Let’s for a moment imagine your species is wildly successful. Through the slow evolutionary process it’s become adapted with high precision to its environmental niche. You and all your fellows are now, perhaps even literally, fat and sassy. But, again, especially when you’re so well adapted, any significant genetic change tends not to be in your best interest—just as a random change in some of the microscopic magnetic domains on an audio tape is unlikely to improve the music recorded there. You can’t stop deleterious mutations from happening, just as you can’t prevent a slow degradation of the recorded music, but those mutations are restrained from spreading through the species. Natural selection sifts through the population and quickly disposes of whatever doesn’t work, or doesn’t work as well. It is not considered an extenuating or mitigating circumstance that, by some remote accident, the mutation might be useful in the future. Darwinian selection is for the here and now. Summary judgment is rendered. With careful discrimination, the scythe of selection swings.

But now, let’s imagine that something changes. A small world hurtling through space finds a blue planet smack in its path, and the resulting explosion sprays enough fine particles into the upper atmosphere to darken and cool the Earth; your lake then freezes over, or the savanna vegetation that sustains you shrivels and dies. Or the tectonic engine in the Earth’s interior creates a new island arc and a flurry of volcanic explosions changes the composition of the air, so now more greenhouse gases are released into the atmosphere, the climate warms, and the tidepools and shallow lakes in which you have been luxuriously wallowing begin to dry up—or a dam of glacial ice is breached, creating an inland sea where your congenial desert habitat used to be.

Perhaps the change comes from a biological direction: The animals you eat are now better camouflaged, or defend themselves with greater obstinacy; or animals that eat you have become more adept at the hunt; or your resistance to a new strain of microorganism turns out to be poor; or some plant you habitually eat has evolved a toxin that makes you ill. There can be a cascade of changes—a relatively small physical alteration leading to adaptations and extinctions in a few directly affected species, and further biological changes propagating up and down the food chain.

Now that your world has changed, your once wildly successful species may be reduced to much more marginal circumstances. Now some rare mutation or an improbable combination of existing genes might be much more adaptive. The once-spurned hereditary information may now be given a hero’s welcome, and we are reminded once more of the value of mutation and sex. Or, it may be, no new and more useful genetic information is generated fortuitously in the nick of time, and your species continues its downward drift.

Omnicompetent organisms do not exist. Breathing oxygen lets you be far more efficient in extracting energy from food; but oxygen is a poison for organic molecules, so arrangements for routine handling of oxygen by organic molecules are going to be expensive. The ptarmigan’s white feathers provide superb camouflage in the Arctic snows; but in consequence it absorbs less sunlight and greater demands are placed on its thermoregulatory system. The peacock’s gorgeous tail makes him nearly irresistible to the opposite sex, but also provides a conspicuous luncheon advertisement for foxes. The sickle-cell trait confers immunity to malaria, but condemns many to debilitating anemia. Every adaptation is a trade-off.

Imagine designing a vehicle that drives off roads, flies through the air, and swims underwater. Such a machine, if it could be built at all, would perform none of its functions well. When we need to travel on “unimproved” land we build all-terrain vehicles, when beneath the water, submarines, and when through the air, airplanes. It’s for good reason that these three kinds of vehicles, while roughly of similar shape, in fact tend not to look very much alike. Even so-called “flying boats” are not very seaworthy, nor are they very easy to fly.

Birds that are superb underwater swimmers, such as penguins, or highly capable runners, such as ostriches, tend to lose their ability to fly. The engineering specifications for swimming or running conflict with those for flying. Most species, faced with such alternatives, are forced by selection into one adaptation or the other. Beings that hold all their options open tend to be eased off the world stage. Overgeneralization is an evolutionary mistake.

But organisms that are too narrowly specialized, that perform exceedingly well but only in a single, restrictive environmental niche, also tend to become extinct; they are in danger of making a Faustian bargain, trading their long-term survival for the blandishments of a brilliant but brief career. What happens to them when the environment changes? Like barrelmakers in a world of steel containers, blacksmiths and buggy-whip tycoons in the time of the motorcar, or manufacturers of slide rules in the age of pocket calculators, highly specialized professionals can become obsolete virtually overnight.

If you’re receiving a forward pass in American football, you must keep your eye on the ball. At the same time you must keep your eye on the opposition tacklers. Catching the ball is your short-term objective; running with it after you have it is your longer-term objective. If you worry only about how to outrun the defenders, you may neglect to catch the ball. If you concentrate only on the reception, you may be flattened the moment you receive the ball, and risk fumbling it anyway. Some compromise between short-term and longer-term objectives is called for. The optimum mix will depend on the score, the down, the time remaining, and the ability of the opposing tacklers. For any given circumstance there is at least one optimum mix. As a professional player you would never imagine that your job as a receiver is solely catching passes or solely running with the ball. You will have acquired a habit of quickly estimating the risks and the potential benefits, and the balance between short-term and long-term goals.

Every competition requires such judgments; indeed, they constitute a large part of the excitement of sport. Such judgments must also be made daily in everyday life. And they’re a central and somewhat controversial issue in evolution.

The danger of overspecialization is that when the environment changes, you’re stranded. If you’re superbly adapted to your present habitat, you may be no good in the long term. Alternatively, if you spend all your time preparing for future contingencies—many of them remote—you may be no good in the short term. Nature has posed life a dilemma: to strike the optimum balance between the short-term and the long, to find some middle road between overspecialization and overgeneralization. The problem is compounded, of course, by the fact that neither genes nor organisms have a clue about what future adaptations are possible or useful.

Genes mutate from time to time, and because the environment is changing, it once in a great while happens that a new gene equips its bearer with improved means of survival. It is now more “fit” for its environmental niche. Its adaptive value, its potential to help the organism that bears it leave many viable offspring, has increased. If a particular mutation secures for its possessor a mere 1% advantage over those who lack it, the mutation will be incorporated into most members of a large, freely interbreeding population in something like a thousand generations²—which is only a few tens of thousands of years even for large, long-lived animals. But what if mutations conferring even so small an advantage occur too rarely; or what if several genes must all, improbably, mutate together, each in the right direction, in order to adapt to the new conditions? Then all members of the population may die.

Is there an evolutionary strategy by which individuals and the species can escape from this trap, some trick by which the extremes of overspecialization and overgeneralization can both be avoided? For major environmental catastrophes there may be no such strategy. The dinosaurs had proliferated into an impressive range of environmental niches, and yet not one of them survived the mass extinctions of 65 million years ago. For quick, but less apocalyptic environmental change there are several ways. It helps to reproduce sexually, as we’ve described, because recombination of genes greatly increases the overall genetic variety. It helps to occupy a large and heterogenous territory, and not be too specialized. And it helps if the population breaks up into many nearly isolated subgroups—as was first clearly described by the population geneticist Sewall Wright, who died almost a centenarian in 1987. What follows is a simplification of a complex subject, some aspects of which are under renewed debate.³ But even if it were no more than metaphor, its explanatory power—for mammals, and especially for primates—is considerable.

——

The genes—the instruction manuals written down in the ACGT alphabet of DNA—are mutating. Some genes, in charge of important matters such as the business end of an enzyme, change slowly; indeed, they may change hardly at all in tens or even hundreds of millions of years—because such changes almost always make some molecular machine tool work more poorly, or not at all. Organisms with the mutated gene die (or leave fewer offspring) and the mutation tends not to be passed on to future generations. The sieve of selection strains it out. Other changes that do no damage—in an untranscribed nonsense sequence, or in the blueprints for structural elements involved in orienting the machine tool, say, or draping it over a molecular jig—can spread through future generations quickly, because an organism bearing the new mutation will not be eliminated by selection: In the code for structural elements, the particular sequence of As, Cs, Gs, and Ts hardly matters at all; what’s needed are placeholders, any sequence that codes for the shape of a subcellular handle, say, never mind which amino acids the handle is made of. Changes in ACGT sequences that are ignored anyway also won’t do any harm. Occasionally an organism hits the jackpot, and a favorable mutation will, in relatively few generations, sweep through the entire population; but the overall genetic change due to favorable mutations is slow, because they happen so rarely.

Some genes will be carried by almost all of the population; others will be present in only a tiny fraction of the population. But not even very useful genes will be carried by everyone, either because the gene is new and there hasn’t been time enough for it to spread through the whole population, or because there are always mutations transforming or eliminating a given gene, even a beneficial one. If the absence of a useful gene isn’t positively lethal, in a big enough population some organisms will always be without it. In general, any given gene is distributed through the population: Some individuals have it, and some don’t. If you divide your species up into smaller, mutually isolated subpopulations, the percent of individuals that carry a given gene will vary from group to group.

There are around ten thousand active genes in a typical “higher” mammal. Any one of them may vary from individual to individual and group to group. A few are extinguished for a time or for all time. A few are spanking new and are being spread quickly through the population. Most are old-timers. How useful any given gene is (in the population of wolves or humans or whatever mammal we have in mind) depends on the environment, and that’s changing too.

Let’s follow one of those ten thousand genes. Maybe it’s for extra testosterone production. But it could be any gene. The fraction of the population possessing this gene, relative to all possible alternative genes, is called the gene frequency.

Imagine now a set of isolated populations of the same species. Maybe they’re troops of monkeys that live in adjacent, nearly identical mountain valleys, separated by impassable mountains. Whatever differences there are in the chances of survival or of leaving descendants in the two groups, it won’t be because one is living in a more favorable physical environment.

Not all values of the gene frequency are equally adaptive. Instead, there’s an optimum frequency in the population. If the gene frequency is too low, maybe the monkeys are insufficiently vigilant in defending themselves against predators. If it’s too high, maybe they kill themselves off in dominance combat. When two isolated populations, in otherwise identical circumstances, have different constellations of active genes, their members will have different Darwinian fitness.

But the optimum frequency of this gene depends on the optimum frequency of other genes, as well as on the fluid and varying environment in which our monkeys must live. There might be more than one optimum frequency, depending on circumstances. The same is true for all ten thousand genes—their optimum frequencies all mutually dependent, all varying as the environment does. For example, a higher frequency of a gene for extra testosterone might be useful in dealing with predators and other hostile groups, provided genes for peacekeeping within the group were also more abundant. And so on. The optima interlace.

So a set of gene frequencies that once made your group superbly adapted may now constitute a marked disadvantage; and gene frequencies that once conferred only marginal fitness may now be the key to survival. What a disturbing concept of existence: Just when you’re most in harmony with your environment, that’s when the ice you’re skating on begins to thin. What you should have been emphasizing, had you been able, is early escape from optimum adaptation—a deliberate fall from grace contrived by the well-adjusted, the elective self-humbling of the mighty. The meaning of “overspecialized” becomes clear. But this is a strategy, we well know from everyday human experience, that privileged populations are almost never willing to embrace. In the classic confrontation between short-term and long, the short-term tends to win—especially when there’s no way to foretell the future.

Yes, they lack foresight. But how could they know? It’s asking a great deal of monkeys to foresee future geological or ecological change. We humans, who with our intelligence ought to be much more capable prophets than monkeys, have difficulty enough foreseeing the future, and still more difficulty acting on our knowledge.⁴ In military operations, ward-heeler politics, much of corporate strategy, and national response to the challenge of global environmental change, the short-term tends to predominate. So offhand, you might think that precautionary maintenance of a collection of gene frequencies that will be optimum for some future circumstance when no one is even aware of this fact is simply too difficult to arrange. You might think that there’s a flaw in the evolutionary process, that life, under some circumstances, might get stranded.

What could possibly cause the gene frequency in different populations to drift to suboptimal values? Suppose the mutation rate went up because of some new chemical in the environment (belched up from the Earth’s interior), or an increase in the flux of cosmic rays (perhaps from some exploding star halfway across the Milky Way). Then the gene frequencies in isolated populations diversify. You might even get a population that, by accident, winds up with the optimum frequencies needed to adapt to a future need. But that will be very rare. More likely, big changes will be lethal. So an increase in the mutation rate tends mainly to spread out the variation in gene frequencies, but not too much.

The population will, through mutation and selection together, tend to follow the changing circumstances, always working toward the optimum adaptation. If the external conditions vary slowly enough, the population might always be close to the optimal adaptation. Gene frequencies are always in slow motion. This gradual movement, driven by mutation and natural selection in a changing physical and biological environment, is just the evolutionary process outlined by Darwin; and Wright’s continuously changing gene frequencies are a metaphor of natural selection.

——

Up to now each isolated subpopulation we’ve been considering has been large, comprising maybe thousands of individuals or more. But now, Wright’s critical step: Let’s think about small groups, with no more than a few dozen individuals. They tend to become closely inbred. After a few generations, who’s available to mate with except relatives? So let’s look at inbreeding for a moment before considering the evolutionary prospects of small populations.

Some human cultures have sex in private and eat in public, some do it the other way around; some live with their aged relatives, some abandon them, and some eat them; some institute rigid rules that even toddlers must obey, and some let children do almost anything they want; some bury their dead, some burn their dead, and some set them out for the birds to eat; some use cowrie shells for money, some use metal, some paper, and some do without money altogether; some have no gods, some have one god, some have many gods. But all of them abominate incest.

Incest avoidance is one of the few invariables common to the spectacular diversity of human cultures. Sometimes, though, exceptions were made for (who else?) the ruling class. Since kings were gods, or near enough, only their sisters were considered of sufficiently exalted status to be their mates. Mayan and Egyptian royal families were inbred for generations, brothers marrying sisters—the process mitigated, it is thought, by unsanctioned and unrecorded couplings with nonrelatives. The surviving offspring were not conspicuously more incompetent than the usual, run-of-the-mill kings and queens, and Cleopatra, Queen of Egypt—officially the product of many consecutive generations of incestuous matings—was gifted by many standards. The historian Plutarch described her as not incomparably beautiful; still,

She was fluent not only in Egyptian, Greek, Latin, and Macedonian, but also in Hebrew, Arabic, and the languages of the Ethiopians, the Syrians, the Medes, the Parthians, “and many others.”⁵ She’s described as “the only human being except Hannibal who [ever] struck fear into Rome.”⁶ She also gave birth to several apparently healthy children—although they were not fathered by her brother. One of them was Ptolemy XV Caesar, son of Julius Caesar and titled King of Egypt (until murdered at age seventeen by the future Emperor Augustus). Cleopatra certainly does not seem to have exhibited marked physical or intellectual deficits, despite the alleged close relation of her parents.

Nevertheless, inbreeding produces a statistical genetic deficit that takes its toll chiefly in the deaths of infants and juveniles (and we don’t have a good record of Mayan and Egyptian royal children who died at birth or were put to death in infancy). There is considerable evidence for this in many—but by no means all—groups of animals and plants. Even in sexual microorganisms, incest causes striking increases in the deaths of the young.⁷ In incestuous unions in zoos, mortality in the offspring increased steeply for forty different species of mammals—although some were much more vulnerable to close inbreeding than others.⁸ In successive brother-sister matings in fruit flies, only a few percent of the offspring survived by the seventh generation.⁹ In baboons, matings between first cousins result in infants that die, within the first month of life, about 30% more often than in baboon matings where the parents are not close relatives.¹⁰ Most normally outbred plants—corn, for example—deteriorate on consistent inbreeding. They become smaller, scrawnier, more withered. That’s why we have hybrid corn. Many plants with both male and female parts are configured, as Darwin first noted, so they cannot easily have sex with themselves (“self-incompatibility” this ultimate incest taboo is called). Many animals, including the primates, have taboos that inhibit mating with close relatives.¹¹

Purebred dogs are prone to deformities and crippling defects. The biologists John Paul Scott and John L. Fuller performed breeding experiments—that is, artificial selection—on five breeds of dogs:

Similar genetic deficits are found in the limited data available on human incest in modern times. The increased infant death rate resulting from first cousin marriages¹³ is only about 60%. But in a Michigan study¹⁴ in the middle 1960s, eighteen children from brother-sister and father-daughter matings were compared with a control group of children from non-incestuous matings. Most of the children of incest (eleven out of eighteen) died within their first six months, or showed serious defects—including severe mental retardation. No history of such defects was found in the parents or their families. The remaining children seemed normal in intelligence and in all other respects, and were recommended for adoption. None of the children in the control group died or was institutionalized. Compared to brother-sister and father-daughter matings in other animals, though, these mortality and morbidity rates seem high; perhaps incestuous unions that produce abnormal children were more likely to come to the attention of the scientists making the study.

The dangers of repeated inbreeding seem so clear that we can safely conclude that unsanctioned sexual unions, impregnations of Queens of Egypt by someone other than the Pharaoh, occurred among Cleopatra’s immediate ancestors. Even a few sibling matings in consecutive generations would probably have led to death, or at least to a Cleopatra very different from the vital individual history reveals to us. But one generation of outcrossing goes far to cancel the previous inbreeding.

Inbreeding is a particular danger in very small groups, because in them it can hardly be avoided. If a new nonlethal mutation occurs in one individual, it either gets lost—because, for example, its bearer has no offspring—or it’s not many generations before it’s in nearly everybody, even if it’s slightly maladaptive. So now most males in the population have, say, a little too much testosterone; the conflicts and the distractions of conflict are taking their toll, and the youngsters are not being cared for as they should. The population has wandered from optimum adaptation; if inbreeding is intense, it may be that eventually none of the members of the group leaves offspring.

If inbreeding weren’t so risky, you might think that small populations are the way to get to gene frequency constellations that are not now especially adaptive, but that will be so at some time in the future. If the population is small, then new mutations or new combinations of letters and sequences in the genetic code can propagate through the entire population in only a few generations. New random experiments in biology are being conducted that could not occur in large populations. As a result, almost always, the group goes hurtling away from optimum adaptation. But comparatively rare genes and gene combinations can be tried out so quickly in a small population that it can swiftly cover a lot of ground in the possible range of gene frequencies.

What’s happening here is described as “accidents of sampling,” which have much more profound consequences in small populations than in large ones: Imagine you’re flipping a coin. Your chance of getting one head in one trial or flip is clearly 50%, one chance in two. The coin has only a head and a tail, and it has to turn up one side or the other. With two flips, the full menu of equally possible outcomes is: two tails, a head and a tail, a tail and a head, or two heads. So your chance of getting two consecutive heads is one in four, or, equivalently, one-quarter, or ½ × ½. With three flips, the chance that they’re all heads is one chance in eight (½ × ½ × ½), or one in 2³. You can flip ten heads in a row once in about a thousand trials (2¹⁰ = 1024). (If we’d witnessed only that trial, we might think you’re phenomenally lucky.) But a hundred heads in a row will take about a billion billion trillion trials (2¹⁰⁰ roughly equals 10³⁰)—which is the same as forever.

In small populations major accidents of sampling are inevitable; in large populations they are nonexistent. Were a national opinion poll to query three people only, there would be little reason to believe the results—that is, to think these three opinions adequately sampled the opinions of most citizens. One of the individuals polled might, by accident, be a Libertarian or a Vegetarian, a Trotskyite or a Luddite, a Coptic or a Skeptic—all with interesting perspectives, but none an accurate reflection of the general population. Now imagine that the opinions of these three were somehow proportionately amplified to become the opinions of the population of the United States as a whole; a major transformation in national attitudes and politics would have been worked. The same can be true genetically when a few individuals from a large population establish a new and isolated community.

Accidents of sampling happen when the population sampled is very small. In many elections, when the pollsters sample five hundred or a thousand randomly chosen people, the results repeatedly prove to be representative of the nation as a whole.* With five hundred or a thousand truthful random samplings, the findings are accurate to within a few percent. (The variation expected is the square root of the sample size.) If you ask a large number of randomly selected people, you will reliably sample the average*; if you ask only a few, you may sample atypical or fringe opinions. Pollsters would gladly sample smaller populations; it would save them money. But they dare not—the errors would be too large, the sampled opinions too unrepresentative.

As in opinion polls, so it is in the genetics of populations: With a small enough group, substantial deviations † from the average can be sampled and become established. With mutually isolated small groups, many different sets of gene frequencies get tried out—most maladaptive, but a few, fortuitously, poised for the future. This is called genetic drift.

Or suppose that your name is Theodosius Dobzhansky and that you live in New York City. Even if you have ten sons, your name will continue to be “rare and outlandish” so long as you continue to reside in the big city. But move the family to a small town, have many descendants, and Dobzhansky will eventually become a common and unremarkable name. Similarly, any extraordinary hereditary predisposition in the Dobzhansky genes will affect only a tiny fraction of the population while you’re in New York, but might in a few generations become a major genetic feature of the citizenry of the town.¹⁵

Is there any way to preserve the accidents of sampling inherent in small groups, while avoiding the slow deterioration intrinsic to incest? Imagine that each group is significantly inbred, but that outbreeding is sometimes indulged in. Individuals from largely isolated subpopulations occasionally find each other and mate, enough to mitigate the more severe genetic consequences of incest. Different constellations of genes will be established in each subpopulation by genetic drift. Each small group will have a different set of hereditary propensities. They will not all, therefore, be optimally adapted to current circumstances. Now that the environment has changed, none of them may be. Being far from optimally adapted, their lives will be hard. Not one of these groups will be as well off as it was earlier. Many groups will die out. Now, though, when the environmental crisis comes, a few of these smaller populations will find themselves, by accident, advantageously situated, “preadapted.”

The trick is to combine the accidents of sampling of small groups (so at least one group will be by chance fortunately poised for the next environmental crisis) with the stability of large groups (so once the new, desirable adaptation is hit upon, it is spread to a substantial population). Because the lucky group—with newly optimal gene frequencies—is also in genetic contact with other groups, its new constellation of adaptive genes is passed on. Other groups acquire the new capabilities, the new mix of traits, the new adaptations; and simultaneously the most dangerous consequences of inbreeding are avoided.

Here then is a trial-and-error mechanism through which a large population can explore the mix of possible gene frequencies. When the adaptations that formerly led to our success now become only marginally useful, we have a way out. Dividing a species into many quite small, fairly inbred populations, but allowing occasional interbreeding among these populations, is the solution Sewall Wright proposed. It avoids both traps, overspecialization and overgeneralization.¹⁶ And to the extent that major evolutionary steps occur relatively quickly in small, semi-isolated populations, the relative paucity of intermediate forms in the fossil record—one of the problems that plagued Darwin—would be explained.¹⁷

——

No organisms have ever sat down and decided, as a matter of conscious species-wide evolutionary policy, to divide themselves up into small populations, amplify accidents of genetic sampling, and at the same time avoid the more flagrant forms of incest. But, as always happens in the evolutionary process, any species that, by accident, makes appropriate arrangements preferentially reproduces. If enough evolutionary experiments are tried over the immense vistas of time available in the history of life, then very improbable adaptations—in group size, say, or in the balance between inbreeding and outbreeding—can be institutionalized. Here we are talking about the evolution of a mechanism to guarantee continuing evolution, a second-order or meta-evolutionary development.¹⁸

What would it feel like from the inside if you were a member of a species that had, through natural selection, made arrangements for genetic drift? You would enjoy living in small groups. You would hate crowds. For accidents of sampling to work on an appropriate time scale, a group might have to comprise no more than one hundred or two hundred individuals, and—according to Wright—would probably be best with only a few dozen members. Groups of six to eight or fewer tend to be unstable; they’re too vulnerable to being wiped out by predators or flood or disease, a different example of accidents of sampling. You would conceive a passionate loyalty to the group, something like intense family feeling, superpatriotism, chauvinism, ethnocentrism. (Especially because most members of your group are close relatives, you might when necessary be moved to something like altruistic or even heroic actions on their behalf.) You would also need to avoid any merger of your group with another, because much bigger groups would inhibit accidents of sampling. So it would be helpful if you conceived a passionate hostility to other groups, a vivid sense of their deficiencies, something like xenophobia or jingoism.

Those other groups are, of course, composed of individuals of the same species as you. They look almost exactly like you. To fan the flames of xenophobia, you must examine them with minute attention and exaggerate whatever differences can be discerned, always to their disadvantage. They have slightly different heredities and slightly different diets, so they don’t smell quite the same as you and yours. If your olfactory powers are sufficiently finely tuned, maybe their scents will render them grotesque, hateful, odious.

It would be even better if you could establish some distinctions. If differences in dress and language are unavailable—having not yet been invented, for example—differences in behavior, posture, or vocalizations would be helpful. Anything that can distinguish your group from the others could work to keep hatreds high and resist merger. Other groups, conveniently, are similarly disposed. These nonhereditary differences between one group and another—even arbitrary differences, only distantly connected with any adaptive advantage, but serving to preserve group independence and coherence—are called, collectively, culture. At a rudimentary level many animals have it.¹⁹ Cultural diversity helps preserve genetic drift.

At the same time, avoiding too much inbreeding and guaranteeing at least occasional outbreeding are essential. So you would feel a revulsion about incest, or at least about the most consanguineous matings. Wherever possible, this revulsion would be reinforced by your copying the attitudes of your fellows, by culture. There would be an incest taboo (relaxed perhaps if the population is reduced to only a few survivors). Outbreeding might be officially proscribed—perhaps, among humans, by young men attacking males from other groups who, even accidentally, wander into the neighborhood, or by fathers mourning, as if dead, daughters who run off with foreigners. But despite the pervasive ethnocentrism and xenophobia, now and then you would find members of other, hostile groups unaccountably attractive. Surreptitious matings would occur. (This is, more or less, the theme of Romeo and Juliet, Rudolph Valentino’s The Sheik, and a vast industry of books on romance, targeted at women.)

A promising survival strategy, in short, is this: Break up into small groups, encourage ethnocentrism and xenophobia, and succumb to the occasional sexual temptations provided by the sons and daughters of enemy clans. Devise your own culture: The more your species is capable of learned behavior, the greater the differences that can be established between one group and another. Behavioral differences eventually lead to genetic differences, and vice versa. Incomplete isolation—just the right mix of aloofness and sexual abandon with other groups—generates diversity. And diversity is the raw material on which selection operates.

There seems to be, then, a reason—at the heart of population genetics and evolution—for small semi-isolated groups as the substructure of larger populations, for xenophobia, ethnocentrism, territoriality, incest avoidance, occasional outbreeding, and migration away from the most successful communities. These mechanisms work especially for those species that find themselves in a swiftly changing environment, biologically or physically. Archaebacteria, ants, and horseshoe crabs have not much been in this category; birds and mammals have. So next time you hear a raving demagogue counseling hatred for other, slightly different groups of humans, for a moment at least see if you can understand his problem: He is heeding an ancient call that—however dangerous, obsolete, and maladaptive it may be today—once benefitted our species.

A solution has been found to the problem of how to arrange for gene frequencies to respond quickly to a volatile, changing environment. And the solution seems eerily familiar. After a journey into an abstract world of population genetics and gene frequencies, we turn a corner and suddenly find ourselves gazing at something that looks very much like … ourselves.