Reinforcement

Definition and historical origins

Reinforcement is the evolutionary process by which natural selection strengthens prezygotic reproductive barriers between diverging populations when they come into secondary contact and hybridisation produces offspring of reduced fitness. The logic is straightforward: if individuals that mate with members of the other population produce fewer viable or fertile offspring than individuals that mate assortatively with their own population, selection will favour any trait — in mate preference, mating signal, or timing of reproduction — that reduces the probability of hybridisation. Over time, this selection increases the degree of prezygotic isolation, potentially completing speciation that began in geographic isolation.^{1, 4}

The concept has deep roots in evolutionary thought. Alfred Russel Wallace argued in the mid-nineteenth century that natural selection could act to prevent hybridisation between divergent forms, and the process is sometimes called the "Wallace effect" in his honour. Theodosius Dobzhansky gave the idea its modern formulation in his 1937 book Genetics and the Origin of Species and in a subsequent 1940 paper in which he proposed that natural selection would favour increased sexual isolation between populations that produce unfit hybrids.^{1, 14} The term "reinforcement" was introduced later to describe this selective strengthening of prezygotic barriers, and R. K. Butlin provided an influential early review of the concept and its controversy in 1987.⁶

Theoretical framework

The theoretical basis of reinforcement rests on the fitness cost of hybridisation. When two partially diverged populations meet in secondary contact, individuals that hybridise produce offspring that suffer reduced fitness due to genetic incompatibilities, ecological maladaptation, or both. These fitness costs create selection against hybridisation itself. Any heritable trait that causes an individual to prefer conspecific mates — or to avoid heterospecific mates — will be favoured because it reduces the production of unfit hybrid offspring. As alleles for stronger mate discrimination spread through the population, prezygotic isolation increases.^{4, 5}

The process requires several conditions to operate. First, there must be a cost to hybridisation: if hybrids are equally fit, there is no selective disadvantage to interbreeding and no pressure for reinforcement. Second, there must be genetic variation for mate preference or other prezygotic traits that can be selected upon. Third, gene flow between the populations must not be so high that it swamps the buildup of divergence. This third condition was the source of much historical scepticism about reinforcement. Critics argued that the very gene flow that brings populations into contact would also homogenise any incipient differences in mate preference, preventing reinforcement from proceeding to completion.^{5, 6}

Servedio and Noor provided a comprehensive review of theoretical models and showed that reinforcement is possible under a wider range of conditions than early critics assumed. Reinforcement is facilitated when the loci controlling mate preference are physically linked to (or identical to) the loci causing hybrid unfitness, because linkage prevents recombination from breaking apart the association between mate preference and hybrid incompatibility alleles. Reinforcement is also more likely when hybrid fitness costs are strong, when populations differ in easily discriminable mating signals, and when the zone of contact is narrow relative to the species ranges.⁴

Evidence from Drosophila

The genus Drosophila has provided much of the empirical evidence for reinforcement. Coyne and Orr conducted a landmark comparative analysis of Drosophila species pairs, examining the relationship between genetic distance (a proxy for divergence time), geographic overlap, and the strength of prezygotic and postzygotic isolation. They found that prezygotic isolation was significantly stronger in sympatric species pairs than in allopatric pairs at the same level of genetic divergence, while postzygotic isolation showed no such difference. This asymmetric pattern — stronger mate discrimination in sympatry but not stronger hybrid inviability — is precisely what reinforcement predicts, because reinforcement acts specifically on prezygotic barriers.⁷

Noor provided more targeted evidence from closely related Drosophila species. In the D. pseudoobscura/D. persimilis species pair, females from sympatric populations showed significantly stronger discrimination against heterospecific males than females from allopatric populations. This pattern of enhanced mate discrimination in sympatry is a hallmark signature of reinforcement and has been documented in multiple Drosophila species pairs.³ Matute extended the evidence to postzygotic mechanisms, demonstrating that conspecific sperm precedence — the preferential utilisation of conspecific sperm over heterospecific sperm by multiply mated females — is stronger in sympatric populations of D. yakuba and D. santomea, providing evidence for reinforcement acting on gametic isolation in addition to behavioural mate choice.⁸

Nosil, Crespi, and Sandoval demonstrated that reinforcement can overcome ongoing gene flow. Working with Timema walking-stick insects rather than Drosophila, they showed that populations in contact zones with a related species displayed stronger sexual isolation than populations distant from the contact zone, even though gene flow between the host-plant-adapted populations was not negligible. The study provided evidence that the strength of selection against hybridisation can exceed the homogenising effect of gene flow under natural conditions.¹²

Reproductive character displacement

Reinforcement is closely related to the phenomenon of reproductive character displacement, in which mating signals, preferences, or other traits involved in mate recognition diverge more in areas where two species co-occur (sympatry) than in areas where they are geographically separated (allopatry). Reproductive character displacement is the observable outcome of reinforcement: if selection in zones of contact favours individuals with stronger conspecific mate preferences, the result is exaggerated divergence in mating-related traits in sympatry relative to allopatry.^{5, 11}

Pfennig and Pfennig provided a comprehensive review of character displacement as a driver of diversification and distinguished between ecological character displacement (driven by resource competition) and reproductive character displacement (driven by selection against hybridisation). In the case of reinforcement, reproductive character displacement is specifically predicted to affect traits involved in species recognition and mate choice, such as mating calls, colour patterns, chemical signals, or the timing of reproductive activity.¹¹

A classic example comes from frog mating calls. Fouquette documented that the advertisement calls of sympatric populations of closely related frog species diverge more in call frequency and temporal pattern than those of allopatric populations, consistent with reinforcement driving reproductive character displacement in acoustic mating signals.⁹ In threespine stickleback fish, Rundle and Schluter showed that females from sympatric lakes, where benthic and limnetic species coexist, exhibited stronger preferences for conspecific males than females from allopatric populations, providing direct evidence that reinforcement had strengthened mate discrimination in the zone of overlap.¹⁰

Evidence from other taxa

While Drosophila research established the empirical foundation for reinforcement, evidence has accumulated from a wide range of organisms, demonstrating that the process is taxonomically widespread. Yukilevich conducted a broad comparative analysis across animal taxa and found that the signature of reinforcement — stronger prezygotic isolation in sympatry than in allopatry at a given level of genetic divergence — was detected in multiple groups including insects, amphibians, fishes, and birds. The analysis further showed that the strength of the reinforcement signal varied among taxa, being strongest in groups where hybrid fitness costs were most severe and where mating signals were most easily discriminated.¹⁵

In mosquitoes, Bargielowski and colleagues documented reinforcement between Aedes aegypti and Aedes albopictus in the southeastern United States. Where the two species overlapped, A. aegypti females showed enhanced discrimination against A. albopictus males compared with allopatric populations. The study was notable because it demonstrated reinforcement in a medically important vector species and showed that the process can operate over relatively short timescales following the recent range expansion of A. albopictus into the New World.¹⁶

In flowering plants, reinforcement has been inferred from patterns of pollen-pistil incompatibility and floral trait divergence in sympatry. Because plants cannot move to avoid heterospecific mating, selection for reinforcement may act on floral morphology, phenology, and pollinator attraction to reduce the receipt of heterospecific pollen. Nosil reviewed the evidence for reinforcement across multiple systems and concluded that while the process is not universal — many zones of secondary contact show stable hybrid zones without evidence of reinforcement — it is sufficiently common to be regarded as an important contributor to the completion of speciation.¹³

Controversies and limitations

Despite substantial empirical support, reinforcement remains a concept with significant theoretical challenges and ongoing debate. The central objection, raised most forcefully by critics in the 1980s and 1990s, concerns the tension between gene flow and the buildup of divergence. If populations are exchanging genes at the rate necessary for hybridisation to occur, those same migrant alleles may disrupt the genetic basis of mate discrimination, preventing reinforcement from going to completion. Butlin argued that this "gene flow problem" made reinforcement an unlikely outcome in most contact zones and that the patterns attributed to reinforcement might often have alternative explanations, such as ecological character displacement or the fusion of populations that fail to complete speciation.⁶

Subsequent theoretical work has shown that the gene flow problem, while real, is not insurmountable. Models that incorporate strong selection against hybrids, tight linkage between preference loci and incompatibility loci, and one-allele mechanisms (in which a single allele for increased choosiness can spread in both populations simultaneously) have demonstrated that reinforcement can proceed to completion even with moderate levels of gene flow.⁴ The empirical evidence from Drosophila, sticklebacks, and other systems has also been broadly consistent with reinforcement, although alternative explanations are difficult to rule out definitively in any individual case.

A further complication is distinguishing reinforcement from other processes that produce similar patterns. Reproductive character displacement can result from selection against costly but non-hybridisation-related mating interactions, such as wasted reproductive effort or interference competition for mates. To attribute a pattern of enhanced prezygotic isolation in sympatry to reinforcement specifically, it is necessary to demonstrate that hybridisation incurs a fitness cost and that the divergence in mating traits is a response to that cost rather than to other forms of selection.^{5, 11}

Role in completing speciation

Reinforcement occupies a distinctive role in the speciation process. Unlike other mechanisms that generate reproductive isolation — such as the accumulation of Dobzhansky-Muller incompatibilities through genetic drift in allopatry — reinforcement is a process in which natural selection directly acts to increase reproductive isolation. It is the one mechanism of speciation in which the completion of species boundaries is itself the target of selection, rather than a byproduct of divergence in other traits.^{4, 5}

Coyne and Orr argued that reinforcement is most likely to contribute to speciation in cases where populations have diverged substantially in allopatry but have not yet evolved complete prezygotic isolation before coming into secondary contact. In such cases, the fitness cost of hybridisation provides the selective pressure needed to close the remaining gap in reproductive isolation. Reinforcement thus acts as a "finishing touch" on speciation that was initiated by other processes, rather than as a primary cause of divergence.^{5, 7}

The relationship between reinforcement and sexual selection is also significant. In many cases, the traits that are displaced in sympatry are those involved in sexual signalling and mate choice — male coloration, courtship song, pheromone composition, or display behaviour. Reinforcement can thus accelerate the divergence of sexually selected traits, potentially contributing to the rapid evolution of species-specific mating signals that characterises many adaptively radiating groups. In this way, reinforcement links the ecology of hybrid zones to the evolution of biodiversity, making it a process of both theoretical and practical importance for understanding how the diversity of life is generated and maintained.^{4, 13, 15}