Milk came first, the gene came late

Milk is a strange thing for an adult mammal to drink. Every mammal is born able to digest lactose and nearly every one switches the enzyme off after weaning, because there is no reason to keep paying for a tool you will never use again. Humans in much of the world are the exception, and for a long time the story of why seemed obvious: people started keeping animals, people started drinking milk, and natural selection did the rest. The evidence has since rearranged itself, and the new order of events is more interesting than the old one.

The residue in the pot

You cannot excavate milk. What you can do is analyse the fats absorbed into the clay walls of a pot thousands of years ago and tell dairy fat from carcass fat by its carbon isotopes. Applied to more than two thousand vessels from the Near East and southeastern Europe, that method pushed confirmed milk processing back into the seventh millennium BC, with northwestern Anatolia standing out. ¹

The same approach settled an older archaeological argument. Perforated ceramic fragments from the Kuyavia region of Poland, roughly seven and a half thousand years old, had long been guessed to be sieves. Lipid analysis found milk fat in the perforations, confirming they were cheese strainers, which means Neolithic farmers were already separating fat-rich curd from lactose-rich whey. ² That is worth dwelling on, because straining is itself a way of removing the part of milk that makes people ill.

One gene, several answers

The European version of lactase persistence is a single-letter change, the variant known as -13910*T, and it does not sit inside the lactase gene at all. It sits in an enhancer within an intron of the neighbouring MCM6 gene, acting as a switch that keeps LCT running into adulthood. It carries one of the strongest signals of recent positive selection anywhere in the human genome. ³

It is also not the only answer. Pastoralist populations in East Africa evolved their own variants, at least three of them, on entirely different haplotype backgrounds from the European one. ⁴ Different populations, facing the same problem, arrived independently at the same solution by different molecular routes. There is no single lactose tolerance gene. There is a recurring evolutionary answer to a recurring question.

The order of events, corrected

In 2022 a large collaboration put the two lines of evidence side by side: roughly seven thousand usable fat residues drawn from more than thirteen thousand pottery sherds across five hundred and fifty-four sites, matched against genotypes from over seventeen hundred ancient Eurasian genomes. The result upended the tidy version. Milk was in the European diet from close to the beginning of farming, yet the lactase persistence allele stayed rare for thousands of years and only became common around a thousand BC. ⁵

More pointedly, how much milk a region consumed did not predict how strongly the gene was selected for. What did predict it were markers of population density and pathogen exposure. The model that fits: people who could not digest lactose drank milk anyway, because it was there. In ordinary years that meant discomfort. In famine or epidemic years, when someone is already malnourished or already losing fluid to disease, lactose-driven diarrhoea stops being uncomfortable and starts being fatal. Selection did not run on the everyday advantage of milk. It ran on the worst years.

What fermentation is doing

Yogurt is what happens when two bacteria are set to work on the problem. Lactobacillus delbrueckii subspecies bulgaricus and Streptococcus thermophilus are used together because they feed each other: the lactobacillus breaks milk protein down into peptides and amino acids the streptococcus needs, and the streptococcus returns formate, pyruvate, and carbon dioxide that the lactobacillus grows on. Neither works as well alone.

Both convert lactose into lactic acid, and that acid does the physical work. Milk starts near pH 6.8. As it falls toward 4.6 it crosses the isoelectric point of casein, where the protein micelles lose the charge that keeps them apart and lock into the three-dimensional network we experience as set yogurt. The tang and the texture are the same event, seen from two directions.

The digestive part is neater still. Lactase-nonpersistent people generally tolerate yogurt better than milk, and the reason is that the bacteria bring their own enzyme. Their beta-galactosidase survives the stomach because it is packaged inside intact bacterial cells, then becomes active again in the less acidic small intestine, and yogurt's slower transit gives it longer to work. A controlled study comparing live yogurt against the heat-killed equivalent in lactase-deficient volunteers measured exactly this, with significantly less lactose reaching the end of the small intestine when the cultures were alive. ^{6 7}

This is one of the very few food claims the European Food Safety Authority has actually authorised: live yogurt cultures improve lactose digestion in people who have difficulty digesting it. The condition attached matters, though. The product has to carry at least a hundred million live starter organisms per gram at the moment you eat it. ⁸ Yogurt that has been heat-treated after fermentation, which is a common shelf-stability move, does not qualify. The bacteria are the mechanism, so killing them removes it.

The dried version

Every dairying culture eventually solves the same storage problem, and jameed is the Levantine and Bedouin answer: yogurt from sheep or goat milk, salted, strained, and dried into hard pale balls that keep for a very long time without refrigeration, then broken up and rehydrated into a sauce. The name comes from the Arabic root for hardening, which is about as direct as naming gets. It is a technology for turning a spring milk surplus into food you can carry into a season that has none.

One caution on the folklore. The story that Bulgarian peasants lived extraordinarily long lives on yogurt, which is where the whole modern health reputation of the stuff comes from, was built on birth records that could not support it. Life expectancy in Bulgaria around 1900 was somewhere near forty. The bacterium itself was first isolated by Stamen Grigorov in 1905; Ilya Metchnikoff, who popularised the longevity theory, did not discover it, and his Nobel Prize was for immunology and had nothing to do with yogurt. The chemistry above is solid. The legend attached to it is not.