Hi Faith. I think part of the problem here is that some of the ways certain terms are being bandied about here may be a bit misleading (or unclear). Maybe I can help out - I don't want you to think we're "ganging up" on you.
1. Genetic drift. This refers to random fluctuations in the frequency of alleles within a population. The fluctuation is not due to selection at all, rather to stochastic (statistical) effects. The alleles do not differ on average in fitness - as was mentioned, this is the heart of "neutral theory". If there's no effect on fitness, the alleles are not acted upon by natural selection. Drift could also be considered sampling error (more on this later in conjunction with founder effects). Think of it this way: if you have a population of "A" alleles, and an "a" single mutation occurs, what happens when the population reproduces? Say only two offspring survive. AA x Aa yields a probability of 25% that the "a" allele will be lost in the first generation (AA = 1/2 x 1/2 = 1/4). However, looking at multiple alleles in multiple populations with multiple offspring, the probability distribution, somewhat counterintuitively, of the neutral allele doesn't vary around the mean distribution. There is an equal probability of the frequency of this allele either increasing or decreasing randomly. This is called the "drunkards' walk" (or "random walk"). You can graph this fairly easily if you're interested. The simple version of the relevant equation is V =
p(1-
p)/2N, where V is variance in frequency,
p is probability distribution for a single population, and N is the number of gene copies (2N indicating haploid, with two alleles for each gene in the adults). It is obvious that the size of the initial population will also have a great deal of effect - genetic drift occurs faster in small populations than large ones. If you plot the drift of multiple neutral alleles over time, the graph comes out looking like a plate of spaghetti as lines criss-cross back and forth, disappear, rise to fixation, etc.
Adding to this, which should be obvious from the equation, a rare allele (
p close to zero) will be more likely to "walk off the end of the pier" and disappear than a more common allele. Of course, things get a bit more complicated when you're dealing with metapopulations (populations of populations), but that's the simplest way of looking at it.
Drift drives evolution by changing the frequency of alleles in a population, irrespective of any selection pressures. It may drive an allele to fixation (
p = 1), but alternatively may cause it to disappear completely (
p = 0). It does NOT automatically reduce variation.
2. Other factors: Something that is taken into consideration in more complex treatments is the effect of other non-deterministic factors, like accidents. If you have a small population which is polymorphic for a neutral allele with a low frequency, accidents can change the frequency of this allele downward. Consider: a group of 10 organisms of which 8 are homozygous (AA), and two are heterozygous (Aa). If one of the Aa individuals is eaten before it reproduces, the frequency of "a" is reduced. OTOH, accidents can increase the frequency of "a" by an AA getting eaten. (Apologies to all for the gross oversimplification).
3. Bottleneck. A bottleneck is simply where a population is reduced to a very few individuals, and hence a very few alleles. If the population survives, we say it has passed through a bottleneck. One effect of bottlenecks is to homogenize a population - every member of the population shares identical alleles, whether through drift, selection, or even just sampling error. As an interesting aside, bottlenecks can actually counteract selection by reducing the population to such an extent that drift may cause fixation of a deleterious allele, causing the population to actually move from one fitness peak to the "valley" in its adaptive landscape - allowing it (assuming it survives) to climb to another adaptive peak. This is something that selection cannot do. IOW, in the context of your argument, reduction in variability caused by a bottleneck can in fact impel an increase in variability or even fitness when the population rebounds. Bottlenecks may be caused by selection pressures, but they are NOT selection pressures - they are a possible result. In addition, bottlenecks are NOT evolution, they are simply a term describing the effect of population reduction where the population subsequently returns to a larger size over time (and hence increased variability).
4. Founder effect. This is actually a special type of bottleneck. It is NOT caused by selection. Founder effects occur when a small (often just a few individuals or even one pregnant female) subset of a population arrives in a new habitat cut off from the parent (or "source") population. In this case, the founding population will likely be unrepresentative of the source population, in terms of the alleles present in the source. IOW, it is a form of sampling error - the probability that a rare allele carried in the source population will appear in the new population is very low (actually, the likeliehood is identical to the
p of the allele in the source population). Just like any other bottleneck, founder effects show similar genetic markers (homogeneity, etc). Indeed, drift can have a significant effect on founder populations because they are so small.
Note, however, founder populations are significant for evolution (just like mass extinctions) - not because of the genetics - but because of the potential for the population to divide and divide again in subsequent generations due to the availability of new, unoccupied niches (see "adaptive radiation").
To sum up: drift has nothing to do with selection, but does drive evolution. Bottlenecks and founder effects have nothing to do with selection, and also have nothing intrinsically to do with evolution - they are descriptions of population dynamics.
Hope this helped.
™ Version 4.2
