17.2.3—Effects on allele frequencies

Allele frequencies can change across generations through non-random selection or through chance. Natural selection changes frequencies when a heritable phenotype gives some individuals greater survival or reproductive success in a particular environment; genetic drift changes frequencies by random sampling, especially when the population is small.

• Natural selection: an environmental pressure consistently favours some heritable phenotypes over others, so the associated alleles tend to increase or decrease directionally across generations. The direction depends on which phenotype is advantageous in that environment.
• Genetic drift: chance differences in which individuals survive or reproduce change allele frequencies, without the change needing to favour a phenotype. The effect is stronger in small populations, where random sampling is a larger fraction of the population.
• Founder effect: a special small-population event in which a few individuals start a new population. Their allele frequencies may differ from the source population simply by chance, so the new population can show reduced variation or unusual allele frequencies.
• Bottleneck: a sharp reduction in population size leaves a small, chance sample of the original alleles. Frequencies can then differ from the original population, and the remaining variation may be reduced.

To interpret a change, ask three questions: (1) did an environmental pressure consistently favour a heritable phenotype, suggesting natural selection; (2) was there a small-population or sampling event, suggesting drift; and (3) did a few colonising individuals or a population reduction define the event, suggesting founder effect or bottleneck? Compare allele frequencies before and after the event, but do not infer a cause from the direction of change alone.

Natural selection is non-random with respect to the environmental pressure, whereas drift is random with respect to which alleles are sampled. Drift is not the same as an allele being useful, and a founder effect is not a purposeful adaptation. These mechanisms can be discussed without Hardy–Weinberg calculations or a specific antibiotic case.