Richard Lewontin: "[T]oo rapid for genetic adaptation"

I have had a New York Review of Books essay by Richard Lewontin, titled, "Why Darwin?" on my desktop for a week without getting to the last section of it.

Like many essays in the NY Review of Books, Lewontin's shoehorns small points from the books into an argument of his own. As you might guess from the title, Lewontin's theme is that Darwin has been overrated -- a result of biologists overemphasizing a "great man" story of the history of their science, and an unjustified belief in the ubiquity and power of natural selection. Lewontin mobilizes his argument against Jerry Coyne's Why Evolution Is True.

I don't really find the "pluralist versus adaptationist" debate very interesting. Despite the vocal complaints of some, I can't ever seem to locate the mythical "adaptationists" who deny that non-adaptive evolution ever happens. So the "debate" always comes down to whether particular adaptive hypotheses are true. Since no scientific hypothesis is true a priori, and since "those adaptationists are always saying stupid things" is not a scientific argument, I don't see the point.

Still, I meant to get to the last section of Lewontin's essay, and this morning I finally read it. To close his case for the weakness of natural selection, Lewontin turns to another new book by Greg Gibson, titled, It Takes a Genome: How a Clash Between Our Genes and Modern Life Is Making Us Sick. The book is an extended account of "diseases of civilization", a topic that I discussed here last week ("Arrested adaptation and the 'diseases of civilization'"). Here's a passage from the book's promotional material (on the Amazon page):

In It Takes a Genome, Greg Gibson posits a revolutionary new hypothesis: Our genome is out of equilibrium, both with itself and its environment. Simply put, our genes arent coping well with modern culture. Our bodies were never designed to subsist on fat and sugary foods; our immune systems werent designed for todays clean, bland environments; our minds werent designed to process hard-edged, artificial electronic inputs from dawn til midnight. And thats why so many of us suffer from chronic diseases that barely touched our ancestors.

Set aside for a moment how "revolutionary" this hypothesis is -- I'll revisit the idea in another post. The question is whether this mismatch between our environments and our genetic variation means that human evolution "stopped" or that we are still "adapted to the Pleistocene". As I pointed out in my earlier post, both propositions are true: human populations are mismatched with their current environments, and human populations have been recently adapting very rapidly to new environments. Here's what I wrote last week:

[M]any of today's chronic diseases reflect the reaction of human biology to novel environments for which our genes are not well adapted. But we don't need to exaggerate the slowness of human evolution to arrive at that conclusion. Recent rapid evolution of humans does not mean that humans are perfectly adapted to the present. Far from it -- if human populations have undergone rapid genetic changes into the past thousand years, it is a strong sign that fitness has not yet maximized in the post-agricultural environment.

I can contrast my point of view with Richard Lewontin's, who perfectly reiterates the "human evolution stopped in the Pleistocene" version of events.

An important property of adaptive evolution is that it is usually a slow process. Certainly there are cases where a single genetic change can mean the difference between life and death in a hostile environment. The classic cases are the mutations that give pathogenic microorganisms the ability to resist antibiotics or mutations that allow crops to resist pathogens, for example insects or herbicides. But these are not representative models for how species adapt, by accumulation of mutations of small effect, to changes in food availability, temperature modifications, and the thousand shocks that flesh is heir to. The usual small differences in fitness among genotypes are therefore manifest as detectable evolutionary change only after thousands of generations.
This deliberate tempo has presented the human species with a problem of adaptation. With a human generation of about twenty-five years, there have been roughly only one hundred generations since the founding of the Roman Republic. Yet the changes in the human environment caused by changes in human activity have been enormous. Changes in diet, habitation, working conditions, the pollution of air and water, and especially the considerable increase of lifespan that result in major alterations and breakdowns in the bodily machinery have all been too rapid for genetic adaptation.

Notice the false premises: Adaptive evolution is "usually a slow process." Species adapt by "accumulation of mutations of small effect." It's as if he were transported back in time to 1908 where no one had heard of the breeder's equation.

There's nothing impossible about long series of small changes. But they are not the only mode of adaptation, or even the most likely one. Populations with additive genetic variation that correlates with fitness will change rapidly under selection. The structure of the additive variation may lead to strong selection on one gene of large effect, or selection in parallel across many genes of varying effects. Series of small changes may be required for some adaptations, but a rapid environmental change (as Lewontin observes for humans) may cause bursts of rapid changes in allele frequencies.

To maintain the slowness of human evolution, Lewontin must do three things:

  1. Assume humans are genetically uniform.

  2. Where humans obviously are not uniform, argue that variations are uncorrelated with fitness.

  3. Ignore any historical or genetic evidence that might contradict 1 and 2.

Keeping in mind the short length of this section of the essay, Lewontin does manage all three of these conditions.

I think it's downright sneaky the way Lewontin reinforces the assumption of human genetic uniformity. He refers to "the human genotype" as if there were only one! By emphasizing that "parts of the human genome are out of correspondence with modern life", he precludes the possibility that some human genomes may be more in correspondence than others. Sure, if humans share a single genome, they can't possibly differ in any adaptive way.

But diversity is the reality. Examples of recent human evolution are fixtures in biology textbooks, from sickle-cell to lactase persistence. These are traits that have rapidly changed in frequency during the last 2500 years, due to changes in recent human environments -- disease for the former, diet for the latter. These rapid transformations in precisely those that Lewontin says are impossible -- environmental changes being "too rapid for genetic adaptation." A number of morphological changes are also evident when comparing archaeological and recent skeletal samples in many parts of the world. Somehow the relevance of these recent changes goes unmentioned in the essay.

One of the best-characterized examples of evolution in recent populations is the rapid Holocene evolution of pigmentation phenotypes. It's a textbook example of human variation, and several adaptive hypotheses may explain it. So pigmentation would seem an unlikely example of how human evolution has been too slow to cope with the environment. But Lewontin finds a way:

[H]igh doses of solar radiation that is experienced by surfers on the California beaches might induce an eventually fatal skin cancer, but the cancer death almost always occurs well after reproductive age, so there is no opportunity for selection to act.

I agree that current patterns of cancer mortality of light-skinned surfers may have little impact on their fitness. In other words, this chronic disease is a sign of an environmental "mismatch" that future genetic evolution is unlikely to erase.

But why turn to false arguments about the speed of evolution to make this point? Surely Lewontin knows that "reproductive age" in humans is not synchronous with reproductive effort? Skin cancer is one of the earliest-killing cancers, with a good fraction of victims dying at ages when they might otherwise be helping raise their kids or grandchidlren. Lewontin must also know that human populations vary greatly in their skin cancer susceptibility, and that some surfers (the dark pigmented ones) have lower skin cancer rates after the same sun exposure. Skin cancer may or may not be the best explanation for dark pigmentation in low-latitude human populations (there are others, none mutually exclusive), but this example works strongly against Lewontin's claims that natural selection is "slow" and that human environmental changes have been "too rapid for genetic adaptation." We aren't perfectly adapted today, and the rate of our evolution in the recent past was very fast.


Lewontin RC. 2009. Why Darwin? New York Review of Books 56(9) May 28, 2009. Online