COLLEGE BIOLOGY • EVOLUTION & NATURAL SELECTION

Variations in Populations

Understanding how genetic diversity within populations fuels evolutionary change through natural selection and drift.

Historical Context & Motivation

The study of variation within populations stands as one of the foundational pillars of evolutionary biology, yet the recognition that heritable differences among individuals drive evolutionary change required centuries of intellectual development. Before Darwin, most naturalists subscribed to typological thinking—the idea that species are defined by an ideal "type" and that individual variation is merely noise around that essence. This Platonic worldview actively discouraged the study of within-population differences, treating them as imperfections rather than raw material for adaptive change.

The shift toward population thinking—where variation itself becomes the central object of study—transformed biology from a descriptive catalog of species into a predictive science. Darwin's great insight was that populations harbor abundant heritable variation, and that differential reproductive success among variants leads to adaptive evolution. However, Darwin lacked a mechanism of inheritance, a gap that would not be filled until the rediscovery of Mendel's work and the eventual synthesis of genetics with natural selection in the early twentieth century.

1859
On the Origin of Species
Charles Darwin publishes his theory of evolution by natural selection, emphasizing the importance of individual variation within populations as the substrate for evolutionary change, though lacking a particulate theory of inheritance.
1866
Mendel's Laws of Inheritance
Gregor Mendel publishes his experiments on pea plants, establishing the particulate nature of inheritance through discrete hereditary factors (later called genes), though his work remained largely unrecognized until 1900.
1908
Hardy-Weinberg Equilibrium
G. H. Hardy and Wilhelm Weinberg independently derive the Hardy-Weinberg principle, establishing a null model for allele frequencies in populations and providing the mathematical foundation for population genetics.
1930–1942
The Modern Synthesis
Fisher, Haldane, Wright, Dobzhansky, and Mayr integrate Mendelian genetics with Darwinian selection, creating the Modern Evolutionary Synthesis. Population-level variation becomes a quantifiable, predictable phenomenon driven by mutation, selection, drift, and gene flow.
1966–Present
Molecular & Genomic Era
Lewontin and Hubby's protein electrophoresis studies reveal unexpectedly high levels of genetic variation in natural populations. The advent of DNA sequencing and genome-wide association studies enables direct measurement of variation at the nucleotide level across entire genomes.

The central question that motivated this entire intellectual trajectory remains remarkably relevant: How much variation exists in natural populations, what mechanisms generate and maintain it, and how does this variation translate into evolutionary change? Understanding population-level variation is prerequisite to grasping natural selection, speciation, adaptation, and conservation genetics.

Core Principles & Definitions

Population variation encompasses all heritable and non-heritable differences among individuals within a defined group of interbreeding organisms. To analyze variation rigorously, biologists distinguish between phenotypic variation—the observable differences in morphology, physiology, and behavior—and genotypic variation—the underlying differences in DNA sequence among individuals. Only the heritable component of phenotypic variation, which has a genetic basis, can respond to natural selection and contribute to evolutionary change across generations.

1

Genetic Variation

Differences in DNA sequence among individuals, including single nucleotide polymorphisms (SNPs), insertions/deletions, copy number variants, and chromosomal rearrangements. This is the ultimate source of all heritable variation in a population.
2

Phenotypic Variation

Observable differences among organisms resulting from the interaction of genotype and environment (VP = VG + VE + VG×E). Includes both continuous traits (e.g., height) and discrete traits (e.g., blood type).
3

Sources of Variation

New genetic variation arises through mutation (the ultimate source), recombination during meiosis (which shuffles existing alleles into new combinations), and gene flow (migration of alleles between populations).
4

Allele Frequency

The proportion of a specific allele among all copies of that gene in a population. Population genetics tracks changes in allele frequencies across generations as the primary measure of evolution. A population is evolving when allele frequencies change over time.
5

Heritability

The proportion of phenotypic variation attributable to genetic differences among individuals ( = VA / VP). High heritability indicates that a trait can respond strongly to selection.
KEY TAKEAWAY
Think of a population's genetic variation as the inventory in a warehouse. Mutation is the factory producing new items, recombination is the packaging department bundling items into novel combinations, and natural selection is the market demand that determines which items sell (increase in frequency) and which are discontinued (decrease in frequency). Without a well-stocked warehouse—without variation—there is nothing for natural selection to act upon, and the population cannot adapt.

Visualizing Population Variation

One of the most powerful ways to understand variation in populations is to visualize how a quantitative trait is distributed across individuals. Most continuous traits—body mass, beak depth, enzyme activity—follow an approximately normal (bell-shaped) distribution when measured across a large population. The shape of this distribution reveals critical information: the mean indicates the population's central tendency, while the variance (or standard deviation) quantifies the spread of variation around that mean. Natural selection acts on this distribution, reshaping it over generations through directional, stabilizing, or disruptive selection.

A normal distribution of a quantitative trait (e.g., beak depth) in a population. The dashed cyan line marks the population mean (x̄), with approximately 68.3% of individuals falling within ±1 standard deviation (σ). The left and right tails contain rare, extreme phenotypes that may be favored under directional selection.

The width of this distribution is directly proportional to the amount of variation in the population—a narrow curve indicates low variation (most individuals are similar), while a broad curve indicates high variation (individuals span a wide range of trait values). When environmental conditions change, selection acts on the tails of the distribution, favoring individuals with previously rare phenotypes. Over many generations, this differential survival and reproduction shifts the entire distribution, changing the population mean and potentially altering the variance as well. The three modes of selection—directional, stabilizing, and disruptive—each reshape this distribution in characteristic ways, which we explore further in Section 5.

Mathematical Framework

Population genetics provides the mathematical scaffolding for understanding how variation is maintained, lost, or altered over time. The Hardy-Weinberg equilibrium establishes the null model: a hypothetical population in which allele and genotype frequencies remain constant across generations in the absence of evolutionary forces. Any deviation from Hardy-Weinberg predictions indicates that one or more evolutionary mechanisms—mutation, selection, drift, gene flow, or non-random mating—are acting on the population.

ALLELE FREQUENCY IDENTITY
p + q = 1
For a diploid locus with two alleles, p = frequency of allele A₁ and q = frequency of allele A₂ in the population. These frequencies must sum to 1 because A₁ and A₂ are the only two alleles at this locus.
HARDY-WEINBERG GENOTYPE FREQUENCIES
p² + 2pq + q² = 1
Under random mating, the expected genotype frequencies are: (homozygous A₁A₁), 2pq (heterozygous A₁A₂), and (homozygous A₂A₂). This equation is simply the binomial expansion of (p + q)².
PHENOTYPIC VARIANCE DECOMPOSITION
V_P = V_A + V_D + V_I + V_E + V_{G×E}
Total phenotypic variance (VP) decomposes into additive genetic variance (VA), dominance variance (VD), epistatic variance (VI), environmental variance (VE), and gene-by-environment interaction (VG×E). Only VA directly predicts response to selection.
BREEDER'S EQUATION (RESPONSE TO SELECTION)
R = h² × S
The breeder's equation predicts the evolutionary response to selection. R = response to selection (change in mean phenotype), = narrow-sense heritability (VA / VP), and S = selection differential (difference between the mean of selected parents and the overall population mean).
💡 Why Does Only Additive Variance Matter?
Dominance and epistatic variance involve non-additive interactions between alleles at the same locus or between loci, respectively. Because offspring inherit single alleles (not genotypes) from each parent, only the additive effects of alleles are reliably transmitted across generations. This is why the breeder's equation uses narrow-sense heritability (h² = VA/VP) rather than broad-sense heritability (H² = VG/VP).

Modes of Natural Selection on Variation

Natural selection acts on the phenotypic variation within a population, but not all selection pressures reshape the distribution of traits in the same way. Three classical modes of selection describe the primary patterns by which natural selection alters the frequency distribution of continuous traits: directional selection shifts the mean toward one extreme, stabilizing selection narrows the distribution around the current mean, and disruptive selection favors both extremes at the expense of intermediate phenotypes, potentially increasing variance.

The three modes of natural selection compared. In each panel, the dashed violet curve represents the original distribution before selection. Directional selection shifts the mean toward one tail. Stabilizing selection narrows the distribution (reduces variance). Disruptive selection produces a bimodal distribution (increases variance), potentially leading to speciation.
Comparison of the three primary modes of natural selection and their effects on population trait distributions
Mode of SelectionEffect on MeanEffect on VarianceClassic Example
DirectionalShifts toward one extremeMay decrease over timePeppered moth darkening during Industrial Revolution; antibiotic resistance in bacteria
StabilizingRemains unchangedDecreases (narrows distribution)Human birth weight: very low and very high birth weights have higher mortality
DisruptiveMay remain similar but distribution becomes bimodalIncreases (broadens distribution)African seed-cracker finch beak size: small or large beaks are favored, but intermediate beaks are not

Worked Example: Hardy-Weinberg & the Breeder's Equation

Consider a population of wildflowers in which flower color is controlled by a single locus with two alleles: A₁ (red, dominant) and A₂ (white, recessive). In a census of 500 individuals, you observe 320 red flowers and 180 white flowers. You also measure petal length, a continuous trait with a population mean of 2.5 cm. A pollinator prefers flowers with longer petals, selectively visiting (and pollinating) flowers whose petals average 3.0 cm. If narrow-sense heritability for petal length is h² = 0.6, predict the allele frequencies and the response to selection on petal length.

Hardy-Weinberg Analysis & Predicting Evolutionary Response
1
Step 1 — Determine q² from phenotype countsWhite flowers have genotype A₂A₂ (homozygous recessive). The frequency of the white phenotype is the frequency of the A₂A₂ genotype: q² = 180 / 500 = 0.36.
q² = 0.36
2
Step 2 — Calculate allele frequenciesTaking the square root of q² gives q = √0.36 = 0.60. Since p + q = 1, we find p = 1 − 0.60 = 0.40. Thus the frequency of A₁ is 0.40 and the frequency of A₂ is 0.60.
p = 0.40, q = 0.60
3
Step 3 — Predict expected genotype frequenciesUnder Hardy-Weinberg equilibrium: p² = (0.40)² = 0.16 (A₁A₁), 2pq = 2 × 0.40 × 0.60 = 0.48 (A₁A₂), q² = (0.60)² = 0.36 (A₂A₂). In a population of 500, we expect 80 A₁A₁, 240 A₁A₂, and 180 A₂A₂ individuals. The 320 red flowers should comprise 80 homozygous dominant and 240 heterozygous individuals.
Expected: 0.16 A₁A₁, 0.48 A₁A₂, 0.36 A₂A₂
4
Step 4 — Calculate the selection differential (S)The selection differential S is the difference between the mean petal length of the selected parents and the overall population mean. Here, S = 3.0 cm − 2.5 cm = 0.5 cm. This represents the intensity of selection imposed by pollinator preference.
S = 0.5 cm
5
Step 5 — Apply the breeder's equation to predict responseUsing R = h² × S, we get R = 0.6 × 0.5 cm = 0.3 cm. This means the offspring generation is predicted to have a mean petal length of 2.5 + 0.3 = 2.8 cm, an increase of 0.3 cm in a single generation of directional selection. The 40% of phenotypic variation that is non-heritable (environmental) does not contribute to the evolutionary response, which is why R < S.
R = 0.3 cm; predicted new mean = 2.8 cm

Sources, Maintenance, and Loss of Variation

A fundamental question in evolutionary biology is not only how variation arises but also how it is maintained in populations over evolutionary time. Directional selection tends to reduce variation by fixing advantageous alleles, and genetic drift erodes variation in small populations through random sampling. Yet natural populations typically harbor substantial genetic diversity—a paradox that has driven decades of theoretical and empirical research. Several mechanisms counterbalance the loss of variation, including mutation-selection balance, heterozygote advantage (overdominance), frequency-dependent selection, and spatial and temporal variation in selection pressures.

Summary of mechanisms that generate, maintain, or reduce genetic variation within populations
MechanismEffect on VariationDetails & Examples
MutationIntroduces new alleles; increases variationPoint mutations, insertions, deletions, duplications. Typical rate ~10⁻⁸ to 10⁻⁹ per base pair per generation in eukaryotes. Mutation alone is a weak force but is the ultimate raw material for all variation.
RecombinationReshuffles existing alleles; increases genotypic variationIndependent assortment and crossing over during meiosis create new allele combinations. Does not generate new alleles but vastly expands the combinatorial space of genotypes.
Gene FlowIntroduces alleles from other populations; can increase or homogenize variationImmigration of individuals carrying novel alleles. Opposes genetic divergence between populations. Even low levels of migration (Nm > 1) can prevent fixation by drift.
Genetic DriftRandomly removes alleles; decreases variationStronger in small populations. Bottlenecks and founder events can drastically reduce variation. Expected heterozygosity declines by 1/(2N) per generation under pure drift.
Balancing SelectionMaintains multiple alleles; preserves or increases variationIncludes heterozygote advantage (e.g., sickle cell in malaria regions), negative frequency-dependent selection (e.g., rare male advantage), and temporally varying selection. Actively opposes fixation.
Directional SelectionFixes advantageous alleles; decreases variation at selected lociSelective sweeps reduce variation at and near the selected locus. Genome-wide, directional selection reduces standing variation unless continually replenished by mutation and gene flow.
KEY TAKEAWAY
Think of population variation as the water level in a reservoir. Mutation and gene flow are the streams feeding water into the reservoir, recombination stirs the water to create new mixing patterns, while directional selection and drift drain water out. Balancing selection acts as a dam that prevents complete drainage at certain loci. The equilibrium water level—the standing variation in a population—depends on the relative strength of these inflows and outflows.

Connections to Advanced Evolutionary Theory

The classical framework of population variation connects directly to several advanced topics in modern evolutionary biology and genomics. Understanding how variation is structured within and among populations is critical for fields ranging from conservation genetics to personalized medicine. The neutral theory of molecular evolution, proposed by Motoo Kimura in 1968, challenged the selectionist paradigm by arguing that most genetic variation at the molecular level is selectively neutral and changes in frequency primarily through drift. This debate between neutralism and selectionism continues to shape how biologists interpret genomic variation data.

How foundational concepts in population variation extend to advanced evolutionary and genomic analyses
Classical FrameworkAdvanced Extension
Hardy-Weinberg equilibrium for a single locusMulti-locus models, linkage disequilibrium (LD), genome-wide association studies (GWAS)
Breeder's equation (R = h²S) for univariate traitsMultivariate breeder's equation (Δz̄ = GP⁻¹s) using the G-matrix for correlated traits
Phenotypic variance decomposition (V_P = V_G + V_E)QTL mapping and GWAS to identify specific loci underlying V_A; missing heritability problem
Three modes of selection (directional, stabilizing, disruptive)Selection gradient analysis (Lande-Arnold method), fitness landscapes, adaptive dynamics
Genetic drift in idealized populationsEffective population size (Nₑ), coalescent theory, demographic inference from genomic data
Variation within a single populationF-statistics (Fₛₜ), population structure, landscape genetics, speciation genomics

Perhaps the most active area of research connecting population variation to human welfare is conservation genetics. Small, isolated populations lose variation through drift and inbreeding, reducing their capacity to adapt to changing environments—a phenomenon called inbreeding depression. Conservation biologists use metrics such as heterozygosity, allelic richness, and effective population size (Nₑ) to assess population viability and design strategies such as genetic rescue through managed gene flow.

Practice Problems

PROBLEM 1CONCEPTUAL
A population of insects exhibits no phenotypic variation in wing length—every individual has wings of exactly 12 mm. Can natural selection drive evolutionary change in wing length in this population? Explain your reasoning in terms of the relationship between variation, heritability, and the breeder's equation.
PROBLEM 2BASIC CALCULATION
In a population of 1,000 diploid organisms, you observe 90 individuals with genotype AA, 420 with genotype Aa, and 490 with genotype aa. Calculate the allele frequencies of A and a, then determine whether this population is in Hardy-Weinberg equilibrium.
PROBLEM 3INTERMEDIATE
A population of lizards has a mean body length of 15.0 cm with a phenotypic variance of 4.0 cm². The additive genetic variance is 1.6 cm², dominance variance is 0.8 cm², and environmental variance is 1.6 cm². A predator selectively eats smaller lizards, so only lizards with a mean body length of 17.0 cm survive to reproduce. (a) Calculate narrow-sense heritability. (b) Calculate the selection differential. (c) Predict the mean body length of the next generation.
PROBLEM 4APPLIED
The Florida panther (Puma concolor coryi) population was reduced to approximately 20–30 individuals by the 1990s and exhibited severe inbreeding depression, including cryptorchidism, heart defects, and kinked tails. In 1995, eight female Texas pumas were introduced. Using your knowledge of population variation, explain: (a) Why did the small population size lead to loss of variation? (b) What evolutionary mechanism does the introduction of Texas pumas represent? (c) Why would this intervention be expected to improve population fitness?
PROBLEM 5CRITICAL THINKING
The neutral theory of molecular evolution (Kimura, 1968) proposes that most genetic variation at the molecular level is selectively neutral and that changes in allele frequency are primarily driven by drift rather than selection. If this is true, what implications does it have for using standing genetic variation as a predictor of a population's adaptive potential? Consider: (a) the distinction between neutral and functional variation, (b) the role of effective population size, and (c) under what circumstances neutral variation could become non-neutral.

Summary: Variations in Populations

Variation within populations is the essential raw material for evolution by natural selection. Genetic variation arises through mutation (the ultimate source of new alleles), is reshuffled by recombination during meiosis, and is redistributed among populations via gene flow. The Hardy-Weinberg equilibrium (p² + 2pq + q² = 1) provides the null model against which evolutionary change is measured. Phenotypic variance decomposes into additive genetic, dominance, epistatic, and environmental components, with only the additive genetic variance (V_A) determining a population's response to selection.

The breeder's equation (R = h² × S) quantifies the expected evolutionary response from narrow-sense heritability and the strength of selection. Three modes of selection—directional, stabilizing, and disruptive—reshape trait distributions in characteristic ways. Variation is maintained by balancing selection (heterozygote advantage, frequency-dependent selection) and eroded by genetic drift and directional selection. Understanding these dynamics is central to fields from conservation genetics to personalized genomic medicine.

Varsity Tutors • College Biology • Variations in Populations