Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2025 Jul 1.
Published in final edited form as: Evolution. 2024 Jul 1;78(7):1338–1346. doi: 10.1093/evolut/qpae073

Selection drives divergence of eye morphology in sympatric Heliconius butterflies

Daniel Shane Wright 1,✉, Juliana Rodriguez-Fuentes 1,b, Lisa Ammer 1, Kathy Darragh 2,3,a, Chi-Yun Kuo 1, W Owen McMillan 3, Chris D Jiggins 2,3, Stephen H Montgomery 3,4, Richard M Merrill 1,3,✉
PMCID: PMC7616201  EMSID: EMS197235  PMID: 38736286

Abstract

When populations experience different sensory conditions, natural selection may favor sensory system divergence, affecting peripheral structures and/or downstream neural pathways. We characterized the outer eye morphology of sympatric Heliconius species from different forest types and their first-generation reciprocal hybrids to test for adaptive visual system divergence and hybrid disruption. In Panama, Heliconius cydno occurs in closed forests, whereas Heliconius melpomene resides at the forest edge. Among wild individuals, H. cydno has larger eyes than H. melpomene, and there are heritable, habitat-associated differences in the visual brain structures that exceed neutral divergence expectations. Notably, hybrids have intermediate neural phenotypes, suggesting disruption. To test for similar effects in the visual periphery, we reared both species and their hybrids in common garden conditions. We confirm that H. cydno has larger eyes and provide new evidence that this is driven by selection. Hybrid eye morphology is more H. melpomene-like despite body size being intermediate, contrasting with neural trait intermediacy. Overall, our results suggest that eye morphology differences between H. cydno and H. melpomene are adaptive, and that hybrids may suffer fitness costs due to a mismatch between the peripheral visual structures and previously described neural traits that could affect visual performance.

Keywords: adaptation, PST, visual system, hybrid disruption

Introduction

Sensory systems mediate the transmission of information between an organism and its surroundings (Stevens, 2013). Natural selection is expected to favor divergent sensory phenotypes across populations exposed to different sensory conditions and/or which exploit different resources, potentially leading to both pre- and post-mating mating reproductive isolation (Dell’Aglio et al., 2024). Habitat-associated variation in sensory traits is well-documented, particularly for vision (Webster, 2015). However, most studies of visual adaptation between populations have focused on color vision in aquatic organisms (Carleton & Yourick, 2020; Cummings & Endler, 2018), whereas other aspects of visual perception are understudied. For example, we know little about how visual signals are evaluated (Rosenthal, 2018) or how adaptive processes operate at different levels within the visual pathway. Moreover, we lack information on the visual capabilities of hybrids despite their potential to contribute to reproductive isolation between species. Here, we characterize the outer eye morphology of Heliconius butterflies, examine evidence for adaptive divergence, and consider how this may impact visual system performance in hybrids.

The compound eyes of insects represent an easily quantifiable sensory structure that directly affects visual perception. The insect compound eye consists of numerous independent photosensitive units, ommatidia, each of which receives visual (light) information and transfers it to the brain. Variation in the total number and size of ommatidia directly affects visual perception and often correlates with diel activity (Greiner, 2006; Land, 1997; Stöckl et al., 2017; Warrant, 2004). For example, nocturnal and crepuscular species often have larger eyes and larger facets, enhancing sensitivity in dim-light environments, as observed across insect taxa (Freelance et al., 2021). Associations between the local environment and the visual systems of diurnal insects are less studied and often included only as a comparison to other nocturnal species (e.g., Frederiksen & Warrant, 2008). Nonetheless, the conditions experienced by diurnal insects can vary greatly (Endler, 1993), and may represent important adaptations to the local environment.

Heliconius butterflies inhabit tropical and subtropical regions of the Americas and rely heavily on vision when foraging for both flowers and hostplants (Dell’Aglio et al., 2016; Gilbert, 1982), as well as finding and choosing suitable mates (Crane, 1955; Estrada & Jiggins, 2008; Hausmann et al., 2021; Jiggins et al., 2001; Merrill et al., 2019). In Panama, the closely related species Heliconius melpomene and Heliconius cydno are broadly sympatric, but occupy different forest types (Estrada & Jiggins, 2002). H. melpomene primarily lives in forest edge habitats, whereas H. cydno occurs deeper within closed-canopy forests, where light intensity is decreased (Fig. 1A) (DeVries, 1987; Estrada & Jiggins, 2002; Seymoure, 2016). Although patterns of opsin expression suggest few differences in spectral sensitivity (McCulloch et al., 2017), recent data on brain morphologies of H. melpomene and H. cydno reported heritable differences in the size of the visual neuropils that exceed expected rates of neutral genetic divergence (Montgomery et al., 2021). Using wild caught individuals, Seymoure et al. (2015) similarly found that i) H. cydno has larger eyes than H. melpomene and ii) H. cydno males have larger eyes than H. cydno females (intraspecific differences in H. melpomene were non-significant). However, these results were based on individuals sampled as adults in their respective habitats and may include effects of environment-induced plasticity. Also, total ommatidia counts were measured for only two individuals for each species and sex, so statistical power to explore different eye traits was limited.

Figure 1.

Figure 1

Open in a new tab

(A) Heliconius cydno chioneus and Heliconius melpomene rosina occur sympatrically in Panama but occupy different habitats: H. cydno is found in closed-forest environments, whereas H. melpomene resides at the forest edge. (B, C, E, F) Estimated marginal means (e.m.m.) of the minimum adequate statistical models for (B, C) facet count and (E, F) corneal area, demonstrating the significant effects of (B, E) species and (C, F) sex, while accounting for other significant terms (tibia length and sex/species). The interaction between species and sex was never significant (FDR-p > 0.2). Different letters indicate significant differences (p < 0.05), and the error bars represent 95% confidence intervals. (D, G) Major axis regressions of (D) facet count or (G) corneal area and body size, measured as hind tibia length. Double-logarithmic plots are presented to explore the allometric relationships between eye morphology and body size. Males are represented by solid shapes and solid lines, and females are represented by open shapes and dashed lines. C = H. cydno; M = H. melpomene.

Given the evidence for selection acting on the visual processing regions of the brain, a more thorough examination of the visual periphery in these species is warranted. In particular, the respective roles of plasticity and selection has not yet been assessed. Neither has the potential for mismatched neural and peripheral traits within the visual pathways of interspecific hybrids. To address this, we characterized the outer eye morphology of H. melpomene and H. cydno to test for patterns of adaptive divergence in the visual system. First, we compared the eye morphology of butterflies (15+ for each species and sex) reared under common garden conditions. We then used a quantitative genetics approach to test if the species-specific eye differences are due to selection using PST-FST analysis. Finally, we report patterns of eye morphology in first-generation (F1) hybrids.

Methods

Butterfly specimens

We collated samples from outbred stocks of Heliconius cydno chioneus (C) and Heliconius melpomene rosina (M), established from wild butterflies caught in Gamboa and the nearby Soberanía National Park, Panama. Most butterflies were sampled in 2015-2017 (9 H. cydno and 20 H. melpomene) or 2018-2019 (16 H. cydno and 10 H. melpomene) versus fewer samples in 2008-2009 (9 H. cydno and 3 H. melpomene). We also exploited previously sampled reciprocal F1 hybrids between H. cydno and H. melpomene, generated by crossing a H. cydno female with a H. melpomene male (CxM) or a H. melpomene female with a H. cydno male (MxC). All hybrids were generated in 2017-2019, involving four CxM crosses and six MxC crosses (unique parents for each cross; see Table S1, Fig. S2). All butterflies were reared under common garden conditions in the Smithsonian Tropical Research Institute insectaries in Gamboa, and all specimens were preserved in DMSO/EDTA/NaCl and stored at -80° C as described in Merrill et al. (2019).

Sample preparation

Samples were prepared following previously published methods (Seymoure et al., 2015; Wright et al., 2023). In brief, we thawed specimens at room temperature, imaged the entire head (minus antennae and proboscis), and dissected out both eyes and the hind legs. The legs were immediately imaged (see below), while the eyes were placed in 20% sodium hydroxide (NaOH) for 18-24 hours to loosen the tissues behind the cuticular cornea. The following day, we cleaned each eye cuticle of excess tissue and mounted it on a microscope slide in Euparal (Carl Roth GmbH). The sample was left to dry overnight before imaging.

Image analysis

We used ImageJ/Fiji (Schindelin et al., 2012) to analyze each mounted cornea for the total number of facets and total corneal area. All slides were imaged at 7.5x on a Leica M80 stereomicroscope fitted with a Leica Flexacam C1 camera and the Leica Application Suite X (LAS X) software. Each image contained a 1mm scale bar for calibration. Facet counts were measured via image thresholding and the Analyze particles function, and corneal surface area was measured with the Freehand selection and Measure options (full protocol provided as supplementary methods). This semi-automated method differs slightly from the approach used by Seymoure et al. (2015) but gives quantitatively similar results (Fig. S1). To account for differences in head and/or body size, we measured the distance between each eye cuticle (inter-ocular width, e.g., (Gaspar et al., 2020; Posnien et al., 2012; Wainwright et al., 2023)) and hind tibia length using the Straight line and Measure options. Finally, we also measured facet diameter as in Seymoure et al. (2015): within each anatomical region of the eye, we measured across ten facets in a row in two separate locations at least ten facets apart (using the Straight line and Measure options). Per facet diameter was calculated by dividing each line segment by ten, and facet diameter per eye region was calculated as the mean of the two locations.

The number of facets (Pearson’s r [95% C.I.]: r (118) = 0.969 [0.956, 0.978]), corneal area (r (118) = 0.986 [0.979, 0.990]) and hind tibia length (r (101) = 0.930 [0.898, 0.952]) on the left vs. right sides of the butterfly were highly correlated. Therefore, for all subsequent analyses, we used only the left eye and left leg unless either was missing, damaged or had poor image quality (i.e., not all facets visible), in which case the right side was substituted.

Statistical analysis

Eye morphology

We used linear models (lm function) in R to explore how facet count and corneal area are influenced by species (H. cydno vs. H. melpomene), sex (male vs. female), body size (hind tibia length), head size (inter-ocular width), and sample period (2008-2009, 2015-2017, or 2018-2019) as: log10(facet count or corneal area) ~ species * sex + log10(tibia length) + log10(inter-ocular width) + sample period. Log10-transformations were used to normalize the residuals around the allometric relationships to meet the assumptions of normality (Thorpe, 1975). We also used linear models to assess i) the relationship between facet count and corneal as: log10(corneal area) ~ log10(facet count) * species * sex and ii) body and head size differences as: log10(tibia length) ~ species + sex or log10(inter-ocular width) ~ species + sex. For facet diameter, we used a linear mixed model [lmer in lme4 (Bates et al., 2015)] to assess how facet size was influenced by species, sex, anatomical region of the eye (lateral, dorsal, posterior, anterior, anterioventral, ventral), tibia length, inter-ocular width, and sampling period as fixed effects and butterfly ID as a random effect (to account for multiple measurements from the same individual). The full model was as follows: log10(facet diameter) ~ species + sex + eye region + log10(tibia length) + log10(inter-ocular width) + sampling period + (1|ID). We did not include interactions between species, sex, and eye region because of high collinearity and because model comparison [AIC; model.sel function in the MuMIn package (Barton & Barton, 2023)] indicated a better fit without interactive predictors (see supplementary analyses for individual analyses per anatomical eye region). For all models, the significance of fixed effect parameters was determined by likelihood ratio tests via the drop1 function, and minimum adequate models (MAM) were selected using statistical significance (Crawley, 2013; Nakagawa & Cuthill, 2007). We used the Anova function in the car package (Fox & Weisberg, 2018) to estimate significant fixed effect parameters and report false discovery rate (FDR; Benjamini & Hochberg, 1995) adjusted p-values (p.adjust function, method = "fdr") to account for multiple testing. Model assumptions were confirmed via visual inspection (residual vs. fitted and normal Q-Q plots). To accurately visualize multiple significant fixed effects, we extracted and plotted the estimated marginal means from each MAM using the emmeans function in the emmeans package (Lenth et al., 2023).

We also explored whether the scaling relationships between eye morphology (facet count and corneal area) and body size (hind tibia length) differed for H. cydno and H. melpomene using major axis regressions via the sma function in the smatr package (Warton et al., 2012). Following the standard allometric scaling relationship, log y = β log x + α, we tested for shifts in the allometric slope (β). Where a common slope was supported, we subsequently tested for differences in α that would indicate ‘grade-shifts’ (test = “elevation”) and for major axis-shifts along a common slope (test = “shift”).

Test of selection

We next used a quantitative genetics approach to test whether eye morphology differences between H. cydno and H. melpomene are due to selection. QST is a quantitative genetic analogue of FST that measures additive genetic variation among populations relative to total genetic variance. However, QST estimates for quantitative traits are difficult, so QST is often replaced with its phenotypic analogue PST (Leinonen et al., 2013). Comparisons between PST and FST can be used as a test of divergent selection, where PST values that exceed genome-wide FST suggest greater phenotypic divergence than expected by neutral genetic divergence.

We calculated PST values using the Pst function in the Pstat package (Da Silva & Da Silva, 2018) for raw, log10-transformed, and body-size corrected measurements of eye morphology (i.e., facet count and corneal area). Allometrically-scaled correlations (using tibia length) per species and sex were performed via the allomr function in the allomr package (Schär, 2023). PST approximation to QST depends on heritability, h2, and a scalar c that expresses the proportion of the total variance that is presumed to be due to additive genetic effects across populations (Brommer, 2011). Heritability estimates (h2) for facet count and corneal area are unknown for these species, so in addition to the default value of 1, we used varying c/h2 ratios [ranging from 0.33 to 4, following Montgomery et al. (2021)], which encompasses a wide range of possible h2 values. Genome-wide FST values between H. c. chioneus and H. m. rosina were obtained from Martin et al. (2013), derived from four wild-caught individuals per species using 100-kb genomic windows. We calculated p-values as the proportion of the FST distribution that was above each PST value (Leinonen et al., 2013); values above the 95th percentile of the FST distribution were interpreted as an indication of selection.

Hybrid phenotypes

We re-ran the linear models described above but included the CxM and MxC hybrids as two additional groups within the species factor. To test if hybrid body size was intermediate to H. cydno and H. melpomene, we also re-tested the linear models: log10(tibia length) ~ species + sex and log10(inter-ocular width) ~ species + sex. In the case of more than two categories per fixed effect parameter (i.e., species), we used post hoc Tukey tests (glht-multcomp package (Hothorn et al., 2008)) to obtain parameter estimates and report Bonferroni adjusted p-values for multiple comparisons.

Results

Heritable shifts in eye morphology between species residing in different forest types

In total, we sampled 15 male and 19 female H. cydno and 18 male and 15 female H. melpomene, all reared under common garden conditions. We found that H. cydno was larger (using tibia length as a body size proxy; F1, 65 = 58.77, FDR-p < 0.001) and had bigger heads than H. melpomene (using inter-ocular width as a head size proxy; F1, 65 = 20.35, FDR-p < 0.001). However, there was no evidence for sexual size dimorphism in either species (FDR-p > 0.8), and body/head size did not differ across collection periods (FDR-p > 0.25). Three fixed effects, tibia length, species, and sex, were retained in our model examining facet count; inter-ocular width and sample period were both non-significant (FDR-p > 0.42; Table S2). After accounting for size (larger butterflies had more facets: F1, 63 = 20.44, FDR-p < 0.001), H. cydno had more facets than H. melpomene (F1, 63 = 7.32, FDR-p = 0.02; Fig. 1B), and males had more facets than females (F1, 63 = 27.35, FDR-p < 0.001; Figs. 1C, S2). The interaction between species and sex was not significant (FDR-p = 0.56; Fig. S3), implying conserved patterns of sexual dimorphism. We found similar results for corneal area (Fig. 1E, F), where only tibia length, species, and sex were retained in the model (Table S2). Larger corneal area appears to be due to an increase in facet number [as evidenced by facet count significantly affecting corneal area in the model log10(corneal area) ~ log10(facet count) * species * sex (F1,64 = 235.38, FDR-p < 0.001)] and not from differences in facet diameter; only tibia length (χ2 = 47.47, df = 1, FDR-p = 0.001) and anatomical eye region (χ2 = 454.45, df = 5, FDR-p < 0.001; Table 1) affected diameter; both species and sex were non-significant (FDR-p > 0.7).

Table 1.

Facet diameter varied across the anatomical regions of the eye, invariantly between species (FDR-p = 0.7) or between males vs. females (FDR-p = 0.8). See table S5 for hybrid facet diameter.

facet diameter (μm ± s.d.)
species sex lateral dorsal posterior anterior anterio-ventral ventral
M 26.74 ± 0.92 23.50 ± 1.70 24.10 ± 1.29 27.23 ± 0.87 24.67 ± 1.85 24.08 ± 1.21
H. cydno
F 27.66 ± 1.22 24.95 ± 2.01 24.00 ± 1.66 27.68 ± 1.02 25.31 ± 1.99 24.00 ± 1.63
M 26.42 ± 1.09 22.64 ± 1.48 23.03 ± 1.29 26.27 ± 1.04 24.71 ± 1.38 23.68 ± 1.47
H. melpomene
F 25.80 ± 1.78 22.99 ± 2.12 22.78 ± 1.09 26.15 ± 0.88 24.37 ± 1.20 22.92 ± 1.93
Open in a new tab

Given the sex-specific differences in eye morphology reported above, we analyzed the scaling relationships between eye morphology and body size for males and females separately (Fig 1D, G). The only significant difference in slope (β) was when comparing the scaling relationship between corneal area and tibia length for H. cydno vs. H. melpomene males (FDR-p = 0.016; Table S3); all other comparisons were non-significant, confirming common slopes (FDR-p > 0.23; Table S3). In isolation, body size and facet count were uncorrelated for H. cydno males (r2 = 0.004, p = 0.8), but there was no statistical difference in scaling between the species suggesting this is potentially due to increased variance in H. cydno (Table S3). For all comparisons with a common slope, tests for grade-shifts (α) were non-significant (FDR-p > 0.4), but there was a significant shift along the common axis (FDR-p < 0.001; Table S3).

Differences in eye morphology are driven by selection

Our PST-FST analyses suggest that the visual systems of H. cydno and H. melpomene have likely diverged as the result of selection rather than genetic drift. PST was significantly higher than FST for both facet count and corneal area (p < 0.001) for all comparisons (Fig. 2; Table S6), where the proportion of phenotypic variance due to additive genetic effects within-populations far exceeds the proportion of phenotypic variance due to additive genetic effects between-populations. Quantitatively similar results were obtained regardless of the phenotypic measurement evaluated (i.e., raw data, log10 transformed, allometrically corrected values; Table S6), when each sex was examined separately (Tables S7, S8), and when accounting for potential pedigree effects in our study population (see supplementary analyses).

Figure 2.

Figure 2

Open in a new tab

Location of the calculated PST values for (A) facet count and (B) corneal area in the distribution of FST values between H. cydno chioneus and H. melpomene rosina (values from Martin et al., 2013). Here, both morphological measurements are allometrically corrected using tibia length as a body-size proxy and presented using varying c/h2 ratios (see Table S6 for PST estimates using raw and log10 transformed values). The dotted line represents the 95th percentile of the FST distribution.

Hybrid eye morphology is H. melpomene-like

We examined the eye morphology of 60 F1 hybrids, including 15 male and 15 female CxM individuals, and 15 female and 15 male MxC individuals (Table S4). Tukey post hoc (with Bonferroni correction) revealed patterns of intermediacy in body size (using tibia length as a body size proxy) for both hybrid directions compared to the parental species: both CxM and MxC hybrids were larger than H. melpomene (p < 0.001), MxC was smaller than H. cydno (t = -3.148, p = 0.0124), and CxM did not differ from H. cydno (p = 1.0; Fig. 3A). Similar patterns were observed for head size (using inter-ocular width as a head size proxy): both CxM and MxC hybrids were larger than H. melpomene (p < 0.018), but neither hybrid type differed from H. cydno (p = 1.0). However, F1 hybrid eye morphology was not intermediate to the parental species: both hybrid types had significantly fewer facets (p < 0.0035) than H. cydno but did not differ from H. melpomene (p = 1.0; Fig. 3B). Results for corneal area were similar (Table S4; Fig. S4, S5). These patterns were also evident when exploring the scaling relationships between eye morphology and body size of the F1 hybrids (Fig. 3C, D; Fig. S5).

Figure 3.

Figure 3

Open in a new tab

(A) Estimated marginal means (e.m.m.) of the minimum adequate statistical model for body size (using tibia length as a proxy) including F1 hybrids, demonstrating the significant effect of species (sex was non-significant, FDR-p = 0.56). E.m.m. for (B) facet count, showing the significant effect of species, while accounting for significant tibia length and sex effects. In both plots, different letters indicate significant differences (Bonferroni adjusted p < 0.05), and the error bars represent 95% confidence intervals. (C-D) Major axis regressions of facet count and body size (tibia length) for males and females separately. Double-logarithmic plots are presented to explore the allometric relationships between eye morphology and body size. C = H. cydno; M = H. melpomene; CxM = F1 hybrid of H. cydno mother crossed with H. melpomene father; MxC = F1 hybrid of H. melpomene mother crossed with H. cydno father.

Discussion

When populations are exposed to different sensory conditions, natural selection may favor divergence in sensory traits, which can contribute to speciation. We compared the outer eye morphology of Heliconius butterflies that occupy different environments to test for patterns of adaptive visual system divergence. Our results show that H. cydno, which occupies closed-canopy forests, has larger eyes. This difference is observed in both wild (Seymoure et al., 2015) and insectary-reared individuals (this study), suggesting heritable differences in facet number. By combining our phenotypic data with genome wide estimates of FST, we additionally provide evidence that selection has driven the divergence of eye morphology in these butterflies. Finally, we show that F1 hybrid eye morphology is not intermediate to the parental species, contrasting with patterns for body size and neuroanatomy. This suggests that hybrid visual performance may suffer due to mismatched trait combinations within the visual periphery and brain regions responsible for processing visual information.

Our results are consistent with previous work by Seymoure et al. (2015), which reported larger eyes for H. cydno, and bigger eyes in H. cydno (but not H. melpomene) males. However, the individuals sampled for our analyses were raised under common garden conditions, reducing the potential for environmental effects and genotype–environment interactions, which may give a distorted picture of the contribution of genetic variation on which selection can act (Brommer, 2011; Pujol et al., 2008). A potential caveat of our results is that we cut each cuticle four times to mount it on the microscope slides, which may have disrupted the semi-automated counts of individual facets. It is possible that more advanced 3D imaging techniques (e.g., Buffry et al., 2024), where cutting is not required, may give slightly higher total counts, though likely at the expense of overall sample size and associated statistical power. Regardless, our facet counts are consistent with prior work (Seymoure et al., 2015) (Fig. S1), and the close correspondence between wild and insectary-reared butterflies further suggests that the differences in eye morphology are largely heritable.

Habitat-associated variation in eye morphology has been reported across taxa (e.g., insects: (Greiner, 2006); mammals: (Veilleux & Lewis, 2011); fish: (Lisney et al., 2020); snakes: (Liu et al., 2012); primates: (Kirk, 2004)), and this variation is generally interpreted as an adaptive response to the local sensory conditions. For example, visual perception in insects is affected by the total number of ommatidia and their size (Greiner, 2006; Land, 1997; Warrant, 2004), and nocturnal and crepuscular species often possess larger eyes and larger facets to enhance sensitivity in dim-light environments (Freelance et al., 2021). However, most studies do not formally evaluate the role of selection, which is a key element to define adaptation (Gould & Vrba, 1982). We are aware of only one study that has attempted to addresses this topic: Brandon et al. (2015) reported that eye size variation in a wild Daphnia population is associated with variation in reproductive output, suggesting that selection is operating, either directly or indirectly, on eye size variation (though the underlying mechanisms remain unknown).

In addition to revealing heritable differences in eye morphology, our results suggest that eye morphology in H. cydno and H. melpomene have evolved as the result of divergent selection, as opposed to genetic drift. Although our PST-FST approach to test for evidence of selection acting on eye morphology is limited by the difficulty in approximating PST to QST (Brommer, 2011), these limitations are partially overcome by rearing our butterflies under common garden conditions (Leinonen et al., 2008, 2013). Moreover, as with most insects, the heritability of facet count and corneal area are unknown for the species used in this study. To account for this, we used a wide range of c/h2 ratios, including the default assumption of c = h2 (i.e., c/h2 = 1), where the proportion of phenotypic variance due to additive genetic effects is the same for between-population variance and within-population variance (Brommer, 2011). In all cases, PST values were higher than the 95th percentile of the genome wide FST distribution (Tables S6-S8).

In insects, eye size may increase due to an increase in the number of ommatidia, an increase in individual ommatidia size, or both. Our results suggest that larger eye size in H. cydno, and in males of both species, is predominantly due to an increase in ommatidia number, as we found no species differences in facet diameter. This contrasts with Seymoure et al. (2015), who reported larger facets in wild H. cydno compared to H. melpomene, but these results stemmed from an analysis of covariance including other species (H. sapho and H. erato), and there were no direct comparisons between H. cydno and H. melpomene. Our more robust analyses demonstrates that the larger eyes of H. cydno, and male Heliconius, are due to an increase in ommatidia number. In insects, increased ommatidia number is thought to contribute to higher visual acuity (Land, 1997) - behavioral and morphological data revealing sexual dimorphism in the visual acuity of H. erato support this prediction (Wright et al., 2023). We note that our allometric analyses suggest part of the variation we observe is associated with body size, within and between species. However, our PST-FST analyses account for body size and still suggest a signal of selection. The pattern we observe in hybrids, where grade-shifts are clearly observed between cydno and cydno x melpomene hybrids (Figure 3C), also suggest a strong genetic component independent of body size. As such, while the behavioural impact of increased eye size likely depends specifically on the raw number of facets, we conclude that selection on body size alone does not explain the increase in H. cydno eye size.

Multiple non-exclusive selective pressures could be driving the differences we report here. First, species differences may be explained by H. cydno occupying closed-forest environments (Estrada & Jiggins, 2002), where more ommatidia are advantageous due to e.g., increased visual acuity (Land, 1997; Wright et al., 2023). More ommatidia in males may also be due to general ecological differences between the sexes: males actively search for and identify mates (Rutowski, 2000), whereas females need to identify hostplants for oviposition (Benson, 1978; Brown, 1981). These differing selective pressures may have resulted in different visual adaptations, such as female-specific UV discrimination in some Heliconius species (Chakraborty et al., 2023; Finkbeiner & Briscoe, 2021; McCulloch et al., 2022), though this trait is not present in the species studied here. Interestingly, we observe species differences for both sexes (Fig. S3) and when exploring the PST-FST results for each sex separately, we still observed evidence of selection (Tables S7, S8). Taken together, our results suggest that the selective pressure on males to have more ommatidia acts in both species in concert with selection for more ommatidia in closed-forest environments. Similar patterns may exist across Heliconius, but to date, few species have been surveyed for eye morphology.

Our results also mirror neuroanatomical comparisons between species across the cydno-melpomene clade, where larger visual neuropils are reported for cydno-clade species occupying closed-forest environments, as opposed to H. melpomene (Montgomery et al., 2021). These neural differences appear to be heritable adaptations, based on similar tests of selection to those reported here. Thus, the combined results on brain (Montgomery et al., 2021) and eye morphology (this study) suggest whole visual system adaptation, from the sensory periphery to the brain. Similar habitat-associated differences in neuroanatomy have been reported in other Neotropical butterflies (Montgomery & Merrill, 2017; Wainwright & Montgomery, 2022), indicating a broader pattern of sensory adaptation in ecologically divergent, but closely related, butterflies.

One notable difference in the results obtained for eye and neural traits, however, is in the pattern of variation among F1 interspecific hybrids. The eye morphology of H. cydno and H. melpomene F1 hybrids tended to be more H. melpomene-like. This contrasts with patterns for body size (Fig. 3A) and neuroanatomy, where hybrids are intermediate for at least some traits (Montgomery et al., 2021). The observation that hybrid eye morphology is melpomene-like, but hybrid body size tends to be intermediate indicates that these two traits have different underlying genetic architectures (i.e., eye size is not simply genetically correlated with increasing body size). Furthermore, evidence suggests that both eye morphology (this study) and neural anatomy (Montgomery et al., 2021) are under divergent selection and in the predicted direction (i.e., bigger eyes and larger visual neuropils in H. cydno). If these observations truly represent adaptations (as our results suggest), then hybrids may have suboptimal visual system functioning, whereby peripheral sensory structures (number of facets) are mismatched to subsequent processing regions (optic lobe neuropils). This might predict that hybrids suffer a fitness deficit due to lower performance in visually oriented tasks, such as foraging and mate detection. However, behavioral experiments are required to test these hypotheses and to further support our evidence for adaptive visual divergence in Heliconius.

Supplementary Material

Supplementary Analysis
Supplementary Materials
Supplementary Methods

Teaser text.

Natural selection may favor sensory system divergence when populations experience different conditions. We reared Heliconius cydno (from closed forests), H. melpomene (from forest edge), and their hybrids under common garden conditions and report larger eyes in H. cydno. We also present evidence that this interspecific divergence is driven by selection, and that hybrid eye morphology is H. melpomene-like despite body size being intermediate, contrasting with prior reports of intermediate neural traits. These results are indicative of adaptive visual divergence and suggest that a mismatch between eye morphology and neural traits could affect visual performance, to the disadvantage of hybrids.

References

  1. Barton K, Barton MK. Package ‘MuMIn’. 2023. https://CRAN.R-project.org/package=MuMIn .
  2. Bates D, Mächler M, Bolker B, Walker S. Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software. 2015;67(1):1–48. doi: 10.18637/jss.v067.i01. [DOI] [Google Scholar]
  3. Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological) 1995;57(1):289–300. [Google Scholar]
  4. Benson WW. Resource partitioning in passion vine butterflies. Evolution. 1978;32(3):493–518. doi: 10.1111/j.1558-5646.1978.tb04593.x. [DOI] [PubMed] [Google Scholar]
  5. Brandon CS, James T, Dudycha JL. Selection on incremental variation of eye size in a wild population of Daphnia. Journal of Evolutionary Biology. 2015;28(11):2112–2118. doi: 10.1111/jeb.12711. [DOI] [PubMed] [Google Scholar]
  6. Brommer JE. Whither Pst? The approximation of Qst by Pst in evolutionary and conservation biology. Journal of Evolutionary Biology. 2011;24(6):1160–1168. doi: 10.1111/j.1420-9101.2011.02268.x. [DOI] [PubMed] [Google Scholar]
  7. Brown KS. The Biology of Heliconius and Related Genera. Annual Review of Entomology. 1981;26(1):427–457. doi: 10.1146/annurev.en.26.010181.002235. [DOI] [Google Scholar]
  8. Buffry AD, Currea JP, Franke-Gerth FA, Palavalli-Nettimi R, Bodey AJ, Rau C, Samadi N, Gstöhl SJ, Schlepütz CM, McGregor AP, Sumner-Rooney L, et al. Evolution of compound eye morphology underlies differences in vision between closely related Drosophila species. BMC Biology. 2024;22(1):67. doi: 10.1186/s12915-024-01864-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Carleton KL, Yourick MR. Axes of visual adaptation in the ecologically diverse family Cichlidae. Seminars in Cell & Developmental Biology. 2020;106:43–52. doi: 10.1016/j.semcdb.2020.04.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chakraborty M, Lara AG, Dang A, McCulloch KJ, Rainbow D, Carter D, Ngo LT, Solares E, Said I, Corbett-Detig RB, Gilbert LE, et al. Sex-linked gene traffic underlies the acquisition of sexually dimorphic UV color vision in Heliconius butterflies. Proceedings of the National Academy of Sciences. 2023;120(33):e2301411120. doi: 10.1073/pnas.2301411120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Crane. Imaginal behaviour of a Trinidad butterfly, Heliconius erato hydara Hewitson, with special reference to the social use of color. Zoologica. 1955;40:167–196. [Google Scholar]
  12. Crawley MJ. The R book. Second edition. Wiley; 2013. [Google Scholar]
  13. Cummings ME, Endler JA. 25 Years of sensory drive: The evidence and its watery bias. Current Zoology. 2018;64(4):471–484. doi: 10.1093/cz/zoy043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Da Silva SB, Da Silva A. Pstat: An R Package to Assess Population Differentiation in Phenotypic Traits. The R Journal. 2018;10(1):447. doi: 10.32614/RJ-2018-010. [DOI] [Google Scholar]
  15. Dell’Aglio DD, Losada ME, Jiggins CD. Butterfly Learning and the Diversification of Plant Leaf Shape. Frontiers in Ecology and Evolution. 2016;4 [Google Scholar]
  16. Dell’Aglio DD, Rivas-Sánchez DF, Wright DS, Merrill RM, Montgomery SH. The Sensory Ecology of Speciation. Cold Spring Harbor Perspectives in Biology. 2024;16(1):a041428. doi: 10.1101/cshperspect.a041428. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. DeVries PJ. The Butterflies of Costa Rica and their natural history. Princeton University Press; New Jersey, US: 1987. [Google Scholar]
  18. Endler JA. The color of light in forests and its implications. Ecological Monographs. 1993 [Google Scholar]
  19. Estrada C, Jiggins CD. Patterns of pollen feeding and habitat preference among Heliconius species. Ecological Entomology. 2002;27(4):448–456. doi: 10.1046/j.1365-2311.2002.00434.x. [DOI] [Google Scholar]
  20. Estrada C, Jiggins CD. Interspecific sexual attraction because of convergence in warning colouration: Is there a conflict between natural and sexual selection in mimetic species? Journal of Evolutionary Biology. 2008;21(3):749–760. doi: 10.1111/j.1420-9101.2008.01517.x. [DOI] [PubMed] [Google Scholar]
  21. Finkbeiner SD, Briscoe AD. True UV color vision in a female butterfly with two UV opsins. Journal of Experimental Biology. 2021;224(18):jeb242802. doi: 10.1242/jeb.242802. [DOI] [PubMed] [Google Scholar]
  22. Fox J, Weisberg S. An R Companion to Applied Regression. SAGE Publications; 2018. [Google Scholar]
  23. Frederiksen R, Warrant EJ. Visual sensitivity in the crepuscular owl butterfly Caligo memnon and the diurnal blue morpho Morpho peleides: A clue to explain the evolution of nocturnal apposition eyes? Journal of Experimental Biology. 2008;211(6):844–851. doi: 10.1242/jeb.012179. [DOI] [PubMed] [Google Scholar]
  24. Freelance CB, Tierney SM, Rodriguez J, Stuart-Fox DM, Wong BBM, Elgar MA. The eyes have it: Dim-light activity is associated with the morphology of eyes but not antennae across insect orders. Biological Journal of the Linnean Society. 2021;134(2):303–315. doi: 10.1093/biolinnean/blab088. [DOI] [Google Scholar]
  25. Gaspar P, Arif S, Sumner-Rooney L, Kittelmann M, Bodey AJ, Stern DL, Nunes MDS, McGregor AP. Characterization of the Genetic Architecture Underlying Eye Size Variation Within Drosophila melanogaster and Drosophila simulans. G3 Genes|Genomes|Genetics. 2020;10(3):1005–1018. doi: 10.1534/g3.119.400877. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Gilbert LE. The coevolution of a butterfly and a vine. Scientific American. 1982;247(2):110–121. [Google Scholar]
  27. Gould SJ, Vrba ES. Exaptation—A missing term in the science of form. Paleobiology. 1982;8(1):4–15. [Google Scholar]
  28. Greiner B. Adaptations for nocturnal vision in insect apposition eyes. International Review of Cytology. 2006;250:1–46. doi: 10.1016/S0074-7696(06)50001-4. [DOI] [PubMed] [Google Scholar]
  29. Hausmann AE, Kuo C-Y, Freire M, Rueda-M N, Linares M, Pardo-Diaz C, Salazar C, Merrill RM. Light environment influences mating behaviours during the early stages of divergence in tropical butterflies. Proceedings of the Royal Society B: Biological Sciences. 2021;288(1947) doi: 10.1098/rspb.2021.0157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hothorn T, Bretz F, Westfall P. Simultaneous Inference in General Parametric Models. Biometrical Journal. 2008;50(3):346–363. doi: 10.1002/bimj.200810425. [DOI] [PubMed] [Google Scholar]
  31. Jiggins CD, Naisbit RE, Coe RL, Mallet J. Reproductive isolation caused by colour pattern mimicry. Nature. 2001;411(6835):302–305. doi: 10.1038/35077075. [DOI] [PubMed] [Google Scholar]
  32. Kirk EC. Comparative morphology of the eye in primates. The Anatomical Record Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology. 2004;281A(1):1095–1103. doi: 10.1002/ar.a.20115. [DOI] [PubMed] [Google Scholar]
  33. Land MF. Visual acuity in insects. Annual Review of Entomology. 1997;42(1):147–177. doi: 10.1146/annurev.ento.42.1.147. [DOI] [PubMed] [Google Scholar]
  34. Leinonen T, McCairns RJS, O’Hara RB, Merilä J. QST-FST comparisons: Evolutionary and ecological insights from genomic heterogeneity. Nature Reviews Genetics. 2013;14(3):179–190. doi: 10.1038/nrg3395. [DOI] [PubMed] [Google Scholar]
  35. Leinonen T, O’hara RB, Cano JM, Merilä J. Comparative studies of quantitative trait and neutral marker divergence: A meta-analysis. Journal of Evolutionary Biology. 2008;21(1):1–17. doi: 10.1111/j.1420-9101.2007.01445.x. [DOI] [PubMed] [Google Scholar]
  36. Lenth RV, Bolker B, Buerkner P, Giné-Vázquez I, Herve M, Jung M, Love J, Miguez F, Riebl H, Singmann H. emmeans: Estimated Marginal Means, aka Least-Squares Means. 2023. (1.8.7) [Computer software]. https://cran.r-project.org/web/packages/emmeans/index.html.
  37. Lisney TJ, Collin SP, Kelley JL. The effect of ecological factors on eye morphology in the western rainbowfish, Melanotaenia australis. Journal of Experimental Biology. 2020;223(10):jeb223644. doi: 10.1242/jeb.223644. [DOI] [PubMed] [Google Scholar]
  38. Liu Y, Ding L, Lei J, Zhao E, Tang Y. Eye size variation reflects habitat and daily activity patterns in colubrid snakes. Journal of Morphology. 2012;273(8):883–893. doi: 10.1002/jmor.20028. [DOI] [PubMed] [Google Scholar]
  39. Martin SH, Dasmahapatra KK, Nadeau NJ, Salazar C, Walters JR, Simpson F, Blaxter M, Manica A, Mallet J, Jiggins CD. Genome-wide evidence for speciation with gene flow in Heliconius butterflies. Genome Research. 2013;23(11):1817–1828. doi: 10.1101/gr.159426.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. McCulloch KJ, Macias-Muñoz A, Mortazavi A, Briscoe AD. Multiple Mechanisms of Photoreceptor Spectral Tuning in Heliconius Butterflies. Molecular Biology and Evolution. 2022;39(4):msac067. doi: 10.1093/molbev/msac067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. McCulloch KJ, Yuan F, Zhen Y, Aardema ML, Smith G, Llorente-Bousquets J, Andolfatto P, Briscoe AD. Sexual Dimorphism and Retinal Mosaic Diversification following the Evolution of a Violet Receptor in Butterflies. Molecular Biology and Evolution. 2017;34(9):2271–2284. doi: 10.1093/molbev/msx163. [DOI] [PubMed] [Google Scholar]
  42. Merrill R, Rastas P, Martin SH, Melo MC, Barker S, Davey J, McMillan WO, Jiggins CD. Genetic dissection of assortative mating behavior. PLoS Biology. 2019;17(2):e2005902-21. doi: 10.1371/journal.pbio.2005902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Montgomery SH, Merrill R. Divergence in brain composition during the early stages of ecological specialization in Heliconius butterflies. Journal of Evolutionary Biology. 2017;30(3):571–582. doi: 10.1111/jeb.13027. [DOI] [PubMed] [Google Scholar]
  44. Montgomery SH, Rossi M, McMillan WO, Merrill RM. Neural divergence and hybrid disruption between ecologically isolated Heliconius butterflies. Proceedings of the National Academy of Sciences. 2021;118(6) doi: 10.1073/pnas.2015102118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Nakagawa S, Cuthill IC. Effect size, confidence interval and statistical significance: A practical guide for biologists. Biological Reviews. 2007;82(4):591–605. doi: 10.1111/j.1469-185X.2007.00027.x. [DOI] [PubMed] [Google Scholar]
  46. Posnien N, Hopfen C, Hilbrant M, Ramos-Womack M, Murat S, Schönauer A, Herbert SL, Nunes MDS, Arif S, Breuker CJ, Schlötterer C, et al. Evolution of Eye Morphology and Rhodopsin Expression in the Drosophila melanogaster Species Subgroup. PLoS ONE. 2012;7(5):e37346. doi: 10.1371/journal.pone.0037346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Pujol B, Wilson AJ, Ross RIC, Pannell JR. Are QST—FST comparisons for natural populations meaningful? Molecular Ecology. 2008;17(22):4782–4785. doi: 10.1111/j.1365-294X.2008.03958.x. [DOI] [PubMed] [Google Scholar]
  48. Rosenthal GG. Evaluation and hedonic value in mate choice. Current Zoology. 2018;64(4):485–492. doi: 10.1093/cz/zoy054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rutowski RL. Variation of eye size in butterflies: Inter-and intraspecific patterns. Journal of Zoology. 2000;252(2):187–195. [Google Scholar]
  50. Schär S. Package allomr: Removing Allometric Effects of Body Size in Morphological Analysis. 2023 doi: 10.1006/jtbi.2000.2043. [DOI] [PubMed] [Google Scholar]
  51. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, Schmid B, Tinevez J-Y, et al. Fiji: An open-source platform for biological-image analysis. Nature Methods. 2012;9(7) doi: 10.1038/nmeth.2019. Article 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Seymoure BM. Heliconius In A New Light: The Effects of Light Environments on Mimetic Coloration, Behavior, and Visual Systems. 2016 [Google Scholar]
  53. Seymoure BM, Mcmillan WO, Rutowski R. Peripheral eye dimensions in Longwing (Heliconius) butterflies vary with body size and sex but not light environment nor mimicry ring. The Journal of Research on the Lepidoptera. 2015;48:83–92. doi: 10.5962/p.266475. [DOI] [Google Scholar]
  54. Stevens M. Sensory ecology, behaviour, and evolution. First edition Oxford University Press; 2013. [Google Scholar]
  55. Stöckl A, Smolka J, O’Carroll D, Warrant E. Resolving the Trade-off Between Visual Sensitivity and Spatial Acuity—Lessons from Hawkmoths. Integrative and Comparative Biology. 2017;57(5):1093–1103. doi: 10.1093/icb/icx058. [DOI] [PubMed] [Google Scholar]
  56. Thorpe R. Quantitative handling of characters useful in snake systematics with particular reference to intraspecific variation in the ringed snake Natrix natrix (L) Biological Journal of the Linnean Society. 1975;7(1):27–43. [Google Scholar]
  57. Veilleux CC, Lewis RJ. Effects of Habitat Light Intensity on Mammalian Eye Shape. The Anatomical Record. 2011;294(5):905–914. doi: 10.1002/ar.21368. [DOI] [PubMed] [Google Scholar]
  58. Wainwright JB, Montgomery SH. Neuroanatomical shifts mirror patterns of ecological divergence in three diverse clades of mimetic butterflies. Evolution. 2022;76(8):1806–1820. doi: 10.1111/evo.14547. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wainwright JB, Schofield C, Conway M, Phillips D, Martin-Silverstone E, Brodrick EA, Cicconardi F, How MJ, Roberts NW, Montgomery SH. Multiple axes of visual system diversity in Ithomiini, an ecologically diverse tribe of mimetic butterflies. Journal of Experimental Biology. 2023;226(24):jeb246423. doi: 10.1242/jeb.246423. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Warrant E. Vision in the dimmest habitats on earth. Journal of Comparative Physiology A. 2004;190:765–789. doi: 10.1007/s00359-004-0546-z. [DOI] [PubMed] [Google Scholar]
  61. Warton DI, Duursma RA, Falster DS, Taskinen S. Smatr 3– an R package for estimation and inference about allometric lines. Methods in Ecology and Evolution. 2012;3(2):257–259. doi: 10.1111/j.2041-210X.2011.00153.x. [DOI] [Google Scholar]
  62. Webster MA. Visual Adaptation. Annual Review of Vision Science. 2015;1(1):547–567. doi: 10.1146/annurev-vision-082114-035509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Wright DS, Manel AN, Guachamin-Rosero M, Chamba-Vaca P, Bacquet CN, Merrill RM. Quantifying visual acuity in Heliconius butterflies. Biology Letters. 2023;19(12):20230476. doi: 10.1098/rsbl.2023.0476. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Analysis
EMS197235-supplement-Supplementary_Analysis.pdf (271.3KB, pdf)
Supplementary Materials
EMS197235-supplement-Supplementary_Materials_.pdf (764.5KB, pdf)
Supplementary Methods
EMS197235-supplement-Supplementary_Methods.pdf (7.4MB, pdf)

RESOURCES