[csense]

Natural selection is a process by which beneficial mutations and recombinations are kept and inherited into future generations, and harmful ones are rejected. Natural selection uses those three processes, and eventually a beneficial mutation/recombination will be shared by the entire population.

Well, if natural selection can describe any given genetic state, of any given population of organisms, then I don't see how it is either testable or falsifiable, let alone useful in a scientific theory.

[Sofa King]

"Well, if natural selection can describe any given genetic state, of any given population of organisms,"

It can't. Natural Selection would be at a loss to describe a species developing a trait that made it less likely that the members of the population who possesed it would be able to survive, breed, and pass on their genes (although the we may not always understand what the advantage of a given trait is).

If, for example, a culture of bacteria was routinely exposed to an antibiotic that it was previously immune to, it would be a violation of natural selection if that culture's future generations slowly started becoming susceptible to it's effects.