Underdominance¶
A population-genetic fitness ordering in which a heterozygote is less fit than either corresponding homozygote in a declared context.
Core Idea¶
Underdominance is the population-genetic arrangement in which a heterozygote fares worse than both corresponding homozygotes on a declared fitness measure in a declared context. For alleles or chromosomal arrangements \(A\) and \(B\), it requires \(w_{AB}<w_{AA}\) and \(w_{AB}<w_{BB}\). The comparison is within the same locus or arrangement, environment and relevant life-cycle measure. A different outcome for one hybrid, or a low value on one assay, is not automatically the complete three-genotype ordering.[1]
The ordering is the identity; the population trajectory is a derived question. In a simple isolated, randomly mating population with fixed genotype fitnesses, heterozygote disadvantage gives an unstable interior allele-frequency equilibrium and different boundary outcomes on its two sides. Láruson and Reed show why that familiar statement must not replace the definition: when demes exchange migrants, network structure can permit differentiated selection–migration equilibria under specified conditions. The same local fitness ordering survives although the simple one-population forecast does not.[1]
Natural chromosome differences can supply the fitness contrast. Stathos and Fishman attributed underdominant male pollen sterility in Mimulus hybrid comparisons to heterozygous reciprocal translocations. That is strong evidence for a particular fertility-component effect, not a measurement of every component of lifetime fitness or a direct observation that either arrangement has fixed in a wild population.[2]
Structural Signature¶
Sig role-phrases: corresponding genotype alternatives — common fitness frame — heterozygote disadvantage — conditional population context — evidence resolution.
- Corresponding genotype alternatives. \(AA\), \(AB\) and \(BB\) designate the two homozygotes and their heterozygote at a common locus or arrangement. This shared genetic frame makes the middle case genuinely comparable with both endpoints; an unrelated hybrid phenotype cannot fill the role.[1]
- Common fitness frame. State the environment, life-cycle interval and whether the values are overall fitness or a component such as male fertility. The inequality is meaningful only when its three terms use a compatible measure.[1][2]
- Heterozygote disadvantage. The constitutive double inequality places \(AB\) strictly below each homozygote. Equal homozygote fitness, sometimes assumed for analysis, is not required; an \(AB\) value below only one endpoint is not this strict pattern.[1]
- Conditional population context. Mating structure, migration and topology govern what allele frequencies do after the ordering is specified. They are essential to a dynamical prediction but not to recognizing the ordering itself.[1]
- Evidence resolution. An assay can establish a component-level underdominant effect without establishing total fitness or an evolutionary outcome. The strength of the name attached to an empirical case must match what was actually compared.[2]
The first three roles constitute the abstraction. The latter two prevent a valid fitness pattern from silently becoming an unsupported assertion about fixation or speciation.
What It Is Not¶
It is not heterozygosity itself: a mixed genotype can have higher, equal or intermediate fitness. Nor is it enough to show that the heterozygote is worse than one homozygote; both inequalities matter.[1]
It is not a universal unstable threshold or guaranteed fixation. Those are consequences in a restricted single-population model, while coupled demes can have stable differentiated outcomes. Drift, migration and context changes cannot be erased from a forecast merely because the local ordering is known.[1]
It is not every chromosomal rearrangement or every hybrid fertility loss. Stathos and Fishman distinguished the underdominant male-sterility effects of particular reciprocal translocations from the examined inversions, which did not show the same direct effect. A measured pollen effect also cannot simply be renamed whole-organism fitness.[2]
Scope of Application¶
In population-genetic models, the abstraction specifies the relative fitness configuration of two alleles and a heterozygote before one asks for trajectories. A single isolated population can display the familiar unstable internal equilibrium, but that result requires the model's mating, selection and constancy assumptions.[1]
In subdivided-population analysis, local underdominance is one input to a larger selection–migration system. Láruson and Reed compare network arrangements of connected demes and find that topology can materially change stability. The abstraction continues to name the local genotype ordering, while the network model—not the word underdominance alone—determines the larger pattern.[1]
In natural cytogenetics, rearrangement heterozygotes can be examined for reduced fertility components. The Mimulus study provides a concrete case of translocation-associated underdominant male sterility, but its scope is the compared hybrid and pollen-fertility evidence, not an all-purpose claim about every rearrangement or the history of fixation.[2]
Clarity¶
The label answers a precise question: which of the three genotype classes occupies the fitness trough? Write both inequalities and name the fitness measure before invoking population dynamics. This separates a structural property of genotype fitness from a model's account of what selection later does with that property.[1]
It also forces the empirical claim to match its observation. In the Mimulus case, pollen fertility is an informative male-fitness component. The statement “underdominant male sterility was attributed to reciprocal translocations” is better supported than “the rearrangement inevitably fixed or caused complete reproductive isolation.” The latter adds unmeasured fitness components and population history.[2]
Manages Complexity¶
The double inequality compresses many genotype measurements into one diagnostic configuration: the mixed genotype is disadvantaged relative to either pure genotype. It does not require the two homozygotes to be equal, so it preserves a useful general pattern while leaving room for asymmetric endpoints.[1]
The compression has a boundary. A fitted three-fitness vector does not also encode migration, demographic stochasticity, network topology or changes in environment. By storing “ordering” separately from “dynamics,” an analyst can reuse the same local configuration across different population models without smuggling in the first model's outcome.[1]
Abstract Reasoning¶
First compare \(w_{AB}\) independently with \(w_{AA}\) and \(w_{BB}\) on one declared scale. If either strict comparison fails, do not call the configuration underdominance on that scale. If the evidence covers only fertility or another component, state the component-level inference and leave overall fitness open.[1][2]
Only after that recognition step select a dynamical model. A simple isolated-population derivation can explain why the initial allele frequency matters; a connected-deme analysis asks whether migration and topology alter the stability picture. Láruson and Reed's comparison demonstrates that recognizing the same local relation does not license transporting every predicted endpoint between models.[1]
Knowledge Transfer¶
Within population genetics, the three-genotype comparison transfers from allelic model fitness to chromosomal-arrangement fertility studies as a bounded diagnostic: identify the two homozygote arrangements, their hybrid heterozygote and a comparable fitness component. Conversely, an empirical fertility result can motivate a model, but the model must declare how the component relates to total fitness and what population structure it assumes.[2][1]
Outside genetics, a “mixed choice worse than either pure choice” may be an analogy to other coordination or compatibility problems. The word Underdominance should not be transferred literally without alleles, heterozygotes and comparable fitness. The abstract endpoint-versus-middle ordering is a possible future-prime question; it is not evidence that this domain-specific identity already operates unchanged in other fields.
Examples¶
Canonical: natural chromosomal-rearrangement hybrid fertility¶
In Mimulus lewisii–M. cardinalis hybrid work, Stathos and Fishman used comparisons of chromosomal arrangements and found that heterozygous reciprocal translocations accounted for underdominant male pollen sterility. Their abstract reports substantially lower pollen fertility in diploid F1 hybrids than in a chromosome-doubled comparison and attributes the effect to specific translocations, not the examined inversions. The result concerns a measured male-fertility component; the paper separately discusses evolutionary models of how such rearrangements might spread.[2]
Mapped back: The genotype alternatives are corresponding arrangement states and their hybrid heterokaryotypic state; the common fitness frame is pollen-fertility performance in the studied comparison; heterozygote disadvantage is the lower hybrid fertility attributed to the translocation heterozygotes. The population context is not an observed frequency trajectory here, and evidence resolution confines the conclusion to male fertility rather than total fitness or fixation.
Applied analysis: connected populations¶
Láruson and Reed specify \(AA\), \(AB\) and \(BB\) fitnesses for an underdominant allele pair and examine multiple demes connected in different network topologies. In their simplified model the two homozygotes are assigned equal fitness, but that equality is a convenience, not a defining condition. Under some migration structures, different demes can maintain differentiated stable configurations even though the heterozygote is disadvantaged locally.[1]
Mapped back: The genotype alternatives are \(AA\), \(AB\) and \(BB\); the common fitness frame is relative fitness within the model; heterozygote disadvantage is the \(AB\) deficit below both homozygotes; population context is the stipulated deme network and migration; evidence resolution identifies a theoretical stability result, not an observed natural-population fate.
Structural Tensions¶
T1: Local simplicity versus spatial fidelity. A fixed-fitness isolated-population model makes the frequency-dependent threshold consequence intelligible, but loses the migration/topology effects that can sustain differentiation. A network model retains those effects while requiring stronger structural and parameter assumptions; one cannot claim both the minimal model's universal endpoint and the richer model's spatial realism at no explanatory cost. Diagnostic: Does the inference concern one effectively isolated mating population, or connected demes whose topology may alter stability?
T2: Precise component attribution versus comprehensive fitness inference. A pollen-fertility assay can isolate an underdominant mechanism and distinguish translocations from inversions, but a whole-fitness claim must include other survival and reproductive components. Broadening immediately sacrifices evidential precision; waiting for all components sacrifices a useful mechanistic conclusion already supported by the assay. Diagnostic: Is the decision about a measured male-fertility component, or does it require a supported total-fitness comparison?
Structural–Framed Character¶
Underdominance sits toward the structural side within its genetic frame: the double inequality is mathematically sharp, while the relata are irreducibly genotypes and contextual fitnesses. Its evaluative weight is limited: “disadvantage” reports lower reproductive performance on the stated scale, not a moral verdict or claim that one genotype ought to prevail. Its human-practice dependence lies in selecting assays and models, but organisms need not be classified by a practitioner for differential fertility to occur. Its institutional origin as a research term and model convention does not constitute the phenomenon; no legal or administrative rule creates it. Its vocabulary travel is weak: the inequalities can be imitated outside genetics, yet “heterozygote,” “homozygote” and fitness remain biological commitments. Finally, apparent import versus recognition in another field must be tested as either a carefully marked analogy or recognition of a more general, still-unadmitted ordering skeleton—not automatically literal underdominance. Its character: a crisp structural relation whose typed population-genetic carrier and evidence conditions keep the named abstraction domain-specific.[1][2]
Structural Core vs. Domain Accent¶
The skeletal relation is a mixed alternative below both matched pure alternatives on one evaluation scale. That relation invites a future-prime question about mixed-state disadvantage, but no such prime is asserted here. The actual identity requires homologous \(AA\), \(AB\) and \(BB\) genotype alternatives and an environment-specific measure of fitness; population-genetic mating and migration assumptions then govern the derived trajectories. Thus the portable inequality alone is thinner than underdominance, while the genetic terms and inferential tests do not transfer literally to unrelated domains. Calling a non-genetic compromise “underdominance” is analogy unless a separate structural-prime identity is independently established.[1]
Instantiates / Related Primes¶
No strict typed parent relation is asserted in the current DAG. Neither live Natural Selection's iterative filter nor Selection Coefficient's relative parameter is a necessary genus of the static three-genotype fitness ordering; broad Comparison is related but not asserted as a strict parent.
Neighborhood in Abstraction Space¶
Underdominance sits in a sparse region of the domain-specific corpus (63rd 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
- Genetic Load — 0.86
- Wallace Effect — 0.86
- Co-adaptation — 0.85
- Kin selection — 0.84
- Trivers–Willard Hypothesis — 0.84
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
Overdominance reverses the relevant comparison: the heterozygote exceeds both homozygotes on the declared fitness scale. Dominance in phenotype or gene expression does not itself say which genotype has the lowest fitness. A selection coefficient is a relative-fitness parameter whose reference and sign must be declared, whereas underdominance asks whether the heterozygote lies below both alternatives.[1][3]
Hybrid sterility can supply evidence for an underdominant fertility component, but it can have other genetic causes and does not alone report whole fitness. Fixation is a possible model or historical outcome, not the constitutive fitness pattern. Stathos and Fishman's distinction among rearrangement classes and Láruson and Reed's network counterexamples make both boundaries consequential.[2][1]
References¶
[1] Aki J. Láruson and Floyd A. Reed, “Stability of Underdominant Genetic Polymorphisms in Population Networks”, original research preprint, v2 (2015), abstract, §1 and §2.1; full PDF inspected 2026-10-01. The network conclusions are conditional on the paper's model. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t ↩u
[2] Angela Stathos and Lila Fishman, “Chromosomal rearrangements directly cause underdominant F1 pollen sterility in Mimulus lewisii–Mimulus cardinalis hybrids”, Evolution 68 (2014), 3109–3119, abstract and introduction; full PDF inspected 2026-10-01. The empirical readout is pollen fertility. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k
[3] Encyclopedia of Abstractions, live domain-specific Selection Coefficient, Core Idea and Structural Signature, inspected 2026-10-01. registry ↩
[4] Encyclopedia of Abstractions, live prime Natural Selection, Core Idea and Structural Signature, inspected 2026-10-01. registry
[5] Encyclopedia of Abstractions, live prime Comparison, Core Idea and Structural Signature, inspected 2026-10-01. registry