Wing interference patterns (WIP) are stable structural colors in insect wings caused by thin-film interference. This study seeks to establish WIP as a stable, sexually dimorphic, species-level character across the four families of Tipuloidea and investigate generic level WIP. Thirteen species of Tipuloidea were selected from museum specimens in the Academy of Natural Sciences of Drexel University collection. One wing from a male and female of each representative species was excised and mounted to a slide with coverslip, placed against a black background, and imaged using an integrated microscope camera. Images were minimally retouched but otherwise unchanged. Descriptions of the WIP for each sex of each species are provided. Twelve of thirteen species imaged had WIP, which were stable and species specific while eight of those twelve had sexually dimorphic WIP. Comparisons of three species of Nephrotoma were inconclusive regarding a generic level WIP. Gnophomyia tristissima had higher intraspecific variation than other species examined. This study confirms stable, species specific WIP in all four families of crane flies for the first time. More research must be done regarding generic-level stability of WIP in crane flies as well as the role sexual and natural selection play in the evolution of wing interference patterns in insects.

Cryptic, Cylindrotomidae, dimorphism, Limoniidae, morphology, Pediciidae, Tipulidae, WIP

Wing interference patterns (WIP) were historically known but only recently revealed to be a cryptic physical character of insect wings first reported by Shevtsova et al. (2011). WIP colors and patterns are stable regardless of the angle viewed, so are not iridescent (Shevtsova et al. 2011), but they exist thanks to the same physical property of light that generates iridescence in soap bubbles and oil slicks: thin film interference (Sun et al. 2013). Thin film interference occurs when light enters one of two parallel, thin, non-absorbing (i.e., light) layers, or films, and then bounces between the upper and lower films until exiting through one film (Sun et al. 2013). Thin film interference is most common in the clear wings of insects, with the two thin, chitinous layers forming the parallel thin films (Sun et al. 2013) and 20% of light passing into the wing being reflected between the planes, and exiting through the upper layer (Shevtsova et al. 2011). When the light exits the upper plane of the wing, the thickness of the chitin, as well as the microtopography of the wing at the point of exit will dictate the color observed (Shevtsova 2012). These patterns are created from stable structural properties of the wing, but our ability to observe them may be obscured against white or light-colored backgrounds or enhanced with a black background and bright light perpendicular to the plane of the wing (Shevtsova et al. 2011).

Research suggests that WIP are a novel morphological character that may be sexually selected for in some groups. Cryptic species have been discovered using WIP in two genera of wasps in the family Eulophidae (Shevtsova and Hansson 2011; Hansson and Hambäck 2013). Several works have used WIP as a character in describing a novel species (Buffington 2012; Mitroiu 2013) and as a character in a dichotomous key (Mitroiu 2013; Zhang et al. 2014). Sexual dimorphism of WIP has been documented in many groups of wasps and flies (Shevtsova et al. 2011; Shevtsova and Hansson 2011; Shevtsova 2012) and sexual selection of male WIP by female Drosophila melanogaster Meigen, 1830 has been observed (Katayama et al. 2014; Hawkes et al. 2019).

Most research on WIP in insects has focused on small Hymenoptera and Diptera with clear wings and reduced venation; the result is a continuous pattern across the wing. The existence of WIP in these groups has been well documented in the Cynipoidea (Buffington and Sandler 2011), Eulophidae (Hansson and Shevtsova 2010, 2012) and Drosophilidae (Shevtsova et al. 2011; Katayama et al. 2014; Hawkes et al. 2019). A few studies have looked at larger wasps (Shevtsova et al. 2011; Kangasniemi 2012) or larger flies (Shevtsova et al. 2011; Zhang et al. 2014) or other orders of insects like Odonata (Brydegaard et al. 2018) and Hemiptera (Simon 2013).

Crane flies (Tipuloidea) are one of the largest groups in Diptera with 15,632 recognized species and a global distribution (Oosterbroek 2021). Crane flies vary in wing length from 3 to 36 mm and occupy a diverse set of habitats (Gelhaus and Podenine 2019) and act as a major food source in aquatic and riparian ecosystems (Baxter et al. 2005; Gelhaus and Podeniene 2019). Furthermore, there is evidence of cryptic species within the superfamily (Ujvárosi et al. 2010; Salmela et al. 2014) and a lack of monophyly within the families (Petersen et al. 2010). Tipuloidea is represented in the WIP literature from a single male specimen of Tipula (Savtshenkia) confusa van der Wulp, 1883 presented in Shevtsova et al. (2011). Before that, though, the patterns were mentioned without additional emphasis in at least a few species descriptions of crane flies as early as a century ago (Enderlein 1912). Given the size, diversity, and unresolved phylogeny below the family level, establishment of a novel character such as WIP is essential to our understanding of the group.

Crane flies have relatively large wings with multiple branches of major veins that form upwards of fifteen cells. Additionally, crane flies can have pigment, setae, folds, and reinforcements of the wing surface and veins; all of which play a role in the overall visibility of the WIP (Shevtsova 2012). Crane flies are poorly represented in the current WIP literature; many of the other taxa examined have wings that are small, clear, and with reduced venation. The largest known species of crane fly in North America is Holorusia hespera Arnaud & Byers, 1990 (Alexander 1920, as rubiginosa) which has a wing length of 40 mm while the smallest species has a wing length of 2.0 mm (de Jong et al. 2007). We suspect given this wide range of wing lengths that WIP will be vastly different among the species of Tipuloidea. Additionally, Shevtsova et al. (2011) showed that in Hymenoptera and Diptera, WIP follow the Newton series: a repeating stable series of color bands known to occur in thin films of between 50 and 1500 nm thickness (Shevtsova et al. 2011). We did not measure wing thickness or map the Newton sequence on our WIP images, but we suspect some of the largest species of the group will have wings too thick to transmit a WIP. If so, we expected to observe opaque, gray wings as described in Shevtsova et al. (2011).

We aim to establish the existence of Wing Interference Patterns across the four families of Tipuloidea using male/female representative pairs and to provide descriptions of the color and pattern of WIP for both sexes for each species. We also seek to establish the stability of WIP within a species and confirm sexual dimorphism in each representative species.

All specimens were selected from the entomology collection at the Academy of Natural Sciences of Drexel University (ANSP) in Philadelphia, Pennsylvania. We examined 45 species (2373 individuals) (Table 1) and selected representative species for each family, with additional species sampled to evaluate generic level WIP relationships (Nephrotoma spp.) and to investigate intraspecific variation (Gnophomyia tristissima Osten Sacken, 1860) (Table 2). In each representative species both a male and female specimen were selected for comparison. Additionally, both pinned specimens and ethanol-preserved specimens of the same sex and species were compared to determine if preservation had an effect on the stability of WIP. No deviations were seen in WIP color or pattern between same sex specimens of the same species and as such specimens from both preservation methods were used in this study. Preservation type of each specimen used in this study is provided (Table 2). Herein we follow the four-family taxonomy of Oosterbroek (2021). We selected taxa based on availability in the collection and the condition of the wings, as the wings of many pinned specimens were exceedingly brittle and often curled or folded, rendering them unusable for imaging. We specifically chose one species, Holorusia hespera, based on size, as we expected this species to have wings too thick to display a WIP. Within each sex for each species, we observed wings across age (1904–2019), geography, and preservation type to confirm stability of the WIP within a species. We found no evidence that any of these metrics had an effect on the stability of WIP. Through this investigation we discovered higher intraspecific variation in G. tristissima and chose to image additional specimens for this species (Fig. 1).

Figure 1. 

Comparison of the variation in WIP of three female and three male specimens of Gnophomyia tristissima. Females examined in this study were found to have a range of WIP from A dark blue/ purple B blue with mottled yellow C green/yellow with hints of blue which appeared most like the male WIP. Males examined also had a range of WIP from D green with mottled blue which appeared most like the female WIP E solidly green F green with mottled magenta. Patterns B and E were the most encountered patterns for females and males, respectively. Scale bars: 1.0 mm.

Specimens were prepared as in Shevtsova et al. (2011) with certain exceptions. We excised one wing from each specimen, and placed each wing on a slide. Next a drop of 70% ethanol was added to help flatten and orient the wing on the slide and while still wet from the ethanol, a cover slip was carefully placed on top of the wing and slide. We applied gentle pressure to the cover slip to remove air bubbles. Once the ethanol had evaporated the cover slip was adhered to the slide by placing a small drop of Euparal at each corner, making sure that none of the Euparal had seeped onto the wing. Initially we had tested adhesives and fixatives placed over the entire cover slip, but this blocked transmission of WIP.

A black background was created using light-absorbing black-out fabric with adhesive backing from Edmund Optics (Item #54–585). This fabric was placed beneath each wing slide prior to imaging. Imaging for all but three species was performed using a Leica S9i stereomicroscope (Model AF6000). Images for three species (Tipula (Yamatotipula) sayi Alexander, 1911, Dolichopeza (Oropeza) obscura (Johnson, 1909), and Nephrotoma macrocera (Say, 1823)) were too distorted to be used and were reshot using a Leica (Model EZ4 D) stereomicroscope with an integrated 3mp camera. Each specimen was imaged using LEICA APPLICATION SUITE X (version 3.0.11.20652). Files were saved in .tiff format and edited in ADOBE PHOTOSHOP (version CS6). Alterations in Photoshop were restricted to increasing the saturation by no more than 10%, reducing brightness by up to 20%, increasing the contrast by up to 20%, cropping the wing from the background, darkening the background, and using the spot-healing tool to remove dust and debris as needed. We follow the four-pattern concept of WIP put forth in Buffington and Sandler (2011): campiform (WIP of mostly one color, usually blue), galactiform (mottled patches and swirls of color like the spirals of galaxies), radiform (radial bands emerging from the medial sector and expanding in concentric bands to the margin), and striatiform (longitudinal bands of color that often follow anal veins). Wing definitions follow those of Saigusa (2006) (Fig. 2).

Figure 2. 

Excised wing of a male specimen of Dolichopeza obscura against a white background with notations of wing veins and cells used in this study. Veins are noted in blue with uppercase letters while cells are noted in red with lowercase letters; naming and notations follow those of Saigusa (2006). Abbreviations: A/a: anal vein/cell, bm: basal medial cell, br: basal radial cell, C/c: costal vein/cell, CuA/cua: anterior cubitus vein/cell, CuP/cup: posterior cubitus vein/cell, d: discal cell, M/m: Medial vein/cell, R/r: radial vein/cell, Rs: radial sector vein, Sc/sc: subcostal vein/cell. Image not to scale.

Descriptions of WIP herein are an attempt to provide a written account of the WIP across the entire wing. This was difficult to capture in a single image, mostly due to the size of many wings as well as the folds and textured surfaces of the wings of many crane flies examined. In the larger species, the veins themselves were thick enough to keep the wing from lying flat on the slide. Rather than making composite images we attempted to provide as complete a WIP as possible in a single image. As such, cells adjacent to the costal margin (especially c, sc₁, and sc₂ that in many taxa are exceedingly narrow) often have WIP that are distorted by corrugation, folds, and pigment. We have still attempted to provide those details in the descriptions and have noted in parentheses behind the given characters when they are not visible in the corresponding figure though WIP should be visible when observed in person.

We confirm stable, structural Wing Interference Patterns (WIP) in the four families of Tipuloidea. Despite small deviations the overall patterns were stable within each sex and/ or species and were not affected by age of specimen, location collected, or method of preservation. Twelve of the thirteen species sampled had a distinct WIP across the entire wing surface. The sole exception was Holorusia hespera, which lacked a WIP, likely due to the size and thickness of the wing as we predicted (Fig. 6C, D). This agrees with the findings of Shevtsova et al. (2011) who noted that the WIP of flies follow the Newton sequence, and wings thicker than 1500 nm will lack a WIP and appear opaque gray. We found eight of the twelve species with a WIP were sexually dimorphic in color if not pattern; Cylindrotoma distinctissima (Fig. 3A, B), Dicranomyia liberta (Fig. 4C, D), Dolichopeza obscura (Fig. 5C, D), and Tipula borealis (Fig. 8A, B) lacked sexually dimorphic WIP.

We have demonstrated that WIP are stable within each sex and each species. In ten of twelve species there were minimal variations in WIP consistent with phenotypic variation. Because WIP color is determined by the nanometer-level thickness of the chiton layer, one could expect that among individuals of the same sex there would be some degree of variation in wing thickness. Indeed, these results support the findings of Shevtsova et al. (2011) and Shevtsova and Hansson (2011) who both noted patterns are more stable than color or hue of WIP. We also provide the first cell-by cell descriptions of WIP as a diagnostic character. One species, Gnophomyia tristissima, showed increased variation in color and pattern relative to the other species we examined. Both males and females showed this variation and there appeared to be a gradient of WIP color. Additionally, some males and females showed an almost inverted WIP. More work is needed to understand if this is simply a gradient of wing thickness or if there is a selection force acting on the WIP in this species.

Sexual dimorphism of WIP in crane flies is clear and common in the species we examined. While there is no documented evidence of sexual selection of WIP in Tipuloidea, female choice of WIP has been demonstrated in various species of Drosophila (Ala-Honkola and Manier 2016; Hawkes et al. 2019) including evidence that females use visual cues in choosing mates and prefer winged males (Watanabe et al. 2018). Further, Katayama et al. (2014) found that females of D. melanogaster preferred wings of mates with a high degree of saturation and a centrally stable hue. Even with our limited sampling of species, many of the sexually dimorphic species in our study had males with bolder, more contrasted colors than females. G. tristissima is a widespread species (Alexander 1920) and as such we may be seeing the results of female choice of WIP or a geographic effect. A larger study of this species should be done to try and gauge the level of variation as well as the existence of sexual selection for WIP in G. tristissima. Butterworth et al. (2021) performed quantitative, viewer independent methods to evaluate WIP in Calliphoridae, a method that is well suited for examining the variation in G. tristissima as well as further WIP studies in Tipuloidea.

Our work suggests that WIP are a stable, reliable species level trait in Tipuloidea and this agrees with the recent WIP literature (Shevtsova et al. 2011; Shevtsova and Hansson 2011; Zhang et al. 2014; and Butterworth et al. 2021). In Tipuloidea, traditional identification of species is often based on male genitalic features, with females in some groups being difficult to identify to species, and even identification keys to the species level are not available for the female stage for many genera. The WIP may provide a useful set of features for separating very similar species in the female stage, e.g., in Tipula (Beringotipula) separation of the over 20 species in North America at present is based on male genitalic features only.

Additionally, we did not see evidence of a generic level pattern among the three species of Nephrotoma examined, although all three species generally had green/magenta striated wings. We are aware of two studies to examine generic level WIP (Buffington and Sandler 2011; Shevtsova and Hansson 2011) both of which found some suggestion of a generic level WIP in certain taxa. We sampled only three of 484 species in the genus Nephrotoma. Given the size of the genus, a more focused study examining a larger number of species within the genus, examining the WIP of Nephrotoma in a phylogenetic context, and use of viewer independent testing would help increase our understanding of WIP at the generic level.

We note that pigmented patterning on the wing appears to reduce the extent of WIP, at least based on our limited survey here. Species with either costal darkening (Tipula sayi, Tricyphona inconstans) or more extensive marmorated pattern (Tipula borealis, Dactylolabis cubitalis), had reduced WIP, at least in the region of the pigment and the pterostigma. Freeman (1968) suggested that marmorated and spotted wing patterns in crane flies are associated with woodland habitats, while clear or striped wings with open habitats; these two different light exposures might impact WIP visibility. Within these broad wing pattern categories, though, there can be a range of wing orientation when at rest which can also impact WIP visibility. For example, Tipula borealis holds the wings outstretched and at a slight angle (Fig. 9A, B), while patterned wing Dactylolabis (Dactylolabis) montana (Osten Sacken, 1860) and Dicranomyia (Dicranomyia) simulans simulans (Walker, 1848) (similar to Dactylolabis cubitalis) rest with the wings closed over the abdomen (Fig. 9C, D; Adler and Adler 1991).

Figure 9. 

Images showing WIP on several species of crane fly in nature A male Tipula (Yamatotipula) aprilina Alexander, 1918 displaying WIP in nature B female Tipula (Yamatotipula) aprilina displaying WIP in nature C pair of Gnophomyia tristissima perched on a leaf in copula. Both flies are displaying their sexually dimorphic WIP. The female (bottom) has a blue WIP while the male (top) displays a green WIP D an individual of Elliptera clausa Osten Sacken, 1877 displaying a WIP with wings folded. Sex unknown. Copyright (A, B) 2021, photograph JK Gelhaus; (C) 2020, photograph Katja Schulz, used with permission by the artist and under a creative commons license (https://creativecommons.org/licenses/by/4.0/) with alterations limited to cropping and resizing of this image; (D) 2016, photograph JK Gelhaus. Images are not to scale.

Wing interference patterns are stable and they are readily visible in nature (Fig. 9A–D), but their visibility depends on the angle at which light hits the wing as well as the angle it is viewed (Shevtsova 2012). A good example is the male wing of Dolichopeza obscura used in this study. When the wing of the male specimen of D. obscura is displayed on a white background with light from many different angles it appears clear or slightly stained (Fig. 2) while the same wing, when placed on a dark background with light parallel to the wing displays a bold and intricate WIP (Fig. 5D). Using a pinned museum specimen, we have demonstrated how quickly the WIP transmission can change with a change in background color (see Suppl. material 1: Movie S1). Given this, it is possible that crane flies are using WIP as a type of dynamic flash coloration (Murali 2018) to avoid predation. Additionally, erratic flight patterns have been found to increase the effectiveness of dynamic flash coloration to avoid predation (Murali and Kodandaramaiah 2020) and Pritchard (1983) notes both the wing pattern and the flight pattern of crane flies act to obfuscate the flies from predators. This is merely an observation, and we recommend a true behavioral study to understand what role, if any, WIP plays in predator avoidance strategies in Tipuloidea.

We do remain curious if the WIP in one individual is recognized by other conspecific crane flies. Although WIP were selected for in Drosophila (Katayama et al. 2014), in most cases crane flies do not exhibit complicated male-female pre-copulatory behavior (Pritchard 1983), and in many crane flies males seem to find females by touching legs during male search of vegetation (some Limoniidae, Tipulidae Stich 1963; Pritchard 1983) or in mating swarms (some Limoniidae, Pediciidae, Alexander 1920; Pritchard 1983). Also, crane flies males search for females or swarm usually during crepuscular periods (Sullivan 1981) or at least early morning and late afternoon (Gelhaus, pers. obs.) when lighting would not be expected to highlight WIPs. Males could potentially recognize a species-specific WIP at close range after initial leg contact, though.

The scope of this study was to establish the existence of WIP in the four families of Tipuloidea and confirm that WIP could exhibit sexual dimorphism. We have confirmed stable, structural WIP in male/female pairs of twelve species of crane fly across the four families of Tipuloidea. Of these, eight species displayed sexually dimorphic WIP between male and female specimens. One species showed high intraspecific variation that may be a result of sexual selection, though more research is required. Our work supports the consensus in the literature that WIP are species-specific. This work provides the basis for further research and documentation of WIP in crane flies. We did not compare subspecies in this study and comparisons at the generic level were inconclusive, though we cannot discount a generic level WIP relationship. We believe WIP could be a useful tool to discern cryptic species in crane flies or as a novel character to identify females that cannot be separated based on the current morphology. Additionally, WIP may be used for predator avoidance by crane flies.

We would like the thank the Malacology Department and Entomology Department of The Academy of Natural Sciences of Philadelphia at Drexel University (ANSP) for use of their imaging equipment and especially Paul Calomon, Dr. Daniel Otte, Isabelle Betancourt, and Joseph Sweeney for training on imaging equipment and software. We would like to also thank Dr. Sigitas Podenas, Vilnius University, Lithuania for early help on imaging of the wings. Thank you to the Richard Engle and Gelhaus lab members (Bolortsetseg Erdenee, Larry Henderson, Steve Mason, Joseph Sweeney) at Drexel University for suggesting edits to an earlier version of this manuscript. Finally, many thanks to Dr. Greg Setliff of Kutztown University of Pennsylvania for first putting the WIP bug in our ears many years ago.

The Department of Biodiversity, Earth & Environmental Sciences of Drexel University and its travel grant, the Teck-Kah Lim travel award, and The William L. McLean III Fellowship for the study of environmental sciences and ornithology all contributed funds used to complete and/or present this research.