Lactase persistence in populations that drink milk: a classic story of human evolution re-evaluated

July 29, 2022 • 9:15 am

The classic tale of “gene-culture coevolution” in humans—the notion that cultural changes in behavior changed the selection pressures that impinged on us—is the evolution of “lactase persistence” (LP) over the past four thousand years.  LP is a trait that allows you to consume, as an adult, lots of milk or dairy products without suffering the side effects of indigestion, flatulence, or diarrhea.

Young children are able to tolerate milk while nursing, of course, but after weaning many of them no longer tolerate milk—they are lactose intolerant (LI). The ability to digest lactose goes away after weaning because the gene producing the necessary enzyme gets turned off.

The gain of LP, which enables you to drink milk and eat dairy products into adulthood without ill effect, rests on single mutations in the control region of the gene producing lactase, an enzyme that breaks down the milk sugar lactose.  These mutations have arisen independently several times, but only after humans began “pastoral” activities: drinking milk from domesticated sheep, goats, and cows. And the mutations act to keep lactase turned on even after weaning. (Why humans turn off the gene after weaning isn’t known, but presumably involved the metabolic cost of producing an enzyme that wasn’t used in our ancestors, who didn’t drink milk after weaning until about about 10,000 years ago—when farming and animal domestication began.)

Based on analysis of fossil DNA, the LP mutations began spreading through Europe (starting from what is now Turkey) about 4000 years ago. And so the classic story—one that I taught my evolution classes—is that humans began drinking milk from captive herds, and that gave an advantage to retaining the ability to digest milk even after weaning. Ergo, natural selection for the nutritional benefits of milk led to the spread of LP mutations, as their carriers may have had better health (ergo more offspring) than individuals who turn off the enzyme at weaning).

This leads to the “coevolution” that is the classic evolutionary tale: a change in human behavior (raising animals for milk) led to selection for the persistence of the milk-digesting enzyme, and thus to genetic evolution. The “coevolution” part is the speculation that being able to digest milk without side effects would cause humans to raise even more dairy animals and drink even more milk, intensifying the selection for LP, and so the gene for LP would keep increasing in frequency.

A new paper in Nature, which is being touted all over social media, argues against this classic story, suggesting that it’s more complex than previously envisioned.  Although the new results are touted as overturning the earlier story, they really don’t. There is still human genetic evolution promoted by a change in culture, and there’s still a reproductive advantage in drinking milk.

The new part of the story is simply that that reproductive advantage comes not constantly (as previously envisioned), but only during times of famine and disease, when those who couldn’t digest lactose were at a severe disadvantage because the diarrhea caused by lactose intolerance would contribute to the death of diseased or malnourished individuals. This is a twist on the main story, but doesn’t overturn it completely. There’s still the connection between culture and human evolution, and there’s still a reproductive advantage to LP that leads to natural selection and genetic evolution of our species.  What’s different is how and when the selection acts (see “the upshot” at the bottom).

Click the title screenshot below to read, or you can download the pdf here. The full reference is at the bottom, and Nature deemed this worthy of two News and Views pieces in the same issue: (here and here).

First, the authors show the spread of dairy use in the figure below (the redder the color, the more milk usage over time in Eurasia. This was estimated from looking at the frequency of pot shards that had milk residue (click to enlarge). By 1500 BC, milk use was widespread.

Caption (from Nature): Interpolated time slices of the frequency of dairy fat residues in potsherds (colour hue) and confidence in the estimate (colour saturation) using two-dimensional kernel density estimation. Bandwidth and saturation parameters were optimized using cross-validation. Circles indicate the observed frequencies at site-phase locations. The broad southeast to northeast cline of colour saturation at the beginning of the Neolithic period illustrates a sampling bias towards earliest evidence of milk use. Substantial heterogeneity in milk exploitation is evident across mainland Europe. By contrast, the British Isles and western France maintain a gradual decline across 7,000 years after first evidence of milk about 5500 BC. Note that interpolation can colour some areas (particularly islands) for which no data are present.

One reason the authors doubt the classical story is that while dairying and milk-drinking by adults began about 10,000 years ago, the gene for LP (determined from sequencing “fossil DNA”) didn’t spread widely until about 4,000 years ago.  Why is that? The mutation for LP is dominant, which means it could have spread widely very quickly, as even carriers of one copy would have a reproductive advantage. This temporal disparity is what led the authors to propose their alternative hypotheses for the spread of the LP alleles (there are several).

Further, when the authors tried to correlate the frequencies of the LP allele with the frequency of milk use (the classical explanation), they found no correlation—that pattern was indistinguishable from a general rise in frequency over Europe regardless of milk use.

One other set of data led to the new hypothesis. That is the observation that LI people in both Britain and China can still drink lots of milk without suffering any measurable health or reproductive effects (milk drinking has recently proliferated in China).  Of course, things are different now from 4000 years ago, but one of the differences led to the authors’ two hypotheses: the spread of the LP allele was promoted especially strongly in prehistoric times by the prevalence of famine and of disease—with the latter coming often from animals, either domesticated or those that hang around settlements. (As the authors note: “about 61% of known and about 75% of emerging human infectious disease today come from animals”).

So the authors erected two hypotheses, the crisis mechanism and the chronic mechanism. I’ll let them describe the hypotheses that they tested (my emphases)

Given the widespread prehistoric exploitation of milk shown here and its relatively benign effects in healthy LNP individuals today, we propose two related mechanisms for the evolution of LP. First, as postulated in ref. 24, the detrimental health consequences of high-lactose food consumption by LNP individuals would be acutely manifested during famines, leading to high but episodic selection favouring LP. This is because lactose-induced diarrhoea can shift from an inconvenient to a fatal condition in severely malnourished individuals and high-lactose (unfermented) milk products are more likely to be consumed when other food sources have been exhausted. This we name the ‘crisis mechanism’, which predicts that LP selection pressures would have been greater during times of subsistence instability. A second mechanism relates to the increased pathogen loads—especially zoonoses—associated with farming and increased population density and mobility. Mortality and morbidity due to pathogen exposure would have been amplified by the otherwise minor health effects of LNP in individuals consuming milk—particularly diarrhoea—due to fluid loss and other gut disturbances, leading to enhanced selection for LP We name this the ‘chronic mechanism’, which predicts that LP selection pressures would have increased with greater pathogen exposure.

In other words, the reproductive advantage of having the LP allele came from the reproductive disadvantage (through death) of lactose-intolerant people during times of famine and disease.

They tested the two hypotheses by correlating indices of famine and of disease deduced from archeological and paleontological evidence:

Crisis mechanism: “Subsistence instability”, or famine, was assessed by prehistoric fluctuations in population size, which, the authors say, is correlated with the likelihood of famine (they provide no evidence for the latter supposition). But the correlation gives a significantly better fit to the pattern of LP allele frequency than just assuming uniform selection over time and space.

Chronic mechanism:  The authors hypothesized that the frequency of disease would correlate with the likelihood of “zoonoses” (diseases caught from animals), which itself would correlate with temporal variation in settlement densities.  These data, which to me would be correlated with “prehistoric fluctuations in population size” above, also explained LP allele frequencies better than an assumption of uniform selection.

Of course, there’s no reason (and the authors say this) that both mechanisms couldn’t operate together. Curiously, though, indices of the density of domestic animals did not support the “chronic mechanism” though measurements of the proportion of wild animals around humans did.  This implies that, if the “chronic mechanism” is correct, people were getting sick not from their horses, dogs, cattle, or sheep, but from wild animals (perhaps from eating them).

Other hypotheses that the authors mention but didn’t test include “drinking milk as a relatively pathogen-free fluid”, allowing “earlier weaning and thus increased fertility.” I would add that if diseases are causal here, they could come not from being around animals, but having drunk contaminated water, giving an advantage to those who prefer milk. But there’s no way of assessing that from the archaeological record.

The upshot: On the last page of the paper the authors say that they’ve debunked the prevailing narrative:

The prevailing narrative for the coevolution of dairying and LP has been a virtuous circle mechanism in which LP frequency increased through the nutritional benefits and avoidance of negative health costs of milk consumption, facilitating an increasing reliance on milk that further drove LP selection. Our findings suggest a different picture. Milk consumption did not gradually grow throughout the European Neolithic period from initially low levels but rather was widespread at the outset in an almost entirely LNP population. We show that the scale of prehistoric milk use does not help to explain European LP allele frequency trajectories and thus it also cannot account for selection intensities. Furthermore, we show that LP status has little impact on modern milk consumption, mortality or fecundity and milk consumption has little or no detrimental health impact on contemporary healthy LNP individuals.

Instead, they say that they find support for the increase of LP alleles through both famine or pathogen exposure.

Well, the data are the data, and their indices comport better with those data than does the classical hypothesis—the “prevailing narrative.” I’m still not convinced that their proxies for famine or disease are actually correlated with famine and disease themselves, but other researchers will undoubtedly dig into that.

What I want to emphasize is that if the work of Evershed et al. is accurate, it still does not overturn the story of gene-culture “coevolution”.  The “coevolution” is still there, the fact that a change in human culture influenced our evolution is still there, and the fact that drinking milk conferred higher reproductive fitness is still there. What has changed is only the nature of selection. Granted, that’s a significant expansion in understanding the story, but to listen to the media—social or otherwise—you’d think that the “classical narrative” is completely wrong. It isn’t. It’s still correct in the main, but the way selection acted may be different from what we used to think. The media love “evolution scenarios are wrong” tales, and that seems to be the cast of at least some stuff I’ve seen in the news and on social media.


Reference: Evershed, R.P., Davey Smith, G., Roffet-Salque, M. et al. 2022. Dairying, diseases and the evolution of lactase persistence in Europe. Nature.

28 thoughts on “Lactase persistence in populations that drink milk: a classic story of human evolution re-evaluated

  1. Oh and I forgot, the Big Question of Big Questions :

    What does lactase persistence in populations that drink milk tell us about God?

    1. I don’t know about God, but I’m sure the woke will somehow tell us that this justifies belief that we are all really trans people from Wakanda, or something like it!

  2. I have followed this story since the early 1970s. I cannot say I am very convinced by the Nature paper from what I have read so far. So far I see holes in the story. For example If people were making cheese – rich in everything but lactose rather than drinking raw milk, the whole thing comes crashing down. And calcium in milk – the big selective advantage in northern Europe – makes temporary disadvantages of lactose (it takes a lot to cause diarrhoea) look like a sideshow.

      1. Spontaneous thoughts as regards calcium – as assisted by searching NIH, etc. – some quotes :

        “In milk, casein undergoes phase separation to form colloidal casein micelles”

        “Calcium is a divalent cation, allowing it to form bonds with several caseinate anions.”

        Not sure what that tells us with regard to lactose and this phrnomenon, but wanted to review my notes and share.
        Also PubMed on casein micelle structure – appears to be a complex topic.

    1. Cheese is apparently a rather old affair, at least 7200 years old. Butter too, probably. Thus, yes, they are holes in this story.

      1. In the color-coded dairy use map, the territory of France is strikingly red in color by ~ 5000 bc. This no doubt tells us the evolutionary history of le plateau de fromage. Could a more fine-scaled investigation of residues on boards reveal a history of fromages régionaux?

    2. I agree with Malcom..cheese is a big issue here. I can’t see (and they don’t say) how their lipid analysis would distinguish fresh milk from any type of fermented milk product in those old pots. It’s maybe not explicit, but the text does seem to equate detection of milk fat residue with consumption of lactose-containing milk whereas it could have been any type of fermented/ripened milk product in the pottery. All these would have had less lactose than fresh milk due to microbial consumption – and some, as Malcom says, would have had almost no lactose (e.g Parmesan….mmmmm). So widespread ancient consumption of milk-products across diverse LI populations might actually have been mainly fermented products lacking harmful amounts of lactose. I wonder if the positive LP selection during the (possibly) more warlike bronze and iron ages was down to the ability to gulp fresh milk on-the-hoof..pun intended. Faster (and less flatulent) armies, in other words.

  3. I don’t buy the idea that there would be a reproductive advantage to LP only, or primarily, during times of scarcity. Milk and its by-products are a significant source of protein. It is much more useful to have a milk cow than to eat it. I would think that dairy animals would be an important source of protein, keeping in mind how uncommon meat was in most pre-modern diets.

  4. The original simple hypothesis of a time-dependent simple correlation between use of dairy cattle and LP is clearly falsified now (see also the low rates of LP in nomadic Asian steppe herders), and I agree with Jerry that this attempt to pinpoint intervening factors is interesting and might well be correct or partly correct, but is still pretty hypothetic.
    Science self correcting here, but it’s strange that the simple original hypothesis held on for so long, resting as it did on somewhat forced assumptions (people consuming milk instead of fermented dairy products, LNP causing more than just bloat and a little discomfort).

    I would add that some pleiotropic effect should not be excluded at this stage. Maybe LP was important for something altogether different, then secondarily high-LP populations in northwestern Europe took to drinking more unprocessed milk as adults.

  5. Haven’t read the paper but it seems to focus only on evolution of LP in Eurasia. Is this added crisis mechanism relevant to the evolution of LP in Africa?

  6. In times of famine the herds of cattle or other milk producing livestock would also be likely to be subject to dietary insufficiency and more likely to stop lactating (or at least struggle to produce as much milk). As you say the data are the data but this seems to me to be another potential drawback to their hypothesis.

  7. To complicate matters, I have some close anecdotal indications that lactose intolerance may be reversible. My first wife, from South-East Asia, was LI when she came to Europe as a post grad. She used milk whenever she was constipated, some cramps, but definitely no more constipation.
    After a few years it did not appear to work anymore, she could drink a liter of milk, but no cramps or nausea and no remedy.
    I suspect the lactase gene can be switched on again, in some cases at least.
    Her friend reported the same, by consistently drinking milk, the side effects, cramps, nausea and the (desired) diarrhea disappeared after some years.
    If this was due to switching a gene on again, I’d be very curious to know what the mechanism was.
    LI and LP is a question that never ceases to intrigue, from many angles.

    Footnote, although I’m lactose tolerant, I rarely drink just milk (I prefer my coffee black and hot – an endless source of jokes about my alleged sexual preferences), but cheeses from Gorgonzola, Roquefort to Parmesan and Reblochon, and so many others, are high on my list of great foods. But I guess the gastronomic merits of cheeses is an altogether different subject.

    1. I don’t know if that is really a complication, as there are several things to consider about how the lives back then. One being that most people did not live that long thousands of years ago. If a gene may reactivate its expression in an aging genome, that wouldn’t matter so much at a time when most people did not live beyond 30. Another is changes in intestinal flora, of which I know nothing, but maybe that too could effect how tolerant one is to lactose.

      1. The idea that many people did not live beyond 30 should be nuanced. Many died before 5, but having survived that, many lived into their 70’s. That 30 is an average, skewed because of high child mortality, and mortality of young adult males due to incessant tribal warfare .

      1. He he he!
        I said alleged. One could be cute and say to have no idea where these allegations come from, of course, which I do herewith.

  8. “But the situation is more complicated than that” could be said about nearly everything in biology. If I ever taught genetics again, every unit would end with “but the situation is more complicated than that.” Or if I taught anyting, really.

    1. However, as a great plus, that makes it even more fascinating.
      I wish I would be alive say 50 years from now (not impossible, but highly unlikely), just to see the new developments.

  9. There’s alot of self assertion about the “nutritional benefits” of animal products.. it’s as if this is being wrote by someone working for the dairy industry..

    And everyone in the comments is problematically ignoring that these dairy products aren’t healthy and promote disease, and a mutation in some people is not some indicative of progress, as well as other assumptions that are being backed up by circular reasoning and bias..

    This is just bad science.

Leave a Reply