that traits travel together in bundles. They do not. They scatter, and they scatter independently.

Long before satellites could measure ultraviolet exposure directly, researchers were already mapping this gradient by hand. In 1969 a now-famous scatter plot related indigenous skin reflectance to latitude across the globe, and the relationship was visible even with the crude tools of the time. Jablonski and her colleague George Chaplin refined that picture dramatically in 2000, assembling reflectance data from 191 populations and matching it against NASA satellite measurements of ground-level UV. The hand-drawn intuition of the 1960s became a quantitative model.

Skin color as a dial between two dangers

So why does sunlight set skin color at all? Jablonski and Chaplin's model treats the pigment in our skin as a tunable compromise between two competing pressures, both of which involve sunlight and both of which can hurt you. The pigment doing the tuning is melanin, the dark molecule that absorbs and scatters ultraviolet radiation before it can reach deeper tissues.

The first pressure pushes toward darker skin. Ultraviolet light, in high doses, destroys folate, a B vitamin essential for making new cells and for healthy fetal development. Populations living under intense equatorial sun face strong selection to protect their folate stores, and heavy melanin does exactly that, acting as a built-in sunscreen. The second pressure pushes the other way. The body manufactures vitamin D using ultraviolet light as the trigger, and vitamin D is required to absorb calcium and build bone. Under the weak, slanting sun of high latitudes, too much melanin would block the little UV available and starve the body of vitamin D, leading to deficiency. The result is a balancing act. Near the equator, the folate threat dominates and selection favors dark skin; far from it, the vitamin D threat dominates and selection favors lighter skin that lets scarce UV through. Melanin is simply the dial that finds the working point for a given dose of sunlight. The cline you see on a map of human skin color is the visible record of that compromise, repeated across every latitude.

Three populations, three answers to thin air

If skin color shows how a single environmental pressure shapes a single trait, high-altitude adaptation shows something even more striking: that evolution can solve the same problem in genuinely different ways. Living permanently above about 3,500 meters is physiologically brutal, because the thin air delivers far less oxygen with every breath. Three groups of humans have settled such heights for thousands of years, the Tibetans of the Himalayan plateau, the Andean peoples of the South American highlands, and the highland populations of Ethiopia, and when researchers examined how their bodies cope, they did not find one shared solution. They found three.

Tibetans manage low oxygen without dramatically thickening their blood, partly through a variant of a gene called EPAS1 that regulates the body's response to low oxygen. The Andean pattern looks different, leaning more on changes in oxygen-carrying capacity. The Ethiopian highlanders show yet another physiological profile, and the genetic signals there point to still other genes. Three populations, three independent routes to the same end. The Tibetan story carries a remarkable twist of its own: the beneficial EPAS1 allele was not invented from scratch in modern humans but inherited through interbreeding with the Denisovans, an archaic human group, and then favored by selection once Tibetans moved up onto the plateau. A piece of someone else's genome, tens of thousands of years old, became the key to breathing in the highest inhabited places on Earth.

Drinking milk, evolved more than once

The third case study is the one you can test at your own breakfast table. Most adult mammals, and most adult humans across history, stop producing the enzyme that digests lactose, the sugar in milk, after they are weaned. The ability to keep producing it into adulthood, called lactase persistence, is the exception, and it is recent. It evolved at least three separate times within roughly the last seven thousand years, each time alongside a culture that had begun keeping dairy animals.

Northern European farmers carry one regulatory variant near the lactase gene. East African Nilotic pastoralists, whose lives revolve around cattle, carry a different one. Arabian camel-herders carry yet another. Three populations, three distinct genetic changes, all sitting near the same gene and all producing the same outcome, the capacity to digest fresh milk as an adult. This is convergent evolution caught in the act, and it is also a vivid example of culture steering biology. The genetic change did not create the habit of drinking milk; the habit of keeping dairy animals created the selective pressure that favored the genetic change. Where dairying spread, milk-tolerance followed, and it followed by different molecular paths in different places.

Why none of this brings race back

It is tempting to look at clean gradients and convergent solutions and conclude that they vindicate the old categories after all, that here at last is the biology underneath race. They do not, and the three case studies are precisely why. Notice the pattern they share. Each adaptation is local, shaped by a specific environment. Each is recent, having arisen within the last few thousand to few tens of thousands of years. Each evolved more than once, by different routes in different places. And not one of them aligns with the borders of any racial category.

The traits do not co-vary. Dark skin does not predict lactase persistence; high-altitude adaptation does not track skin color; the cline for one trait crosses the cline for another at every angle. A person's value on the skin-pigmentation gradient gives you essentially no information about where they fall on the lactose gradient or the altitude gradient. The gradients run in different directions and bundle different populations together depending on which trait you choose. This is the heart of the matter. Real biological variation, far from rescuing the racial boxes, dissolves them, because the variation refuses to sort the same people the same way twice.

It helps to remember how new all of this is. Almost all of the population-level variation that nineteenth-century race science tried to systematize arose within the last fifty thousand years, after modern humans dispersed out of Africa and encountered new climates, new altitudes, and new diets. On the scale of human evolution, which stretches back hundreds of thousands of years, these differences are brand new, a thin and recent overlay on a deeply shared inheritance.

Bodies shaped after birth, and ancestry that is always mixed

Two final pieces complete the picture, and both push against any notion of fixed, pure types. The first is that genes are not the only thing shaping a body. The human body responds to its developmental environment, especially early in life. The work of David Barker in the 1980s showed that nutrition in the womb and in infancy influences the risk of adult diseases like heart disease and diabetes, and we now understand part of the mechanism through epigenetics, chemical marks that adjust how genes are read and that can carry an environmental signal across cell generations. Two people with similar genomes can develop measurably different bodies depending on the conditions of their early growth. The body is plastic, not stamped from a fixed mold.

The second is that when geneticists actually read whole genomes, the supposedly pure populations of older race science turn out to be thoroughly mixed. Genome-wide ancestry methods routinely reveal admixture where the old categories predicted none. In 2017, work led by Pontus Skoglund documented a component of archaic ancestry in West African populations, a contribution from some deeply divergent human lineage that no continental box anticipated. Everyone, examined closely enough, is a blend. There is no genomic line you can draw that cleanly partitions humanity into the categories of the nineteenth century, because the data underneath those categories is gradient, recent, convergent, plastic, and admixed all the way down.

Key Takeaways

Human biological variation is real and well documented, but it looks nothing like race: it is clinal rather than categorical, organized trait by trait into gradients that do not line up, and almost all of it arose within the last fifty thousand years. Skin color is a melanin dial balancing two sunlight-driven dangers, the destruction of folate under high UV and the loss of vitamin D under low UV, producing a clean latitude gradient. High-altitude adaptation in Tibetan, Andean, and Ethiopian populations shows the same problem solved three independent ways, with the Tibetan EPAS1 variant inherited from Denisovans. Lactase persistence evolved at least three separate times in the last seven thousand years among dairying peoples, a case of culture driving biology. None of these traits co-vary, none align with continental borders, and your position on one cline tells you almost nothing about another. Layer on developmental plasticity, where early-life environment shapes adult health through epigenetic mechanisms, and genome-wide ancestry showing that every population is admixed, and the conclusion is firm: real human variation does not restore the racial categories, it dissolves them.