A narrow hybrid zone between the grasshoppers Stenobothrus clavatus and Stenobothrus rubicundus (Orthoptera: Gomphocerinae): female preferences for courtship songs

Two grasshopper species, Stenobothrus rubicundus and Stenobothrus clavatus, were previously shown to hybridize in a narrow contact zone on Mount Tomaros in northern Greece. The species are characterized by complex and completely different courtship songs. In the present study, we investigated female preferences for the courtship songs of S. rubicundus, S. clavatus and hybrids in playback experiments. Playback of the courtship songs revealed assortative preferences in females of the parental species: they significantly more often preferred the songs of conspecific males. Hybrid females showed a lower selectivity than parental females, responding somewhat equally eager to playback of the songs of S. clavatus, S. rubicundus, and natural hybrid song, although less actively to the F₁ hybrid song. The results suggest that hybrid males may lose to males of parental species, whereas hybrid females would even have an advantage over parental females. Comparison of responses of females from allopatric populations and Mount Tomaros to different song types shows no evidence for reinforcement. Asymmetry found in female preferences may have implications for the structure of the hybrid zone. © 2013 The Linnean Society of London     

A narrow hybrid zone between the grasshoppers Stenobothrus clavatus and Stenobothrus rubicundus: courtship song analysis

The closely related grasshopper species Stenobothrus rubicundus and Stenobothrus clavatus are known to hybridize in a very narrow contact zone on Mt. Tomaros in northern Greece. These species produce very different and complex courtship songs accompanied with visual display. We analyzed the courtship songs and underlying stridulatory movements of the hind legs in natural hybrids from Mount Tomaros. The two species were also hybridized in the laboratory and their songs were compared with the songs of the natural hybrids. Some hybrid songs were shown to have intermediate features between parental songs, whereas other hybrid songs comprised completely new elements. The clavatus‐like song elements were found to dominate in hybrid songs. These song features may influence the mating success of hybrid males in the contact zone. A comparison of hybrid songs with the song pattern of the north European S. rubicundus populations allowed us to suggest a scenario of S. rubicundus and S. clavatus origin. © 2012 The Linnean Society of London, Biological Journal of the Linnean Society, 2012, ?? , ??–??.     

Variation in complex courtship traits across a hybrid zone between grasshopper species of the Chorthippus albomarginatus group

Closely related grasshopper species of the Chorthippus albomarginatus group are notable for their extremely complex courtship songs, accompanied by a visual display. Two species of this group, Ch. albomarginatus and Ch. oschei, were previously shown to hybridize in a wide mosaic hybrid zone in Ukraine and Moldova. In this paper, variation in five courtship song characters, one character of visual display and the number of stridulatory pegs were analysed across the hybrid zone to estimate selection against hybrids and strength of assortative mating. Comparison of cline width and position across the hybrid zone showed concordant and coincident clines in four traits, such as three song characters and one morphological character, and discordant and non‐coincident clines in two other song characters and the character of visual display. Concordance of clines in different characters suggests an equal strength of selection acting on underlying loci. Increase of variance and covariance between phenotypic traits at the cline centre could more likely result from assortative mating than from selection against hybrids. Most pairwise cases showed the highest covariance for the oschei‐like, than for the albomarginatus‐like hybrid populations. This indicates that introgression of the oschei genes into the albomarginatus genome is stronger than vice versa, and may be evidence of the movement of the hybrid zone in favour of Ch. albomarginatus. Analysis of associations between phenotype and local vegetation showed that mosaic structure of the hybrid zone is explained to a great extent by habitat–phenotype associations. © 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102 , 275–291.     

Testing an hypothesis of hybrid zone movement for toads in France

Hybrid zone movement may result in substantial unidirectional introgression of selectively neutral material from the local to the advancing species, leaving a genetic footprint. This genetic footprint is represented by a trail of asymmetric tails and displaced cline centres in the wake of the moving hybrid zone. A peak of admixture linkage disequilibrium is predicted to exist ahead of the centre of the moving hybrid zone. We test these predictions of the movement hypothesis in a hybrid zone between common (Bufo bufo) and spined toads (B. spinosus), using 31 nuclear and one mtDNA SNPs along a transect in the northwest of France. Average effective selection in Bufo hybrids is low and clines vary in shape and centre. A weak pattern of asymmetric introgression is inferred from cline discordance of seven nuclear markers. The dominant direction of gene flow is from B. spinosus to B. bufo and is in support of southward movement of the hybrid zone. Conversely, a peak of admixture linkage disequilibrium north of the hybrid zone suggests northward movement. These contrasting results can be explained by reproductive isolation of the B. spinosus and B. bufo gene pools at the southern (B. spinosus) side of the hybrid zone. The joint occurrence of asymmetric introgression and admixture linkage disequilibrium can also be explained by the combination of low dispersal and random genetic drift due to low effective population sizes.     

Evidence for concerted movement of nuclear and mitochondrial clines in a lizard hybrid zone

Moving hybrid zones provide compelling examples of evolution in action, yet long‐term studies that test the assumptions of hybrid zone stability are rare. Using replicated transect samples collected over a 10‐year interval from 2002 to 2012, we find evidence for concerted movement of genetic clines in a plateau fence lizard hybrid zone (Sceloporus tristichus) in Arizona. Cline‐fitting analyses of SNP and mtDNA data both provide evidence that the hybrid zone shifted northward by approximately 2 km during the 10‐year interval. For each sampling period, the mtDNA cline centre is displaced from the SNP cline centre and maintaining an introgression distance of approximately 3 km. The northward expansion of juniper trees into the Little Colorado River Basin in the early 1900s provides a plausible mechanism for hybrid zone formation and movement, and a broadscale quantification of recent land cover change provides support for increased woody species encroachment at the southern end of the hybrid zone. However, population processes can also contribute to hybrid zone movement, and the current stability of the ecotone habitats in the centre of the hybrid zone suggests that movement could decelerate in the future.     

Formation of a recent hybrid zone offers insight into the geographic puzzle and maintenance of species boundaries in musk turtles

Speciation is the result of an accumulation of reproductive barriers between populations, but pinpointing the factors involved is often difficult. However, hybrid zones can form when these barriers are not complete, especially when lineages come into contact in intermediate or modified habitats. We examine a hybrid zone between two closely related riverine turtle species, Sternotherus depressus and S. peltifer, and use dual‐digest RAD sequencing to understand how this hybrid zone formed and elucidate genomic patterns of reproductive isolation. First, the geographical extent and timing of formation of the hybrid zone is established to provide context for understanding the role of extrinsic and intrinsic reproductive isolating mechanisms in this system. The strength of selection on taxon‐specific contributions to maintenance of the hybrid zone is then inferred using a Bayesian genomic cline model. These analyses identify a role for selection inhibiting introgression in some genomic regions at one end of the hybrid zone and promoting introgression in many loci at the other. When selective pressures necessary to generate outliers to the genomic cline are considered with the geographical and temporal context of this hybrid zone, we conclude that habitat‐specific selection probably limits introgression from S. depressus to S. peltifer in the direction of river flow. However, selection is mediating rapid, unidirectional introgression from S. peltifer to S. depressus, which is probably facilitated by anthropogenic habitat alteration. These findings indicate a potentially imminent threat of population‐level genomic extinction for an already imperiled species due to ongoing human‐caused habitat alteration.     

Steep,coincident, and concordant clines in mitochondrial and nuclear‐encoded genes in a hybrid zone between subspecies of Atlantic killifish,Fundulus heteroclitus

Steep genetic clines resulting from recent secondary contact between previously isolated taxa can either gradually erode over time or be stabilized by factors such as ecological selection or selection against hybrids. We used patterns of variation in 30 nuclear and two mitochondrial SNPs to examine the factors that could be involved in stabilizing clines across a hybrid zone between two subspecies of the Atlantic killifish, Fundulus heteroclitus. Increased heterozygote deficit and cytonuclear disequilibrium in populations near the center of the mtDNA cline suggest that some form of reproductive isolation such as assortative mating or selection against hybrids may be acting in this hybrid zone. However, only a small number of loci exhibited these signatures, suggesting locus‐specific, rather than genomewide, factors. Fourteen of the 32 loci surveyed had cline widths inconsistent with neutral expectations, with two SNPs in the mitochondrial genome exhibiting the steepest clines. Seven of the 12 putatively non‐neutral nuclear clines were for SNPs in genes related to oxidative metabolism. Among these putatively non‐neutral nuclear clines, SNPs in two nuclear‐encoded mitochondrial genes (SLC25A3 and HDDC2), as well as SNPs in the myoglobin, 40S ribosomal protein S17, and actin‐binding LIM protein genes, had clines that were coincident and concordant with the mitochondrial clines. When hybrid index was calculated using this subset of loci, the frequency distribution of hybrid indices for a population located at the mtDNA cline center was non‐unimodal, suggesting selection against advanced‐generation hybrids, possibly due to effects on processes involved in oxidative metabolism.     

Ambiguous species boundaries: Hybridization and morphological variation in two closely related Rubus species along altitudinal gradients

Although hybridization frequently occurs among plant species, hybrid zones of divergent lineages formed at species boundaries are less common and may not be apparent in later generations of hybrids with more parental‐like phenotypes, as a consequence of backcrossing. To determine the effects of dispersal and selection on species boundaries, we compared clines in leaf traits and molecular hybrid index along two hybrid zones on Yakushima Island, Japan, in which a temperate (Rubus palmatus) and subtropical (Rubus grayanus) species of wild raspberry are found. Leaf sinus depth in the two hybrid zones had narrower clines at 600 m a.s.l. than the molecular hybrid index and common garden tests confirmed that some leaf traits, including leaf sinus depth that is a major trait used in species identification, are genetically divergent between these closely related species. The sharp transition in leaf phenotypic traits compared to molecular markers indicated divergent selection pressure on the hybrid zone structure. We suggest that species boundaries based on neutral molecular data may differ from those based on observed morphological traits.     

Cross‐decades stability of an avian hybrid zone

Hybrid zones are particularly valuable for understanding the evolution of partial reproductive isolation between differentiated populations. An increasing number of hybrid zones have been inferred to move over time, but in most such cases zone movement has not been tested with long‐term genomic data. The hybrid zone between Townsend's Warblers (Setophaga townsendi) and Hermit Warblers (S. occidentalis) in the Washington Cascades was previously inferred to be moving from northern S. townsendi southwards towards S. occidentalis, based on plumage and behavioural patterns as well as a 2000‐km genetic wake of hermit mitochondrial DNA (mtDNA) in coastal Townsend's Warblers. We directly tested whether hybrid zone position has changed over 2–3 decades by tracking plumage, mtDNA and nuclear genomic variation across the hybrid zone over two sampling periods (1987–94 and 2015–16). Surprisingly, there was no significant movement in genomic or plumage cline centres between the two time periods. Plumage cline widths were narrower than expected by neutral diffusion, consistent with a ‘tension zone’ model, in which selection against hybrids is balanced by movement of parental forms into the zone. Our results indicate that this hybrid zone is either stable in its location or moving at a rate that is not detectable over 2–3 decades. Despite considerable gene flow, the stable clines in multiple phenotypic and genotypic characters over decades suggest evolutionary stability of this young pair of sister species, allowing divergence to continue. We propose a novel biogeographic scenario to explain these patterns: rather than the hybrid zone having moved thousands of kilometres to its current position, inland Townsend's met coastal Hermit Warbler populations along a broad front of the British Columbia and Alaska coast and hybridization led to replacement of the Hermit Warbler plumage with Townsend's Warbler plumage patterns along this coastline. Hence, hybrid zones along British Columbia and Alaska moved only a short distance from the inland to the coast, whereas the Hermit Warbler phenotype appears stable in Washington and further south. This case provides an example of the complex biogeographic processes that have led to the distribution of current phenotypes within and among closely related species.     

Clinal variation in a brown lemur (Eulemur spp.) hybrid zone: Combining morphological,genetic and climatic data to examine stability

Studies of hybrid zones can inform our understanding of reproductive isolation and speciation. Two species of brown lemur (Eulemur rufifrons and E. cinereiceps) form an apparently stable hybrid zone in the Andringitra region of south‐eastern Madagascar. The aim of this study was to identify factors that contribute to this stability. We sampled animals at 11 sites along a 90‐km transect through the hybrid zone and examined variation in 26 microsatellites, the D‐loop region of mitochondrial DNA, six pelage and nine morphological traits; we also included samples collected in more distant allopatric sites. Clines in these traits were noncoincident, and there was no increase in either inbreeding coefficients or linkage disequilibrium at the centre of the zone. These results could suggest that the hybrid zone is maintained by weak selection against hybrids, conforming to either the tension zone or geographical selection‐gradient model. However, a closer examination of clines in pelage and microsatellites indicates that these clines are not sigmoid or stepped in shape but instead plateau at their centre. Sites within the hybrid zone also occur in a distinct habitat, characterized by greater seasonality in precipitation and lower seasonality in temperature. Together, these findings suggest that the hybrid zone may follow the bounded superiority model, with exogenous selection favouring hybrids within the transitional zone. These findings are noteworthy, as examples supporting the bounded superiority model are rare and may indicate a process of ecologically driven speciation without geographical isolation.     

Tracing species replacement in Iberian marbled newts

Secondary contact between closely related species can lead to the formation of hybrid zones, allowing for interspecific gene flow. Hybrid zone movement can take place if one of the species possesses a competitive advantage over the other, ultimately resulting in species replacement. Such hybrid zone displacement is predicted to leave a genomic footprint across the landscape in the form of asymmetric gene flow (or introgression) of selectively neutral alleles from the displaced to the advancing species. Hybrid zone movement has been suggested for marbled newts in the Iberian Peninsula, supported by asymmetric gene flow and a distribution relict (i.e., an enclave) of Triturus marmoratus in the range of T. pygmaeus. We developed a panel of nuclear and mitochondrial SNP markers to test for the presence of a T. marmoratus genomic footprint in the Lisbon peninsula, south of the enclave. We found no additional populations of T. marmoratus. Analysis with the software Structure showed no genetic traces of T. marmoratus in T. pygmaeus. A principal component analysis showed some variation within the local T. pygmaeus, but it is unclear if this represents introgression from T. marmoratus. The results may be explained by (a) species replacement without introgressive hybridization and (b) displacement with hybridization followed by the near‐complete erosion of the footprint by purifying selection. We predict that testing for a genomic footprint north of the reported enclave would confirm that species replacement in these marbled newts occurred with hybridization.     

Phenotypic and genetic introgression across a moving woodpecker hybrid zone

In hybrid zones in which two divergent taxa come into secondary contact and interbreed, selection can maintain phenotypic diversity despite widespread genetic introgression. Red‐breasted (Sphyrapicus ruber) and red‐naped (S. nuchalis) sapsuckers meet and hybridize along a narrow contact zone that stretches from northern California to southern British Columbia. We found strong evidence for changes in the structure of this hybrid zone across time, with significant temporal shifts in allele frequencies and in the proportions of parental phenotypes across the landscape. In addition to these shifts, we found that differences in plumage predict genetic differences (R² = 0.80), suggesting that plumage is a useful proxy for assessing ancestry. We also found a significant bimodal distribution of hybrids across the contact zone, suggesting that premating barriers may be driving reproductive isolation, perhaps as a result of assortative mating based on plumage differences. However, despite evidence of selection and strong patterns of population structure between parental samples, we found only weak patterns of genetic divergence. Using museum specimens and genomic data, this study of sapsuckers provides insight into the ways in which phenotypic and genetic structure have changed over a 40‐year period, as well as insight into the mechanisms that may contribute to the maintenance of the hybrid zone over time.     

Clinal genetic variation and the 'rare allele phenomenon' in random mating populations of <Emphasis Type="Italic">Urophora cardui</Emphasis> (Diptera: Tephritidae)

In the present study we investigate a contact zone between two population groups of the tephritid fly Urophora cardui. We investigate scenarios that may have produced the genetic differentiation of the two groups, and we describe the 'rare allele phenomenon' from the contact zone. The rare allele phenomenon refers to alleles that are found at high frequency in contact zones but are rare or lacking outside the contact zone. The phenomenon is often observed in hybrid zones between subspecies of limited reproductive compatibility, but seldom in populations with random mating. Clinal genetic variation was observed at three loci in the contact zone. Three alleles at the locus Aat showed steep clines, between 20--70 km wide. A rare Aat-A allele occurred at high frequency in the centre of the contact zone. Two further loci, Hk and Pgd, showed less steep clinal genetic variation, the transition being in and slightly south of the centre of the Aat cline. Populations showed Hardy--Weinberg proportions and there was no evidence for linkage dis-equlibrium. These findings suggest random mating and gradual introgression between the population systems, which may originate from at least two range expansions. Aat's steep clines and rare allele may indicate selection on Aat alleles, although we presently can not quantify any agents. Because U. cardui experiences random mating in the contact zone with no apparent 'hybrid' incompatibility, mating experiments offer the possibility for future enquiries about the genetic basis of the rare allele phenomenon.     

Hybridization and mitochondrial genome introgression between Rana chensinensis and R. kukunoris

Mitochondrial genome (mito‐genome) introgression among metazoans is commonplace, and several biological processes may promote such introgression. We examined two proposed processes for the mito‐genome introgression between Rana chensinensis and R. kukunoris: natural hybridization and sex‐biased dispersal. We sampled 477 individuals from 28 sites in the potential hybrid zone in the western Tsinling Mountains. Mitochondrial gene (cyt‐b) trees were used to examine the introgression events. Microsatellite DNA loci, cyt‐b and morphological data were used to identify hybrids and to examine the extent of natural hybridization. We detected rampant bidirectional introgressions, both ancient and recent, between the two species. Furthermore, we found a wide hybrid zone, and frequent and asymmetric hybridization. The hybrid zone cline analysis revealed a clear mitochondrial–nuclear discordance; while most nuclear markers displayed similar and steep clines, cyt‐b had a displaced cline centre and a more gradual and wider cline. We also detected strong and asymmetric historical maternal gene flow across the hybrid zone. This widespread hybridization and detected low mito‐nuclear conflicts may, at least partially, explain the high frequency of introgression. Lastly, microsatellite data and population genetic methods were used to assess sex‐biased dispersal. A weak pattern of female‐biased dispersal was detected in both species, suggesting it may not play an important role in the observed introgression. Our data are consistent with the hybridization hypothesis, but support for the sex‐biased dispersal hypothesis is weak. We further suggest that selective advantages of the R. kukunoris‐type mito‐genome in thermal adaptation may also contribute to the introgression between the two species.     

High gene flow despite opposite chirality in hybrid zones between enantiomorphic door snails

We studied differentiation and geneflow patterns between enantiomorphic door‐snail species in two hybrid zones in the Bucegi Mountains (Romania) to investigate the effects of intrinsic barriers (complications in copulation) and extrinsic selection by environmental factors. A mitochondrial gene tree confirmed the historical separation of the examined populations into the dextral Alopia livida and the sinistral Alopia straminicollis in accordance with the morphological classification, but also indicated gene flow between the species. By contrast, a network based on amplified fragment length polymorphisms (AFLP) markers revealed local groups of populations as units independent of their species affiliation. Admixture analyses based on AFLP data showed that the genomes of most individuals in the hybrid zones are composed of parts of the genomes of both parental taxa. The introgression patterns of a notable fraction of the examined markers deviated from neutral introgression. However, the patterns of most non‐neutral markers were not concordant between the two hybrid zones. There was also no concordance between non‐neutral markers in the two genomic clines and markers that were correlated with environmental variables or markers that were correlated with the proportion of dextral individuals in the populations. Neither extrinsic selection by environmental factors nor intrinsic barriers resulting from positive frequency‐dependent selection of the prevailing coiling direction were sufficient to maintain the distinctness of A. straminicollis and A. livida. Despite being historically separated units, we conclude that these taxa now merge where they come into contact.     

Movement of a Heliconius hybrid zone over 30 years: A Bayesian approach

Hybrid zones have long been of interest to biologists as natural laboratories where we can gain insight into the processes of adaptation and speciation. Repeated sampling of individual hybrid zones has been particularly useful in elucidating the dynamic balance between selection and dispersal that maintains most hybrid zones. Here, we revisit a hybrid zone between Heliconius erato butterflies in Panamá for a third time over more than 30 years. We combine a novel Bayesian extension of stepped‐cline hybrid zone models with environmental data to understand the genetic and environmental causes of cline dynamics in this species. The cline has continued to move west, likely due to dominance drive, but has slowed and broadened. Environmental analyses suggest that widespread deforestation in Panamá could be leading to decreased avian predation and relaxed selection, causing the observed changes in cline dynamics.     

Discordant patterns of introgression across a narrow hybrid zone between two cryptic lineages of an Iberian endemic newt

The study of natural hybrid zones can illuminate aspects of lineage divergence and speciation in morphologically cryptic taxa. We studied a hybrid zone between two highly divergent but morphologically similar lineages (south‐western and south‐eastern) of the Iberian endemic Bosca's newt (Lissotriton boscai) in SW Iberia with a multilocus dataset (microsatellites, nuclear and mitochondrial genes). STRUCTURE and NEWHYBRIDS analyses retrieved few admixed individuals, which classified as backcrosses involving parental individuals of the south‐western lineage. Our results show asymmetric introgression of mtDNA beyond the contact from this lineage into the south‐eastern lineage. Analysis of nongeographic introgression patterns revealed asymmetries in the direction of introgression, but except for mtDNA, we did not find evidence for nonconcordant introgression patterns across nuclear loci. Analysis of a 150‐km transect across the hybrid zone showed broadly coincident cline widths (ca. 3.2–27.9 km), and concordant cline centres across all markers, except for mtDNA that is displaced ca. 60 km northward. Results from ecological niche modelling show that the hybrid zone is in a climatically homogenous area with suitable habitat for the species, suggesting that contact between the two lineages is unlikely to occur further south as their distributions are currently separated by an extensive area of unfavourable habitat. Taken together, our findings suggest the genetic structure of this hybrid zone results from the interplay of historical (biogeographic) and population‐level processes. The narrowness and coincidence of genetic clines can be explained by weak selection against hybrids and reflect a degree of reproductive isolation that is consistent with cryptic speciation.     

Preliminary analysis of a hybrid zone between two subspecies of Zootermopsis nevadensis

For species largely allopatric in distribution, zones of contact provide an opportunity for hybridization, testing grounds for species boundaries, and may result in the formation of a new species. Thus, hybrid zones have the potential to provide important insights into speciation. In this study, we performed a preliminary analysis of a hybrid zone between two subspecies of the dampwood termite, Zootermopsis nevadensis (Z. n. nuttingi and Z. n. nevadensis) near Bartle, CA, using 12 microsatellite loci and a mitochondrial gene. Fifty-seven colonies collected in 36 locations were analyzed. The analysis of genetic markers revealed a large hybrid zone approximately 104 km in width. Although stepped clines best explained the data, we are unable to rule out the existence of a mosaic hybrid zone. We inferred weak selection (s* < 0.05%) against hybrids, but the data also suggested the existence of a barrier to gene flow from Z. n. nevadensis to Z. n. nuttingi, but not in the other direction. Given the large zone of contact, extensive sampling is needed to obtain a more complete characterization of the hybrid zone. However, the results of this study suggest that despite the accumulation of phenotypic and genetic differences and time since divergence (~2 million years), Z. n. nuttingi and Z. n. nevadensis are capable of extensive hybridization.     

Introgression between invasive and native blue mussels (genus Mytilus) in the central California hybrid zone

The ecological and genetic factors determining the extent of introgression between species in secondary contact zones remain poorly understood. Here, we investigate the relative importance of isolating barriers and the demographic expansion of invasive Mytilus galloprovincialis on the magnitude and the direction of introgression with the native Mytilus trossulus in a hybrid zone in central California. We use double‐digest restriction‐site‐associated DNA sequencing (ddRADseq) to genotype 1337 randomly selected single nucleotide polymorphisms and accurately distinguish early and advanced generation hybrids for the first time in the central California Mytilus spp. hybrid zone. Weak levels of introgression were observed in both directions but were slightly more prevalent from the native M. trossulus into the invasive M. galloprovincialis. Few early and advanced backcrossed individuals were observed across the hybrid zone confirming the presence of strong barriers to interbreeding. Heterogeneous patterns of admixture across the zone of contact were consistent with the colonization history of M. galloprovincialis with more extensive introgression in northern localities furthest away from the putative site of introduction in southern California. These observations reinforce the importance of dynamic spatial and demographic expansions in determining patterns of introgression between close congeners, even in those with high dispersal potential and well‐developed reproductive barriers. Our results suggest that the threat posed by invasive M. galloprovincialis is more ecological than genetic as it has displaced, and continues to displace the native M. trossulus from much of central and southern California.     

Müllerian mimicry of a quantitative trait despite contrasting levels of genomic divergence and selection

Hybrid zones, where distinct populations meet and interbreed, give insight into how differences between populations are maintained despite gene flow. Studying clines in genetic loci and adaptive traits across hybrid zones is a powerful method for understanding how selection drives differentiation within a single species, but can also be used to compare parallel divergence in different species responding to a common selective pressure. Here, we study parallel divergence of wing colouration in the butterflies Heliconius erato and H. melpomene, which are distantly related Müllerian mimics which show parallel geographic variation in both discrete variation in pigmentation, and quantitative variation in structural colour. Using geographic cline analysis, we show that clines in these traits are positioned in roughly the same geographic region for both species, which is consistent with direct selection for mimicry. However, the width of the clines varies markedly between species. This difference is explained in part by variation in the strength of selection acting on colour traits within each species, but may also be influenced by differences in the dispersal rate and total strength of selection against hybrids between the species. Genotyping‐by‐sequencing also revealed weaker population structure in H. melpomene, suggesting the hybrid zones may have evolved differently in each species, which may also contribute to the patterns of phenotypic divergence in this system. Overall, we conclude that multiple factors are needed to explain patterns of clinal variation within and between these species, although mimicry has probably played a central role.