Artificial Selection - AP Biology

SECTION 1

Historical Context & Motivation

For thousands of years, humans have shaped the organisms around them—long before anyone understood the genetic mechanisms responsible. Ancient farmers saved seeds from the most productive plants, and herders bred their strongest livestock, incrementally shifting the characteristics of entire populations over generations. This practice, which Charles Darwin would later term artificial selection, provided the conceptual scaffold upon which Darwin built his theory of evolution by natural selection. By studying how pigeon breeders could produce astonishingly diverse breeds from a single wild ancestor, Darwin recognized that nature itself could act as a selecting agent, favoring individuals best suited to their environment. Understanding artificial selection therefore illuminates not only modern agriculture and biotechnology but also the fundamental evolutionary principles tested on the AP Biology exam.

~10,000 BCE • Neolithic Agricultural Revolution

Early farmers in the Fertile Crescent begin selectively planting seeds from wild grasses with larger, non-shattering seed heads, initiating the domestication of wheat (Triticum spp.) and barley.

1859 • On the Origin of Species

Charles Darwin publishes his landmark work, devoting the first chapter to 'Variation Under Domestication' and using artificial selection as a powerful analogy for natural selection.

1900–1910 • Rediscovery of Mendelian Genetics

De Vries, Correns, and von Tschermak independently rediscover Mendel's laws, providing the particulate-inheritance framework that explains how selective breeding changes allele frequencies.

1930s–1940s • Modern Synthesis & Quantitative Genetics

Fisher, Wright, and Haldane integrate Mendelian genetics with Darwinian selection. The breeder's equation (R = h²S) quantifies the response to selection in agricultural populations.

1990s–Present • Marker-Assisted & Genomic Selection

Advances in DNA sequencing allow breeders to select directly on genotype rather than phenotype, dramatically accelerating genetic gain in crops and livestock.

The central question artificial selection poses for biology is deceptively simple: if humans can so dramatically reshape organisms in mere centuries—producing Great Danes and Chihuahuas from the same wolf ancestor—what might millions of years of environmental pressure accomplish? Answering this question requires understanding how heritable variation, differential reproduction, and allele frequency change interact—the very mechanisms at the heart of the AP Biology curriculum.

SECTION 2

Core Principles & Definitions

Artificial selection operates through the same evolutionary logic as natural selection, but with one critical substitution: the selective agent is a human decision-maker rather than the environment. A breeder identifies individuals with a desired phenotype—higher milk yield, brighter plumage, disease resistance—and permits only those individuals to reproduce. Over successive generations, the alleles underlying the favored trait increase in frequency, while those associated with disfavored traits decline. This deceptively straightforward process rests on several foundational principles that connect genetics, heredity, and population biology.

Phenotypic Variation

Individuals in a population must differ in the trait of interest. Without variation—whether in kernel size, coat color, or growth rate—there is nothing for the breeder to select among. This variation arises from genetic differences (allelic variation), environmental effects, and gene × environment interactions.

Heritability

A portion of the observed phenotypic variation must be attributable to genetic differences that can be transmitted from parent to offspring. Heritability (h²) quantifies this fraction. If h² = 0, selection cannot shift the population mean because offspring do not resemble selected parents genetically.

Differential Reproduction

The breeder must impose a selection differential by allowing only individuals above (or below) a phenotypic threshold to breed. This non-random reproductive success is what drives allele frequency change across generations.

Cumulative Generational Change

A single round of selection produces only a modest shift in the population mean. Sustained, directional selection over many generations compounds these small changes, producing dramatic phenotypic divergence from the ancestral population—as seen in the transformation of teosinte into modern maize.

Trade-offs & Genetic Constraints

Intense selection on one trait can reduce genetic diversity (a genetic bottleneck effect), increase the frequency of deleterious alleles via linkage, or compromise other fitness-related traits through pleiotropy. Bulldogs bred for flat faces, for example, suffer chronic respiratory distress.

✦

KEY TAKEAWAY

Think of artificial selection like a music producer curating a playlist: the producer (breeder) chooses only certain songs (phenotypes) to make the final cut. Over many albums (generations), the style of the playlist (population) shifts to reflect the producer's taste. However, just as over-reliance on one genre can make a catalog monotonous, over-selection on a single trait can erode genetic diversity and create unforeseen vulnerabilities—a principle that echoes the founder effect and genetic drift concepts in population genetics.

SECTION 3

Visual Explanation — From Wild Ancestor to Domesticated Varieties

This branching diagram illustrates how a single wild ancestor species, Brassica oleracea, was artificially selected in different directions to produce broccoli (selection on flower clusters), cabbage (selection on terminal bud), kohlrabi (selection on stem), and kale (selection on leaves). All four varieties remain the same species, underscoring that phenotypic divergence does not require speciation.

The Brassica oleracea example is a favorite on the AP exam because it crystallizes several major ideas simultaneously. First, it demonstrates that phenotypic diversity can arise without speciation—all of these crops can still interbreed. Second, it shows that selection acts on existing variation: the wild ancestor already possessed alleles for variable leaf size, stem thickness, and inflorescence architecture. Humans did not create new mutations; they merely altered the frequencies of pre-existing alleles by choosing which individuals to propagate. Third, the diagram highlights that selection on different structures (leaves vs. stems vs. flowers) generates divergent lineages from a common ancestor—an analogy for adaptive radiation in natural populations.

Wild coastal cabbage, broccoli, cabbage, purple kohlrabi above its roots and kale accompany an authored organ-selection branch diagram. Artificial selection has emphasized different organs of Brassica oleracea: flower clusters in broccoli, the terminal bud in cabbage, an enlarged stem in kohlrabi and leaves in kale.

Selection emphasizes different organs. Artificial selection has emphasized different organs of Brassica oleracea: flower clusters in broccoli, the terminal bud in cabbage, an enlarged stem in kohlrabi and leaves in kale. Artificial selection has emphasized different organs of Brassica oleracea: flower clusters in broccoli, the terminal bud in cabbage, an enlarged stem in kohlrabi and leaves in kale.

SECTION 4

The Genetic Mechanism — Allele Frequency Shifts

At the molecular level, artificial selection does not introduce new genetic material. Instead, it reshapes the allele frequency distribution within a population. Consider a locus with two alleles, A (associated with a desired trait) and a (associated with an undesired trait). If the breeder consistently prevents aa individuals from reproducing and favors AA and Aa genotypes, the frequency of allele A rises each generation while the frequency of a declines. When this selective pressure is sustained, the population can approach fixation for the favored allele—a state where p(A) ≈ 1.0 and p(a) ≈ 0. The quantitative framework for predicting how rapidly this shift occurs integrates heritability, selection intensity, and phenotypic variance.

THE BREEDER'S EQUATION

R = h² × S

where R = response to selection (change in population mean between generations), h² = narrow-sense heritability (proportion of phenotypic variance due to additive genetic effects), and S = selection differential (difference between the mean phenotype of selected parents and the overall population mean).

SELECTION DIFFERENTIAL

S = μ_selected − μ_population

The selection differential measures the intensity of the breeder's choice. A larger S indicates that only extreme phenotypes are permitted to breed—a more stringent selection regime that, given adequate heritability, accelerates the evolutionary response.

HERITABILITY

h² = V_A / V_P

where V_A = additive genetic variance and V_P = total phenotypic variance (V_A + V_D + V_E). Traits with high h² respond rapidly to selection; traits dominated by environmental variance (low h²) respond slowly or not at all.

🎯

CONNECTING TO THE AP EXAM

The College Board expects you to recognize that artificial selection is an evolutionary mechanism that violates the Hardy-Weinberg assumption of no selection. Any question about changing allele frequencies in a breeding program connects back to the conditions for Hardy-Weinberg equilibrium. If a selective agent—human or environmental—favors certain phenotypes, allele frequencies will shift, and the population is evolving.

SECTION 5

Types of Artificial Selection & Classic Examples

Just as natural selection can be directional, stabilizing, or disruptive, artificial selection takes different forms depending on the breeder's goals. Understanding these modes is essential because AP Biology exam questions frequently ask students to identify the type of selection acting on a population based on shifts in phenotypic distributions.

The three panels compare how directional, stabilizing, and disruptive selection reshape a normally distributed trait. Dashed curves represent the original population; solid colored curves represent the post-selection distribution. In directional selection, the mean shifts toward one tail. In stabilizing selection, the variance narrows around the same mean. In disruptive selection, the distribution becomes bimodal as extremes are favored.

Selection Mode	Effect on Mean	Effect on Variance	Artificial Selection Example
Directional	Shifts toward one extreme	Often decreases (alleles fixed)	Breeding dairy cows for higher milk yield; selecting wheat for shorter stalk (Green Revolution)
Stabilizing	Stays approximately the same	Decreases (extremes removed)	Selecting tomatoes for uniform fruit size to fit standardized packaging
Disruptive	May split into two sub-means	Increases (bimodal distribution)	Dog breeding producing Chihuahuas and Great Danes from the same ancestor; ornamental goldfish varieties

Comparison of the three modes of selection and their effects on phenotypic distributions

SECTION 6

Worked Example — Applying the Breeder's Equation

A plant breeder wants to increase the average seed oil content in a sunflower population. The current population mean is 40% oil, the narrow-sense heritability for oil content is h² = 0.50, and the breeder selects only the top-performing individuals with a mean oil content of 48%. Using the breeder's equation, we can predict the expected mean oil content in the next generation.

Predicting Response to Selection in Sunflowers

1
Step 1 — Identify Given Values

Population mean (μ_pop) = 40% oil. Mean of selected parents (μ_selected) = 48% oil. Narrow-sense heritability (h²) = 0.50.
2
Step 2 — Calculate the Selection Differential (S)

S = μ_selected − μ_pop = 48% − 40% = 8 percentage points. This value quantifies how extreme the breeder's cutoff was—only parents that were 8 percentage points above average were permitted to reproduce.

S = 8 percentage points
3
Step 3 — Apply the Breeder's Equation (R = h² × S)

R = 0.50 × 8 = 4 percentage points. The response to selection tells us by how much the offspring generation's mean is expected to exceed the parental generation's mean. Because only half of the phenotypic variation is heritable (h² = 0.50), only half of the selection differential is realized as genetic gain.

R = 4 percentage points
4
Step 4 — Predict the Next Generation's Mean

New population mean = μ_pop + R = 40% + 4% = 44% oil. After one generation of selection, the expected mean oil content rises from 40% to 44%. If heritability remains stable and selection pressure is maintained, this gain can be compounded across successive generations.

Predicted next-generation mean = 44% oil
5
Step 5 — Interpret Biological Significance

Note that the offspring mean (44%) is intermediate between the population mean (40%) and the selected parents' mean (48%), weighted by heritability. This result embodies a central principle: selection can only act on heritable variation. The remaining 50% of the variance (due to environmental factors, dominance, and epistasis) does not respond to selection and cannot be transmitted to offspring.

SECTION 7

Strengths, Limitations & Ethical Considerations

Artificial selection has delivered spectacular successes—modern maize produces ears roughly 1,000 times larger than its teosinte ancestor, and broiler chickens now reach market weight in half the time they did in the 1950s. Yet these achievements carry inherent biological costs and raise questions that frequently appear on the AP exam. The following table contrasts the major advantages and limitations.

Strengths	Limitations
Rapid phenotypic change—can achieve in decades what natural selection requires millennia to accomplish	Reduced genetic diversity (genetic bottleneck)—purebred populations often have very low heterozygosity
Predictable outcomes when heritability data are available (breeder's equation)	Selection plateau—as favorable alleles approach fixation, additive genetic variance ( $V_A$ ) declines and R → 0
Addresses human needs (food security, medicine, companionship)	Unintended pleiotropic effects—selecting for one trait can compromise another (e.g., brachycephalic dog breeds with respiratory distress)
Provides direct evidence that heritable variation + differential reproduction = evolution	Inbreeding depression—small breeding populations increase homozygosity, exposing deleterious recessive alleles
Can be combined with modern genomic tools for precision (marker-assisted selection)	Vulnerability to environmental change—genetically uniform monocultures are susceptible to epidemics (e.g., Irish Potato Famine, Panama disease in bananas)

Advantages and limitations of artificial selection in biological populations

✦

KEY TAKEAWAY

Artificial selection is essentially an accelerated evolutionary experiment with a human experimentalist. Its power lies in demonstrating the core Darwinian mechanism: if traits are heritable and some individuals reproduce more than others, the population will evolve. Its danger lies in the same logic: by narrowing which alleles persist, breeders inadvertently create populations that are evolutionarily brittle—well-adapted to current conditions but poorly equipped to respond to new selective pressures. This trade-off mirrors the engineering concept of over-optimization: a bridge designed to bear one specific load will fail under unexpected forces.

SECTION 8

Connections to Natural Selection & Modern Biotechnology

Artificial selection does not exist in a vacuum—it connects directly to the broader AP Biology framework of evolution, genetic engineering, and biodiversity. Understanding how artificial selection parallels and diverges from natural selection is a frequent exam target. Similarly, modern biotechnological approaches extend the logic of artificial selection into the molecular domain, making the comparison between traditional breeding and genetic engineering an important conceptual bridge.

Feature	Artificial Selection	Natural Selection	Genetic Engineering
Selective agent	Human breeder	Environment (biotic + abiotic factors)	Scientist manipulating DNA directly
Source of variation	Pre-existing alleles (sometimes induced mutation)	Random mutation, recombination, gene flow	Transgenes, CRISPR-edited alleles, or synthetic sequences
Speed	Fast (years to decades)	Slow (generations to millennia)	Fastest (single generation possible)
Direction of selection	Human-chosen trait (may reduce fitness)	Favors traits that increase fitness in current environment	Targeted trait (precise locus)
Effect on diversity	Generally reduces within-population diversity; increases between-breed diversity	Maintains or increases diversity via balancing selection; reduces via directional	Can introduce novel diversity (transgenes from other species)
Fitness outcome	Often decreases wild fitness; optimizes for human utility	Increases fitness relative to current environment	Variable; depends on the engineered trait

Comparison of artificial selection, natural selection, and genetic engineering

A critical distinction for the AP exam: artificial selection and natural selection both change allele frequencies through differential reproduction, making them both mechanisms of evolution. However, artificial selection is the only one that is goal-directed—natural selection has no foresight or intention. This teleological distinction is philosophically important and a frequent distractor in multiple-choice questions. Meanwhile, genetic engineering (including CRISPR-Cas9) represents the logical extension of artificial selection: rather than waiting for desired alleles to arise through recombination, scientists now introduce or edit them directly. As you study for the AP exam, recognize that the same underlying principle—heritable variation leading to differential allele transmission—unites all three approaches.

💡

AP Exam Connection

Questions about artificial selection frequently appear in the context of Evidence of Evolution (Big Idea 1). Darwin's use of artificial selection as evidence for natural selection is a classic FRQ topic. Be prepared to explain why domesticated organisms provide evidence that populations can change over time and how this supports the inference that similar mechanisms operate in nature.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A dog breeder selects only the fastest individuals from each litter over many generations. Which of the following best describes the evolutionary mechanism at work and its effect on the population?

A. Genetic drift is reducing genetic diversity because only a few dogs reproduce each generation.
B. Directional artificial selection is shifting allele frequencies toward alleles that contribute to running speed.
C. Stabilizing selection is maintaining the average running speed in the population while reducing variance.
D. Gene flow from wild wolf populations is introducing new alleles that increase running speed.

Show answer (B)

Correct answer: B. The breeder imposes directional artificial selection by consistently allowing only the fastest individuals to reproduce, shifting allele frequencies toward alleles associated with speed. A is incorrect because genetic drift is a random, non-directional process; the breeder's choice is intentional and trait-specific. While reduced population size can cause drift concurrently, drift alone does not explain the consistent improvement in speed. C is incorrect because stabilizing selection favors the intermediate phenotype and removes extremes, not what is described here. D is incorrect because there is no indication of gene flow from wild populations; the change is driven by the breeder's selection decisions within the breeding population.

Question 2: BASIC CALCULATION

In a population of tomato plants, the mean fruit mass is 150 g. A breeder selects only plants with a mean fruit mass of 170 g as parents for the next generation. If the narrow-sense heritability (h²) of fruit mass is 0.60, what is the expected mean fruit mass in the F₁ generation?

A. 158 g
B. 160 g
C. 162 g
D. 170 g

Show answer (C)

Correct answer: C. Using the breeder's equation: S = 170 − 150 = 20 g; R = h² × S = 0.60 × 20 = 12 g; predicted mean = 150 + 12 = 162 g. A is incorrect (158 g corresponds to h² = 0.40, not 0.60). B is incorrect (160 g corresponds to h² = 0.50). D is incorrect because it assumes h² = 1.0 (all phenotypic variation is heritable), which would mean environment contributes zero variance—an unrealistic assumption for fruit mass.

Question 3: INTERMEDIATE

After many generations of artificial selection for increased egg production in a chicken population, the breeder notices that each generation's improvement has become negligibly small despite maintaining the same selection intensity. Which of the following best explains this observation?

A. The chickens have reached the biological maximum for egg production determined by fundamental physiological constraints.
B. Favorable alleles have approached fixation, depleting the additive genetic variance ( $V_A$ ) and reducing h², so R approaches zero despite a constant S.
C. The environment has become more variable over time, causing the chickens to express a wider range of phenotypes unrelated to genotype.
D. New deleterious mutations are being introduced at a rate that exactly offsets any gains from selection.

Show answer (B)

Correct answer: B. This describes a selection plateau. After many generations of directional selection, alleles that increase egg production become fixed or near-fixed. Since

V_A

decreases, h² =

V_A

V_P

also decreases, and R = h² × S shrinks toward zero even when S is held constant. A is incorrect because while physiological limits exist, the plateau is better explained by the exhaustion of heritable variation rather than an absolute limit. C is incorrect because increased environmental variance would increase

V_P

but does not explain the halt in genetic gain—it would actually reduce h² in a different way without implying allele fixation. D is incorrect because mutation rates are far too low per generation to quantitatively offset selection-driven gains in a large breeding program.

Question 4: APPLIED

4 pts

A researcher hypothesizes that artificial selection for drought tolerance in a crop species will reduce the population's genetic diversity at neutral loci across the genome (i.e., a genome-wide selective sweep effect). Design an experiment to test this hypothesis. In your response, include: (a) the independent and dependent variables, (b) the experimental and control groups with appropriate replication, (c) the data that should be collected and a method for measuring genetic diversity, and (d) a prediction of expected results if the hypothesis is supported.

Show answer

Point 1 — Variables (1 pt): The independent variable is the selection regime (artificial selection for drought tolerance vs. no selection). The dependent variable is genetic diversity at neutral microsatellite or SNP loci across the genome, measured as expected heterozygosity (

H_e

) or allelic richness. Point 2 — Experimental design with controls (1 pt): Establish at least three replicate populations from the same base population (to control for genetic drift). For each replicate, maintain a selected line (only drought-tolerant individuals breed) and a control line (random mating, no selection imposed). Both lines should be the same effective population size. Conduct selection for a minimum of 10 generations. Point 3 — Data collection method (1 pt): At the start and after every 2–3 generations, genotype a random sample of ~30 individuals per line using genome-wide SNP markers or microsatellites. Calculate expected heterozygosity (

H_e

), number of alleles per locus, and F_ST between selected and control lines. Compare these metrics between selection and control lines using appropriate statistical tests (e.g., paired t-test or ANOVA across replicates). Point 4 — Predicted results (1 pt): If the hypothesis is supported, the selected lines should show significantly lower

H_e

and allelic richness at neutral loci compared to control lines after multiple generations of selection. F_ST between selected and control lines should increase over time, and the magnitude of diversity loss should exceed that expected from genetic drift alone (as estimated from the control lines). If neutral loci near the selected loci show greater diversity loss than distant neutral loci, this would provide further evidence of a selective sweep.

Question 5: CRITICAL THINKING

4 pts

A plant geneticist conducted five consecutive generations of directional selection for increased stem height in a wildflower population. The data below show the population mean stem height, selected parents' mean, and offspring mean for each generation. Generation 1: Pop mean = 20 cm, Selected mean = 26 cm, Offspring mean = 23.6 cm Generation 2: Pop mean = 23.6 cm, Selected mean = 29.6 cm, Offspring mean = 27.2 cm Generation 3: Pop mean = 27.2 cm, Selected mean = 33.2 cm, Offspring mean = 29.6 cm Generation 4: Pop mean = 29.6 cm, Selected mean = 35.6 cm, Offspring mean = 30.8 cm Generation 5: Pop mean = 30.8 cm, Selected mean = 36.8 cm, Offspring mean = 31.4 cm (a) Calculate the heritability (h²) for generation 1 and generation 5. (b) Describe the trend in heritability across the five generations. (c) Explain the biological mechanism responsible for this trend. (d) Predict what would happen if the geneticist introduced individuals from a different, unselected population and continued selection.

Show answer

Point 1 — Heritability calculations (1 pt): For Generation 1: S = 26 − 20 = 6 cm; R = 23.6 − 20 = 3.6 cm; h² = R/S = 3.6/6 = 0.60. For Generation 5: S = 36.8 − 30.8 = 6 cm; R = 31.4 − 30.8 = 0.6 cm; h² = R/S = 0.6/6 = 0.10. Point 2 — Trend description (1 pt): Heritability declines progressively across the five generations, from h² = 0.60 in generation 1 to h² = 0.10 in generation 5. The response to selection (R) decreases from 3.6 cm to 0.6 cm even though the selection differential (S) remains constant at 6 cm. The population is approaching a selection plateau. Point 3 — Biological mechanism (1 pt): Sustained directional selection progressively fixes alleles that increase stem height, depleting the additive genetic variance (

V_A

) within the population. Since h² =

V_A

V_P

, as

V_A

decreases, h² declines. The remaining phenotypic variance is increasingly dominated by environmental variance (

V_E

) and non-additive genetic variance (dominance, epistasis), which do not contribute to the response to selection. Point 4 — Prediction with gene flow (1 pt): Introducing individuals from an unselected population would introduce new alleles (including alleles for increased height that may not be present in the selected population). This gene flow would increase

V_A

and thereby increase h², breaking through the selection plateau. The response to selection (R) would increase in subsequent generations as the breeder could once again select from a broader range of heritable variation. This is analogous to how outcrossing can restore heterozygosity in inbred populations.

SUMMARY

Lesson Summary — Artificial Selection

Summary

Artificial selection is the process by which humans act as the selective agent, choosing which individuals in a population reproduce based on desired phenotypic traits. This process requires three prerequisites: phenotypic variation in the trait, heritability (h²) ensuring that trait differences have a genetic basis, and differential reproduction imposed by the breeder. The breeder's equation (R = h² × S) quantifies the expected response to selection, where R is the change in the population mean, h² is narrow-sense heritability, and S is the selection differential.

Artificial selection can be directional (shifting the mean), stabilizing (reducing variance), or disruptive (favoring extremes). Classic examples include Brassica oleracea crops and dog domestication. Sustained selection can lead to a selection plateau as additive genetic variance is depleted through allele fixation, reduced genetic diversity, and inbreeding depression. For the AP exam, remember that artificial selection violates the Hardy-Weinberg equilibrium assumption of no selection, and that Darwin used it as foundational evidence for evolution by natural selection.