Whence comes death? How natural selection has fashioned aging and death; why "natural" isn't the last word in healthy foods and lifestyles; and some observations on objective science pursued by subjective humans.

One doesn't have to entertain faith in an omnipotent deity to seek comfort from the idea that nature is benevolent. The notion that our bodies are programmed to self-destruct with age is profoundly unsettling to many. As we recognize that all creatures fall victim to decrepitude and death, we imagine our human fate to derive from some immutable physical law to which our bodies eventually succumb despite all efforts to keep us strong and healthy.

But aging is no physical necessity, nor is it an accident; it is a part of life's developmental program, following birth, growth, maturation, and reproduction as night follows day. This fact is becoming clear to developmental biologists, as the mechanisms of aging are elucidated in the laboratory. Some of these mechanisms can act precipitously, as when the salmon dies after spawning or a marigold plant dries up after its flowers go to seed. Some mechanisms are as old as sexual reproduction--which is to say they have been selected and preserved over 500 million years of biological evolution. Most can be modulated at will by metabolic signals, as when half-starved mice live to extraordinary lifespans. And some recent laboratory investigations have located single genes that seem to have no other purpose than to act as a time bomb, assuring the bearer's demise.

Humanity's earliest conception of aging was that the process was akin to physical wear and tear. Knives get dull--why shouldn't our teeth? Wheels get rusty and squeak when they turn--isn't that what happens to our joints? It's a theory with a great deal of intuitive appeal. But after the mid-nineteenth century, when the intuition that things wear out was codified as the second law of thermodynamics, this position became untenable. The second law is indeed related to the reason why knives gets dull and bearings rust; but living matter is not subject to the same constraint. A living organism extracts energy from its environment and thereby is able to maintain--even to increase-- order within itself, while dumping its entropy back into the environment. The total entropy--organism plus environment--is bound by the second law to increase; but as long as it can take in food and unload its waste, the living organism performs a magic that non-living things cannot match: it can grow, it can repair, it can even build itself from scratch, starting with a DNA blueprint and necessary nutrients. Continually replenished by food or by sunlight, there is no physical necessity for anything about it to deteriorate with age.

Every living metabolism includes highly developed machinery for repair and maintenance. You know that when you break a leg the bone grows back together. It's less obvious that the DNA in your cells is constantly being checked for errors: specialized molecules creep along the length of the DNA strand searching for pieces out of place and repairing them on the spot. You can recover completely from a cut in your finger, but if the finger is severed from your hand, you'll never grow a new one; this limitation is standard in mammals. But a squid can replace a severed tentacle, and if a starfish loses an arm not only does the starfish grow another arm but the arm will grow a whole new starfish!

Imagine how complete and how robust is the starfish's system of regeneration. Then consider that the starfish ages and dies, with a lifespan of about eight years. Now the inadequacy of the theory that things wear out becomes apparent: evolution has devised for all of us a system of maintenance and repair that is remarkably efficient. It's a genuine curiosity that evolution, after creating such intricate and comprehensive systems of repair, refuses to deploy them in a way that would maintain the body in a state of peak performance. This is the classical problem of aging, the problem that has baffled and confused generations of evolutionary theorists.

In the 1890s, when the two sciences of thermodynamics and evolution were yet new, August Weismann realized that there was no physical necessity for the body to degrade over time. Weismann was the first to speculate that the evolutionary reason for aging and death has something to do with the good' of the species. He called it "making room" in the environment. When their lease was up, the old would have the good grace to pack their bags and vacate the premises, leaving space for the young. The constant turnover of the population would help to maintain its diversity, and make the species more adaptable to changing circumstances.

But as evolution grew up as a quantitative science, Weismann's hypothesis began to seem untenable. In the early part of the twentieth century, the concept of fitness was first quantified as a time-weighted measure of an individual's rate of reproduction. How could aging and death make a positive contribution to an individual's reproductive success? If there is any benefit to aging, it accrues not to the individual but to the population as a whole.

In the 1960s, evolutionary theorist George Williams argued that benefits to the group are too weak a force to be invoked as an evolutionary explanation. Evolution is in the business of testing random mutations that first arise in a single individual. The first test for any new trait is whether it can progress from the individual to achieve prevalence in a group; only traits that passed this test can ever be tested in competition group against group. Hence individual selection will trump group selection every time, Williams said. Aging and the imposition of a finite lifespan may well have the potential to help maintain population diversity, but this benefit accrues only to the group. Aging harms the individual who ages and puts that individual at a selective disadvantage; hence aging can never be selected as an adaptation.

This interdict on arguments from group effects in evolution became a powerful current of evolutionary thought from the 1970s onward. It was difficult for a generation of naturalists trained in observation and classification to assimilate the quantitative methods of the new biology; but it was perhaps too easy to absorb, undigested, the message of Williams' book: selection sees only individual reproductive rate. Group selection is not a viable evolutionary force.

How, then, could the anti-group selectionists explain the evolution of aging? How would they account for its prevalence in the world of higher plants and animals? Williams himself put forth a theory that aging was an epiphenomenon, a side-effect of selection going for broke in its quest for faster reproduction. His theory of pleiotropy posited that reproduction and aging were genetically linked. In 1977 it was Thomas Kirkwood who put that hypothesis into its most common sensical and appealing form, which remains popular among evolutionists to this day. Everything the body does requires energy: activity, metabolism, reproduction, repair and maintenance. The body must ration the limited supply of available food energy, and myriad generations of natural selection have taught the body how to balance its energy budget to best effect.

In any such tradeoff, no one demand can be fully met. It follows that the body does a pretty good job of repair and maintenance, an adequate job to get the body through the period of intense reproduction. But thereafter, the evolutionary payoff for keeping the body in good repair becomes smaller, so damage is permitted to accumulate. Eventually this compromises the capacity of the system to get anything done--including repair. This theory of aging was called by Kirkwood the disposable soma--a name which derives from the fact that the individual's reproductive output and not the physical body is the basis of the target function of natural selection. The individual is willing to trade even her or his own body for enhanced reproductive prospects.

The disposable soma is a good theory. Straightforward and intuitive, it invokes nothing abstruse or mysterious, and it accounts for the universality of aging within the context of pure individual selection. The disposable soma has been king of the roost, the prevailing theory among experts in the evolution of aging for almost a generation. But there are deep problems with the theory that are becoming difficult to ignore. There are indications that aging is not a side-effect but a full-blown program of nature. The evidence may be leading us right back to Weismann's abandoned idea that the old are removing themselves to make way for the young.

These new hints that proponents of the disposable soma and other related theories may be on a dead-end path come from three places. First are the intriguing examples of ways that the body can circumvent aging when it is ecologically appropriate, as when the system is stressed. Second is the universality of aging mechanisms; some of the biochemical pathways of aging are common to such diverse species as humans and baking yeast! Third is the smoking gun, the discovery in recent years of "aging genes" that can be disabled in laboratory tests, causing animals to live long beyond their natural lifespans.

The body's capacity to forestall aging under stress suggests that in "normal" times the body may actually be engaged in a strategic surrender. We're all so familiar with the idea that physical exertion promotes a healthier heart, lower risk for some cancers, and a longer life that we may no longer pause to think how curious a phenomenon this is. When it is not distracted by the rigors of aerobic exercise, why is the body less able to deal with the ravages of age?

Plants succumb to aging just as animals do, and nowhere are the results more visible than in the annuals we plant in our gardens. At summer's end a marigold bush will wither and die after its flowers have all gone to seed, but if the flowers are snipped, the plant is reinvigorated to blossom again and again and again. It seems almost as though the plant finds security in the knowledge that it has passed on its genetic legacy and only then can it surrender to death's beckoning. The kind of accelerated aging that we observe in marigolds is typical of semelparous organisms whose life histories culminate in a single act of reproduction following growth. There are semelparous animals as well, including Pacific salmon and octopi, whose lifespans are subject to experimental manipulation. After laying her eggs, the female octopus stops eating and slowly starves to death.

Lest anyone doubt this is an example of programmed death, the locus of the program has been discovered in the animal's "optic gland." This organ can be surgically removed, after which the animal no longer knows it is supposed to starve itself; hence it survives to breed another season.

Recently, however, an even more striking experimental fact about aging has emerged from the backwaters of gerontological laboratories into the popular science press: the less an animal is fed, the longer it lives. With a diet rich in vitamins and minerals, while drastically reduced in caloric content, laboratory animals from crabs to monkeys have been observed to live almost twice as long as their fully-fed siblings. These animals not only live longer, they are more active and more resistant to disease, and the closer they are to starvation, the more resilient and robust they are.

It isn't difficult to guess the evolutionary provenance of this adaptation. It is essentially that populations under stress avail themselves of every possible advantage to avoid the finality of extinction. But the body's behavior under caloric stress highlights the converse question: how is it that when stress is absent, the body cares less well for itself? When resources are plentiful, why should the metabolism be trying less hard to forestall death?

Experiments in caloric restriction (CR) are a challenge for any of the "tradeoff" theories of aging, but the disposable soma theory is particularly vulnerable, exactly because it is premised on a tight calorie budget. The essence of the theory is that there's never enough food energy to do a perfect job of everything, so the body's allocation for repair and maintenance will always be shortchanged. How can the theory accommodate the fact that when energy is least available, repair and maintenance are at their most efficient--even as expenditures for activity and for immune function are also at their peak?

The second source of evidence that aging is an adaptation --a deliberate "creation" of natural selection--comes from ewly discovered data on the universality of aging mechanisms. Bacteria and algae don't age because their life cycle is so primitive that there is nothing to distinguish young from old cells. Higher, one-celled organisms like the amoeba and paramecium do experience aging, however. Much larger than bacteria, single-celled protists are distinguished by a more complex metabolism and DNA that is concentrated in a nucleus. Protists carry the mother of all aging mechanisms: a replication counter that will kill a cell line after it has procreated a fixed number of times. The tally is maintained in a repetitive tail on the end of every DNA strand, called the telomere. The DNA replication that takes place whenever a cell divides omits a stretch of that tail, and the chromosome becomes shorter with each cell division. The telomere is a buffer zone, completely expendable, so for awhile the shortening has no effect on the generational cells' viability. But eventually the telomere is used up, the chromosome's integrity is in danger, and the cell, in fact, refuses to replicate.

One curious thing about this process is that it appears to be entirely avoidable. All eukaryotes (cells with a nucleus, including protists and all animals and plants) carry the gene for the enzyme called telomerase. In the presence of telomerase, the entire chromosome is faithfully replicated, and the tail doesn't lose any length. A bit of reflection makes it clear that this capability is absolutely essential to the long-term continuity of life. But in replication of protist cells, the gene for telomerase is silenced just as it is needed most; instead, the telomeres are permitted to shrink with each replication. They are only replenished when the protist is joined with another in sexual union.

The primitive form of sex enacted by protists is called conjugation, and it is entirely separate from the reproductive process. Two protists of the same species come together, their cell walls dissolve at the boundary, and they join temporarily as one unit. Within the double cell, the two nuclei merge, and each chromosome bonds along its full length to its counterpart from the other cell. Then the chromosomes come apart, the two nuclei separate, and the double cell once again becomes two. The two individuals have lost their identity; the cells that emerge each contain a blend of cytoplasm from the two original cells, and, what is more important, the genes in the nuclei have been thoroughly mixed and exchanged. And, by the way, the telomeres are restored to their full and original length in this process, so each organism is authorized another several hundred rounds of asexual reproduction.

The purpose of this process--identical to the purpose of sex in higher plants and animals--is to exchange genetic material, giving selection an opportunity t6 work on new combinations of traits. The telomeres must be seen as a policing agent, enforcing the injunction to share. Each cell is endowed with the mandate to go forth and multiply, but subject to the proviso that it must find a mate and commingle its genetic legacy, at least every few hundred generations, on pain of death. There can be no doubt that this is an adaptation, a mechanism that has evolved because it serves a purpose.

Since conjugation creates new gene combinations, the purpose appears to be the imposition of diversity on every colony of protists via an imperative of genetic exchange. Reproductive success is rewarded up to a point, but no single cell line is permitted to dominate the culture, no matter how much better adapted it happens to be to the specific set of circumstances that happen to prevail at present.

Were the colony to lose that genetic diversity, it might be sorry tomorrow. The environment is constantly in flux, and if some novel adaptation is universally adopted in response to a transient change in the environment, the danger is that the environment may return to its former condition at a time when the adaptations perfect for that condition have been lost forever. The need for diversity in a population may seem to be a technicality, a detail too small to support a phenomenon as ancient and ubiquitous as aging, but diversity is the fuel for evolutionary change. Without diversity there is no natural selection, indeed there are no choices from which to select. There is deep appreciation in the scientific community for the importance of diversity, going all the way back to Charles Darwin.

(Darwin understood the ongoing need for population diversity but was mystified about the mechanism that was able to sustain it. This is a celebrated footnote in the history of science: unaware of the concurrent researches of the Austrian monk Gregor Mendel, Darwin had no concept of the genetic laws of heredity. He thought the traits of each offspring were some kind of average between those of the father and the mother. But in an averaging process, the extremes would be lost and each generation would tend to be more homogeneous than the last. How was it that diversity in nature did not collapse from all that averaging, bringing evolution to a screeching halt? This was a subject of passionate concern to Darwin, and a frequent refrain of his early scientific critics. Mendel's revelations would have resolved the issue handily, but busy and successful Darwin never found time to read the postal missives from the modest Franciscan.)

The story of telomeres may be a hint that population diversity is the fundamental meaning, the evolutionary purpose of aging and obligatory death. All higher life forms evolved from the protists and retain their reproductive counting mechanism as the first aging mechanism. In higher life forms, there are other mechanisms of aging, and it is uncertain whether telomeres contribute substantially to the process; but every cell in our bodies, every animal and plant cell in the world, is still programmed by its telomeres to reproduce only a finite number of times before it dies. Evolution has evinced an extraordinary capacity to generate variety while conserving that which is essential. If an aging mechanism has been conserved since the very first eukaryotes emerged 500 million years ago, this may be taken as a sign that aging is essential to some basic life function. But the fact that aging is fundamentally destructive, and offers no benefit to the individual presents us with a paradox and a mystery. What is the nature of this essential benefit? It is a good guess that it has to do with the viability of populations and the maintenance of diversity as fuel for the ongoing process of evolution.

The third storyline in support of aging as an adaptation comes from the gene splicing technology that is just now maturing. The tiny roundworm C. elegans embodies an aging process that is uniquely flexible: when food and water are plentiful its normal lifespan is just ten days, but under stress it can enter a state akin to hibernation and survive as a dauer for many weeks. Ten years ago, experiments with these worms first hinted at the genetic basis for their aging mechanism and their means of long-term preservation. A single gene was found that could be silenced, with the result that the worm lived to several weeks without becoming a dauer. The animals seem to suffer no ill effects from the change.

So here is direct evidence for the existence of aging genes in nature. The meaning of this discovery was not lost on the community of genetic scientists from whence it emerged, but evolutionary biologists were more skeptical. Perhaps the existence of this gene has something to do with the peculiar two-phase life strategy of the roundworm and shouldn't be. ascribed a general significance. But soon a gene was discovered that is connected with the insulin metabolism of yeast cells, the removal of which caused the cells to live longer. Remarkably, the insulin metabolism of yeast and of humans are clearly related, and the analogous human gene was quickly identified.

Then, in 1998, a Cal Tech laboratory found an aging gene in fruit flies, the workhorses of experimental evolution. The next year, discovery of the first aging gene in a mammal was announced by an Italian genetic laboratory. Mice were reported to live 30 percent longer when the p66 gene was disabled, and, like calorically restricted mice, the genetically altered mice evinced an enhanced resistance to stress.

What is more provocative yet, the function of the p66 gene is known to be related to programmed cell death--or apoptosis. Apoptosis is such an orderly and well-behaved process that it is widely recognized as an adaptation. But it had always been assumed that apoptosis was triggered only in infected cells or in cancerous cells in order to protect the rest of the organism from spread of disease. The suggestion that apoptosis of normal, healthy cells could be a mechanism of aging is tantamount to acceptance of aging as a purposeful life function, an evolutionary adaptation. And the existence of any single gene that extends lifespan without deleterious side effects poses an essential difficulty for the tradeoff theories, including disposable soma.

All these indications that aging is a developmental program have been well received and gradually integrated into the thought of developmental biologists, geneticists, and ecologists. Evolutionary biologists, however, still resist the message because it essentially contradicts their beliefs about group selection. A new theoretical framework is needed in which the subtle power of natural selection is acknowledged, in which populations and perhaps entire ecosystems may be conceived as engaging in coordinated evolution.

The solution may soon be at hand. In a developing paradigm shift, the whole rationale for the primacy of individual selection is coming under attack from two directions. Evolutionary ecologist David Wilson has for nearly thirty years been building a case that group selection is more viable than naive theory would indicate, and of late his followers seem to be gathering force. His multilevel selection theory offers a framework within which individual and group selection may share the stage. Year after year in the evolutionary literature, the circumstances under which group selection may make its effects felt seem to be expanding. To ecologists, this must come as no surprise; indeed, if the breadth and the ubiquity of cooperative phenomena in nature is a reliable indication, then group selection must be a primary force of nature.

The other attack on the proscription of group selection comes from computer simulations. With the rise in power and availability of computers in recent years has come an explosion in computer modeling of evolutionary processes. It was a tide of rigorous mathematical thinking that brought the population geneticists to supremacy over the old school of descriptive naturalists a generation ago, and it may be that the next wave will elevate the computer modelers, with their messages from the school of chaos theory.

In the old-style analysis, the spread of a gene or a lineage was described as a smooth and gradual process, incrementally taking over a population. The computer simulations are better able to model the randomness and the uniqueness of genetic mutations, as well as the complexity of interactions among individuals spread out on a landscape. In computer models, individuals have been observed to organize spontaneously into loose, temporary groupings, enhancing the prospects of any trait that benefits others at the expense of the self. This is the essential insight of chaos theory: that a small number of discrete, random events may determine the outcome of a process, and that outcome may be different from what one would conclude from tracking a smooth and gradual progression of population averages.

It may turn out that analysis and simulation are complementary tools, each offering a different set of insights into evolutionary dynamics. Will next year's computer model demonstrate ways in which populations with limited lifespans are more diverse and robust, so that they win out in competition with hypothetical ageless populations? This would signal a renaissance for Weismann's century-old idea that the old are bowing out to make room for the young.

If we come to realize that evolution works not just on individuals but on communities, if we deeply absorb the message that our bodies have evolved to degrade with age and to die--martyrs all to the communal cause--how does that affect the way we think about ourselves and our relationship to nature?

For one thing, we may revise downward our opinion of "natural" foods and remedies. We may never have thought about it in this way, but at root the appeal of the natural comes from faith in evolution: what is natural is part of the environment in which humans and their ancestors evolved; hence we are presumed to be well adapted to it. If natural foods are better for us, it is because they are the foods that evolution has equipped our bodies to assimilate.

Sometimes nature wants what's best for us individually, but sometimes it conspires with our bodies to do us in (all for the common good, of course). To extend human life beyond our "natural" lifespan--even to address classic ailments of old age--may require drugs or treatments far removed from anything found in nature. If you're a nonsmoker, eating less may be the most effective step you can take to enhance your health and extend your life; but you'll seek in vain for the "natural" instinct that supports your willpower when you're denying the body's appetite. We have evolved to enjoy food so that we will put away a comfortable layer of fat for the lean winter which may ever be just around the corner. Nature wants us to eat when there's food to be eaten, but nature wants us to die after a programmed lifespan. If you or I have our own ideas about how long we want to live, we may have to play some very unnatural tricks or even to do battle with nature in order to get there.

The story of evolution and aging offers another moral, an uplifting and hopeful message of cooperation, The economic culture of our generation has embraced unbridled individual competition. For a few, this has unleashed a restless quest for the accumulation of wealth while the many suffer a disquieting economic decline. Perhaps the parallel trends in scientific and political thought are no accident: just as our nation has discovered that pure competition is the one perfect economic system, the academic establishment of evolutionary biology has decreed that natural selection knows only what's good for the individual, that there is no such thing as "group selection," and that "evolutionary altruism" is really an illusion. Both these intellectual trends may prove ephemeral. Then perhaps we can learn from evolution's insistence on tempering pure competition with universal mortality that diversity of the population is essential to the long-term health of a community. Perhaps this message can help us to strip away the illusion that our life's purpose is individual achievement and restore to us a sense of our common destiny.

The cycle of life and death can only have meaning in the context of a great evolutionary progression that will carry our descendants into realms of being and experience far beyond our present imaginings. It is inspiring to reflect that we have evolved to live communally, to share and to love, and, when our time is up, to sacrifice our very lives that the community might continue to change and to evolve.

Joshua Mitteldorf lives in Philadelphia and studies evolution of group behaviors with the aid of computer simulation for the Biology Department of the University of Pennsylvania. In private life he is a musician and environmental activist.