Varying and Unchanging Whiteness on the Wings of Dusk-Active and Shade-Inhabiting Carystoides Escalantei Butterflies

Home , Delias nigrina, Morpho, Morpho cypris, Morpho rhetenor

Varying and unchanging whiteness on the wings of dusk-active and shade-inhabiting Carystoides escalantei butterflies

Dengteng Gea,b, Gaoxiang Wua, Lili Yangc, Hye-Na Kima, Winnie Hallwachsd, John M. Burnse, Daniel H. Janzend,1, and Shu Yanga,1

aDepartment of Materials Science and Engineering, University of Pennsylvania, Philadelphia, PA 19104; bState Key Laboratory for Modification of Chemical Fibers and Polymer Materials, Institute of Functional Materials, Donghua University, Shanghai 201620, People’s Republic of China; cState Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai 201620, People’s Republic of China; dDepartment of Biology, University of Pennsylvania, Philadelphia, PA 19104-6018; and eDepartment of Entomology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20013-7012

Contributed by Daniel H. Janzen, May 23, 2017 (sent for review January 19, 2017; reviewed by May R. Berenbaum, Jun Hyuk Moon, and John Shuey) Whiteness, although frequently apparent on the wings, legs, Natural whiteness is often attributed to the scattering or dif- antennae, or bodies of many species of moths and butterflies, along fusion of light from (sub)micrometer-sized textures including with other colors and shades, has often escaped our attention. Here, ribs, ridges, pores, and others (25, 26) as seen as white spots and we investigate the nanostructure and microstructure of white spots stripes on the wings of Morpho cypris and Morpho rhetenor helena on the wings of Carystoides escalantei, a dusk-active and shade- (18, 27) in the Nymphalidae. Other species, such as Pieris inhabiting Costa Rican rain forest butterfly (Hesperiidae). On both brassicae and Delias nigrina in the Pieridae, have 100- to 500-nm males and females, two types of whiteness occur: angle depen- beads randomly packed in-between ribs and cross-ribs, enhanc- dent (dull or bright) and angle independent, which differ in the ing the whiteness (19). Metallic to silvery whiteness occurs on microstructure, orientation, and associated properties of their some butterfly wings as the result of mixing structure colors scales. Some spots on the male wings are absent from the female reflected from the membrane scale. For example, the thickness EVOLUTION wings. Whether the angle-dependent whiteness is bright or dull between ribs and scale membrane on Argyrophorus argenteus depends on the observation directions. The angle-dependent (Nymphalidae) wings changes across the wing; although multiple scales also show enhanced retro-reflection. We speculate that reflective colors can be seen at the microscopic scale, bright Carystoides the biological functions and evolution of spot pat- metallic white appears at the macroscopic scale (28). The ease terns, scale structures, and their varying whiteness are adapta- with which shining metallic whiteness could be distinguished ’ tions to butterfly s low light habitat and to airflow experienced from dull whiteness and other colors could be important for on the wing base vs. wing tip. communication between butterflies and for the butterfly’s evolution–devolution. butterfly wings | whiteness | angle dependent | retro-reflection | Here, we investigate a dusk-active and shade-inhabiting Costa Area de Conservación Guanacaste Rican hesperiid butterfly, also known as a “skipper butterfly,” Carystoides escalantei (janzen.sas.upenn.edu/) (29), which has utterflies and moths display on their wings a dazzling array of different types of whiteness, including angle dependent (dull vs. Bcolors (1, 2)—from jet black, looking like a hole in the wing, — to bright white like a bicycle reflector in automobile headlights Significance and with every imaginable color chart hue in-between. Not only are the colors there, but they occur in hundreds of thousands of Whiteness, although frequently apparent on the wings, legs, patterns. Wings are flying billboards and exceedingly complex antennae, or bodies of many species of moths and butterflies, answers to the optimization that past and present natural se- has often escaped our attention. Here, we investigate the lection imposes, whether to some aspect of courtship or escape nanostructure and microstructure of white spots on the wings from predators (3, 4), and even at times in regulating body of Carystoides escalantei, a dusk-active and shade-inhabiting temperature or as serendipitous outcomes of selection for Costa Rican rain forest butterfly (Hesperiidae). We identify two texture, toughness, waterproofing, light weight, and air friction types of whiteness: angle dependent and angle independent. – (5 7). To date, most attention has been on the mechanics and We speculate that the biological functions and evolution of chemistry of a particular color of some easily accessible species Carystoides spot patterns, scale structures, and their varying that are particularly attractive to humans. For example, the whiteness are adaptations to the butterfly’s low light habitat bright blue morpho butterflies (8–12) have inspired many phys- and to airflow experienced on the wing base vs. wing tip icists, materials scientists, and engineers to investigate how the during flight. Sex and species differences in the location of color is produced (9, 13–15), and then to mimic the color and its angle-dependent white spots on the wings may function in mechanisms, or the combination of color and water repellency both intraspecific and interspecific communication. (11, 16, 17). Whiteness, although frequently apparent on the wings, legs, or Author contributions: D.G., W.H., D.H.J., and S.Y. designed research; D.G., G.W., L.Y., and bodies of many species of moths and butterflies, along with other H.-N.K. performed research; J.M.B. and D.H.J. identified the butterflies; D.G., G.W., L.Y., – H.-N.K., W.H., D.H.J., and S.Y. analyzed data; and D.G., W.H., J.M.B., D.H.J., and S.Y. wrote colors and shades (18 21), has received relatively little attention the paper. and is easily ignored. Although “white” may often be simply one Reviewers: M.R.B., University of Illinois at Urbana–Champaign; J.H.M., Sogang University; more color among many within a larger complex pattern, there and J.S., The Nature Conservancy. are times when the whiteness itself appears to be a key signal. “ ” The authors declare no conflict of interest. Note that there is no pigment for whiteness. Indeed, structural 1To whom correspondence may be addressed. Email: [email protected] or whiteness is technologically important in systems ranging from [email protected]. power-efficient computer displays, to sensors, to energy-efficient This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. buildings, windows, and vehicles (22–24). 1073/pnas.1701017114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1701017114 PNAS Early Edition | 1of6 Downloaded by guest on September 27, 2021 bright) and angle independent, on its wings and antennae. We (janzen.sas.upenn.edu/) (30). It likely serves three broad func- tested the white spots on all of the species of Carystoides ex- tions: part of courtship under low or limiting light conditions amined here, and none of them is UV reflective. Some of the (poor visibility at dusk or dawn), part of aposematic warning white spots on the male Carystoides wing are not present on the coloration, and when the skipper is perched, some mimetic re- female forewing, suggesting an intersexually different biological semblance to rotting and fungus-attacked inedible foliage. function. We characterize the micromorphology of each white Carystoides escalantei (Fig. 1) (Hesperiidae), a medium-small spot and infer the relationship between the optical properties denizen of Neotropical rain forest, is common in ACG mid- and structures. The wing scales from angle-independent white elevation rain forests. Its caterpillars, feeding on the leaves of spots are laid down and stacked on top of each other on the wing understory palms (janzen.sas.upenn.edu/caterpillars/database. membrane. These scales have high aspect ratio (AR = length/ lasso), have often been found and reared by the ACG Lepi- width) of ∼3–5. The scales on angle-dependent white spots, doptera inventory (31, 32). The adult butterflies, however, are however, stand vertically on the membrane at different scale rarely seen because they fly at dusk and in the deep shade, and angles. It appears that the hierarchically structured contours their colors are dull, drab, and inconspicuous compared with the consisting of ridges, nanoribs, and scale membranes, all con- brilliant colors of many other more sun-loving butterflies. When tribute to the appearance of the whiteness on Carystoides escalantei. a pinned, dried specimen of Carystoides escalantei with its wings The scales in angle-independent white spots have micrometer-sized spread open in a horizontal plane is placed on a flat surface for periodically occurring ridges with micropores in-between, whereas routine perpendicular examination, the large spots on the wing the scales in angle-dependent white spots have vertically tilted look plain white; but when viewed at a low angle, the white spots scales with undulated, periodic ridges, and the ribs are perpen- look brilliant. This in turn prompted reexamination of living dicular to the ridges on both sides of the scale. Whether the butterflies when they perched on plants (Fig. 1 A–C). In the angle-dependent whiteness is bright or dull depends on whether image of a male Carystoides, ventral hindwing spots are bright the scales are observed from the wing base or the wing tip. white in an oblique frontal view (Fig. 1B) but inconspicuous in a Furthermore, the angle-dependent scales show enhanced retro- perpendicular side view (Fig. 1A). The female Carystoides, on the reflection. We postulate that the different kinds of whiteness other hand, has nearly lost the white spots on the underside of its resulting from microstructure and orientation of Carystoides are hindwing (Fig. 1C). adaptations to their low light habitat and interspecific and in- On the midsection of the upper side of the male Carystoides traspecific communication. Whiteness differs according to the wing, there are four white spots, labeled a1, a2, and a3 on the angle at which the wing is seen, which varies with perching vs. forewings and a4 on the hindwing vs. three white spots, a1, a2, flying, and with airflow experienced on the wing base vs. wing tip and a3 on the forewings of the female Carystoides; they are bright as the wing moves up and down. or dull (Fig. 1 D and E and Fig. S1), depending on the viewing angle. Besides the difference in intensity of the whiteness, the Results and Discussion location and quality of whiteness on the wings also vary between White coloration occurs on the wings of thousands of species males and females. The white spots on the wing tips (c1 in Fig. of moths and butterflies including many of those in Area de 1D) and antennae (b1 in Fig. 1 D and E) are angle independent. Conservación Guanacaste (ACG) in northwestern Costa Rica Immediate questions are, Why are there different kinds of white

Fig. 1. A living male (A and B)andfemale (C) Carystoides escalantei, photographed at a perpendicular angle (A and C) and a low angle (B) under the light illumination from the camera lens side. Photographs of dried and pinned specimens of male (D) and female (E) Carystoides escalantei from dorsal and ventral sides under room fluorescent lamp. A and B are 09-SRNP-30081, C is 08-SRNP-33167, D is 08-SRNP-65732, and E is 12-SRNP-31560 [voucher codes are for individual specimens as in janzen.sas. upenn.edu/caterpillars/database.lasso; all four are called Carystoides escalanteiDHJ02 (29) until the taxonomy of the two sibling species is worked out].

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1701017114 Ge et al. Downloaded by guest on September 27, 2021 Fig. 2. (A–D) Photographs of the male (A and C) and female (B and D) Carystoides escalantei forewing on the dorsal side as taken from the body (A and B) and wing tip (C and D). Illumination light source and the camera are on the same side, and the arrows show the illumination direction. (E) Retro-reflectance spectra of spot a1 (male, dorsal) taken from the vertical direction (β = 0°, black line), wing tip side (β = 40°, red line), and wing body side (β = −40°, green line). (Inset) Schematic illustration of the measurement setup.

spots in different places, and for different sexes? What aspects of have a high aspect ratio (AR = length/width) of ∼3–5. The scales the wing scales contribute to the differences in operational on angle-dependent white spots, however, have AR of ∼1 and whiteness? Are the micromechanisms of whiteness the same in stand vertically on the membrane with different scale angles different spots? Does whiteness originate from the same basic (Fig. 3 D and E). kind of wing scale within a species and among species, or does it Closer examination of the scales on angle-independent white have different evolutionary starting points? spots shows micrometer-sized periodic ridges with micropores in- To address some of these questions, we characterize the between (Fig. 3C and Fig. S5B). The scales on the angle- morphology of the Carystoides scales and their corresponding dependent white spots, on the other hand, are undulated like optical properties of different white spots. As seen in Fig. 2 and ripples, consisting of periodic ridges and ribs that are perpen- Fig. S1, the angle-dependent white spots (a1–a3) on the forewing dicular to the ridges on both sides of the scale (Fig. 3F; see EVOLUTION show different reflectance when observed from the light illumi- definitions in Fig. 3G). From the 3D topography profile of the nation side. In the male, when viewed from the wing base on the scale obtained by atomic force microscopy (AFM) (Fig. S6), the dorsal side, a1 and a2 appear bright white, whereas a3 is dull periodicity and height of the ridges are ∼2 μm and 400–500 nm, (Fig. 2A). When viewed from the wing tip, a3 is bright white, respectively, and the periodicity and the height of the ribs are whereas a1 and a2 appear dull (Fig. 2C). For the female, the ∼165 and 50–100 nm, respectively. The thickness of the scale three forewing spots are either bright white or dull white on the membrane is ultrathin, ∼150 nm (Fig. 3F, Inset), and has low AR dorsal side when observed from the wing base and the wing tip, (AR of ∼1, estimated from the optical image in Fig. 3H). Such respectively (Fig. 2 B and D). Here, the retro-reflection intensity of a1 (male, dorsal) was tested using the setup shown in the Inset in Fig. 2E. Retro-reflection is the special type of reflection, where light striking the surface is redirected back to the source of light. As specified in Fig. S2, α is the incident angle, β is the detecting angle, and we set α to be equal to β. Typically, the intensity of the bright white spot (wing tip side, β = −40°, green line in Fig. 2E) is two times higher than that of the dull white spot (wing body side, β = 40°, red line in Fig. 2E). This suggests that retro-reflection intensity differs with varying incident angles. The whiteness of the spots a1–a3 on the ventral side is shown in Fig. S1. All of these spots are bright white when viewed from the wing base. For comparison, another four species of Carystoides—C. hondura, C. Janzen03, C. basoches,andC. Burns01, which were reared from wild-caught caterpillars by the biodiversity inventory of the rain forests of ACG in northwestern CostaRica(janzen.sas.upenn.edu/caterpillars/database.lasso)(29– 32)—were also examined. As seen in Fig. S3, the spot a3 (male, dorsal) always has the opposite reflection direction to a1 and a2. However, there is inconsistency of spots a4 (male, ventral, Fig. S4); whereas some showed bright white from the wing tip, others showed bright white from the wing base. We hypothesize that the difference in locations of the white spots and the resulting optical properties on male and female Carystoides are adaptations to some aspect of courtship and both interspecific and Fig. 3. (A) Optical microscopy image of the scales, and (B) low- and (C)high- intraspecific recognition. resolution SEM images of the angle-independent white spot (c1, male, dorsal) on the forewing tip. (D) Optical microscopy image and (E and F)the Unique Hierarchical Scales at the White Spots. To reveal the mor- corresponding SEM images of angle-dependent spot a1 (male, dorsal). (Inset) phology of the various white spots, we first characterized the The cross-sectional image shows the thickness of scale. (G) Schematic illustrations of the scale’s microstructure on angle-dependent white spots, and scales at these locations macroscopically. For angle-independent the light transmission at perpendicular (i) and horizontal (ii) directions to white spots at wing tips (Fig. 3 A and B) and antennae (Fig. S5), the scale. n1 and n2 are refractive index of chitin and air, respectively. (H and the scales are laid down in stacks on the wing membrane much I) Optical microscopy images of single scale from spot a1 (male, dorsal) when like the ones we typically see on butterfly wings. These scales placed on a glass slide (H) and partially on a PDMS membrane (I).

Ge et al. PNAS Early Edition | 3of6 Downloaded by guest on September 27, 2021 Fig. 4. (A) Top-view SEM images of three typical types of scales from angle-dependent white spots. (B) Side-view SEM images of angle-dependent spots from (i) a1 male, (ii) a1 female, and (iii) a3 male, respectively. (Insets) SEM images of the follicles (marked by the white dashed boxes) in the corresponding scales. The red arrows indicate the joints between the scale and the follicle. (C) Schematic illustrations of the change of scale angles θ along the wing body due to the airflow when flapping up and down.

structure seems to provide a stiff lightweight scale that resists scale tilting direction (from wing base to wing tip). When the gravity and alleviates aerodynamic stresses. wing is flapping down, scales are pushed down because the air- Based on the full-scale observation on different white spots, flow is in the same direction as the tilted scales. As the push of we categorize the standing scales into three types as summarized airflow gradually increases in the direction of airflow (see bold in Fig. 4A, including wavy scales (type A), lightly curled scales arrows in Fig. 4C), the scale angle decreases when the airflow is (type B), and highly curled scales, like cut-out cones (type C) against the scales’ tilting direction but increases when the airflow with different tilting angles. As illustrated in Fig. S2, the tilting follows the scales’ tilting direction. This could explain why the angle of the scales, marked as scale angle θ, indicates the in- scales of a3 (male, dorsal) tilt to the opposite side of the scale tersection angle from the normal direction of wing membrane to holder because there should be stronger airflow near the wing the central axis of the scale. θ is set as positive if the scale tilts base region. However, the scale angles of a4 on the hind wings of toward the wing tip. θ values from angle-dependent white spots male Carystoides escalantei vary from species to species (Table 1), were obtained and summarized in Table 1. Nearly all scales on which is consistent with their optical appearance on the wings the male Carystoides escalantei wings are wavy (type A), whereas (Fig. S4). We believe it may be attributed to the complex airflow those on the female Carystoides escalantei wings are curved (type on the hindwing created by the flapping of both forewing and B and type C). Furthermore, the scale angle decreases from the hindwing. wing tip to the wing base on both male and female butterfly’s Typically, whiteness is the result of scattering by random mi- forewings. On the dorsal side, the scales typically stand vertically crostructures (Fig. 3C) accompanied by no pigment or dye be- on the wing membrane with smaller θ, whereas those on the neath the scales. When the size of the features of the structures ventral side tend to have a larger θ. Interestingly, the scale of a3 on the scales decreases to submicrometer sizes and becomes (male, dorsal) is tilted to a negative angle (e.g., −15° as shown in more regular (Fig. S4 A and B), the scattering is decreased and Fig. 4Biii). In other words, the scale is tilted to the opposite di- whiteness becomes the result of mixing multiple grating colors rection of the scale holder at the root (see Insets and red arrows as appears to be the case in the standing scales with angle- in Fig. 4B), and θ varies from specimen to specimen. dependent white spots. As seen under the optical microscope However, why do they present such different scale angles? The at the microscale, a1 (male, dorsal) appeared as purple head-on tilting angle of follicle and scale root are all ∼45° and scales start from the standing scales (Fig. 3D) but multicolored when viewed to bend at the connection between the root and the main scale. from the scale surface (Fig. 3H). The purple color could be at- Our hypothesis is that the variety of scale angles is the response tributed to the interference of the periodic nanoribs (see sche- of the standing scales to the airflow across the surface of the matics in Fig. 3Gi). Supporting this is the prediction from wing. As seen in Fig. 4C, the scales on the dorsal side could be multibeam interference (see detailed discussion in Supporting pushed up by the airflow when the wing is flapping up because Information). We believe the multicolors to be the combination the airflow direction (from wing tip to wing base) is against the of thin film interference at air/scale membrane, air/(membrane + rib) and air/(membrane + ridge) interfaces of different thick- nesses (see schematics in Fig. 3Gii). This is consistent with the Table 1. Scale angles of three types of scales on angle- prediction shown in Supporting Information (Table S1) and fur- dependent white spots ther supported when we put a thin poly(dimethylsiloxane) Male Female (PDMS) film in contact with the bottom side of the scale and the Location of bright color was lost (see the dark region in Fig. 3I). This could the white Scale be attributed to the decreased refractive index contrast, where spots Scale type Scale angle, θ Scale type angle, θ air is replaced with PDMS, and thus decreasing the overall interference intensity. It is also consistent with the observed angle a1 Dorsal Type A 25° Type B 20° dependence of the white spots. Ventral Type A 52° Type B 40° a2 Dorsal Type A 20° Type B 10° Enhanced Retro-Reflection on Angle-Dependent White Spots. To Ventral Type A 45° Type C 35° study the angle-dependent whiteness at a single scale level, we a3 Dorsal Type A −15° Type B 5° used backscattering, reflection probe (Fig. 2E, Inset), and fiber Ventral Type A 35° Type C 35° spectrometer to characterize the reflection of the white spots. a4 Dorsal Type A −12° to 5° —— Fig. 5A presents the retro-reflection intensity of two typical white Ventral Type A/type B ∼−10° or ∼10° —— spots with θ = −15° (a3, male, dorsal) and 40° (a1, female,

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1701017114 Ge et al. Downloaded by guest on September 27, 2021 angles θ in Table 1, we plotted the distribution of γ as seen in Fig. 5D. For scales with θ > 0°, γ is mainly in the bottom-left region of the α–γ coordinates, but at θ = −15° it shifts to the up-right region (Fig. 5D). This suggests that the tilted scales are retro- reflective and the reflection light only can be observed in the opposite direction of the scale tilting direction. The dotted line α = γ represents the incident light in the normal direction of the scale. The calculated values shown in Fig. 5D are in good agreement with the experimental observation in Fig. 5 A and C.

Speculation on the Biological Functions of the White Spots. Given different kinds of whiteness on Carystoides wings, it is natural to ask how does the Carystoides habitat select for wing scale structure and thus the difference in the appearance of the white coloration. Carystoides escalantei appears dark with some notable white spots that seem to flicker on and off during flight, pre- sumably because of change in the angle of vision to the observer as the wings move up and down. The white spots do not signal Fig. 5. (A) Retro-reflection intensity of two typical angle-dependent white aposematism (bad tastes or venom), but they increase the visi- θ = − θ = spots of a3 (male, dorsal, 15°) and a1 (female, ventral, 40°) at in- bility of the butterfly to some observers. The white spots on cident light of 550 nm. The setup for this test is seen in Fig. 2E, Inset. antennae and wing tips could be used for interspecific or in- (B) Optical microscopy images of scales and (C) retro-reflection intensity at 550 nm of a single scale from a3 (male, dorsal, θ = −15°) at various detecting traspecific visual recognition, and also the white spots on the angles (β) in the reflection mode of the optical microscope. (Inset) The setup antennae may be signaling, when the butterfly is waving its an- for retro-reflection measurement of a single scale. Here, α = β. (D) Distri- tennae. Because the dark/white system has high contrast, it can bution of the specular reflection angle (γ) of scales with various scale angles be well recognized even at a low light intensity and by color-blind (θ) and incident angles (α) under a hemispherical illumination. The relation species, for example, in the late evening (dusk) and early of incident angle (α), detecting angle (β), specular reflection angle (γ), and morning hours. On the contrary, color-sensitive receiving sys- scale angle (θ) is shown in Fig. S2. tems, which are common in butterflies with pigments or struc- EVOLUTION tural colors, do not work well for identification or warning at such a low light environment. Moreover, the dark/white visual ventral), respectively, at the incident light of 550 nm. For scale system is highly efficient, with low energy consumption for de- with θ = 40°, the reflection is the highest at the detecting angle β = − γ β = γ livering and receiving information using a single kind of eye re- 50°, reaching the specular reflection angle ( ), that is, . ceptor. The white patterns on Carystoides probably evolved over For scale with θ = −15°, the reflection also increases when β is γ = β = the long term in the dark rainforest understory (where its approaching the specular reflection angle ( 75°). At 0°, caterpillar hosts, palms, also live). It is striking that the angle- the reflection intensity is high for both scales, which we think is dependent white spots show strong bright reflection at a partic- mainly due to the specular reflection from the wing membrane, ular low angle (seen in Fig. 2). The bright white against the dark whereas for the angle-independent white spot (c1, male, dorsal), wings will constitute a complex blur and on–off–on visual vo- the reflection intensity is low and the distribution is rather broad cabulary among individuals of both sexes when in flight. This fits (Fig. S7A). The retro-reflection intensity of the angle-dependent well with the environment at dusk or dawn with its low light spot a1 at the specular reflection angle is comparable to that of intensity and low sunlight illumination angle. the traditional retro-reflection materials such as the glass Like many startlingly bright colors (e.g., red, yellow, and blue) microbead array (Fig. S7), although the latter has uniform in- as patches and stripes on butterfly wings, the white spots tensity across the angles. Accordingly, the intensity contrast are hidden when a butterfly or moth is resting with its wings θ = (Ihighest/Ilowest)ofa1 ( 40°), 10, is much higher than that of c1, “folded,” but displayed when in flight. Their function is to hold 2.64, whereas the intensity contrast of the glass microbead array the pursuit image of a predator and yet abruptly “disappear” ∼ is 1. The combination of high angle dependence and high in- when the moth or butterfly lands on a generally nondescript tensity contrast leads to the brilliant white appearance of white background and folds its wings, thereby leading the predator to spots at a1–a4. think that the insect has “left” the scene. The white spots on the We then investigated the relationship of the reflection in- wings of Carystoides may be functioning in this role, as well as in tensity and tilting angles without the optical influence from the courtship. wing membrane. To do so, we used an optical microscope cou- Another apparent feature of the white spots in Carystoides pled with a fiber spectrometer to obtain optical microscopy im- escalantei is that male and female butterflies have different ages (Fig. 5B) and reflection intensity of several scales at numbers and locations of the white spots, and thus different different β (Fig. 5C). On spot a3 (male, dorsal) with a scale angle optical properties. For example, the white spot on the underside θ = −15°, when tilting the wing to increase the detecting angle of the hindwing is conspicuously present on the male Carystoides (here α = β), the scale color changed from purple (at scale edge) escalantei but appears tiny and “residual” on the female hindwing to bright multicolors (at the scale surface), similar to that seen in underside (Fig. 1E vs. Fig. 1D). It is either part of his courtship Fig. 3 D and H. display or allows the female to recognize him as a male, or both, To better summarize the shiny directions of white spots with when males fly or perch on a twig. Both sexes displayed con- different scale angles, here we introduced the specular reflection spicuous white spots on their forewings, but the location and angle γ, which is the intersection angle from the specular re- sequence are different. The male Carystoides escalantei shows a flection direction of the scale to the normal plane of the wing 2–1 pattern (see a1 to a3 in Fig. 1) from the tip to base, whereas membrane (Fig. S2). Based on the geometric relationship shown the female Carystoides escalantei shows a 1–2 pattern, with an- in Fig. S8, γ = −π + 2θ − α when θ is positive, and γ = π + 2θ − α other small spot near the lower edge of the forewing. Sex rec- when θ is negative. Because light can only be shown and viewed ognition could also come from the different sequence of the from one side of the wing membrane, α and γ are limited to white spots on the forewings, which will generate different blurry −90 ≤ α ≤ 90° and −90° ≤ γ ≤ 90°. Using the measured scale white lines in flight.

Ge et al. PNAS Early Edition | 5of6 Downloaded by guest on September 27, 2021 The structure and optical characterization showed that spot a3 Evolution of the two kinds of white spots, one of which (male, dorsal) had the opposite scale tilting angle and retro- brightens when viewed at certain angles, is probably an adapta- reflection direction (Fig. 2 and Fig. 5D) in comparison with tion to the dusky period of butterfly activity combined with the the spots a1, a2 on male, dorsal wing, or a1, a2, a3 on female, shady habitat of Carystoides. Altogether, the spots may function dorsal wing. As we speculate in Fig. 4, the main possibility to in communicative behavior such as camouflage, species and sex form opposite tilting scales on spot a3 (male, dorsal) was the recognition, competition, and courtship. push force of airflow when Carystoides flies. In the evolution of these traits, Carystoides may recognize spot a3 (male, dorsal) for Materials and Methods courtship and male–male antagonism interactions. Carystoides butterflies, including Carystoides escalantei [now known to be two extremely similar sympatric species, one undescribed (29)], Carystoides Conclusions hondura, Carystoides Janzen03, Carystoides basoches, and Carystoides White spots on the wings of the skipper butterfly Carystoides Burns01, were reared from wild-caught caterpillars by the biodiversity in- escalantei differ intersexually in number and location, and inventory of the rain forests of ACG in northwestern Costa Rica. Morphologies trasexually in the constancy and variability of their whiteness. of the wing scales of C. escalantei were characterized by a field emission Two types of white spots, angle dependent (dull or bright) and high-resolution scanning electron microscope (FEI Strata DB235) operated at angle independent, appear on both male and female wings, but 5.0 kV after sputter coating of a thin layer of gold/palladium (Au/Pd). The scale tilting angles were measured by tilting the sample stage in the scan- differ from each other in the microstructure, orientation, and ning electron microscope. Optical microscopy images were taken by an associated properties of their scales. The scales of angle- Olympus BX 61 in the reflection mode. The surface topography of the scales, independent spots are long, flat against the wing membrane, which were picked up by a 1-mm-thick PDMS film, were imaged by AFM

and stacked atop one another; each has micrometer-sized peri- (Bruker Icon) using a Si3N4 cantilever in the tapping mode. The angle- odic ridges with micropores between them. However, the scales resolved reflection spectra were collected from a USB4000 fiber optical of angle-dependent spots are vertical and variably tilted; and spectrometer (Ocean Optics) with home-built rotating stage, or connected numerous closely set and slightly curved thin ribs are perpen- to an optical microscope (BX 61; Olympus). Fig. S2 illustrates the definition of dicular to periodic ridges formed by overlapping segments. The incident angle α, detecting angle β, and specular reflection angle γ. These whiteness appears to be a mixing of grating colors coming from angles are marked as positive if they move toward the wing tip side. the hierarchical structure and the tilting of the scales. The angle- dependent scales enhance retro-reflection. The bright vs. dull ACKNOWLEDGMENTS. We acknowledge the Penn Singh Center for Nano- technology for access to SEM. We also acknowledge the ACG parataxono- aspect of whiteness depends on whether the more or less verti- mists (32) for finding and rearing the Carystoides caterpillars that produced cally oriented scales are observed from the direction of the wing the adults studied here. The work was in part supported by NSF/INSPIRE base or the wing tip. Scale structure also relates to differing Grant IOS-1343159 (to S.Y.), Grant DEB-0515699 (to D.H.J.), the Wege Foun- airflow at the wing base vs. wing tip as the wing moves up and dation (D.H.J.), the Guanacaste Dry Forest Conservation Fund (D.H.J.), the Smithsonian Institution (J.M.B.), and the Fundamental Research Funds for down during flight. This analysis may catalyze the development the Central Universities, Natural Science Foundation of China Grant of materials mimicking structural whiteness for use in sensing 11604045, Shanghai Pujiang Program Grant 16PJ1400100, and Natural Sci- and in displays with low energy consumption. ence Foundation of Shanghai Grant 17ZR1440000 (to D.G.).

1. Parchem RJ, Perry MW, Patel NH (2007) Patterns on the insect wing. Curr Opin Genet 18. Yoshioka S, Kinoshita S (2006) Structural or pigmentary? Origin of the distinctive Dev 17:300–308. white stripe on the blue wing of a Morpho butterfly. Proc R Soc Lond B Biol Sci 273: 2. Vukusic P, Sambles JR (2003) Photonic structures in biology. Nature 424:852–855. 129–134. 3. Weatherbee SD, et al. (1999) Ultrabithorax function in butterfly wings and the evo- 19. Stavenga DG, Stowe S, Siebke K, Zeil J, Arikawa K (2004) Butterfly wing colours: Scale – lution of insect wing patterns. Curr Biol 9:109–115. beads make white pierid wings brighter. Proc Biol Sci 271:1577 1584. 4. Shevtsova E, Hansson C, Janzen DH, Kjærandsen J (2011) Stable structural color pat- 20. Reed RD, Nagy LM (2005) Evolutionary redeployment of a biosynthetic module: Ex- terns displayed on transparent insect wings. Proc Natl Acad Sci USA 108:668–673. pression of eye pigment genes vermilion, cinnabar, and white in butterfly wing de- – 5. Wasserthal LT (1975) The role of butterfly wings in regulation of body temperature. velopment. Evol Dev 7:301 311. 21. Habel JC, Lens L, Rödder D, Schmitt T (2011) From Africa to Europe and back: Refugia J Insect Physiol 21:1921–1930. and range shifts cause high genetic differentiation in the Marbled White butterfly 6. Chai P, Srygley RB (1990) Predation and the flight, morphology, and temperature of Melanargia galathea. BMC Evol Biol 11:215. neotropical rain-forest butterflies. Am Nat 135:748–765. 22. García PD, Sapienza R, López C (2010) Photonic glasses: A step beyond white paint. 7. Kingsolver JG, Koehl MAR (1985) Aerodynamics, thermoregulation, and the evolution Adv Mater 22:12–19. – of insect wings: Differential scaling and evolutionary change. Evolution 39:488 504. 23. Kuo HP, Chuang MY, Lin CC (2009) Design correlations for the optical performance of 8. Vukusic P, Sambles JR, Lawrence CR, Wootton RJ (1999) Quantified interference and the particle-diffusing bottom diffusers in the LCD backlight unit. Powder Technol 192: – diffraction in single Morpho butterfly scales. Proc R Soc Lond B Biol Sci 266:1403 1411. 116–121. 9. Kinoshita S, Yoshioka S, Kawagoe K (2002) Mechanisms of structural colour in the 24. Harrison AW, Walton MR (1978) Radiative cooling of TiO2 white paint. Sol Energy 20: Morpho butterfly: Cooperation of regularity and irregularity in an iridescent scale. 185–188. Proc R Soc Lond B Biol Sci 269:1417–1421. 25. Yu K, Fan T, Lou S, Zhang D (2013) Biomimetic optical materials: Integration of na- 10. Prum RO, Quinn T, Torres RH (2006) Anatomically diverse butterfly scales all produce ture’s design for manipulation of light. Prog Mater Sci 58:825–873. structural colours by coherent scattering. J Exp Biol 209:748–765. 26. Burresi M, et al. (2014) Bright-white beetle scales optimise multiple scattering of light. 11. Kinoshita S, Yoshioka S, Miyazaki J (2008) Physics of structural colors. Rep Prog Phys Sci Rep 4:6075. 71:076401. 27. Yoshioka S, Kinoshita S (2006) Single-scale spectroscopy of structurally colored but- 12. Sato O, Kubo S, Gu ZZ (2009) Structural color films with lotus effects, superhydrophilicity, terflies: Measurements of quantified reflectance and transmittance. J Opt Soc Am A – and tunable stop-bands. Acc Chem Res 42:1–10. Opt Image Sci Vis 23:134 141. 13. Potyrailo RA, et al. (2013) Discovery of the surface polarity gradient on iridescent 28. Vukusic P, Kelly R, Hooper I (2009) A biological sub-micron thickness optical broad- band reflector characterized using both light and microwaves. J R Soc Interface 6: Morpho butterfly scales reveals a mechanism of their selective vapor response. Proc S193–S201. Natl Acad Sci USA 110:15567–15572. 29. Janzen DH, et al. (2011) Reading the complex skipper butterfly fauna of one tropical 14. Mason CW (1926) Structural colors in insects. II. J Phys Chem 31:321–354. place. PLoS One 6:e19874. 15. Pillai PKC (1968) Spectral reflection characteristics of Morpho butterfly wing. J Opt 30. Chen PC, Xu ZK (2013) Mineral-coated polymer membranes with superhydrophilicity – Soc Am 58:1019 1022. and underwater superoleophobicity for effective oil/water separation. Sci Rep 3:2776. 16. Li J, Liang G, Zhu X, Yang S (2012) Exploiting nanoroughness on holographically 31. Janzen DH, et al. (2009) Integration of DNA barcoding into an ongoing inventory of – patterned three-dimensional photonic crystals. Adv Funct Mater 22:2980 2986. complex tropical biodiversity. Mol Ecol Resour 9:1–26. 17. England G, et al. (2014) Bioinspired micrograting arrays mimicking the reverse color 32. Janzen DH, Hallwachs W (2011) Joining inventory by parataxonomists with DNA diffraction elements evolved by the butterfly Pierella luna. Proc Natl Acad Sci USA barcoding of a large complex tropical conserved wildland in northwestern Costa Rica. 111:15630–15634. PLoS One 6:e18123.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1701017114 Ge et al. Downloaded by guest on September 27, 2021