Skip to content

Biology · Ch 6 — Evolution

Natural Selection — Types and Mechanism

6.9

Natural Selection — Types and Mechanism

When any measurable, continuously varying character within a population — human birth weight, or the depth of a bird's beak, for instance — is carefully recorded across a large number of individuals and plotted as a frequency distribution (the number of individuals showing each particular value of the character, plotted against the character value itself), the resulting graph typically takes the shape of a bell-shaped, roughly symmetrical (normal) curve, with the largest number of individuals clustering around some central, average value of the character, and progressively fewer individuals found as one moves further towards either extreme, low or high, end of the distribution. Natural selection can act upon such a distribution of a heritable character in three fundamentally different ways, depending on precisely which portion, or portions, of the existing distribution happen to be favoured by the prevailing environmental conditions, and which portion, or portions, are instead selected against.

Stabilising selection favours individuals possessing the average, intermediate value of the character, while simultaneously selecting AGAINST individuals found at both extremes of the distribution — those with unusually low values of the character, and those with unusually high values, are both disfavoured relative to individuals nearer the population average. The overall effect of sustained stabilising selection, generation after generation, is that the population's own average value of the character tends to stay roughly constant over time, while the total spread, or variance, of the character across the population is gradually narrowed and reduced, since individuals at both extremes are continually being removed from the population by selection. Human birth weight provides a commonly cited, well-documented real example of stabilising selection: careful epidemiological studies have repeatedly shown that babies born with either an unusually LOW birth weight or an unusually HIGH birth weight have, historically, faced somewhat elevated risks of mortality, compared with babies born at an intermediate, average birth weight, so that selection has consistently favoured the intermediate value of this particular character.

Directional selection, by clear contrast, favours individuals found at just ONE extreme end of the existing distribution, over both the population's average value and the opposite extreme, and the cumulative, generation-by-generation effect of sustained directional selection is that the population's own average value of the character progressively SHIFTS, generation after generation, in the direction of whichever extreme is being favoured by selection. The repeated, well-documented evolution of antibiotic resistance within bacterial populations subjected to sustained, repeated exposure to a particular antibiotic drug is a clear, medically important, and directly observable real-world example of directional selection occurring within an easily observable, human timescale: before antibiotic exposure begins, a bacterial population typically shows a wide range of natural variation in its susceptibility to the antibiotic, with the great majority of individual bacterial cells being fully susceptible and a small minority, purely by chance mutation, being somewhat more resistant; once the antibiotic is repeatedly applied, the susceptible majority of the population is killed off while the more resistant minority survives and reproduces preferentially, so that, generation after generation of the fast-reproducing bacteria, the population's average level of antibiotic resistance rises steadily, precisely the pattern directional selection predicts. …

Figure 6.5Stabilising, Directional and Disruptive Selection

What this figure shows. A row of three bell-curve graphs, each plotting the number of individuals in a population (vertical axis) against a continuously varying character value (horizontal axis), with the original, pre-selection distribution shown as a dashed outline and the resulting post-selection distribution shown as a solid shaded curve in each panel. The left panel (stabilising selection) shows the solid curve narrower than the dashed curve but centred at the same average value, with shaded arrows indicating selection pressure removing individuals from both tails. The middle panel (directional selection) shows the solid curve shifted noticeably to one side of the dashed curve, with a shaded arrow indicating selection pressure pushing the whole distribution towards one extreme. The right panel (disruptive selection) shows the solid curve split into two separate peaks positioned at the two extremes of the orig …