Modes of Selection

In evaluating the way natural selection impacts populations, it is generally convenient to recognize a number of different modes in which selection may act.

Stabilizing selection is characteristic of relatively stable environments with high biological diversity. The latter means that most niches are filled. Under such conditions, the overall pattern of selection is against the extremes in the frequency distribution for most traits. The result is that the population tends to maintain the status quo with respect to prevailing adaptations. Another perspective you might take is to assume that most populations are optimally adapted for the niche they occupy and that there is little opportunity to expand into adjacent niches that are already occupied by similarly well-adapted organisms. The analysis of wing-length in sparrows (Bumpus) is an example of this type of selection. Stabilizing selection is the one mode that does not result in adaptive change and/or evolution.

Directional selection occurs when the environment of a population is changing in some definitive way. This usually means that one extreme for the frequency distribution for some traits gradually becomes less adaptive as the environment changes and is selected against. Provided the population has sufficient variability and the change is not too rapid, the result will normally be to shift the frequency distribution in a direction that is adaptive relative to the changing conditions. The development of insecticide-resistant populations of insects or antibiotic resistance in bacteria is two example of directional selection. If the change is very rapid, the population has low genetic/phenotypic variability, or another species is available that is already well-adapted to the changing environment, the population is question may well be excluded or restricted as a result of the changing conditions. Directional selection can cause the evolution of a species with time, but does not increase bio-diversity.

If a population is exposed to different patterns of selection in different parts of its range, the sub-populations may well diverge with time, each responding to the local selection pattern. Such disruptive selection can also be viewed as different patterns of directional selection occurring within the range of a species. If gene flow is restricted between the sub-populations, they may well evolve into different species, given time for the development of isolating mechanisms that would serve to keep the populations distinct, even if they later come into contact. Disruptive selection can increase bio-diversity. In the extreme, where one species evolves into several, the process is known as adaptive radiation.

Disruptive selection can also operate within a species to produce distinctive morphotypes. The example of hooknose and jack phenotypes in male coho salmon is one such case. Where such selection generates pronounced differences between the sexes, as in the plumage and singing behavior of birds, the process is known as sexual selection. Sexual selection is probably best viewed as a special case of disruptive selection.

This photograph shows the normal form of the moth (upper left), an excellent example of cryptic coloration with respect to the light, lichen-covered bark of the birch trees. A classic series of mark and release studies by Kettleworth, an English biologist, indicate that the melanistic form of the moth (lower right) is twice as likely to be preyed upon by birds than the lighter phenotype. This pattern of selection keeps the wild-type quite common and depresses the number of melanistic moths to a very low level. Most of the textbook analyses would now rush on the advent of industrial pollution, but let's look at the type of selection operating here! It is stabilizing selection, in the absence of other factors, which keeps the numbers of the light-colored variants high will reduce the number of melanistic moths.

The situation gets even more complex if we look at the situation with respect to the melanistic population. Once selection has reduced the numbers of dark-colored moths to a very low level, what tends to operate with respect to their populations is something called frequency-dependent selection. This is a type of selection that tends to maintain rare alleles in a population due to the fact that other forms of selection tend to act on the organisms that are present in larger numbers. In effect, at some low population value, depending on selection pressures, the rarity of an allele can effectively confer some protection against selection.

Now we come to the situation emphasized by most textbooks - the advent of pollution, resulting in the death of the lichens and the darkening of the tree trunks. Now the light-colored variant (left) stands out against the darker background while the melanistic moths (right) blend in. Kettleworth's experiments showed that, under these conditions, the pattern of selection is almost precisely reversed, with the light-colored variants almost twice as likely as the melanistic moths to be subject to bird predation. This results in a rapid shift in gene frequencies so that the melanistic allele is now common and the light-colored moths become rare. Thus, the industrialized areas provide a good case study of directional selection. What the textbooks don't emphasize is that once the melanistic moths become common and the light-colored forms are rare, the situation in the industrial areas now represents a combination of stabilizing and frequency-dependent selection.

Of course, once pollution is reversed in specific areas, directional selection operates to return the light-colored form to dominance, but this will be followed by stabilizing and frequency-dependent selection. Just at the point where you think you understand what is happening, you might back off and consider the whole of England, where the pattern of selection varies, depending on the degree of industrialization. This clearly represents disruptive selection!

Thus, while the case of Biston betularia (often referred to as industrial melanism) is usually used to illustrate directional selection, real-world situations are far more complex. The type of selection operating on populations is subject to dynamic change and the specific type of selection is always relative to populations in some specific context.