How Does Artificial Selection Work?

Artificial selection is the process by which humans change a population by choosing which organisms reproduce. Instead of environmental conditions determining which individuals leave the most offspring, people deliberately favor traits such as larger fruit, higher milk production, docile behavior, or particular body shapes.

The process is a form of evolution because the genetic composition of the population changes across generations. It is not the same as creating a trait from nothing: selection can increase the frequency of inherited variants that already exist, and new mutations can also provide variation for later selection.

The basic mechanism

Artificial selection requires three conditions:

1. Variation: Individuals differ in a trait.
2. Heritability: At least part of the difference is genetically transmitted to offspring.
3. Differential reproduction: Humans allow some individuals to produce more offspring than others.

Suppose a farmer wants sheep with more wool. Some sheep naturally produce more wool than others. If wool production is partly heritable, the farmer selects the high-producing sheep as parents. Their offspring tend, on average, to produce more wool than the original population. Selecting the best wool producers again in the next generation can increase the average further.

In population-genetic terms, genetic variants that contribute to the preferred phenotype tend, on average, to increase in frequency when selection is effective. For a polygenic or environmentally influenced trait, however, not every allele associated with the phenotype will necessarily increase.

The population evolves even though the selection pressure comes from human decisions rather than predators, climate, disease, or competition.

A step-by-step example

Imagine a tomato population with fruit weights ranging from 80 to 180 grams. A grower wants larger tomatoes.

Generation 1: choose parents

The grower saves seeds only from plants producing tomatoes near the upper end of the range, perhaps those averaging 170 grams. This does not guarantee that every offspring will be large, because fruit size is influenced by several genes and by environmental factors such as water, nutrients, and sunlight.

Generation 2: measure the offspring

The selected seeds are planted. If fruit size is heritable, the offspring will have a higher average fruit weight than the original population. The grower again keeps seeds from the largest healthy plants.

Later generations: repeat

After several generations, the population may produce substantially larger tomatoes on average. The change is gradual because each generation contains genetic variation, and selection shifts the frequencies of many alleles rather than simply switching one gene on.

The grower can also select for several traits at once, such as large fruit, disease resistance, and good flavor. However, selecting for one trait may cause an unwanted change in another if the traits are genetically correlated. For example, very large fruit might be associated with weaker stems or later maturation.

Examples of artificial selection

Artificial selection has shaped many familiar organisms:

Organism	Selected characteristics
Dogs	Size, body shape, coat, behavior, and working ability
Maize	Larger ears and more abundant edible seeds
Broiler chickens	Rapid growth and increased meat production
Dairy cattle	Higher milk production
Brassica crops	Different structures selected from related plants, including cabbage, broccoli, kale, and Brussels sprouts
Domestic horses	Speed, strength, size, temperament, and endurance

The many dog breeds are a particularly visible example. Humans selected among early domesticated dog populations, whose ancestors were ancient wolf populations, for behavior and appearance. Repeated breeding produced forms ranging from very small companion dogs to large working breeds. Their common ancestry does not mean every modern breed appeared in a single generation; breed differences accumulated through many generations of selection.

Artificial selection versus natural selection

Artificial and natural selection use the same evolutionary ingredients: variation, inheritance, and unequal reproductive success. The main difference is who or what determines reproductive success.

Feature	Artificial selection	Natural selection
Selective agent	Humans	Environment and ecological interactions
Main criterion	Human preference or usefulness	Survival and reproduction in a particular environment
Typical direction	Often rapid and strongly directed	Determined by changing conditions and trade-offs
Outcome	Crops, breeds, or production traits	Adaptations to environmental conditions

The distinction is about the source of selection, not whether genetics is involved. Artificial selection can be powerful because humans may prevent most individuals from breeding or may control which individuals mate.

Why selection works only sometimes

A trait must have a heritable component for selective breeding to produce a lasting genetic change. If a plant is tall only because it received more water, choosing that plant as a parent may not produce taller offspring in the same conditions.

The amount of available genetic variation also matters. If every individual has nearly identical alleles for a trait, there may be little for selection to change. Breeders may introduce variation by crossing different lines, preserving rare individuals, or using mutation and chromosome-based breeding methods in some crops.

The response also depends on how strongly selection is applied and how accurately the trait is measured. Selecting one exceptional individual can be less effective than selecting several individuals whose performance is consistently high across environments.

Under standard quantitative-genetic assumptions, the expected response in the population mean in one generation can be summarized by the breeder’s equation:

$$R = h^2S$$

Here, $R$ is the expected change in the population mean in one generation, $h^2$ is the narrow-sense heritability of the trait, and $S$ is the difference between the selected parents’ average and the population average. Higher heritability and stronger selection generally produce a larger response, although real breeding programs must also consider health, fertility, and genetic diversity.

Costs and limits of artificial selection

Selecting intensely for a narrow goal can reduce genetic diversity. That may increase inbreeding, making harmful recessive conditions more likely and reducing a population’s ability to respond to new diseases or environmental changes.

Some traits also have biological trade-offs. Chickens bred for extremely rapid muscle growth may experience leg or heart problems. Dogs selected for particular skull shapes may have breathing difficulties. These outcomes are not evidence that selection failed; they show that maximizing one characteristic can affect other traits and overall fitness.

Modern breeding programs often reduce these risks by recording pedigrees, using genetic testing, maintaining multiple breeding lines, and selecting an index of traits rather than a single feature. Ethical breeding therefore considers animal welfare and long-term population health, not just the desired appearance or production measure.

Common mistakes

• “Artificial selection is not evolution.” It is evolution when inherited allele frequencies change over generations.
• “Breeders directly change an organism’s genes in every case.” Traditional selective breeding usually changes gene frequencies by controlling reproduction; it does not directly edit DNA.
• “The selected trait must be controlled by one gene.” Many selected traits, including size, yield, behavior, and milk production, are influenced by many genes and the environment.
• “The largest or most extreme individual is always the best parent.” An extreme phenotype may result partly from temporary environmental conditions, and extreme selection can create health or fertility problems.
• “Artificial selection always makes organisms better.” It makes them more suited to a human-defined goal, which may not improve survival, health, or performance in other environments.