I've been extremely busy over the past week...hence the anemic blogging output. But in the meantime I've been having a pretty involved discussion in the neural network discussion group I belong to about the definition of evolution, not just the colloquial definition, but a good, working scientific definition.

Among laymen, the term often connotes progress, or direction of some sort. But most biologists simply define evolution like this:

Genetic change in a population over time.

No direction, no intent, no particular optimization. Just change.

And most people who do know a little about evolution incorrectly assume that natural selection is the only pressure that brings about this change. In fact, genetic drift plays a very large role in evolution, but not many people understand the concept very well.

Here's an article that explains it fairly well, but the key point is that it is a skewing of the population because of the effect of sampling.

Sampling error is actually pretty easy to understand. If I randomly call 5 people and ask them who they're going to vote for this year, and I get the following results:

Dean: 3
Bush: 1
Kerry: 1

Does this mean that Howard Dean will be our next president? How reliable is my information? And what could I do to make it more reliable?

The answer is obvious. Because it's such a small sample, the chances that my polling reflect widespread views are very small. The obvious solution is to increase the sample size. If I call 1000 people and get the same ratios, my generalization is more likely to be accurate.

It's the same with biological populations. Some populations are very large, and genetic drift (or sampling effects) are rather small. But some populations, such as demes of reptiles, may only be 30-40 in size. Who survives and reproduces to propagate their genes into the next generation is partly a function of how well suited they are to their environment (whether they blend in to the desert sand well enough not to be eaten by hawks, whether their ability to resist parasites is better than others, etc.), but in large part what determines whether or not a given individual survives and reproduces is just dumb luck, and has absolutely nothing to do with their fitness.

If only a small number of our desert lizards survive and reproduce, the genetic makeup of the new population is going to be skewed toward the types of genes inherent in the survivors. Thus, change may trend in a particular direction, that has nothing to do with helping the organisms in the population adapt to their environment.

The point is, populations can evolve without adapting. So evolution is simply change, not necessarily adaptational change.