Sexual Selection
A peacock drags a five-foot train of iridescent feathers across the forest floor. The train costs energy to grow. It slows escape from tigers. It serves no purpose in finding food or surviving winter. By the logic of ordinary Natural Selection and the Concept of Fitness, such a burden should have been eliminated long ago. Yet peacocks, birds of paradise, deer with antlers the size of canoe paddles, and fiddler crabs with one grotesquely oversized claw are common across the tree of life. The puzzle is not an exception but a rule. Sexual selection is the branch of evolutionary theory that explains these apparent contradictions. It shows how a trait can be favored not because it helps an organism survive. Instead, it helps that organism obtain mates or fertilizations. This reframes fitness. Survival is only half the currency of evolution. The other half — reproductive access — has its own logic. It is driven by competition among individuals of one sex. It is also driven by choice exercised by the other. The concept matters because it explains an enormous swath of biological diversity. Natural selection for survival alone cannot explain it. Consider bright colors in otherwise cryptic animals. Consider weapons that endanger their bearers, and elaborate songs and dances. Consider also pervasive differences between males and females of the same species (sexual dimorphism). It also matters because sexual selection can pull traits in directions that reduce a population's mean survival. This is a genuine tension within the broader process of adaptation. It foreshadows later questions about Adaptation and Its Limits. This article works through the mechanism of sexual selection at the level of genes, gametes, and behavior. It covers the classic experiments that established it as real and testable. These include Bateman's fly experiments and their modern replication. It also presents two worked examples with quantitative reasoning. Finally, it explores living applications in fields from conservation genetics to human medicine.
- Sexual selection favors traits that increase mating or fertilization success, even at a cost to survival.
- It arises whenever variance in reproductive success is higher in one sex than the other — usually males.
- Two mechanisms: intrasexual selection (competition, usually male-male) and intersexual selection (mate choice, usually female choice of males).
- Costly ornaments can be honest signals of quality because only high-quality individuals can afford them (handicap principle).
- Bateman's 1948 fruit fly experiments and Angus Bateman's/Tim Birkhead's, Trivers's, and modern DNA-fingerprinting studies established the anisogamy-driven asymmetry.
- Fisherian runaway and good-genes models both explain how female preference and male ornament can co-evolve to extremes.
Sexual selection is natural selection acting on variance in mating success rather than variance in survival, so a trait can spread even while shortening its bearer's life, as long as it sufficiently increases the number of mates or fertilizations obtained.
The core idea
Charles Darwin introduced sexual selection in 1859 and developed it fully in The Descent of Man, and Selection in Relation to Sex (1871) precisely because ordinary natural selection could not explain traits like the peacock's train — he reportedly wrote that the sight of a peacock's feather "made him sick," so badly did it seem to violate his own theory of adaptation by utility. His resolution was to separate two kinds of fitness component. Natural selection favors traits that improve survival and general viability. Sexual selection favors traits that improve access to mates or fertilization of eggs, and it can do so even when the same trait actively reduces survival — a large antler is a target for predators and a caloric drain, but if it wins a mate it wins a chance to reproduce, which is all evolution "counts." Formally, sexual selection is a component of total selection acting specifically on variation in mating success, which decomposes further into (1) success in gaining access to mates at all, and (2) success in fertilizing eggs once mating occurs (a distinction that becomes important once we consider sperm competition and cryptic female choice). It is conventional to split sexual selection into two mechanisms:
Intrasexual selection is competition among members of one sex — typically males — for access to the other sex. This produces weapons: antlers, horns, tusks, enlarged canines, oversized claws, and the sheer body size differences seen in many polygynous mammals. The currency here is winning contests, whether physical combat, sperm competition, or territory defense.
Intersexual selection is mate choice: members of one sex — typically females — discriminate among potential mates and preferentially mate with those bearing particular traits. This produces ornaments and displays: song, color, courtship dance, elaborate structures with no combat function. The currency here is being chosen. Why is competition usually among males and choice usually exercised by females? The answer traces to a fundamental biological asymmetry called anisogamy — the near-universal difference in gamete size between the sexes. Eggs are large and metabolically expensive; sperm are small, cheap, and produced in vast numbers. This asymmetry means that, in most species, one female's reproductive output is limited by the number of eggs she can produce and provision, while one male's potential output is limited mainly by the number of mates he can obtain. Robert Trivers formalized this in 1972 as parental investment theory: the sex that invests more per offspring (usually female) becomes the limiting resource over which the lower-investing sex (usually male) competes, and becomes selective because a poor mating choice is more costly to the higher-investing parent.
Mechanism
Step 1 — Anisogamy sets the stage. Because eggs are costly and limited in number while sperm are cheap and abundant, a male's fitness is generally capped by mate number, and a female's fitness is generally capped by egg number and offspring survival, not mate number. This alone predicts asymmetric variance in reproductive success between the sexes, before any behavior is invoked.
Step 2 — Variance in reproductive success creates the "operational sex ratio." At any given moment, the ratio of sexually available males to sexually available females (the operational sex ratio, OSR) is often skewed toward males, because females are frequently occupied with pregnancy, egg-laying, lactation, or incubation while males remain ready to mate again. A skewed OSR intensifies competition among the more abundant, more available sex for access to the scarcer, more selective sex.
Step 3 — Intrasexual competition selects for weapons and dominance traits. Where males fight, larger body size, weapons, and aggression increase mating success directly. Selection here is straightforwardly analogous to ordinary Directional Selection except the fitness component under selection is number of matings rather than survival to reproductive age. Over generations, mean weapon size, body size, or aggressiveness shifts in the direction that wins more contests, producing sexual size dimorphism (e.g., male elephant seals ~4 times the mass of females).
Step 4 — Intersexual selection selects for ornaments via female choice, and choice itself must be explained. Two non-exclusive models explain why females evolve preferences for costly, seemingly non-utilitarian traits: Fisherian runaway (R. A. Fisher, 1930): Suppose females have even a slight, initially arbitrary preference for a male trait (say, tail length). Females who mate with long-tailed males produce sons with long tails and daughters who share the preference (because both the trait allele and the preference allele are transmitted together and become genetically correlated through assortative mating). This creates positive feedback — the more common the preference becomes, the more advantageous the trait becomes (because more females want it), and the more advantageous the trait becomes, the more advantageous the preference becomes (because a female who prefers it gets more attractive, more successful sons). The runaway can proceed until opposed by natural selection against ever-more-extreme, ever-more-costly ornaments — an equilibrium set by the balance of sexual selection pulling outward and natural selection pulling inward. Good genes / handicap models (Amotz Zahavi, 1975; Alan Grafen formalized mathematically, 1990): Ornaments are costly to produce and maintain, and only individuals in genuinely good condition (good genes, low parasite load, efficient metabolism) can afford full expression of the ornament without dying or losing further fitness. The cost is what makes the signal honest — a low-quality male cannot fake a large train without paying a fatal price, so the trait reliably indicates quality, and females who choose elaborately ornamented males get better genes (or better direct benefits, like healthier sperm or better-defended territories) for their offspring. William Hamilton and Marlene Zuk's 1982 parasite hypothesis is a specific version: bright ornaments signal parasite resistance because only parasite-free individuals can maintain bright, uncompromised coloration.
Step 5 — Selection continues after mating, via sperm competition and cryptic female choice. In species where females mate with multiple males, selection acts on sperm traits (sperm number, motility, longevity) that increase fertilization success relative to rival males' sperm inside the female tract — this is sperm competition. Females can also bias which sperm fertilize their eggs through physiological mechanisms of the reproductive tract — cryptic female choice — extending sexual selection beyond the visible, precopulatory phase into a hidden postcopulatory arena.
Step 6 — Sexual conflict can arise. Because male and female optimal mating rates or offspring investment strategies often differ, traits that benefit one sex's reproductive success can harm the other's — for instance, seminal proteins in Drosophila that increase a male's paternity share but shorten the female's lifespan. This introduces an evolutionary "arms race" between the sexes, distinct from but related to ordinary mate choice.
The evidence
The single most important empirical foundation for sexual selection theory is Angus Bateman's 1948 experiment on Drosophila melanogaster</strong>. Bateman set up populations of fruit flies with equal numbers of males and females, each carrying distinct visible genetic markers (mutations like curly wings or scarlet eyes), and allowed them to mate freely for several days. By counting the mutant phenotypes among the offspring, he could reconstruct — indirectly, via genetic markers rather than direct observation of copulation — how many mates each individual had actually fertilized or been fertilized by, and how many offspring each individual produced. His key finding: variance in reproductive success (number of offspring) was much greater among males than among females, and male reproductive success increased steeply and continuously with number of mates, while female reproductive success barely increased with number of mates beyond the first one or two. In other words, a male who mated with four females sired roughly four times as many offspring as a male who mated with one female, but a female who mated with four males produced approximately the same number of offspring as a female who mated with just one, because her output was capped by egg number, not sperm access. This asymmetric slope — famously plotted as "Bateman's gradient" (the regression of offspring number on mate number, steep in males, flat in females) — became the empirical signature of anisogamy's evolutionary consequence: males benefit reproductively from multiple mating; females, in most systems, do not, or benefit far less. This is the mechanistic root of why males in most species compete more intensely for mates and why females are the choosier sex. Bateman's original study, while historically foundational, had real methodological limitations recognized by later researchers — small sample sizes, the indirect genetic-marker method (which likely undercounted offspring especially from rare mating combinations), and use of a lab population under artificial density. Patricia Adair Gowaty and colleagues (2012) attempted a direct replication using Bateman's own design and found that his statistical methods were flawed in ways that inflated the sex difference in some analyses, though a real (if smaller) asymmetry persisted. This is itself a valuable lesson in the self-correcting nature of science: the qualitative pattern (steeper male gradient) has since been confirmed in many other species using modern molecular paternity analysis (DNA fingerprinting of broods to assign true parentage), even as the original numbers were revised. A second decisive line of evidence comes from field experiments manipulating ornaments directly. Malte Andersson's 1982 study of the long-tailed widowbird (Euplectes progne) in Kenya is a landmark test of intersexual selection. Male widowbirds have tails roughly 50 cm long, far exceeding what is needed for flight. Andersson caught males, then experimentally shortened the tails of some, elongated the tails of others (by cutting and gluing on extra feather segments), and left a control group unmanipulated. Territories and initial mating success were controlled for. The result: males with artificially elongated tails attracted significantly more females to nest on their territories than controls, while males with shortened tails attracted fewer. Because the manipulation was random with respect to the male's underlying quality, this proved that tail length itself, not just correlated male quality, causally affects female mate choice — a direct experimental demonstration of intersexual selection acting in the wild. A third strand of evidence supports the honesty/handicap framework: studies of barn swallows (Møller, 1988) similarly used tail-manipulation experiments and found elongated-tail males gained mating advantages, while studies linking ornament brightness to parasite load (Hamilton & Zuk's comparative analysis across 109 bird species, 1982) found brighter, more elaborately ornamented species tend to have higher blood parasite loads, consistent with ornaments serving as costly, condition-dependent signals rather than arbitrary decorations.
Worked examples
Example 1 — Reading a Bateman gradient from data. Suppose a field study of a fictional but realistic lizard species records, for 20 males and 20 females, the number of mates and number of offspring sired/produced over one season. Summary regression statistics:
| Sex | Slope (offspring per additional mate) | R² |
|---|---|---|
| Males | 3.8 | 0.71 |
| Females | 0.4 | 0.06 |
Interpretation: each additional mate increases a male's offspring count by nearly 4, a strong, statistically robust relationship (R² = 0.71 means 71% of variance in male reproductive success is explained by mate number). For females, the slope is shallow and R² is low — mate number explains almost none of the variance in female reproductive success, meaning female output is capped by something else (likely clutch size or resource access, not mate number). This pattern predicts, before even observing behavior directly, that males in this species should show more intense competition for mates (intrasexual selection, e.g., male-male combat or territory defense) and that females should be the choosier sex — testable predictions that a researcher could now go verify by direct behavioral observation.
Example 2 — Handicap-principle cost-benefit reasoning. Consider a male bird species where a "high-quality" male (strong immune genes, efficient metabolism) can grow a bright orange breast patch at a survival cost of 5% (i.e., his annual survival probability drops from 0.80 to 0.76 due to the metabolic burden and predator conspicuousness). A "low-quality" male attempting the same patch size would face a survival cost of 25% (dropping from 0.60 to 0.45), because he has fewer resources to spare and a weaker immune system, making the ornament riskier for him. Suppose growing the ornament increases mating success from an average of 1.0 mates/season to 2.5 mates/season for any male who displays it, regardless of quality (the ornament itself is what females assess, not directly the underlying genotype). Compute expected offspring for each strategy, assuming 4 offspring per mate on average: High-quality male, no ornament: 0.80 × 1.0 × 4 = 3.2 expected offspring. High-quality male, with ornament: 0.76 × 2.5 × 4 = 7.6 expected offspring. → Ornament strongly favored. Low-quality male, no ornament: 0.60 × 1.0 × 4 = 2.4 expected offspring. Low-quality male, with ornament: 0.45 × 2.5 × 4 = 4.5 expected offspring. → Ornament still favored numerically here, but the relative survival cost (25% vs 5%) is what matters over repeated seasons and under realistic multiplicative lifetime fitness, and in most real systems additional factors (inability to fully express patch size/brightness with fewer resources, condition-dependent display quality, or higher variance/risk of complete reproductive failure from early death) push the balance so that only high-quality males can sustain repeated ornamentation across a lifetime. This is the essence of Zahavi's handicap logic: the cost need not make ornamentation never pay off for low-quality individuals in a single season, but differential costs across quality classes keep the signal statistically honest on average, because low-quality males who chronically over-invest in ornament are disproportionately removed from the breeding population by mortality — which is precisely what field data on condition-dependent ornaments (e.g., carotenoid-based coloration, dependent on foraging efficiency and parasite load) generally show.
Common misconceptions & traps
| Misconception | Reality |
|---|---|
| "Females always choose; males always compete." | Sex roles reverse when investment is reversed — e.g., in pipefish, seahorses, and phalaropes, males invest more (pregnancy/brooding) and females compete for and display to males. |
| "Sexual selection and natural selection are opposing forces that cancel out." | They are both components of total selection; sexual selection can align with or oppose survival selection, and the observed trait reflects the balance, not a war with one fixed winner. |
| "A costly ornament proves the female preference is 'irrational.'" | Under good-genes and handicap models, the preference is adaptive: it tracks a genuinely informative, hard-to-fake signal of quality. |
| "Bateman's original numbers are the final word." | Later reanalysis (Gowaty et al. 2012) found statistical flaws in his method; the qualitative sex-difference pattern holds broadly but the specific 1948 estimates were revised. |
| "Sexual dimorphism is always about mate choice (ornaments)." | Much dimorphism (body size, weapons) arises from intrasexual combat, not female choice, and the two mechanisms can act on different traits in the same species simultaneously. |
| "Sperm competition ends the sexual selection story at copulation." | Selection continues post-copulation via sperm competition and cryptic female choice, hidden processes that can override precopulatory mate choice. |
The scientist behind it
Sexual selection's modern experimental foundation owes most directly to Angus John Bateman (1919–1996), a British geneticist and horticulturist working largely in isolation at the John Innes Horticultural Institution. Bateman was not primarily an evolutionary theorist by career — he worked chiefly on plant genetics and crop breeding — but his 1948 paper on Drosophila mating systems, published in the journal Heredity, became one of the most cited and most contested papers in the history of evolutionary biology. Working with visible marker mutations and simple population cages rather than any advanced instrumentation, he produced data that, decades later, DNA-based paternity studies would broadly (if imperfectly) confirm: that reproductive variance among males vastly exceeds that among females, giving a mechanistic, quantitative anchor to Darwin's largely qualitative 1871 arguments about male competition and female choice. Bateman's paper sat in relative obscurity for over two decades until Robert Trivers revived and extended it in his enormously influential 1972 chapter "Parental Investment and Sexual Selection," which explicitly linked Bateman's gradient to the theory of differential parental investment and cemented both into the mainstream of evolutionary biology.
Origins · 1 minAngus John Bateman and how this was discovered
Where this matters today
Sexual selection theory is not a museum piece; it is actively used across biology and medicine. In conservation genetics, understanding mate choice and sexual selection dynamics matters for captive breeding programs — poorly managed captive populations can inadvertently select against naturally important ornaments or dominance behaviors, reducing reintroduction success, and small population sizes can distort the operational sex ratio in ways that accelerate loss of genetic diversity through skewed male reproductive success. In agriculture and animal breeding, artificial selection programs must account for sexually selected traits (e.g., competitive mating behavior in livestock) that can conflict with production traits, and understanding sperm competition dynamics informs artificial insemination and breeding-stock selection in cattle, poultry, and fish farming. In evolutionary medicine, parental investment asymmetries and sexual conflict models help explain patterns of sex-specific disease risk, the evolution of genomic imprinting (parent-of-origin gene expression conflicts rooted in differential investment), and even aspects of human mating psychology studied in evolutionary psychology, though extrapolation from animal models to human behavior requires caution given the overwhelming role of culture. In pest control and vector biology, sperm competition theory informs sterile-insect-technique programs (used against agricultural pests and malaria-vector mosquitoes) by predicting how many sterile males must be released to outcompete wild males for fertilizations. Finally, sexual selection remains a live research frontier in understanding speciation: divergent mate-choice preferences between populations can drive rapid reproductive isolation, making sexual selection one of the fastest engines of biological diversification documented in nature, particularly in cichlid fish radiations and Hawaiian Drosophila.
Ask the tutor: How would the predictions of Fisherian runaway versus good-genes models differ if you experimentally removed all natural-selection costs of an ornament — which model predicts the trait would still stabilize, and which predicts truly unbounded growth?
Is sexual selection a type of natural selection, or something separate?
Sexual selection is best understood as a specific component of total selection, operating through the same underlying process as natural selection — differential reproductive success of heritable variants — but partitioned to focus on variance arising from mating and fertilization success rather than general survival. Darwin originally treated them as conceptually distinct because sexually selected traits can actively reduce survival, which seemed paradoxical under a pure "survival of the fittest" framing. Modern evolutionary biology unifies them: total fitness is the product (or sum, depending on the model) of survival and mating success, and any trait's fate depends on its net effect across both components. Sexual selection is therefore a subset of natural selection in the broadest technical sense, distinguished because of its distinct causal mechanism (competition and choice) and its capacity to oppose viability selection.
Why is female choice usually considered the more common form of intersexual selection rather than male choice?
This asymmetry follows from anisogamy and parental investment theory: because females typically invest more per offspring (larger, costlier eggs, and often gestation or brooding), a poor mating decision is more costly to a female than to a male, who can typically recover reproductively by seeking another mate quickly. This makes discrimination more valuable for the higher-investing sex. It is not a universal rule — species with reversed parental investment (seahorses, pipefish, some shorebirds like phalaropes) show reversed sex roles, with males being choosier and females competing and displaying. The underlying rule is investment asymmetry, not maleness or femaleness per se.
Do humans show evidence of sexual selection?
Human sexual dimorphism in body size, the existence of secondary sexual characteristics, and cross-cultural patterns in mate preferences are consistent with some history of sexual selection acting during human evolution, and researchers in evolutionary psychology and human behavioral ecology study this actively. However, human mating behavior is heavily shaped by culture, social institutions, contraception, and individual variation in ways that make direct extrapolation from animal experiments risky; claims about specific human preferences being sexually selected adaptations require the same standard of evidence (heritability, fitness consequences, cross-cultural replication) as any other evolutionary hypothesis, and many popular claims in this area remain actively debated rather than settled.
What stops the Fisherian runaway process from continuing forever, producing infinitely long tails?
The runaway is eventually checked by natural selection acting against the ornament's survival costs. As the trait grows more extreme, its burden on survival — energetic cost, predation risk, mechanical impairment — increases, until the marginal gain in mating success from further elaboration is exactly offset by the marginal loss in survival. This produces a stable equilibrium trait size, not infinite growth, and predicts that ornament size should correlate with the strength of countervailing survival costs across species and environments — a pattern generally supported by comparative data (e.g., populations with fewer predators tend to evolve more extreme ornaments).
How do biologists distinguish a trait shaped by intrasexual selection from one shaped by intersexual selection?
Researchers use several converging lines of evidence: direct behavioral observation (does the trait function in male-male combat, or is it displayed to females during courtship without contact?), experimental manipulation (does altering the trait change contest outcomes, mate choice outcomes, or both — as in Andersson's tail-manipulation study), and correlational analyses linking trait variation to either winning fights or attracting mates. Many traits serve both functions simultaneously (a large deer antler both wins fights and may itself be assessed by females as a quality signal), so the two mechanisms are not mutually exclusive and are often studied together as complementary components of a single trait's evolutionary history.
Why did Gowaty's 2012 reanalysis of Bateman's data matter, and does it undermine sexual selection theory?
Gowaty and colleagues physically repeated Bateman's exact experimental protocol and found statistical and methodological issues in his original 1948 analysis, including problems with how offspring from marker-gene crosses were counted, which likely inflated the apparent male-female difference in some of his data sets. This does not undermine sexual selection theory itself, because the broader qualitative pattern — greater variance in male reproductive success linked to anisogamy — has since been independently confirmed across many species using modern molecular paternity testing (DNA fingerprinting), which did not exist in Bateman's era. It is, instead, a textbook example of science self-correcting a foundational study while the theoretical framework it helped establish survives on stronger, more diverse evidence.
Can sexual selection drive speciation?
Yes, and this is one of its most active current research areas. When mate-choice preferences diverge between populations — whether by drift, adaptation to different sensory environments, or runaway co-evolution with local male traits — individuals from different populations may cease to recognize each other as attractive mates, creating reproductive isolation even without any geographic barrier or ecological difference. This mechanism is implicated in rapid, spectacular radiations such as African cichlid fish (differing male coloration and matching female preference across hundreds of species in the same lakes) and Hawaiian Drosophila, making sexual selection a leading hypothesis for some of the fastest known instances of speciation in nature.