Frequency-Dependent Selection¶

Biological selection in which a type's relative fitness depends on how common it or other types are in the population.

Version: v1 · 2026-10-03 · History
Domain-specific #: 13253
Domain group: Natural Sciences
Origin domain: Biology & Ecology
Subdomain: Evolutionary Biology → Biology & Ecology

Core Idea¶

Frequency-dependent selection occurs when the relative survival or reproductive success of a biological type depends on its frequency, or on the composition of other types, in the relevant population. A type therefore cannot be assigned one context-free selective value for all population compositions. In a model this is expressed by fitness functions \(w_i(p)\) for type \(i\) and frequency vector \(p\); what matters is that the relative selective difference changes with \(p\), not that every field study uses this exact notation. Trotter and Spencer's original population-genetic model treats genotype fitness as composition-dependent rather than fixed.^[1]

Negative frequency dependence describes a range in which a type does relatively better when rare or worse as it becomes common; positive dependence describes a range in which commonness raises relative success. These are local, context-dependent signs, not promises of stable polymorphism or inevitable fixation. Hughes and colleagues report rare-color male guppies with higher reproductive fitness; Chouteau and colleagues find protection of warning signals increasing with local frequency among defended butterflies, with saturation for the most common signals. Those are unlike biological channels with the same frequency-to-relative-fitness structure.^[2]^[3]

Structural Signature¶

Sig role-phrases: biological type population — current type frequencies — frequency-conditioned relative fitness — interaction channel — differential continuation — outcome qualifications.

Biological type population. Heritable genotypes or phenotypes provide alternatives. Without alternatives there is no comparative frequency-dependent selection among types.^[1]
Current type frequencies. Their proportions are indexed in a defined locality and time; “rare” and “common” have no meaning outside that reference population. The warning-signal study explicitly uses local frequencies rather than one global frequency.^[3]
Frequency-conditioned relative fitness. A type's survival or reproductive success varies with its own prevalence or the composition of competitors, mates or other interactors. Fixed relative fitness across frequency changes would be frequency-independent selection.^[1]
Interaction channel. Mate-mediated success in the guppy study and predator responses to warning signals in the butterfly study are different channels. A channel explains a case; no single channel defines all FDS.^[2]^[3]
Differential continuation. Frequency matters evolutionarily because types survive or reproduce unequally, shifting their contributions to subsequent population composition. Mere prevalence correlation without a selection difference is insufficient.^[1]^[4]
Outcome qualifications. The local sign and the long-run dynamics are separate questions. Invasion conditions, other fitness components, spatial structure and model details affect coexistence or loss.^[1]

What It Is Not¶

FDS is not any change in selection pressure over time. A season or environmental event can change a type's fitness independently of its current frequency; the diagnostic here is that changing the composition itself changes relative success, under a suitable comparison. Nor is it simply “the rare type wins”: positive FDS has the reverse local tendency, and the sign can differ by range, trait, place or mechanism.^[1]^[3]

It is not protected polymorphism by definition. That neighboring concept requires every relevant alternative to be able to invade when rare under specified dynamics. A rare advantage in a limited comparison does not establish reciprocal invasion or a stable interior state. Likewise positive dependence need not force universal fixation: Trotter and Spencer report model exceptions to the simplistic sign-to-outcome inference.^[1]^[5]

It is not prey switching. A predator reallocating encounters toward common prey can create a frequency-dependent selective effect, but mate choice, signal learning and competitive interactions show that the FDS identity is broader than one predator behavior.^[6]^[2]^[3]

Scope of Application¶

The abstraction belongs to evolutionary biology and population genetics wherever biological alternatives have frequency-conditioned relative fitness. It applies to genotypes, heritable phenotypes and, when modeled explicitly, interacting strategies. The reference scale matters: a warning color may be common within one local assemblage and rare elsewhere. A measured fitness association has to be interpreted in relation to the population and component of fitness actually observed.^[3]^[1]

The original examples here are descriptive. The guppy study concerns rare male color patterns and reproductive success in studied populations; the butterfly study concerns local warning-signal prevalence and protection from predation. Neither example licenses a universal claim about all guppies, all mimicry, or an intervention procedure. The draft does not infer cycles or chaos from frequency dependence alone.^[2]^[3]^[1]

Clarity¶

FDS separates a type's properties from the selective value those properties have at a given prevalence. For a frequency-independent comparison one can hold each type's relative fitness fixed while varying composition. For FDS, that move can reverse a conclusion: the same type can be favored at one frequency and disadvantaged at another. This is why one cross-sectional ranking of types cannot stand for a complete evolutionary prediction.^[1]

It also clarifies two often-confused claims: identifying a conditional fitness relation and deriving a population outcome. Hughes et al.'s rare-male advantage supports negative dependence in the studied reproductive setting; it is not by itself a proof of permanent polymorphism. Chouteau et al.'s local common-signal advantage supports positive dependence over their sampled frequency spectrum, with nonlinear saturation; it is not proof that all populations will become monomorphic.^[2]^[3]

Manages Complexity¶

The framework converts many organism-specific stories into a smaller set of roles: types, their current frequencies, relative fitness and the channel connecting prevalence to success. This allows a mating example and a warning-signal example to be compared without pretending their mechanisms are identical. The sign labels provide a compact first contrast, while the fitness-as-function-of-composition relation preserves the feature that fixed-fitness models omit.^[1]^[2]^[3]

The compression must not erase dependencies. A negative sign in one measured range may coexist with directional selection or other forces; a positive sign can saturate. Trotter and Spencer's pairwise-interaction study shows why the endpoint cannot be read off from a one-word sign: different fitness structures can maintain or fail to maintain variation.^[1]^[3]

Abstract Reasoning¶

If type \(i\) has relative fitness \(w_i(p)\), then a change in composition \(p\) can change the ranking \(w_i(p)-w_j(p)\) even when type identities are unchanged. That means an increase in a successful type can alter the very selection pressure that favored it. Under some negative-frequency cases this opposes further increase; under some positive-frequency cases it reinforces it. The direction of this local response is a structural clue, not a theorem about global stability.^[1]

To infer stable maintenance of multiple types, ask the stronger question whether each can increase from rarity against the others and whether the relevant dynamical assumptions hold. To infer fixation, ask whether a favored type remains favored across the route to high frequency, in the face of other fitness components and spatial effects. The original pairwise-interaction analysis finds outcomes more varied than a simple negative-equals-polymorphism/positive-equals-fixation rule.^[1]^[5]

Knowledge Transfer¶

The roles transfer literally across the two studies: alternatives have local frequencies; those frequencies affect relative fitness through an interaction; differential survival or reproduction changes continuation. In guppies the measured channel is reproductive/mating success for rare male patterns. In the defended-butterfly study it is protection associated with locally common warning signals and predator response. One is negative and one positive, but both are FDS.^[2]^[3]

The broader prime Selection captures a population, a pressure and unequal continuation. FDS adds the evolutionary condition that the pressure's relative fitness values depend on type composition. Feedback is a useful higher-level analogy when frequency changes return to influence later fitness, but its live signature is not asserted as an extra strict parent; a local FDS finding does not by itself establish a complete closed-loop dynamic model.^[4]^[7]

Examples¶

Rare male color patterns in guppies¶

Hughes et al.'s original 2013 study reports higher reproductive fitness for males whose color patterns were rare in the studied wild guppy populations. Rare males acquired more mates and sired more offspring than common-pattern males in its abstracted findings. This is a bounded negative-frequency example; the original abstract does not, alone, establish a universal long-run equilibrium or isolate every contributing fitness component.^[2]

Mapped back: The type population is male color-pattern alternatives; current frequencies distinguish rare from common patterns in the studied populations; frequency-conditioned relative fitness is their reported reproductive-success difference; the interaction channel is mate-mediated success; differential continuation is greater offspring contribution by rare males; the outcome qualification is that stable maintenance is a further dynamical claim, not contained in the observed sign.

Locally common warning signals among defended butterflies¶

Chouteau, Arias and Joron compare warning-signal prevalence across local butterfly assemblages with attacks on artificial butterfly models. Their original paper reports fewer attacks as local signal frequency increases, leveling for the most common signals, consistent with predator avoidance knowledge. This is direct evidence about model attack rates and an inference about protection of natural butterflies, not a direct measurement of their survival or a universal law of every predator or mimicry system.^[3]

Mapped back: The type population is alternative warning patterns; current frequencies are local signal proportions; frequency-conditioned relative fitness is the inferred protection associated with common signals, proxied by artificial-model attack rates; the interaction channel is predator response/learning as interpreted by the authors; differential continuation is a predicted survival advantage if that attack-rate difference carries over to natural butterflies; the outcome qualification is saturation, locality and the proxy-to-survival inference, so no unconditional fixation prediction follows.

Structural Tensions¶

T1 — Easy sign summary versus warranted endpoint inference. Calling a case negative or positive succinctly reports a local frequency–fitness slope, but it invites the costly overreach of treating that slope as a global prediction. Requiring full invasion and dynamical conditions reduces this overreach but demands more evidence and model specificity. Diagnostic: Is the claim only about measured relative success at tested frequencies, or about a stable polymorphism/fixation over all relevant frequencies?^[1]^[2]

T2 — Cross-system relation versus mechanism specificity. A common \(w_i(p)\) representation reveals what guppy mating and butterfly warning signals share, but it hides whether mates, predators or competitors produce the dependence. A mechanism-rich account explains one system more closely but carries less directly to another. Diagnostic: Does the current question ask whether fitness depends on frequency across settings, or which interaction actually causes the dependence in this particular setting?^[2]^[3]

Structural–Framed Character¶

Frequency-dependent selection is structural in its population relation but framed by evolutionary biology: it requires biological types, relative fitness and differential continuation. Its evaluative weight is low; rare advantage or common advantage is not morally good or necessarily adaptive in a broader sense. Its human-practice dependence is low for the process but meaningful for how investigators define type, locality and fitness component. Its institutional origin is evolutionary and population-genetic theory, while the process itself needs no institution. Its vocabulary travels across mating, predation and genotype-interaction research with literal fitness and frequency meanings; beyond evolution, “frequency-dependent value” can be an analogy or higher-order abstraction, not automatically the same selection identity. For import versus recognition, locating the same fitness–frequency relation in a new biological population is recognition of the process, whereas importing its name to fashion or markets without biological differential reproduction changes the object. Its character: a domain-specific evolutionary selection regime that instantiates the broad Selection prime and contains a portable prevalence-dependent skeleton worth keeping distinct from the full biological identity.^[1]^[4]

Structural Core vs. Domain Accent¶

The structural core is a selection pressure that changes as the selected alternatives become more or less prevalent. Live Selection is the admitted portable genus: candidate population, pressure and unequal continuation. FDS adds a frequency-conditioned biological relative-fitness function. A broader prevalence-dependent-advantage pattern may recur elsewhere, but this entry does not pretend that market popularity is literally genetic or ecological selection without the biological roles.^[4]^[1]

The accents are the interaction channels and reference scales. Mate-mediated success of guppy color patterns and predator-mediated protection of butterfly warning signals do not have the same cause; both instantiate the core only because local frequency modifies relative survival/reproduction. Protected Polymorphism is a distinct, stronger boundary condition, while prey switching is one possible behavioral channel, not the general FDS identity.^[2]^[3]^[5]^[6]

This entry is a kind of Selection.

DAG parent — Selection (Selection). Frequency-dependent biological fitness supplies a selection pressure over alternatives; differential survival or reproduction changes their continuation and composition. Selection does not require that its pressure depend on current frequencies, making it the broader genus. Independent DAG review remains pending.^[4]

Related, not asserted direct parents — Natural Selection and Feedback. Repeated heritable FDS can be analyzed within natural selection, and changing frequencies may feed back into future fitness, but the current record does not infer cumulative adaptation or a particular dynamical loop from every bounded FDS observation.^[8]^[7]

Relationships to Other Abstractions¶

Current abstraction Frequency-Dependent Selection Domain-specific

Parents (1) — more general patterns this builds on

Frequency-Dependent Selection is a kind of Selection Prime

A frequency-conditioned biological fitness difference is a special selection basis.

Hierarchy path (1) — routes to 1 parentless root

Frequency-Dependent Selection → Selection

Neighborhood in Abstraction Space¶

Frequency-Dependent Selection sits in a sparse region of the domain-specific corpus (71^st percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Selection, Speciation & Experimental Evolution (22 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With¶

Protected Polymorphism: reciprocal rare-type invasion condition, not a synonym for one observed rare advantage.^[5]
Prey Switching/Apostatic selection: a predator behavior or its negative-selection consequence, narrower than frequency dependence across mating, competition and signaling.^[6]
Frequency-independent directional selection: fixed relative fitness can change frequencies, but its advantage is not itself altered by prevalence.^[1]
Automatic outcomes: negative dependence need not always maintain all types, positive dependence need not always fix one, and cycles or chaos are not constitutive claims here.^[1]
Simple rarity or commonness correlation: selection requires a differential fitness/contribution relation, not merely a difference in counts.^[4]

References¶

[1] Meredith V. Trotter and Hamish G. Spencer, “Frequency-Dependent Selection and the Maintenance of Genetic Variation”, Genetics 176 (2007), 1729–1740, DOI 10.1534/genetics.107.073072; original full paper, Abstract, Introduction and pairwise-interaction analysis. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s

[2] Kimberly A. Hughes, Anne E. Houde, Anna C. Price and F. Helen Rodd, “Mating advantage for rare males in wild guppy populations”, Nature 503 (2013), 108–110, DOI 10.1038/nature12717; original research abstract and metadata inspected 2026-10-01 (full article not inspected). registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k

[3] Mathieu Chouteau, Mónica Arias and Mathieu Joron, “Warning signals are under positive frequency-dependent selection in nature”, PNAS 113 (2016), 2164–2169, DOI 10.1073/pnas.1519216113; original author-hosted full paper, especially pp.2164–2167. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n

[4] Encyclopedia of Abstractions, live prime_abstractions/v2/selection.md, Core Idea, Structural Signature and What It Is Not, inspected 2026-10-01. registry ↩a ↩b ↩c ↩d ↩e ↩f

[5] Encyclopedia of Abstractions, live domain_specific_abstractions/v2/protected_polymorphism.md, Core Idea and Structural Signature, inspected 2026-10-01. registry ↩a ↩b ↩c ↩d

[6] Encyclopedia of Abstractions, live domain_specific_abstractions/v2/prey_switching.md, Core Idea and What It Is Not, inspected 2026-10-01. registry ↩a ↩b ↩c

[7] Encyclopedia of Abstractions, live prime_abstractions/v2/feedback.md, Core Idea and Structural Signature, inspected 2026-10-01. registry ↩a ↩b

[8] Encyclopedia of Abstractions, live prime_abstractions/v2/natural_selection.md, Core Idea and Structural Signature, inspected 2026-10-01. registry ↩