The Declining Force of Natural Selection

The “declining force of natural selection” is the foundation of two classical theories of aging. The phrase means that what happens when an individual is young contributes more to its fitness than what happens later in life. A new study seeks to test whether aging tends to evolve in the direction predicted by this theory, and the results have raised some eyebrows.

Setting the stage

For more than a century, it has been recognized that aging presents a conundrum for the theory of evolution. Aging reduces individual competitiveness and cuts off reproduction via senescence or death. Aging decreases fitness. How could aging have evolved?

There have been two sorts of answers, one based in the fitness of the individual, the other in the fitness of the community. Since aging is irredeemably bad for the individual, the community-fitness idea might seem to be a natural candidate. But most evolutionary biologists today are skeptical of the concept of community fitness; they believe that everything in evolution should be explainable in terms of the selfish gene. The classical hypothesis (from individual fitness alone) is based on the declining force of natural selection. This idea was first articulated by the British Nobel laureate Peter Medawar in a seminal book in 1952. Medawar started from the idea that death rates in nature tend to be high, just because it’s a rough-and-tumble world out there. For reasons that have nothing to do with aging, very few individuals actually live long enough to die of old age. So natural selection has rewarded animals that fight hard and reproduce fast. Even if that causes them to age more rapidly, it won’t matter much in the end because almost everyone will die of something else before they ever get to the age where aging could affect them.

An alternative answer which I and others have promoted is that aging isn’t about individual fitness at all, but rather the fitness of communities and populations. In nature, death rates tend to be clumpy – meaning that either everyone is dying or hardly anyone. If there is a famine or an epidemic, then the death rate is very high and everyone is at risk. But in conditions of plenty, when no new diseases threaten, the death rate can be very low. This clumpiness leads to boom-or-bust population cycles, in which population can grow to unsustainable conditions of crowding, until, too late, the population comes crashing down and extinction is a real risk. The population-based theory is called the Demographic Theory of Aging, and it states that the evolutionary significance of aging is to help level the population death rate in good times and in bad, avoiding population overshoot and reducing the risk of extinction.

Two theories face off

How can we tell which is correct? There is one area where the two ideas clearly make opposite predictions: How ought life span to evolve in response to the “background” death rate? When many animals are dying young, for whatever reason, the individual-based theory says that life span should evolve to be shorter, because the declining force of natural selection is declining so much the faster. But the theory from community fitness predicts the opposite effect: Aging responds as a complement to the background death rate. When the background death rate is low, then life span evolves shorter so that the death rate from aging can take up the slack; when the background death rate is high, then longer life span evolves as aging takes a vacation because there is no need for death from old age when other natural forces are already trimming the population.

(The Caloric Restriction effect is an easy prediction from the Demographic Theory. When there is a famine and everyone is starving, aging slows way down because the community needs to save every individual life. But when food is plentiful, that is the time when the risk of overpopulation looms, so aging occurs more rapidly, in order to thin the population.)

The individual-based theory and the communal-fitness theory make different predictions. Where does nature weigh in?

Two studies, two answers

The canonical experiment was done by Steven Austad as his dissertation project in 1993.  He compared two populations of opossums, one on the mainland that was ravaged by predators, and the other on an island where the opossums had lived without predators for thousands of years. The island opossums lived twice as long as the mainland opossums, supporting the classical idea of Medawar and the individual model.

But a decade later, David Reznick published a study of guppies in the river pools of Trinidad that reached the opposite conclusion. He identified adjacent pools, one with big fish that ate most of the guppies before they could grow up, and one with no larger fish at all, where the guppies grew unmolested. He captured guppies from the pool with low background death rate and from the pool with high background death rate, and brought them all back to his lab in Riverside, California where they could live out their lives in the safety of a fish tank. He found that the guppies from the site with the higher background death rate lived much longer than those from the lower death rate. Those fish that had the shorter life spans in the wild had longer life spans in the aquarium. This is the prediction of the community-fitness theory.

Reznick’s study got lots of attention, but wasn’t enough to overturn the strong presumption in favor of the individual-based theory.

New study from Sweden

Just last year, two researchers from Sweden threw their particular gasoline on this controversy with a lab study of roundworms. In one branch of their experiment, 85% of the worms were selected at random and killed. In the other branch, the temperature was raised gradually until 85% of the worms died of overheating. Hwei-yen Chen and Alexei Maklakov found that the first procedure led to agreement with the classical, individual theory. Life span evolved to be shorter. But under the second regimen, the worms evolved a longer life span.

Interpreting the results, David Dowling remarks that nature is more likely to emulate the second procedure. Individuals don’t tend to die at random, but rather because some environmental challenge takes out those individuals that can’t hack it. He notes that given the positive correlation between heat resistance and longevity, it is inevitable that the second procedure should lead to longer life spans.

But this is a genetic argument, not an evolutionary argument. From an evolutionary perspective, we must never expect that heat resistance is positively associated with longevity. Quite the opposite: the individual theory predicts that it should not be possible to extend life span without paying a price. If some worms live longer than others, they should be paying for it with a greater vulnerability to heat, or with some other weakness.

George Williams (whose name is associated with the classical theory for evolution of aging even more than Peter Medawar) predicted in 1957 that “Low adult death rates should be associated with low rates of senescence, and high adult death rates with high rates of senescence.” He was confident this must be true because he believed that individual fitness trumps group fitness every time, and so every gene that evolves must make a positive contribution to individual fitness. The only way a gene that shortens life span could evolve (in the Williams theory) would be if the primary function of the gene is to increase early survival or fertility. Every aging gene must entail a tradeoff, strong enough so that the net result is to benefit the individual.

Dowling, typically, doesn’t mention the alternative Demographic Theory. But he does duly remark that the genetic correlation is unexpeced and difficult to understand in terms of the classical theory. (Sometimes the cost is found as lower fertility, but Chen and Maklakov were careful to look for incidental effects on fertility, and found that fertility of the long-lived worms was not significantly different from the short-lived worms.)

The bottom line

The key question is whether there is a longevity cost to surviving and reproducing. Selfish gene theory demands absolutely that there must be a cost, but Chen and Maklakov suggest that sometimes there is none. The worms that were more robust to heat shock also lived longer. The more realistic branch of their experiment showed the result predicted by the theory of communal-fitness.

Score one for the Demographic Theory of Aging.

 For basic information about healthy living for a long life,
see the author’s permanent page at

9 thoughts on “The Declining Force of Natural Selection

  1. Hey Josh….
    My vision is terrible ! ! ! (+3.25 glasses) I couldn’t find SkQ anywhere, to purchase. Can you send me some info on those drops? Thanks so much……… ( =

    • SkQ is available by special order at for research only. It comes in a powder, but is soluble in DMSO. Must be kept refrigerated on a daily basis, or frozen if kept for a long time.
      I have not ordered any and don’t know the cost. Let us know if you order some and how it goes.

  2. I support the classical selfish theory. Sure we can bring some wild animals into the lab and see how long they can live, and chart their longevity. However, in the wild, due to predators, disease and accidents, no individuals ever live long enough to grow old. Go out into the woods; you won’t find any “old” animals there. In digging up a 5,000 year old grave site in Europe, they found no individual bodies who lived over 30 years of age. Therefore, evolution cannot select for individuals with a longer lifespan.

    • This is actually the mother of all accepted aging theories today, an idea put forward by Peter Medawar in a series of lectures made into a monograph in 1952, “An Unsolved Problem in Biology.”

      Decades later, the idea was thoroughly disproven by surveys of ages at death in the wild of fish, birds, and insects. Indeed, aging causes enough animals die earlier than they would have without it that there is a large fitness cost to aging that evolutionary biology is on the hook to explain.

  3. Does an individual have to pay a price to live longer? I think not. If aging is caused by some aging genes turning on after reproduction, then the only cost is to turn off the aging genes, which probably would be an energy savings, not a cost.
    As I see it, we should not be looking at the individual; we should be looking at the species. Since 99.999…% of all species that have ever lived are now extinct; evolution may be more interested in the fittest species as a whole, than in the fittest individual. If we accept as a fact that each species is in competition with one or more other species, then we must think of survival of the fittest species. So if lengthening or shortening the lifespan makes the species more fit, in comparison to other species, then it survives, otherwise it may become extinct. It does no good for an individual to be the fittest, if its species becomes extinct. By requiring sexual reproduction, evolution mostly ignores the fittest individual. The only way to save the fittest would be to clone it. What evolution wants is a species with a large variety of characteristics; because there are many micro environments and micro climates available. A species with a large variety of adaptations will fill more micro environments and micro climates, than it would if it just cloned the one fittest individual; since that individual is not the fittest in all the various micro environments and micro climates.
    For example: A species of Panda bear that only eats one species of bamboo, may contain many very fit individuals for living off of that one bamboo; but the species itself is ripe for extinction.

  4. >”How ought life span to evolve in response to the “background” death rate? When many animals are dying young, for whatever reason, the individual-based theory says that life span should evolve to be shorter, because the declining force of natural selection is declining so much the faster.”

    George Williams said that, and it’s still a really common impression, but apparently it’s false. Age-independent changes to extrinsic mortality don’t affect the rate of decline in Hamilton’s indicators; only age-dependent ones do.

    (I haven’t fully grokked it, but the reason has something to do with: an age-independent extrinsic mortality increase, has affects the lx curve, but it also affects r, and these effects cancel each other out.)

    Caswell, H. “Extrinsic Mortality and the Evolution of Senescence”, TREE (2007).

    • “Apparently it’s false” is an overstatement. This paper was written after the fact, when situations were discovered where higher intrinsic mortality leads to longer lifespans. For the simplest situations, George Williams’s prediction is valid.

Leave a Reply

Your email address will not be published. Required fields are marked *