Hominin History / Articles / Lactose tolerance
DNA & Genetics

Why Can Some Adults Drink Milk?

Digesting milk after childhood is an evolutionary oddity. It evolved at least four separate times, spread faster than almost any other human trait, and still has not been fully explained.

The short answer

Most mammals, and about two-thirds of adult humans, lose the ability to digest lactose after weaning. Lactase persistence is caused by a handful of genetic switches near the lactase gene that evolved independently in Europe, East Africa and the Middle East. Ancient DNA shows the European variant was rare until about 3,000–4,000 years ago — millennia after dairying began — suggesting famine and disease, not everyday milk drinking, drove its spread.

Every baby mammal drinks milk, and every baby mammal makes lactase, the enzyme that splits milk sugar (lactose) into glucose and galactose so it can be absorbed. Then, after weaning, the gene that makes lactase is switched down. For a wolf, a horse or a chimpanzee, there is no point in producing an enzyme for a food it will never taste again. (Lactase is one of the molecules you can spin in our 3D protein viewer.)

Most humans follow the same mammalian rule. Roughly two-thirds of adults on Earth are lactase non-persistent: drink a large glass of milk and undigested lactose ferments in the gut, causing bloating, cramps and diarrhoea. The oddity that needs explaining is the other third — people whose lactase stays on for life.

VariantWhere it is commonApprox. ageNotes
−13910*TNorthern & western Europe; also South AsiaArose ~5,000–10,000 years agoRare until the Bronze Age
−14010*CEast Africa (Kenya, Tanzania)~3,000–7,000 years agoAssociated with pastoralists such as the Maasai
−13915*GArabian Peninsula, Middle EastSeveral thousand yearsLinked to camel herding
−13907*GEthiopia, SudanSeveral thousand yearsOne of several African variants

A switch, not a new gene

The mutations that make lactase persistent do not change the lactase gene (LCT) at all. They sit about 14,000 DNA letters upstream, inside a neighbouring gene called MCM6, in a region that acts as an enhancer — a dimmer switch for LCT. A single-letter change there keeps the switch from turning down after childhood.

The European version, a C-to-T change known as −13910*T, is carried by the large majority of people in Scandinavia, Britain and Ireland, but only a small minority in southern Italy or Greece, and is almost absent in East Asia. Because one copy is enough, the trait is dominant.

It evolved at least four times

In 2007 Sarah Tishkoff’s team studied lactase persistence in East African herding peoples and found that most did not carry the European mutation. They carried different ones — including −14010*C — in the same enhancer region, all producing the same result. Further variants turned up in Arabia and Ethiopia.

This is convergent evolution: separate human populations with a heavy reliance on livestock independently evolved the same trait, by different mutations, in the same small piece of DNA. It is one of the clearest examples of recent natural selection anywhere in the human genome.

The speed is extraordinary

Geneticists measure the strength of selection by how quickly a variant spreads. The signal around the European lactase variant is among the strongest ever measured in humans: a long, unbroken block of DNA surrounding it shows it rose to high frequency too quickly for recombination to break it up. Some estimates suggest carriers had several percent more surviving children per generation than non-carriers — enormous by evolutionary standards.

Ancient DNA scrambled the story

The obvious explanation used to be the “culture-historical” one: farmers began milking cattle, sheep and goats around 10,000–8,000 years ago in the Near East, milk became a staple, and people who could digest it thrived. Gene and culture evolved together.

The ancient DNA revolution broke the neat timing. Early Neolithic farmers in Europe — the very people who brought dairying — almost never carried −13910*T. Nor did most Copper Age individuals. Ötzi the Iceman, who lived about 5,300 years ago, was lactose intolerant. The variant only becomes common in European genomes during the Bronze Age, and in many regions only within the last 3,000 years or so.

Meanwhile, chemists were finding milk fats in the pottery. Richard Evershed’s group has detected dairy residues on pots from northwest Anatolia about 9,000 years old, and cheese-making strainers from Poland about 7,000 years old. People were using milk for thousands of years before most of them could digest it.

How did they manage? Cheese and yoghurt

Fermentation is the answer. Bacteria in yoghurt, kefir and aged cheese consume much of the lactose. A hard cheese contains only a trace of the lactose in the milk it was made from. Early farmers could exploit dairy without the enzyme — exactly as many Mongolian herders still do today. They consume large quantities of dairy, mostly fermented, yet the large majority lack lactase persistence, and ancient proteins show their ancestors were already dairying in the Bronze Age.

So why did lactase persistence spread?

In 2022 Evershed, Mark Thomas and a large team published a study in Nature combining nearly 7,000 pottery residues from more than 550 European sites with ancient DNA from over 1,700 individuals. Milk use was widespread for 9,000 years and did not track the rise of the gene closely. What did track it were proxies for famine (population swings) and disease (larger, denser settlements).

Their argument: for a healthy, well-fed person, lactose intolerance is an inconvenience. For a malnourished person during a crop failure, forced to drink fresh milk, diarrhoea could be lethal — especially in a crowded settlement with gut infections circulating. Selection acted hardest in bad years. Supporting this, the team found that in the UK Biobank, milk drinking among non-persistent adults today has little measurable effect on health.

That is not the only hypothesis. Others emphasise calcium and vitamin D absorption at northern latitudes, where sunlight is weak, or the simple calorie value of milk to pastoralists. The different explanations may each have mattered in different places.

Why it matters

Lactose tolerance shows that human evolution did not stop with the Stone Age. A trait now familiar to billions of people became common within the last few thousand years, reshaped by culture, climate and catastrophe. It is also a reminder that the “normal” human condition, globally, is to be unable to digest milk as an adult — the milk drinkers are the mutants.

Explore the ancient genomes behind studies like this in the AADR Genome Atlas.

Open the Genome Atlas →

Frequently asked questions

Why are most adults lactose intolerant?

Like other mammals, humans normally switch down the lactase gene after weaning. About two-thirds of adults worldwide lack lactase persistence, so this is the ancestral, typical condition.

What gene causes lactose tolerance?

Lactase persistence is caused by variants in an enhancer inside the MCM6 gene that keep the neighbouring lactase gene, LCT, active into adulthood. In Europeans the main variant is called -13910*T.

When did lactose tolerance evolve?

The European variant arose within roughly the last 10,000 years but stayed rare until the Bronze Age. Ancient DNA shows it became common in most of Europe only in the last 3,000 to 4,000 years.

Did lactose tolerance evolve more than once?

Yes. Different mutations with the same effect arose independently in Europe, East Africa, Ethiopia and the Arabian Peninsula, a classic example of convergent evolution.

How did early farmers use milk without the gene?

Mainly by fermenting it. Cheese, yoghurt and other fermented dairy products contain far less lactose, which bacteria have already broken down.

Sources & further reading
  1. Evershed, R. P. et al. (2022). 'Dairying, diseases and the evolution of lactase persistence in Europe.' Nature 608. nature.com
  2. Tishkoff, S. A. et al. (2007). 'Convergent adaptation of human lactase persistence in Africa and Europe.' Nature Genetics 39.
  3. Salque, M. et al. (2013). 'Earliest evidence for cheese making in the sixth millennium BC in northern Europe.' Nature 493.
  4. Jeong, C. et al. (2018). 'Bronze Age population dynamics and the rise of dairy pastoralism on the eastern Eurasian steppe.' PNAS 115.