Examples of evolvability traits include
- Sex, the mixing of genes, permits evolution to experiment in parallel, trying many different combinations of traits at once.
- The hierarchical organization of the genome allows for modularity in development, so that organs and appendages can be shuffled, added and deleted without having to re-invent them in their entirety each time.
- Different rates of mutation in different parts of the genome mean that the core metabolism can be protected from disastrous tinkering, while more contingent details of biochemistry are subject to experimentation.
- Population diversity is an evolvability trait. A population with no diversity at all is not subject to natural selection. Back in the 1920s, R.A. Fisher proved the Fundamental Theorem of Natural Selection which says that the evolutionary rate of increase of fitness in a population is proportional to the variance of the fitness within the population.
- Aging (and a shorter life span) contribute to evolvability because (1) generation time is shorter; the population turns over more rapidly; all evolutionary change happens that much more quickly, and (2) diversity is promoted because no individual gets to go on reproducing for too long, dominating the next generation with its own progeny.
It is undeniable that evolvability has been a product of evolution. And yet, this doesn’t jive with the mainstream of evolutionary science. The body of evolutionary theory developed in the 20th century implies that natural selection ought to be nearsighted. The success of a gene in penetrating a population ought to depend largely on its consequences for the viability and reproductive success of its individual bearer. The gene’s long-term effect for progeny in an entire community or species ought to be a far less potent influence on natural selection. This is because mutations appear first in a single individual. The first test it must pass: can it spread to dominate a local population deme*? Only a gene that succeeds at this level can ever be tested for its long-range effect on the population.
It’s a fact that evolvability has evolved. We may not be able to describe the mechanism, or explain how evolvability traits survived the “first test”, but somehow the process must have worked.
Much about aging in the biosphere points to the inference that aging has been positively selected**. Nevertheless, evolutionary biologists have resisted this interpretation because it is theoretically implausible. Aging is bad for the individual fitness, and the individual counts for more than the community when it comes to natural selection. It is considered inconceivable that aging could have passed the “first test” in order to come to dominate a local population.
It’s even worse than that. For models based on kin selection, a trait can be selected despite a cost to the individual if its benefits are focused on others that are likely to bear the same gene. But the only benefit of aging is that it leads to death that creates a vacancy in the niche. This vacancy could be filled by a close relative or a distant relative or no relative at all, an animal that doesn’t age or even an animal of a different species that shares some of the same food species. Kin selection and even MLS models are not promising for evolution of aging.
But when we realize aging contributes to evolvability, and that other evolvability attributes managed to evolve, we may ask, “why not aging, too?” We may not be able to imagine exactly how aging was affirmatively selected, but the same may be said for sex and organization of the genome. Whatever mechanism served to evolve these things might have worked equally well to evolve aging.
This lends credibility to the oldest hypothesis (attributed to Weismann, 1892) about how aging might have evolved. It doesn’t resolve the solve the “first test” objection, but discredits the objection by association. Several of my colleagues have promoted this theory of aging, for example V. Skulachev, G. Libertini, and J. Bowles. A. Martins has published a computer simulation of how it might work
What do I think of this idea? I think it’s part of the picture, but not the first part. Let me explain.
The need for demographic stability
I wrote a few weeks ago about the Demographic Theory of Aging. The punch line is that no community of animals can afford to trash its own ecosystem by eating everything in sight and reproducing without restraint. Simulations, theory, and field observations all agree. The consequences of overpopulation are swift and devastating. I told you the story of the Rocky Mountain Locust. Here’s another story, about reindeer introduced to the Isle of St Matthew in the Bering Sea in 1944: There were no large animals on the island until introduced by man.
The reindeer flourished, their population growing by about a third from each season to the next. That may sound like an extraordinary rate, but the ability of the population to expand rapidly is beneficially adaptive in an empty niche, and may be a life-saver after a natural disaster. So the reindeer population followed a trajectory typical of an exotic species that is successfully introduced, growing on an exponential trajectory. Naturalists estimate the carrying capacity of the island at about 2,000 reindeer, and the population crossed that threshold around 1960.
Such is the relentless logic of exponential growth that just four years later, the population was 6,000 reindeer. The winter of 1964 was severe – not a dramatic departure from what the reindeer expected, but more snow than usual. By the end of the winter, the entire population had starved to death. An expedition the following year counted 42 stragglers (and shot 10 of them in the name of sport and science). Reindeer live typically 18-22 years, so the entire saga had unfolded within the lifetime of a single reindeer.
(from Suicide Genes, forthcoming by Josh Mitteldorf)
Population overshoot can wipe out a population swiftly and efficiently. It is the most potent, most credible and most direct form of group selection. It is also the perfect counterpoise to the “selfish gene”, which measures fitness according to individual reproduction rate.
So here’s the story, as I see it: Population control is an essential function of animal life. (Less so for plants – the higher up the food chain you go, the stronger is the pressure to preserve the ecosystem that you are sitting on.) Population control is essentially a group function. It is the reason that the selfish gene provides such a distorted picture of reality. Some plants may be evolved for maximal reproduction, but animals are evolved for a flexible rate of reproduction matched to the overall death rate.
Aging fits well within this picture. Aging tempers population growth and does so in a way that responds flexibly to demographic conditions. In other words, when everybody is starving, no one is dying of old age. But even better: the body responds to conditions of starvation by becoming stronger and more robust, slowing the aging process, doing everything the metabolism can do to survive through the famine.
Putting it all together
It is easy to understand population control, and how it evolved. It is harder to understand how evolvability arose, and how natural selection has favored it – but we know for a fact that it did evolve. My hypothesis is that both are involved in the evolution of aging. First to arise was population control. The race to reproduce as fast as possible was regulated and reined in. Individuals learned to temper their predation and their reproduction in order to protect a common food supply.
Population control can be achieved either by limiting fertility or by limiting life span, or any combination of the two. So stable ecosystems might have been achieved without any aging at all, but solely via a flexibly responsive birth rate. But the choice between lowering birth rate and raising death rate is aided by the need for evolvability. From the standpoint of evolvability, short life span with high fertility is much better than long life span with low fertility. In fact, the pace of evolutionary change is directly proportional to the rate of population turnover. More births with a shorter life span is much to be preferred.
So population control evolved as a mixture of limited fertility and limited life span. But once this was established, and the rules were stamped into the genome that prevented unrestrained reproduction, then there was room for natural selection to be responsive to more subtle considerations, including those that act on a long time. This resolves the mystery of why evolvability and other group-selected adaptations have been so effective. Once the tyranny of the selfish gene is tempered by the powerful and immediate need for stable ecosystems, there was room for more subtle selective forces, acting over a longer time frame. There was room for evolvability to emerge, in a self-reinforcing positive feedback loop that has made the evolutionary process itself so surprisingly effective.
* A deme is a local breeding population, a community of a single species that is mutually sharing genes together.
**For example, the body seems to be able to slow down aging when stressed, indicating that aging is metabolically avoidable. For example, there are affirmative mechanisms for self-destruction at the cellular level (apoptosis and telomere attrition ) that are associated with aging of the whole body. For example, there are genes that regulate aging that have been preserved by evolution at least since the Cambrian Explosion half a billion years ago. I’ve written a great deal on this subject, including a forthcoming book and a recent book chapter.
This post previously appeared on Josh’s blog here: http://joshmitteldorf.scienceblog.com/2013/07/22/evolution-of-evolution-and-evolution-of-death/