While I’m away, here’s a post in the spirit of the “basic concepts in science” thingee that’s been floating around ScienceBlogs.
I received an email about a recent post that asked why pesticide resistance in insects isn’t an instance of artificial selection as opposed to natural selection. The difference between artificial selection and natural selection isn’t that the selective agent (e.g., pesticides) is a result of human activity. The difference is in what determines what is the ‘fittest’: a person’s decision as to what traits are preferable, or differential survival and reproduction. Artificial selection occurs when organisms with a certain trait or traits, such as looking like a Pomeranian, are chosen and allowed to survive and to reproduce. The ‘fitness criterion’ involved isn’t survival and reproduction: it’s the judgement of the human running the experiment that defines what is the most fit (I don’t think Pomeranians in the wild would be very fit, but rather, a yummy snack for some nasty carnivore).
In the case of pesticide resistance, farmers who use pesticides aren’t intentionally breeding those individuals that happen to be resistant to the pesticide (in fact, they’re trying to kill all of the pests). While the selective pressure, pesticides, is of human-origin, which insects survive and reproduce is not based on an artificial criterion (e.g., I like bugs with red wings), so this is natural selection.
True. But hard to properly explain at the abstract level. That is, how do you represent what humans are doing when they artificially select (intentionally breed) in terms of the Fitness Landscape model?
In a vague sense, humans are themselves “natural” and the product of natural selection and sexual selection (is that natural or artificial?). Hence humans doing the selection are in a second-order way part of natural selection.
Darwin was very influenced, in co-inventing (with Wallace, simultaneously but independently) Evolution by Natural Selection, by his many extended discourses with breeders of commercial plants and animals.
Darwin brilliantly created a synthesis of the existing literature, and his field observations from the Beagle, with what horse breeders and cattle breeders and the like had been doing by ad hoc trial and error over centuries.
Or, really, since the dawn of civilization itself. Or before. It has been suggested that we coevolve with our domesticated flora and fauna, including major food crops and animals, of which people argue whether dogs (bred from wolves) or chickens (bred from southest asian jungle game birds) came first. We keep finding earlier and earlier examples of first domestications.
I appreciate the subtlties to which Jonathan points, but think that Mike has correctly located the conceptual divide between natural and artificial selection. In cases of co-evolution via natural selection, for example, we expect the traits of species1 that underpin the relationship with species2 to have been selected _for_ (in Sober’s tereminology) their contribution to the fitness of species1, but not species2. Where co-evolution grades into symobiosis things get murky, indeed, but that seems a very different situation from artificial selection.
Essentially, Jonathan’s post makes the point that in some underlying philosophical sense we can’t draw a sharp distinction between artificial and natural selection unless we have a sharp distinction between humankind and the rest of nature.
In the meantime, practicing scientists will continue to use the terms natural and artificial selection, because they are useful. And when practicing scientists do so, they will be doing so in precisely the sense that Mike has laid out in his very nice post.
I don’t think the distinction between artificial and natural selection hinges on the distinction between humans and the rest of nature. What it does depend on is the causal role of intentions to bias relative rates of survival and reproduction.
Thanks You So Much.. Have Nice Days..