5.1 Types of natural selection: directional, stabilizing, and disruptive

Types of Natural Selection

Natural selection shapes populations in different ways depending on which phenotypes have the highest fitness. Directional, stabilizing, and disruptive selection each produce distinct changes in a population's trait distribution, shifting the mean, narrowing variation, or splitting it apart.

Understanding these three modes helps you read trait data and figure out what kind of selection is acting on a population. That skill connects raw observations to the bigger picture of how evolution works.

Types of Natural Selection

Directional Selection

Directional selection favors one extreme of a phenotypic trait. Over time, the population mean shifts toward that extreme, and allele frequencies change accordingly.

A classic example: during drought years on the Galápagos Islands, finches with larger, deeper beaks survived at higher rates because they could crack tougher seeds. The mean beak size in the population shifted upward across generations. Other examples include antibiotic resistance in bacteria (where resistant individuals are strongly favored) and industrial melanism in peppered moths (where darker moths had better camouflage on soot-covered trees).

Stabilizing Selection

Stabilizing selection favors intermediate phenotypes and selects against both extremes. This reduces variation around the mean without shifting it.

Human birth weight is the go-to example. Babies that are too small face survival challenges, and babies that are too large create complications during delivery. The result is strong selection for a middle range of birth weights. You also see this pattern in the number of offspring per litter in many mammals and in enzyme efficiency at normal body temperature, where deviations in either direction reduce function.

Disruptive Selection

Disruptive selection favors phenotypes at both extremes of the distribution while selecting against intermediate forms. This increases variation and can produce a bimodal distribution (two peaks instead of one).

In African seedcrackers (Pyrenestes ostrinus), birds with either very large or very small beaks do well because each size is suited to cracking a different type of seed. Intermediate beak sizes are less efficient at both. Other examples include shell color polymorphism in land snails (where different colors are camouflaged in different microhabitats) and coloration patterns in guppies.

If disruptive selection persists and mating becomes assortative (individuals preferring mates with similar phenotypes), it can eventually contribute to speciation.

Effects on Phenotype Distribution

Each type of selection leaves a distinct signature on the shape of the trait distribution:

• Directional: The entire curve shifts toward one extreme. Variation on the non-favored side shrinks, and alleles for the disfavored phenotype can eventually be lost from the population.
• Stabilizing: The curve gets narrower and taller. Both tails shrink as extreme phenotypes are selected against, concentrating individuals around the mean.
• Disruptive: The curve gets wider and can split into two peaks. The middle of the distribution loses individuals while both tails gain them.

Analyzing Selection in Data

When you're given phenotype data and asked to identify the type of selection, follow these steps:

1. Plot the trait distribution at each time point. Look at the shape: Is it a normal bell curve? Shifted? Bimodal?
2. Compare across generations. Has the mean shifted (directional)? Has the spread narrowed (stabilizing)? Have two peaks appeared (disruptive)?
3. Calculate descriptive statistics. Track the mean, variance, and skewness of the distribution. A change in mean suggests directional selection. A decrease in variance suggests stabilizing. An increase in variance (or a shift toward bimodality) suggests disruptive.
4. Consider the environmental context. What selective pressures are present? How does the trait relate to survival and reproduction? A drought favoring large beaks tells a different story than predation favoring intermediate coloration.
5. Test statistically. Perform tests for normality and modality to confirm whether the distribution has genuinely changed shape, not just shifted by chance.

Quantifying Selection

The breeder's equation connects the strength of selection to the evolutionary response:

$R = h^2 S$

• $R$ = response to selection (how much the mean changes in the next generation)
• $h^2$ = narrow-sense heritability (the proportion of phenotypic variation due to additive genetic effects)
• $S$ = selection differential (the difference between the mean of selected individuals and the overall population mean)

This equation tells you that even strong selection (large $S$) won't produce much evolutionary change if heritability is low. Both ingredients matter.

When interpreting results, consider alternative explanations for observed patterns. Genetic drift, gene flow, or environmental changes unrelated to selection can also shift trait distributions, especially in small populations. Strong, consistent patterns across multiple generations provide the best evidence that selection is genuinely at work.