Dunking on medawar selection shadow
Peter Mernyei asks:
Would there really be any evolutionary pressure to keep the body alive longer in old age at the cost of increased frailty? I’d have thought in an ancestral environment you’re likely to die anyway if you’re making either choice, and in any case you’re very likely not reproducing anymore. So my naive expectation would be that evolution just programs whatever works for younger bodies that it actually “sees clearly” and it “doesn’t think about” what happens to older bodies very much.
What you’re describing is known in biogerontology as Medawar’s “selection shadow”. It’s a very intuitive concept and widely used in evolutionary modeling, but I personally find its explanatory power very limited for looking at how evolution actually selects for longer lifespans.
Selection shadow is pretty good at explaining why organisms don’t live literally forever under heavy extrinsic mortality, but we know that our early mammal ancestors couldn’t make it past 10 years even without extrinsic mortality, and we can now live for 100 years. Clearly, at some point evolution did in fact care about “not dying of cancer at 10 years old”, and it there was some feature under selection pressure that naturally alters “in-shadow” lifespan”
What was that feature? One likely candidate is the chance of surviving to age of reproductive maturity.
Assume you have a species with no age-related “defensive frailty” at all. They’re as youthful-looking, strong and sharp at 60 as they are at 20.
Under the lens I’m describing, their main intrinsic cause of death will be cancer. One day you’re looking good, and next day an unlucky mitotic event made one cell a perma-replicator that took the entire organism out.
Assuming one-hit cancers (no extra tumor suppressor defences at all), you’ll get a species with a constant cancer + extrinsic mortality rate. Each year there’s some constant chance to get cancer and die. And a constant chance to be wounded by predator so bad you die. You can model the population like radioactive particles with a known half-life. A very small group of them could have an extremely long lifespan (100, 1,000, even 10,000 years) by luck, but most will die soon after birth.
Is this an evolutionarily stable state for this population?
I don’t know. For some hypothetical populations, it could be, if their reproduction is extremely fast (on the order of cell division speed). This might be the case for some actual primitive multicellular colonies.
But if getting to reproductive age takes a few years, a likely course of events is that one organism in this population will get itself a mutation that produces a tumor suppressor with antagonistically pleiotropic effects: it will reduce cancer rate (now each cell needs two hits instead of one) in exchange for late-life regeneration loss (frailty). Think How could this mutation improve fitness?
Well, now this organism’s odds of getting an early life (childhood) cancer is lower. This means it has better chances of getting to the age of maturity and reproducing. But the mature organisms are now more susceptible to wounds and predators taking them out the longer they live. So we have an evolutionary pressure shifting causes of death from 100% constant 0% age-dependent to 99% constant 1% age-dependent.
And notice that this rebalancing of mortality rates decreases your chances to live 1,000+ years even as it increases your chances to live 10 years.
Organisms adapt to new ecological niches, and if these niches reward slow life strategies, then evolution will select for longer lifespans as part of these strategies (you can’t run a slow life strategy as a rodent that gets cancer at 2 years old).
Slow life strategies here are things like memory formation, sociality, caring for the young, predator avoidance, niche construction, basically anything where benefits to your kin compound over organismal time. These adaptations tend to be associated with longer lifespans nature-wide.