Selective Pressure Initiates Change
I remember a scary building, now long since demolished, in the sciences complex at the University of Oregon in Eugene. This must have been the late 1970's -- I had a key to most of the science buildings, and would explore the nooks and crannies in my off hours, to clear my head. The scary one was abandoned, and had a "danger: radiation" sign on almost every door from the outside all the way through to the bathrooms. It was filled with sinks and tables that made a Geiger counter sing.
This was a biology laboratory. What on earth was all this unnaturally concentrated radiation for? What did it have to do with biology? The answer, as it turns out, is "nothing".
These labs were for drosophila research. Generations of fruit flies were bred with different artificial levels of background radiation in order to 1) quantify the genetic damage induced by the different levels and 2) track the journeys of mutations over generations.
Although much was learned about the structure of the genetic transmission, the impetus for the research was a theory, which is now discredited, but which is still found in the "background dogma". It asserts that the genetic variation which fuels natural selection emerges through x-ray induced genetic mutation. This was, with hindsight, a convenient interpretation of Darwinism in the Age of the Atom. In much the same way "Social Darwinism" -- more-or-less preaching the survival of the strong individual -- was a convenient interpretation of Darwin in the age of the Robber Barons.
The idea that "hopeful monsters provide this necessary variation" flew in the face of a basic observation: living things tend to be quite healthy, but genetic damage tends to kill or maim. The mutated drosophila certainly confirmed this observation.
I saw the answer to this conundrum presented quite clearly, although not generalized, on a NOVA program the other day, about dogs. Under selective pressures -- for example "friendliness towards humans" -- foxes undergo changes in metabolic and hormonal pathways, and developmental timing, that radically change their morphology. They look like domesticated dogs in the subsequent generation. [From experiments by Dmitri Belyaev]
Although the specific change in visible morphology may not be 'shaped' by selective pressures, the change itself is still the direct result of selective pressures. This doesn't require a separate mechanism, such as random mutation introduced by x-rays. The selective pressure itself creates the change, and hence the subsequent variation, simply by 'recruiting' the existing variation and the manifold indirect combinatoric effects of the mechanisms of genetic and epigenetic inheritance.
As is often the case, our cognition tends to separate "roles" in nature before recognizing that they are not so separate. Externally-mutated DNA is not required to act in the role of 'variation fuel' for evolution. I don't mean that DNA mutation doesn't exist. It's just not as fundamental as previously imagined. In some sense, it's a perfect human semantic mistake, emerging from the capture of the broad idea 'genetic variation' by the specific notion of 'randomly mutated DNA'.
Note that much of 'genetic programming' in computer science is based on this superceded notion of random mutation's role. It will need to evolve -- to a newer theory -- if the goal of this technique is to cultivate adaptive power equivalent to that found by evolutionary biologists.