Evolution can be measured directly by tracking genetic change in populations. This page explains how biologists document allele frequency shifts over generations and why those shifts constitute strong evidence for evolution.

What it means to measure evolution genetically

Natural selection and other evolutionary processes act on heritable variation, but evolution is defined at the population level. The most direct way to demonstrate evolution is to show that the genetic composition of a population changes across time.

Because allele frequencies are numerical, they can be compared across sampling dates (e.g., year-to-year) to test whether a population’s gene pool has shifted.

Hardy–Weinberg genotype proportions shown as areas: $p^2$ (homozygous dominant), $2pq$ (heterozygous), and $q^2$ (homozygous recessive). The diagram emphasizes that genotype frequencies are determined by allele frequencies ($p$ and $q$), which is why tracking $p$ over time provides a direct genetic measure of evolutionary change. Source

Why allele frequency change is evidence for evolution

To meet the syllabus focus, the key idea is that documented changes in allele frequencies are measurable and therefore can be used as evidence that evolution occurred.

Core logic

• If a population does not evolve at a locus, its allele frequencies remain stable from generation to generation (within expected sampling error).

• If allele frequencies do change across generations beyond sampling error, then the population has undergone microevolution (evolution on a short timescale).

This evidence is especially compelling because it does not rely on inference from anatomy or fossils; it is a direct comparison of genetic data over time.

How allele frequencies are documented in real populations

Choosing what to measure

Biologists can track allele frequencies using:

• Genotypes at a single gene (e.g., AA, Aa, aa)

• DNA sequence variants (e.g., SNPs) across many loci

• Marker alleles linked to traits of interest

Calculating allele frequency from genotype counts

When genotypes are known for a diploid population, allele frequency can be computed from counts of each genotype.

A parallel expression can be written for the alternative allele (commonly $q$), but the essential skill is converting observed genotype data into an allele-frequency estimate that can be compared across timepoints.

Designing time-series evidence

To credibly claim evolution using allele frequencies, studies typically:

• Sample the same population (or well-defined populations) at multiple times

• Use consistent sampling methods and adequate sample sizes

• Report uncertainty, recognising that estimates vary due to finite sampling

• Replicate across sites or years when possible to verify patterns

Interpreting allele frequency shifts

What counts as “evidence” rather than noise?

Allele frequency estimates can differ slightly between samples even if the true population frequency is unchanged. Strong evidence comes from:

Computer-simulated genetic drift trajectories showing allele frequency $p$ changing over 100 generations in multiple replicate populations. The jagged, divergent lines illustrate how allele frequencies can shift across time due to random sampling effects (drift), sometimes reaching fixation ($p=1$) or loss ($p=0$). Source

• Consistent directional change across multiple generations

• Changes that are larger than expected from sampling variation

• Patterns reproduced across independent samples or cohorts

Connecting genetic change to evolutionary change

When allele frequencies shift, the population’s genetic makeup has changed. That satisfies the syllabus statement: documented changes in allele frequencies over time provide measurable evidence that evolution has occurred in that population, regardless of which mechanism caused the shift.

Common pitfalls in using allele frequencies as evidence

• Changing the sampled population: mixing subpopulations can mimic a time trend.

• Non-representative sampling: sampling only one habitat patch or age group can bias estimates.

• Trait-only claims: phenotypic change alone may reflect environmental effects; allele frequency data ties change to heritable genetic variation.

• Single timepoint comparisons: two timepoints can be misleading; multiple timepoints better support a genuine evolutionary trajectory.