The central idea

Natural selection is an evolutionary process in which individuals with heritable traits that give higher survival or reproductive success leave a greater proportion of offspring than individuals lacking those traits. The key word is heritable. Selection doesn’t reward effort; it changes populations by changing which inherited variants get passed on.

Darwin’s argument can be put simply:

• individuals in a population vary;
• more offspring are produced than the environment can support;
• individuals compete, die, survive and reproduce unequally;
• if the useful differences are inherited, the next generation contains more of the alleles associated with those differences.

Evolution is a change in the heritable characteristics of a population over generations. Natural selection, then, is one mechanism of evolutionary change. It sorts existing variation and changes what becomes common in later generations. It operates all the time — gently in stable conditions, strongly when environments change — and over billions of years it has contributed to the enormous biodiversity of life on Earth.

A useful diagram here would show a simple flow: variation, then differential survival and reproduction, then a change in trait frequency in the population.

Why this was a paradigm shift

A paradigm is a framework of ideas and assumptions that scientists use to interpret evidence. A paradigm shift is a major replacement of one explanatory framework by another because the new framework explains observations more convincingly.

In Darwin’s time, many educated people accepted that species could change. The argument was over the mechanism. Lamarckism is a hypothesis of evolution in which traits acquired during an individual’s lifetime are inherited by its offspring. Darwin’s theory replaced that idea with a mechanism based on inherited variation already present among individuals and unequal reproductive success. This is a classic paradigm shift: people were looking at the same living world, but now they explained it through selection among heritable variants rather than inheritance of acquired improvements.

This connects to the wider question of what changes allele frequencies. Natural selection gives one answer: if some inherited variants help their carriers leave more offspring, the alleles associated with those variants tend to increase in frequency.

Mutation supplies new alleles

Mutation is a change in the base sequence of DNA that can create a new allele. It is the original source of new genetic variation. Without mutation, selection would have no new alleles to favour or remove. In a particular environment, most mutations are neutral or harmful, though occasionally a mutation produces a phenotype that gives an advantage under current conditions.

An allele is a version of a gene that differs from other versions at the same locus. Natural selection acts on phenotypes, but over the long term, evolution depends on whether the alleles affecting those phenotypes are passed to offspring.

Sexual reproduction reshuffles alleles

Sexual reproduction is a reproductive process in which gametes from two parents fuse to form genetically varied offspring. It does not usually create brand-new alleles; mutation does that. What it mainly does is create new combinations of alleles.

During meiosis, crossing over and independent assortment produce gametes with different allele combinations. Fertilization then brings together alleles from two parents. Over generations, a beneficial mutation that first appeared in one individual can end up combined with beneficial alleles that first appeared in other individuals.

This is why reproduction matters in natural selection. Selection can only favour some individuals over others if variation exists, and sexual reproduction keeps rearranging that variation into new genotypes. In an asexual population, mutation can still happen, but variation accumulates and combines more slowly because offspring are usually genetic copies except for new mutations.

More offspring are produced than can survive

Overproduction of offspring means producing more young than can survive to reproductive age in the available environment. It isn’t only seen in species with huge broods or large numbers of seeds. Even organisms that breed slowly may, across a lifetime, produce more offspring than the habitat could support if every one survived.

The result is a struggle for existence. Some seeds miss out on light and water. Some juvenile animals fail to get enough food, shelter or a territory. Some adults do not find a mate. Variation matters here: when individuals differ in ways that affect their success in gaining limiting resources, selection can occur.

Carrying capacity and limiting resources

Carrying capacity is the maximum population size of a species that an environment can support sustainably with its available resources. The resource in shortest supply often sets this limit; this is called a limiting resource.

Examples of resources that may limit carrying capacity include:

• food for herbivores, predators or decomposers;
• water for desert plants and animals;
• light for plants growing under a forest canopy;
• mineral ions in soil for plant growth;
• nesting sites, territories or shelter;
• suitable mates in sparse populations.

Intraspecific competition is competition between individuals of the same species for the same limited resource. It is different from interspecific competition, where individuals of different species compete. Overproduction often leads to natural selection through intraspecific competition, since members of the same species usually need very similar resources. Competition can be reduced when organisms occupy different niches, disperse seeds or young away from parents, or use different resources at different life stages.

Selection pressures can be non-living features of the environment

A selection pressure is an environmental factor that affects the survival or reproductive success of individuals with different phenotypes. Selection pressures can be biotic, such as predation or competition for food, or abiotic, such as temperature, salinity or drought.

An abiotic factor is a non-living physical or chemical component of the environment that affects organisms. High temperature, low temperature, water availability, pH, salinity, light intensity, fire, flooding and pollution can all act as abiotic selection pressures.

Density-dependent and density-independent factors

A density-dependent factor is a factor whose effect on birth rate or death rate changes as population density changes. Food shortage shows this clearly: in a crowded population, competition for food becomes more intense.

A density-independent factor is a factor whose effect on individuals is not mainly determined by population density. A severe frost may kill cold-sensitive plants whether they grow close together or far apart. A heat wave, drought, flood or sudden pollution event can also affect survival without first depending on how many individuals are present per unit area.

Low temperature does not “try” to select. It simply kills or weakens individuals unequally. If cold tolerance varies and is heritable, individuals with alleles contributing to cold tolerance are more likely to survive and reproduce. The abiotic factor has become a selection pressure.

Adaptation and fitness

An adaptation is a heritable characteristic that increases an organism’s chance of survival or reproduction in a particular environment. Adaptations can be structural, behavioural or physiological. Beak shape, courtship behaviour and tolerance to salt can all count as adaptations, if they improve reproductive success in the relevant environment.

Fitness is the relative reproductive success of a genotype or phenotype in a particular environment. In everyday use, fitness often means strength or health. In evolution, it means contribution to the next generation. A genotype with high fitness leaves more surviving, fertile offspring than alternative genotypes in the same population.

Fitness is always tied to the environment. A trait that helps in one setting may be neutral or harmful in another. Thick fur helps in cold climates but carries a cost in hot ones. Camouflage depends on the background. Large horns may help males win mates, but they may also make them more likely to be targeted by hunters.

Survival value and reproductive potential

Survival value is the advantage a trait gives by increasing the chance that an individual survives long enough to reproduce. Reproductive potential is the capacity of an individual or genotype to produce surviving offspring. Natural selection depends on both: surviving without reproducing does not pass alleles on, and producing many offspring that do not survive has little evolutionary effect.

Intraspecific competition matters here. Members of the same species may compete for food, space, water, mates, nesting sites or light. Individuals with more suitable heritable traits gain a larger share of these resources, survive more often, or reproduce more successfully. Their alleles then form a larger fraction of the next generation.

Illustrative cohort showing how higher survival and reproduction make one heritable phenotype contribute more to the next generation.

This section answers one of the linking questions directly: intraspecific interactions occur within one species, while interspecific interactions occur between species. Natural selection within a population often depends on intraspecific differences — some individuals of the same species are better adapted than others under the same pressures.

Acquired traits are not the same as inherited traits

A heritable trait is a characteristic shaped by genetic information and passed from parents to offspring. Natural selection can cause evolutionary change only when the differences being selected are heritable.

An acquired characteristic is a trait that appears during an individual’s lifetime because of environmental influence or use and disuse, not because the base sequence of genes in gametes has changed. A scar, a missing limb, a suntan or larger muscles from training may affect that individual, but they are not encoded as new base sequences in sperm or eggs.

This distinction matters. If an animal becomes stronger through exercise, its gametes do not acquire DNA sequences for “stronger muscles”. If a plant grows taller because it received more light and minerals, that extra height is not automatically written into the genes passed to its seeds.

Gene expression can change without changing DNA sequence

The environment can affect phenotype by changing gene expression. Temperature, for example, may affect pigment production in some mammals. But changing gene expression in body cells is not the same as mutating the DNA of gametes. Unless the genetic difference is present in cells that give rise to gametes, it cannot be inherited in the usual Mendelian sense and cannot drive long-term evolutionary change by natural selection.

Sometimes chemical markers that affect gene expression can pass to offspring for a short time. That still isn’t the same as changing the base sequence of genes. For this syllabus point, keep the main rule clear: natural selection needs heritable variation, not simply environmentally acquired differences.

Mate choice can change populations

Sexual selection is a form of natural selection where differences in mating success make some heritable traits more common. It happens when physical or behavioural traits affect how well an individual attracts a mate or competes for mates.

In many animal species, mating is not random. A potential mate may judge signals such as colour, size, song, display vigour, territory quality or courtship behaviour. These signals can indicate overall fitness: an animal able to grow bright plumage, perform repeated displays or defend a good territory may also carry alleles linked to health, stamina or parasite resistance.

Birds of paradise make a lovely example. In several species, males have elaborate plumage and perform complex courtship displays. If females tend to mate with males that show the most impressive displays, alleles contributing to those displays are passed on more often. Over generations, the population can evolve more exaggerated plumage and behaviour.

There is often a trade-off. A bright colour or long feather may attract mates, but it may also attract predators or make movement harder. Sexual selection can therefore push traits in a direction that looks costly if we only consider survival. In evolutionary terms, the trait persists if the mating advantage outweighs the survival cost.

Modelling selection by controlling the pressure

A model is a simplified representation of a biological process used to test or explain how the process works. In selection models, you control the selection pressure, then check whether trait frequencies shift over generations.

In class, you might model this with artificial “organisms” of different shapes placed in water. Choose the fastest-sinking shapes as survivors, use them to make the next generation, and the shapes should gradually move toward whichever form performs best under that rule. Clay shapes aren’t evolving. The model shows how repeated non-random survival, combined with inheritance of the successful form, can change a population.

Endler’s guppy experiments

John Endler’s guppy work is the classic data example here because he experimentally controlled predation pressure. Male guppies differ in the number and area of their coloured spots. Bright, spotted males may be preferred by females, so colour can increase mating success. Those same conspicuous males may also be easier for predators to see.

In experimental ponds or streams with no predators or weak predators, male colouration tended to increase. With strong predators, conspicuous colouration tended to decrease. That’s the neat balance in the story: sexual selection can favour showier males, while natural selection by predation can favour less conspicuous males.

When you interpret Endler-style data, look for three things: the selection pressure being changed, the trait being measured, and the direction of change over time. If predator type is the controlled variable and mean spot number or spot area changes across generations, the data support the conclusion that selection pressure changed reproductive success of different male phenotypes.

What a gene pool includes

A population is a group of organisms of the same species living in the same area and able to interbreed. A gene pool means the complete set of genes and all their alleles present in a population.

The gene pool belongs to the population, not to one individual. Any one individual carries only a sample of the alleles that exist in that population. When individuals reproduce, they pass alleles into the next generation’s gene pool. Individuals that do not reproduce may have lived successfully for a while, but their alleles are not represented through their offspring.

The gene pool idea is useful because evolution is measured at the population level. A single organism does not evolve during its lifetime. A population evolves when allele frequencies in its gene pool change from one generation to the next.

Isolation allows allele frequencies to diverge

Allele frequency is the proportion of all copies of a gene in a population that are a particular allele. When populations of the same species become geographically isolated, interbreeding may no longer pass alleles between them. Over time, mutation, natural selection, genetic drift and migration history can shift their allele frequencies apart.

Geographical isolation is physical separation of populations by distance or barriers such as oceans, mountains, deserts or ice sheets, reducing or preventing gene flow between them. With less gene flow, each population follows its own evolutionary history.

Database-style comparison of ADH1B rs1229984-A (His48) allele frequencies in separated human populations.

Using databases

Allele frequency databases let biologists compare populations using real genetic data. One human example is variation in the alcohol dehydrogenase gene family. Some alleles of alcohol dehydrogenase are linked with faster ethanol metabolism, and their frequencies differ among human populations in different regions.

When using a database, a sensible workflow is:

• search by gene name, locus name or SNP identifier;
• check which allele is being reported;
• compare allele frequencies among named populations;
• note the sample size and geographic origin;
• avoid assuming that a frequency pattern proves a cause unless there is supporting evidence.

A database map gives descriptive evidence: it shows where alleles are common or rare. Possible explanations include selection, founder effects, drift, migration or cultural history, but each one needs separate evidence.

Natural selection translated into genetics

Darwin described evolution by natural selection before anyone knew the genetic basis of inheritance. Mendelian genetics later filled that gap: genes influence traits, genes have alleles, and gametes pass those alleles on.

Neo-Darwinism is the modern synthesis of Darwin’s natural selection with genetics. It explains evolution as changes in allele frequencies caused by processes such as selection, mutation, gene flow and genetic drift. In this view, natural selection acts on phenotypes, while evolution is measured as allele frequency change in the gene pool.

Put genetically, selection works like this. Individuals with different heritable traits do not survive and reproduce equally. If a genotype gives higher fitness in a particular environment, alleles associated with that genotype are passed on more often. By the next generation, those alleles make up a larger proportion of the gene pool.

This gives the clearest answer to the guiding question about processes that change allele frequencies. Natural selection changes allele frequencies when survival or reproduction differs between genotypes. Mutation, migration and genetic drift can also change allele frequencies, but natural selection is distinctive because it is non-random with respect to fitness.

Three patterns of selection

Directional selection is a pattern of natural selection where one extreme phenotype has the highest fitness, so the population mean shifts in that direction. For example, if larger body size improves survival or reproduction, alleles linked with larger size may become more common.

Stabilizing selection is a pattern of natural selection where intermediate phenotypes have the highest fitness, while extreme phenotypes are selected against. Human birth mass is a common example: very low and very high birth masses can have higher mortality than intermediate masses.

Disruptive selection is a pattern of natural selection where two or more extreme phenotypes have higher fitness than intermediate phenotypes. Variation can be maintained or increased if different strategies work well at the extremes.

All three patterns change allele frequencies. Say that clearly, because students sometimes think stabilizing selection “does nothing”. The mean may stay similar, but alleles associated with extreme phenotypes can still become less frequent.

The equations

The Hardy–Weinberg model uses allele frequencies to predict genotype frequencies in a population that meets specific equilibrium conditions.

$p + q = 1$, where $p$ is the frequency of one allele in the population (dimensionless proportion) and $q$ is the frequency of the other allele in the population (dimensionless proportion).

For two alleles, the predicted genotype frequencies are:

$p^2 + 2pq + q^2 = 1$, where $p^2$ is the predicted frequency of the homozygous genotype for the $p$ allele (dimensionless proportion), $2pq$ is the predicted frequency of the heterozygous genotype (dimensionless proportion), and $q^2$ is the predicted frequency of the homozygous genotype for the $q$ allele (dimensionless proportion).

For a recessive condition, affected individuals are usually homozygous recessive, so the affected genotype frequency is $q^2$. Take the square root to find $q$, then use $p = 1 - q$ to find $p$. Once you know $p$ and $q$, the genotype frequencies come from $p^2$, $2pq$ and $q^2$.

Worked pattern without getting lost

If a recessive phenotype has frequency 0.0004, start with $q^2 = 0.0004$. So $q = 0.02$. Then $p = 1 - 0.02 = 0.98$. The predicted heterozygote frequency is $2pq = 2(0.98)(0.02) = 0.0392$.

A common mistake is to treat the recessive phenotype frequency as $q$ instead of $q^2$. For a recessive phenotype, affected individuals have two recessive alleles, so start with $q^2$.

Genetic equilibrium

Genetic equilibrium is the state where allele and genotype frequencies in a population stay the same from one generation to the next. The Hardy–Weinberg model gives the genotype frequencies expected for a two-allele gene when a population is in this state.

For Hardy–Weinberg equilibrium, these conditions must hold:

• no mutation creating or changing alleles;
• random mating with respect to the gene being studied;
• no immigration or emigration changing allele frequencies;
• a very large population, so chance effects are negligible;
• no natural selection, so survival and reproductive success do not differ between genotypes.

Genetic drift is a change in allele frequency caused by chance events, especially in small populations. That’s why the “large population” condition is included.

What if the data do not fit?

If observed genotype frequencies do not match Hardy–Weinberg predictions, at least one condition is not being met. Non-random mating, mutation, migration, drift or natural selection could be the cause. For example, when survival rates differ between genotypes, the adult population may have fewer of one genotype than expected.

Hardy–Weinberg is not just a calculation trick. It works as a null model. If real data match the model, those data give no evidence that allele frequencies are changing. If real data differ from the model, biologists then ask which evolutionary process is disturbing equilibrium.

Humans as the selecting agent

Artificial selection is selective breeding where humans choose individuals with desired heritable traits to reproduce. You see it often in crop plants and domesticated animals. Farmers, breeders and growers may select for yield, taste, growth rate, docility, disease resistance, milk production, egg production, seed size or flower colour.

It still counts as evolution, because allele frequencies change over generations. The key difference is the selecting agent. In natural selection, environmental pressures decide which individuals leave more offspring. In artificial selection, humans deliberately choose which individuals breed.

Crop domestication shows this clearly. A wild grass with small seed heads can give rise to crop varieties with much higher yield, after many generations of selecting plants with larger and more useful seed heads. Domesticated animals follow the same pattern: breeding again and again from individuals with preferred traits can produce large changes over surprisingly short timescales.

Natural selection can also result from human actions

Not every evolutionary change linked to humans is artificial selection. When humans deliberately choose parents for breeding, that is artificial selection. When humans create a selection pressure without intending to select particular parents, natural selection still does the sorting.

Antibiotic resistance in bacteria is the usual example. When an antibiotic is used, susceptible bacteria are killed more often, while resistant bacteria survive and reproduce. Humans created the environment, but they did not choose particular bacteria as parents in a breeding programme. So this is natural selection under a human-imposed selection pressure, not artificial selection.