Why is Sex Fun?: the evolution of human sexuality

CHAPTER 6. MAKING MORE BY MAKING LESS: The Evolution of Female Menopause

Most wild animals remain fertile until they die, or until close to that time. So do human males: although some men become infertile or less fertile at various ages for various reasons, men experience no universal shutdown of fertility at any particular age. There are innumerable well-attested cases of old men, including a ninety-four-year-old, fathering children.

But human females undergo a steep decline in fertility from around age forty, leading to universal complete sterility within a decade or so. While some women continue to have regular menstrual cycles up to the age of fifty-four or fifty-five, conception after the age of fifty was rare until the recent development of medical technologies using hormone therapy and artificial fertilization. For example, among the American Hutterites, a strict religious community that is well nourished and opposed to contraception, women produce babies as fast as is biologically possible for humans, with a mean interval of only two years between births, and a mean final number of eleven children. Even Hutterite women stop producing babies by age forty-nine.

To laypeople, menopause is an inevitable fact of life, albeit often a painful one anticipated with foreboding. But to evolutionary biologists, human female menopause is an aberration in the animal world and an intellectual paradox. The essence of natural selection is that it promotes genes for traits that increase the number of one's descendants bearing those genes. How could natural selection possibly result in every female member of a species carrying genes that throttle her ability to leave more descendants? All biological traits are subject to genetic variation, including the age of human female menopause. Once female menopause somehow became fixed in humans for whatever reason, why did not its age of onset gradually become pushed back until it disappeared again, because those women who experienced menopause later in life left behind more descendants?

To evolutionary biologists, female menopause is thus among the most bizarre features of human sexuality. As I shall argue, it is also among the most important. Along with our big brains and upright posture (emphasized in every text of human evolution), and our concealed ovula-tions and penchant for recreational sex (to which texts devote less attention), I believe that female menopause was among the biological traits essential for making us distinctively human-a creature more than, and qualitatively different from, an ape.

Many biologists would balk at what I have just said. They would argue that human female menopause does not pose an unsolved problem, and that there is no need to discuss it further. Their objections are of three types.

First, some biologists dismiss human female menopause as an artifact of a recent increase in human expected life span. That increase stems not just from public health measures within the last century but possibly also from the rise of agriculture ten thousand years ago, and even more likely from evolutionary changes leading to increased human survival skills within the last forty thousand years. According to this view, menopause could not have been a frequent occurrence for most of the several million years of human evo-lution, because (supposedly) almost no women or men survived past the age of forty. Of course, the female reproductive tract was programmed to shut down by age forty, because it would not have had the opportunity to operate thereafter anyway. The increase in human life span has developed much too recently in our evolutionary history for the female reproductive tract to have had time to adjust-so goes this objection.

However, this view ignores the fact that the human male reproductive tract, and every other biological function of both women and men, continue to function in most people for many decades after age forty. One would therefore have to assume that every other biological function was able to adjust quickly to our new long life span, leaving unexplained why female reproduction was uniquely incapable of doing so. The claim that formerly few women survived until the age of menopause is based on paleode-mography, that is, on attempts to estimate age at time of death in ancient skeletons. Those estimates rest on un-proven, implausible assumptions, such as that the recovered skeletons represent an unbiased sample of an entire ancient population, or that ancient adult skeletons really can be aged accurately. While paleodemographers' ability to distinguish the ancient skeleton of a ten-year-old from that of a twenty-five-year-old is not in question, the ability they claim to distinguish an ancient forty-year-old from a fifty-five-year-old has never been demonstrated. One can hardly reason by comparison with skeletons of modern people, whose different lifestyles, diets, and diseases surely make their bones age at different rates from the bones of ancients.

A second objection acknowledges human female menopause as a possibly ancient phenomenon but denies that it is unique to humans. Many or most wild animals exhibit a decrease in fertility with age. Some elderly individuals of a wide variety of wild mammal and bird species are found to be infertile. Many elderly female individuals of rhesus macaques and certain strains of laboratory mice, living in laboratory cages or zoos where their lives are considerably extended over expected spans in the wild by gourmet diets, superb medical care, and complete protection from enemies, do become infertile. Hence some biologists object that human female menopause is merely part of a widespread phenomenon of animal menopause. Whatever that phenomenon's explanation, its existence in many species would mean that there is not necessarily anything peculiar about menopause in the human species requiring explanation.

However, one swallow does not make a summer, nor does one sterile female constitute menopause. That is, detection of an occasional sterile elderly individual in the wild, or of regular sterility in caged animals with artificially extended life spans, does nothing to establish the existence of menopause as a biologically significant phenomenon in the wild. That would require demonstrating that a substantial fraction of adult females in a wild animal population become sterile and spend a significant portion of their life spans after the end of their fertility.

The human species does fulfill that definition, but only one or possibly two wild animal species are definitely known to do so. One is an Australian marsupial mouse in which males (not females) exhibit something like menopause: all males in the population become sterile within a short time in August and die over the next couple of weeks, leaving a population that consists solely of pregnant females. In that case, however, the postmenopausal phase is a negligible fraction of the total male life span. Marsupial mice do not exemplify true menopause but are more appropriately considered an example of big-bang reproduction, alias semelparity-a single lifetime reproductive effort rapidly followed by sterility and death, as in salmon and century plants. The better example of animal menopause is provided by pilot whales, among which one-quarter of all adult females killed by whalers proved to be postmenopausal, as judged by the condition of their ovaries. Female pilot whales enter menopause at the ago of thirty or forty years, have a mean survival of at least fourteen years after menopause, and may live for over sixty years.

Menopause as a biologically significant phenomenon is thus not unique to humans, being shared at least with one species of whale. It would be worth looking for evidence of menopause in killer whales and a few other species as possible candidates. But still-fertile elderly females are often encountered among well-studied wild populations of other long-lived mammals, including chimpanzees, gorillas, baboons, and elephants. Hence those species and most others are unlikely to be characterized by regular menopause. For example, a fifty-five-year-old elephant is considered elderly, since 95 percent of elephants die before that age. But the fertility of fifty-five-year-old female elephants is still half that of younger females in their prime.

Thus, female menopause is sufficiently unusual in the animal world that its evolution in humans requires explanation. We certainly did not inherit it from pilot whales, from whose ancestors our own ancestors parted company over fifty million years ago. In fact, we must have evolved it since our ancestors separated from those of chimps and gorillas seven million years ago, because we undergo menopause and chimps and gorillas appear not to (or at least not regularly).

The third and last objection acknowledges human menopause as an ancient phenomenon that is unusual among animals. Instead, these critics say that we need not seek an explanation for menopause, because the puzzle has already been solved. The solution (they say) lies in the physiological mechanism of menopause: a woman's egg supply is fixed at her birth and not added to later in her life. One or more eggs are lost by ovulation at each menstrual cycle, and far more eggs simply die (termed atresia). By the time a woman is fifty years old, most of her original egg supply has been depleted. Those eggs that remain are half a century old, increasingly unresponsive to pituitary hormones, and too few in number to produce enough estra-diol to trigger the release of pituitary hormones.

But there is a fatal counterobjection to this objection. While the objection is not wrong, it is incomplete. Yes, depletion and aging of the egg supply are the immediate causes of human menopause, but why did natural selection program women such that their eggs become depleted or unresponsive in their forties? There is no compelling reason why we could not have evolved twice as large a starting quota of eggs, or eggs that remain responsive after half a century. The eggs of elephants, baleen whales, and possibly albatrosses remain viable for at least sixty years, and the eggs of tortoises are viable for much longer, so human eggs could presumably have evolved the same capability.

The basic reason why the third objection is incomplete is because it confuses proximate mechanisms with ultimate causal explanations. (A proximate mechanism is an immediate direct cause, while an ultimate explanation is the last in the long chain of factors leading up to that immediate cause. For example, the proximate cause of a marriage breakup may be a husband's discovery of his wife's extramarital affairs, but the ultimate explanation may be the husband's chronic insensitivity and the couple's basic incompatibility that drove the wife to affairs.) Physiologists and molecular biologists regularly fall into the trap of overlooking this distinction, which is fundamental to biology, history, and human behavior. Physiology and molecular biology can do no more than identify proximate mechanisms; only evolutionary biology can provide ultimate causal explanations. As one simple example, the proximate reason why so-called poison-dart frogs are poi-sonous is that they secrete a lethal chemical named batra-chotoxin. But that molecular biological mechanism for the frogs' poisonousness could be considered an unimportant detail because many other poisonous chemicals would have worked equally well. The ultimate causal explanation is that poison-dart frogs evolved poisonous chemicals because they are small, otherwise defenseless animals that would be easy prey for predators if they were not protected by poison.

We have already seen repeatedly in this book that the big questions about human sexuality are the evolutionary questions about ultimate causal explanation, not the search for proximate physiological mechanisms. Yes, sex is fun for us because women have concealed ovulations and are constantly receptive, but why did they evolve that unusual reproductive physiology? Yes, men have the physiological capacity to produce milk, but why did they not evolve to exploit that capacity? For menopause as well, the easy part of the puzzle is the mundane fact that a woman's egg supply gets depleted or impaired by around the time she is fifty years old. The challenge is to understand why we evolved that seemingly self-defeating detail of reproductive physiology.

The aging (or senescence, as biologists call it) of the female reproductive tract cannot be profitably considered in isolation from other aging processes. Our eyes, kidneys, heart, and all other organs and tissues also senesce. But that aging of our organs is not physiologically inevitable-or at least it's not inevitable that they senesce as rapidly as they do in the human species, because the organs of some turtles, clams, and other species remain in good condition much longer than ours do.

Physiologists and many other researchers on aging tend to search for a single all-encompassing explanation of aging. Popular explanations hypothesized in recent decades have invoked the immune system, free radicals, hormones, and cell division. In reality, though, all of us over forty know that everything about our bodies gradually deteriorates, and not just our immune systems and our defenses against free radicals. Although I have had a less stressful life and better medical care than most of the world's nearly six billion people, I can still tick off the aging processes that have already taken their toll on me by age fifty-nine: impaired hearing at high pitch, failure of my eyes to focus at short distances, less acute senses of smell and taste, loss of one kidney, tooth wear, less flexible fingers, and so on. My recovery from injuries is already slower than it used to be: I had to give up running because of recurrent calf injuries, I recently completed a slow recovery from a left elbow injury, and now I have just injured the tendon of a finger. Ahead of me, if the experience of other men is any guide, lies the familiar litany of complaints, including heart disorders, clogged arteries, bladder trouble, joint problems, prostate enlargement, memory loss, colon cancer, and so on. All that deterioration is what we mean by aging.

The basic reasons behind this grim litany are easily understood by analogy to human-built structures. Animal bodies, like machines, tend to deteriorate gradually or become acutely damaged with age and use. To combat those tendencies, we consciously maintain and repair our machines. Natural selection ensures that our body unconsciously maintains and repairs itself.

Both bodies and machines are maintained in two ways. First, we repair a part of a machine when it is acutely damaged. For example, we fix a car's punctured tire or bashed-in fender, and we replace its brakes or tires if they become damaged beyond repair. Our body similarly repairs acute damage. The most visible example is wound repair when we cut our skin, but molecular repair of damaged DNA and many other repair processes go on invisibly inside us. Just as a ruined tire can be replaced, our body has some capac-ity to regenerate parts of damaged organs such as by mak-ing new kidney, liver, and intestinal tissue. That capacity for regeneration is much better developed in many other animals. If only we were like starfish, crabs, sea cucumbers, and lizards, which can regenerate their arms, legs, intestines, and tail, respectively!

The other type of upkeep of machines and bodies is regular or automatic maintenance to reverse gradual wear, regardless of whether there has been any acute damage. For example, at times of scheduled maintenance we change our car's motor oil, spark plugs, fan belt, and ball bearings. Similarly, our body constantly grows new hair, replaces the lining of the small intestine every few days, replaces our red blood cells every few months, and replaces each tooth once in our lifetime. Invisible replacement goes on for the individual protein molecules that make up our bodies.

How well you maintain your car, and how much money or resources you put into its maintenance, strongly influence how long it lasts. The same can be said of our bodies, not only with respect to our exercise programs, visits to the doctor, and other conscious maintenance, but also with respect to the unconscious repair and maintenance that our bodies do on themselves. Synthesizing new skin, kidney tissue, and proteins uses up a lot of biosynthetic energy. Animal species vary greatly in their investment in self-maintenance, hence in the rate at which they senesce. Some turtles live for over a century. Laboratory mice, living in cages with abundant food and no predators or risks, and receiving better medical care than any wild turtle or the vast majority of the world's people, inevitably become decrepit and die of old age before their third birthday. There are aging differences even among us humans and our closest relatives, the great apes. Well-nourished apes living in the safety of zoo cages and attended by veterinarians rarely (if ever) live past age sixty, while white Americans exposed to much greater danger and receiving less medical attention now live to an average of seventy-eight years for men, eighty-three years for women. Why do our bodies unconsciously take better care of themselves than do apes' bodies? Why do turtles senesce so much more slowly than mice?

We could avoid aging entirely and (barring accidents) live forever if we went all out for repair and changed all the parts of our bodies frequently. We could avoid arthritis by growing new limbs, as crabs do, avoid heart attacks by periodically growing a new heart, and minimize tooth decay by regrowing new teeth five times (as elephants do, instead of just once, as we do). Some animals thus make a big investment in certain aspects of body repair, but no animal makes a big investment in all aspects, and no animal avoids aging entirely.

Analogy to our cars again makes the reason obvious: the expense of repair and maintenance. Most of us have only limited amounts of money, which we are obliged to budget. We put just enough money into car repair to keep our car running as long as it makes economic sense to do so. When the repair bills get too high, we find it cheaper to let the old car die and buy a new one. Our genes face a similar tradeoff between repairing the old body that contains the genes and making new containers for the genes (that is, babies). Resources spent on repair, whether of cars or of bodies, eat away at the resources available for buying new cars or making babies. Animals with cheap self-repair and short life spans, like mice, can churn out babies much more rapidly than can expensive-to-maintain, long-lived animals like us. A female mouse that will die at the age of two, long before we humans achieve fertility, has been producing five babies every two months since she was a few months old.

That is, natural selection adjusts the relative invest ments in repair and reproduction so as to maximize the transmission of genes to offspring. The balance between re-pair and reproduction differs between species. Some species stint on repair and churn out babies quickly but die early, like mice. Other species, like us, invest heavily in repair, live for nearly a century, and can produce a dozen babies in that time (if you are a Hutterite woman), or over a thousand babies (if you are Emperor Moulay the Bloodthirsty). Your annual rate of baby production is lower than the mouse's (even if you are Moulay) but you have more years in which to do it.

It turns out that an important evolutionary determinant of biological investment in repair-hence of life span under the best possible conditions-is the risk of death from accidents and bad conditions. You don't waste money maintaining your taxi if you are a taxi driver in Teheran, where even the most careful taxi driver is bound to suffer a major fender-bender every few weeks. Instead, you save your money to buy the inevitable next taxi. Similarly, animals whose lifestyles carry a high risk of accidental death are evolutionarily programmed to stint on repair and to age rapidly, even when living in the well-nourished safety of a laboratory cage. Mice, subject to high rates of predation in the wild, are evolutionarily programmed to invest less in repair and to age more rapidly than similar-sized caged birds that in the wild can escape predators by flying. Turtles, protected in the wild by a shell, are programmed to age more slowly than other reptiles, while porcupines, protected by quills, age more slowly than mammals comparable in size.

That generalization also fits us and our ape relatives. Ancient humans, who usually remained on the ground and defended themselves with spears and fire, were at lower risk of death from predators or from falling out of a tree than were arboreal apes. The legacy of the resultant evolutionary programming carries on today in that we live for several decades longer than do zoo apes living under comparable conditions of safety, health, and affluence. We must have evolved better repair mechanisms and decreased rates of senescence in the last seven million years, since we parted company from our ape relatives, came down out of the trees, and armed ourselves with spears and stones and fire.

Similar reasoning is relevant to our painful experience that everything in our bodies begins to fall apart as we grow older. Alas, that sad truth of evolutionary design is cost-efficient. You would be wasting biosynthetic energy, which otherwise could go into making babies, if you kept one part of your body in such great repair that it outlasted all your other parts and your resultant expected life span. The most efficiently constructed body is the one in which all organs wear out at approximately the same time.

The same principle, of course, applies to human-built machines, as illustrated in a story about that genius of cost-efficient automobile manufacture, Henry Ford. One day, Ford sent some of his employees to car junkyards, with instructions to examine the condition of the remaining parts in Model T Fords that had been junked. The employees brought back the apparently disappointing news that almost all components showed signs of wear. The sole exceptions were the kingpins, which remained virtually unworn. To the employees' surprise, Ford, instead of expressing pride in his well-made kingpins, declared that the kingpins were overbuilt, and that in the future they should be made more cheaply. Ford's conclusion may violate our ideal of pride in workmanship, but it made economic sense: he had indeed been wasting money on long-lasting kingpins that outlasted the cars in which they were installed.

The design of our bodies, which evolved through natural selection, fits Henry Ford's kingpin principle with only one exception. Virtually every part of the human body wears out around the same time. The kingpin principle even fits men's reproductive tract, which undergoes no abrupt shutdown but does gradually accumulate a varinty of problems, such as prostate hypertrophy and decreasing sperm count, to different degrees in different men. The kingpin principle also fits the bodies of animals. Animals caught in the wild show few signs of age-related deterioration because a wild animal is likely to die from a predator or accident when its body becomes significantly impaired. In zoos and laboratory cages, however, animals exhibit gradual age-related deterioration in every body part just as we do.

That sad message applies to the female as well as the male reproductive tract of animals. Female rhesus macaques run out of functional eggs around age thirty; fertilization of eggs in aged rabbits becomes less reliable; an increasing fraction of eggs are abnormal in aging hamsters, mice, and rabbits; fertilized embryos are increasingly unvi-able in aged hamsters and rabbits; and aging of the uterus itself leads to increasing embryonic mortality in hamsters, mice, and rabbits. Thus, the female reproductive tract of animals is a microcosm of the whole body in that everything that could go wrong with age may in fact go wrong— at different ages in different individuals.

The glaring exception to the kingpin principle is human female menopause. In all women within a short age span, it shuts down decades before expected death, even before the expected death of many hunter-gatherer women. It shuts down for a physiologically trivial reason-the exhaustion of functional eggs-that would have been easy to eliminate just by a mutation that slightly altered the rate at which eggs die or become unresponsive. Evidently, there was nothing physiologically inevitable about human female menopause, and there was nothing evolutionarily inevitable about it from the perspective of mammals in general. Instead, the human female, but not the human male, has become specifically programmed by natural selection, at some time within the last few million years, to shut down reproduction prematurely. That premature senescence is all the more surprising because it goes against an overwhelming trend: in other respects, we humans have evolved delayed rather than premature senescence.

Theorizing about the evolutionary basis of human female menopause must explain how a woman's apparently counterproductive evolutionary strategy of making fewer babies could actually result in her making more babies. Evidently, as a woman ages, she can do more to increase the number of people bearing her genes by devoting herself to her existing children, her potential grandchildren, and her other relatives than by producing yet another child.

The evolutionary chain of reasoning rests on several cruel facts. One is the human child's long period of parental dependence, longer than in any other animal species. A baby chimpanzee starts gathering its own food as it becomes weaned by its mother. It gathers the food mostly with its own hands. (Chimpanzee use of tools, such as fishing for termites with grass blades or cracking nuts with stones, is of great interest to human scientists but of only limited dietary significance to chimpanzees.) The baby chimpanzee also prepares its food with its own hands. But human hunter-gatherers acquire most of their food with tools, such as digging sticks, nets, spears, and baskets. Much human food is also prepared with tools (husked, pounded, cut up, et cetera) and then cooked in a fire. We do not protect ourselves against dangerous predators with our teeth and strong muscles, as do other prey animals, but, again, with our tools. Even to wield all those tools is completely beyond the manual dexterity of babies, and to make the tools is beyond the abilities of young children. Tool use and tool making are transmitted not just by imitation but by language, which takes over a decade for a child to master.

As a result, a human child in most societies does not become capable of economic independence or adult economic function until his or her teenage years or twenties. Until then, the child remains dependent on his or her parents, especially on the mother, because, as we saw in previous chapters, mothers tend to provide more child care than do fathers. Parents are important not only for gathering food and teaching tool making but also for providing protection and status within the tribe. In traditional societies, the early death of either the mother or the father prejudiced a child's life even if the surviving parent remarried, because of possible conflicts with the stepparent's genetic interests. A young orphan who was not adopted had even worse chances of surviving.

Hence a hunter-gatherer mother who already has several children risks losing some of her genetic investment in them if she does not survive until the youngest is at least a teenager. That one cruel fact underlying human female menopause becomes more ominous in the light of another cruel fact: the birth of each child immediately jeopardizes a mother's previous children because of the mother's risk of death in childbirth. In most other animal species, that risk is insignificant. For example, in one study encompassing 401 pregnant female rhesus macaques, only one died in childbirth. For humans in traditional societies, the risk was much higher and increased with age. Even in affluent, twentieth-century Western societies, the risk of dying in childbirth is seven times higher for a mother over the age of forty than for a twenty-year-old mother. But each now child puts the mother's life at risk not only because of the immediate risk of death in childbirth but also because of the delayed risk of death related to exhaustion by lactation, carrying a young child, and working harder to feed more mouths.

Yet another cruel fact is that infants of older mothers are themselves increasingly unlikely to survive or be healthy because of age-related increases in the risks of abortion, stillbirth, low fetal weight, and genetic defects. For instance, the risk of a fetus carrying the genetic condition known as Down's syndrome increases with the mother's age, from one in two thousand births for a mother under thirty, one in three hundred for a mother between the ages of thirty-five and thirty-nine, and one in fifty for a forty-three-year-old mother, to the grim odds of one in ten for a mother in her late forties.

Thus, as a woman gets older, she is likely to have accumulated more children; she has also been caring for them longer, so she is putting a bigger investment at risk with each successive pregnancy. But her chances of dying in or after childbirth, and the chances that the fetus or infant will die or be damaged, also increase. In effect, the older mother is taking on more risk for less potential gain. That's one set of factors that would tend to favor human female menopause and that would paradoxically result in a woman ending up with more surviving children by giving birth to fewer children. Natual selection has not programmed menopause into men because of three more cruel facts: men never die in childbirth and rarely die while copulating, and they are less likely than mothers to exhaust themselves caring for infants.

A hypothetically nonmenopausal old woman who died in childbirth, or while caring for an infant, would thereby be throwing away even more than her investment in her previous children. That is because a woman's children eventually begin producing children of their own, and those children count as part of the woman's prior investment. Especially in traditional societies, a woman's survival is important not only to her children but also to her grandchildren.

That extended role of postmenopausal women has been explored by Kristen Hawkes, the anthropologist whose re-search on men's roles I discussed in chapter 5. Hawkes and her colleagues studied foraging by women of different ages among the Hadza hunter-gatherers of Tanzania. The women who devoted the most time to gathering food (especially roots, honey, and fruit) were postmenopausal women. Those hardworking Hadza grandmothers put in an impressive seven hours per day, compared to a mere three hours for teenagers and new brides and four and a half hours for married women with young children. As one might expect, foraging returns (measured in pounds of food gathered per hour) increased with age and experience, so that mature women achieved higher returns than teenagers, but, interestingly, the grandmothers' returns were still as high as those of women in their prime. The combination of more foraging hours and an unchanged foraging efficiency meant that the postmenopausal grandmothers brought in more food per day than any of the younger groups of women, even though their large harvests were greatly in excess of what was required to meet their own personal needs and they no longer had dependent young children to feed.

Hawkes and her colleagues observed that the Hadza grandmothers were sharing their excess food harvest with close relatives, such as their grandchildren and grown children. As a strategy for transforming food calories into pounds of baby, it would be more efficient for an older woman to donate the calories to grandchildren and grown children rather than to infants of her own (even if she still could give birth) because the older mother's fertility would be decreasing with age anyway, whereas her own children' would be young adults at peak fertility. Naturally, this food-sharing argument does not constitute the sole reproductive contribution of postmenopausal women in traditional societies. A grandmother also baby-sits her grandchildren, thereby helping her adult children churn out more babies bearing the grandmother's genes. In addition, grandmothers lend their social status to their grandchildren, as to their children.

If one were playing God or Darwin and trying to decide whether to make older women undergo menopause or remain fertile, one would draw up a balance sheet, contrasting the benefits of menopause in one column with its costs in the other column. The costs of menopause are the potential children that a woman forgoes by undergoing menopause. The potential benefits include avoiding the increased risk of death due to childbirth and parenting at an advanced age, and gaining the benefit of improved survival for one's grandchildren and prior children. The sizes of those benefits depends on many details: How large is the risk of death in and after childbirth? How much does that risk increase with age? How large would the risk of death be at the same age even without children or the burden of parenting? How rapidly does fertility decrease with age before menopause? How rapidly would it continue to decrease in an aging woman who did not undergo menopause? All these factors are bound to differ between societies and are not easy to estimate. Hence anthropologists remain undecided whether the two considerations that I have discussed so far-investing in grandchildren and protecting one's prior investment in existing children-suffice to offset menopause's foreclosed option of further children and thus to explain the evolution of human female menopause.

But there is still one more virtue of menopause, one that has received little attention. That is the importance of old people to their entire tribe in preliterate societies, which constituted every human society in the world from the time of human origins until the rise of writing in Mesopotamia around 3300 b.c. Textbooks of human genet-ics regularly assert that natural selection cannot weed out mutations tending to cause damaging effects of age in old people. Supposedly there can be no selection against such mutations because old people are said to be “postrepro-ductive.” I believe that such assertions overlook an essen-tial fact that distinguishes humans from most animal species. No human, except a hermit, is ever truly postre-productive in the sense of being unable to benefit the survival and reproduction of other people bearing one's genes. Yes, I grant that if any orangutans lived long enough in the wild to become sterile, they would count as postre-productive, since orangutans other than mothers with one young offspring tend to be solitary. I also grant that the contributions of very old people to modern literate societies tend to decrease with age-a new phenomenon at the root of the enormous problems that old age now poses, both for the elderly themselves and for the rest of society. Today, we moderns get most of our information through writing, television, or radio. We find it impossible to conceive of the overwhelming importance of elderly people in preliterate societies as repositories of information and experience.

Here is an example of that role. In my field studies of bird ecology on New Guinea and adjacent Southwest Pacific islands, I live among people who traditionally had been without writing, depended on stone tools, and subsisted by farming and fishing supplemented by much hunting and gathering. I am constantly asking villagers to toll me the names of local species of birds, animals, and other plants in their local language, and to tell me what they know about each species. It turns out that New Guineans and Pacific islanders possess an enormous fund of traditional biological knowledge, including names for a thousand or more species, plus information about each habitat, behavior, ecology, and usefulness to humans. All that information is important because wild plants and animals traditionally furnished much of the people's food and all of their building materials, medicines, and decorations.

Again and again, when I ask a question about some rare bird, I find that only the older hunters know the answer, and eventually I ask a question that stumps even them. The hunters reply, “We have to ask the old man.[1]” They then take me to a hut, inside of which is an old man or woman, often blind with cataracts, barely able to walk, toothless, and unable to eat any food that hasn't been prechewed by someone else. But that old person is the tribe's library. Because the society traditionally lacked writing, that old person knows much more about the local environment than anyone else and is the sole source of accurate knowledge about events that happened long ago. Out comes the rare bird's name, and a description of it.

That old person's accumulated experience is important for the whole tribe's survival. For instance, in 1976 I visited Rennell Island in the Solomon Archipelago, lying in the Southwest Pacific's cyclone belt. When I asked about consumption of fruits and seeds by birds, my Rennellese informants gave Rennell-language names for dozens of plant species, listed for each plant species all the bird and bat species that eat its fruit, and stated whether the fruit is edible for people. Those assessments of edibility were ranked in three categories: fruits that people never eat; fruits that people regularly eat; and fruits that people eat only in famine times, such as after-and here I kept hearing a Rennell term initially unfamiliar to me-after the hungi kengi. Those words proved to be the Rennell name for the most destructive cyclone to have hit the island in living memory-apparently around 1910, based on people's references to datable events of the European colonial administration. The hungi kengi blew down most of Ren-nell's forest, destroyed gardens, and drove people to the brink of starvation. Islanders survived by eating the fruits of wild plant species that normally were not eaten, but doing so required detailed knowledge about which plants were poisonous, which were not poisonous, and whether and how the poison could be removed by some technique of food preparation.

When I began pestering my middle-aged Rennellese informants with my questions about fruit edibility, I was brought into a hut. There, in the back of the hut, once my eyes had become accustomed to the dim light, was the inevitable, frail, very old woman, unable to walk without support. She was the last living person with direct experience of the plants found safe and nutritious to eat after the hungi kengi, until people's gardens began producing again. The old woman explained to me that she had been a child not quite of marriageable age at the time of the hungi kengi. Since my visit to Rennell was in 1976, and since the cyclone had struck sixty-six years before, around 1910, the woman was probably in her early eighties. Her survival after the 1910 cyclone had depended on information remembered by aged survivors of the last big cyclone before the hungi kengi. Now, the ability of her people to survive another cyclone would depend on her own memories, which fortunately were very detailed.

Such anecdotes could be multiplied indefinitely. Traditional human societies face frequent minor risks that threaten a few individuals, and they also face rare natural catastrophes or intertribal wars that threaten the lives of everybody in the society. But virtually everyone in a small traditional society is related to each other. Hence it is not only the case that old people in a traditional society are essential to the survival of their own children and grandchildren. They are also essential to the survival of the hundreds of people who share their genes.

Any human societies that included individuals old enough to remember the last event like a hungi kengi had a better chance of surviving than did societies without such old people. The old men were not at risk from childbirth or from the exhausting responsibilities of lactation and child care, so they did not evolve protection by menopause. But old women who did not undergo menopause tended to be eliminated from the human gene pool because they remained exposed to the risk of childbirth and the burden of child care. At times of crisis, such as a hungi kengi, the prior death of such an older woman also tended to eliminate all of her surviving relatives from the gene pool-a huge genetic price to pay for the dubious privilege of continuing to produce another baby or two against lengthening odds. That importance to society of the memories of old women is what I see as a major driving force behind the evolution of human female menopause.

Of course, humans are not the only species that lives in groups of genetically related animals and whose survival depends on acquired knowledge transmitted culturally (that is, nongenetically) from one individual to another. For instance, we are coming to appreciate that whales are intelligent animals with complex social relationships and complex cultural traditions, such as the songs of humpback whales. Pilot whales, the other mammal species in which female menopause is well documented, are a prime example. Like traditional hunter-gatherer human societies, pilot whales live as “tribes” (termed pods) of 50 to 250 individuals. Genetic studies have shown that a pilot whale pod constitutes in effect a huge family, all of whose individuals are related to each other, because neither males nor females resettle from one pod to another. A substantial percentage of the adult female pilot whales in a pod are postmenopausal. While childbirth is unlikely to be as risky to pilot whales as it is to women, female menopause may have evolved in that species because nonmenopausal old females tended to succumb under the burdons of lactation and child care.

There are also other social animal species for which it remains to be established more precisely what percentage of females reach postmenopausal age under natural conditions. Those candidate species include chimpanzees, bono-bos, African elephants, Asian elephants, and killer whales. Most of those species are now losing so many individuals to human depredations that we may already have lost our chance to discover whether female menopause is biologically significant for them in the wild. However, scientists have already begun to gather the relevant data for killer whales. Part of the reason for our fascination with killer whales and all of those other big social mammal species is that we can identify with them and their social relationships, which are similar to our own. For just that reason, I would not be surprised if some of those species too turn out to make more by making less.