Theory of evolution unfortunately does not give advice about how to avoid going extinct, just as "Buy low sell high" is not advice about picking stocks.
Evolutionary theorists have found a collection of rules of thumb that could help at avoiding extinction. One rule of thumb for animals is to be individually small. Large animals have a harder time maintaining large numbers. A given biome can support fewer of them, so they are likely to survive in smaller numbers, and smaller populations are in general less able to respond to natural selection. Human beings are among the largest animals in existence. If you look at the number of species that on average grow larger versus the number of species that on average is smaller, there are few larger animals. Long generation times is also a problem. The slower you reproduce the slower you adapt, and again we are among the slowest. However, in the short run we have managed to create a tremendous population considering our size and metabolic requirements -- we are doing very very well ... in the short run.
Another rule of thumb is to divide a large population into smaller subpopulations with limited genetic exchange among them. A lot of species do this, though for some if them it may be a constraint imposed by their environment. There are various theoretical reasons why this could be good. For one, populations can afford it. In large populations with random mating, it takes only twice as long for a favorable mutation to approach fixation in a population of a trillion as in a population of a million. So reduced exchange between subpopulations won't slow things down too drasticly. But also, there are potential problems with segregation distorters. If a gene can arrange to be present in all of a female's offspring instead of half, that's a giant selective advantage -- enough to overcome a hefty disadvantage in survival. Imagine a gene like that which is lethal when homozygous -- in a large population it spreads fast and then when it is very commen it causes a population crash. Ouch! Better it gets the chance to do that to a single subpopulation before it spreads.
A gene is selected when its frequency in the population increases relative to alternative genes. But just like "buy low sell high" is not the only way to win at the stock market .... (Some buy bankrupt companies for far less than book value and then sell the remnants at a hefty profit after the legalities are settled. Some buy stock that produces large tax-free dividends for them. Etc etc.) A gene is not necessarily selected by improving the survival or fecundity of individuals that carry it. It can increase for other reasons. Evolution is not necessarily progressive. It appears that some species are better than others at making their evolution work for them. Perhaps insects are so wildly successful on land because they are good at evolving into new species that fit new ecological niches, while others that perhaps start with a better body plan are simply not as adaptible.
And as we see with some invasive species introduced to new environments, it is not enough to spread easily in the new environment. To win a species must also be able to survive well in the environment it creates by its initial success. The Black Death adapted to new rodent hosts and from there to human hosts. Each succeeding epidemic wave individually evolved to spread by breath instead of requiring fleas to transmit it. But each succeeding wave probably left no survivors from the germs of the epidemic. The plague which was endemic in some rodent populations reseeded from scratch for each epidemic. After all the sound and fury, nobody won that evolutionary race.
Evolution can in fact force species into evolutionary dead ends and into extinction. That's just something to live with. Or to go extinct with.
Evolutionary theory can describe what happens in evolution at some level of abstraction. Similarly, statistical mechanics can describe what happens in a gas, at a level of abstraction. It would be nice if evolution theory could tell you how to maximise your reproductive success just as it would be nice if your stockbroker could and would tell you how to make millions of dollars. But it does not work that way. And yet it can have some value.
Evolution
Evolutionary theorists have found a collection of rules of thumb that could help at avoiding extinction. One rule of thumb for animals is to be individually small. Large animals have a harder time maintaining large numbers. A given biome can support fewer of them, so they are likely to survive in smaller numbers, and smaller populations are in general less able to respond to natural selection. Human beings are among the largest animals in existence. If you look at the number of species that on average grow larger versus the number of species that on average is smaller, there are few larger animals. Long generation times is also a problem. The slower you reproduce the slower you adapt, and again we are among the slowest. However, in the short run we have managed to create a tremendous population considering our size and metabolic requirements -- we are doing very very well ... in the short run.
Another rule of thumb is to divide a large population into smaller subpopulations with limited genetic exchange among them. A lot of species do this, though for some if them it may be a constraint imposed by their environment. There are various theoretical reasons why this could be good. For one, populations can afford it. In large populations with random mating, it takes only twice as long for a favorable mutation to approach fixation in a population of a trillion as in a population of a million. So reduced exchange between subpopulations won't slow things down too drasticly. But also, there are potential problems with segregation distorters. If a gene can arrange to be present in all of a female's offspring instead of half, that's a giant selective advantage -- enough to overcome a hefty disadvantage in survival. Imagine a gene like that which is lethal when homozygous -- in a large population it spreads fast and then when it is very commen it causes a population crash. Ouch! Better it gets the chance to do that to a single subpopulation before it spreads.
A gene is selected when its frequency in the population increases relative to alternative genes. But just like "buy low sell high" is not the only way to win at the stock market .... (Some buy bankrupt companies for far less than book value and then sell the remnants at a hefty profit after the legalities are settled. Some buy stock that produces large tax-free dividends for them. Etc etc.) A gene is not necessarily selected by improving the survival or fecundity of individuals that carry it. It can increase for other reasons. Evolution is not necessarily progressive. It appears that some species are better than others at making their evolution work for them. Perhaps insects are so wildly successful on land because they are good at evolving into new species that fit new ecological niches, while others that perhaps start with a better body plan are simply not as adaptible.
And as we see with some invasive species introduced to new environments, it is not enough to spread easily in the new environment. To win a species must also be able to survive well in the environment it creates by its initial success. The Black Death adapted to new rodent hosts and from there to human hosts. Each succeeding epidemic wave individually evolved to spread by breath instead of requiring fleas to transmit it. But each succeeding wave probably left no survivors from the germs of the epidemic. The plague which was endemic in some rodent populations reseeded from scratch for each epidemic. After all the sound and fury, nobody won that evolutionary race.
Evolution can in fact force species into evolutionary dead ends and into extinction. That's just something to live with. Or to go extinct with.
Evolutionary theory can describe what happens in evolution at some level of abstraction. Similarly, statistical mechanics can describe what happens in a gas, at a level of abstraction. It would be nice if evolution theory could tell you how to maximise your reproductive success just as it would be nice if your stockbroker could and would tell you how to make millions of dollars. But it does not work that way. And yet it can have some value.