How flies are flirting on the fly

Mate-seeking animals typically rely on sexual communication signals that facilitate mate encounters [1]. While these signals can be diverse (olfactory, visual, acoustic, vibratory, tactile) and may involve multiple sensory modalities [2], certain taxonomic groups use specific primary modalities of communication. Flies (Diptera) have some of the most advanced visual systems in the Insecta [3], often featuring large, sexually dimorphic eyes with specialized “bright zones” [4–6] that may have a functional role during mate-seeking behavior. The fast visual system of flies [7] is considered an adaptation that evolved in support of their advanced flight abilities [3], demanding superb visual acuity to gauge distance traveled and to avoid collisions [3]. Blow flies (Calliphoridae) exploit visual cues when they forage [8], seek oviposition resources [9], or pursue prospective mates [10]. Males have larger eyes than females [6], suggesting that females send and males perceive the visual signals or cues. Occupying vantage points in their territories, males survey rapid fly-bys of females and males, and then decide whether to fend off rival males or pursue prospective female mates.

The design and processing speed of the flies’ compound eyes allow us to infer functional linkage [11] of the visual communication signals that females send and males perceive. Blow flies possess rapid temporal visual discrimination; the flicker fusion threshold (the frequency at which blinking lights are perceived to be constant [12]) of the common green bottle fly Lucilia sericata (Diptera: Calliphoridae) exceeds 180 Hz [13], and may double at temperatures above 30 °C, as shown for the blue bottle fly Calliphora vicina [14], enabling these flies to resolve extremely fast or brief visual stimuli. This ability has been interpreted as an adaptation to support the flies’ advanced flight and collision-avoidance capabilities [3]. If this adaptation were to function also in mate recognition, one would expect an extremely fast and specific visual signal produced by females. We show that L. sericata males distinguish between the rates of light flashes reflected off the moving wings of female and male flies, and are most strongly attracted to flash frequencies characteristic of young females that are prospective mates.

Our search for visual mate recognition cues took into account that mate-seeking males pursue flying females. To test whether wing movement of females affects mate recognition by males, we mounted two live females side by side (Fig. 1a), immobilized the wings of one randomly assigned female and recorded the number of alighting responses by males on or near each female. Significantly more alighting responses by males on or near females that could move their wings than on females that could not (mean ± SE: 34.10 ± 3.76 vs. 20.60 ± 3.79; n = 10, t = –4.43, t crit. two-tail = 2.26, P = 0.002; Fig. 2a; Additional file 1: Data S1) revealed that wing movements by females contribute to mate selection by males.

To visualize optic effects associated with moving wings of L. sericata females, we video-recorded the wing movement of abdomen-mounted flies under direct light at 15,000 frames per second (Fig. 1b), and found that wing movements produce a single, reflected, light flash per wing beat (Fig. 3a–d). Such light flashes (strong light reflections) were not evident (Fig. 3e–h) when we video-recorded wing movements under diffuse light (Fig. 1c), or when we took photographs of L. sericata wings outdoors under a cloudy sky (Fig. 4a). Butterflies and blow flies ([15], this study) exhibit low mating propensity on overcast days when otherwise direct illumination from the sun becomes diffuse [11, 16], and thus reduces flash effects [17]. These observations support the hypothesis that light reflected from the wings produces beacons that contribute to mate recognition.

We tested this hypothesis using two approaches. First, we mounted two live female flies on an aluminum T-bar (Fig. 1d), immobilized their wings, and illuminated each female by a light emitting diode (LED) (Additional file 2: Figure S1a), one that produced light pulses at 190 Hz approximating the wing flash frequency of a flying female, and the other that produced constant light at the same intensity. Second, we isolated the pulsed-light effects from phenotypic traits of female flies by mounting one live female fly and one live male fly side by side (Fig. 1d), immobilizing their wings and illuminating the wings of the male by the 190-Hz light pulses while keeping the wings of the paired female under constant illumination. In both experiments, we placed the T-bar with the two mounted flies into a bioassay cage containing 50 male flies and recorded the numbers of alighting responses on either fly in each pair. In both experiments, the female or male fly exposed to pulsed light (190 Hz) received many more alighting responses (mean ± SE) by males than did the fly illuminated by constant light (Exp. 2: 27.8 ± 4.32 vs. 0.9 ± 0.31; n = 10, t crit. two-tail = 2.26, t = 6.44, P < 0.001; Exp. 3: 18.1 ± 2.98 vs. 0.8 ± 0.29; n = 10, t crit. two-tail = 2.26, t = 5.54, P < 0.001; Figs. 2b, c; Additional file 1: Data S1).

While these results support the hypothesis that light pulses contribute to mate recognition, such signals or cues can be functional in mate recognition only if they differ in frequency by sex and age of free-flying individuals. Only then would a male fly be able to distinguish between rival males and prospective mates traversing his territory. We tested this hypothesis by filming young and old male and female flies in free flight, using a Phantom Miro 3 high-speed camera at a rate of 15,325 frames per sec. For each recording event, we placed 50 young or 50 old male or female flies into a wire mesh cage fitted with a cool-white LED. Following each recording, which typically captured 1–6 flies in free flight, we proceeded to the next recording with a new set of flies in another cage. Analyzing the video-recorded data files (e.g.; Additional file 3: Video S1; Additional file 4: Video S2), we found differences in the frequency of light flashes reflected off the wings of free-flying young and old females and young and old males (one-way ANOVA; F _3,46  = 96.22, P < 0.001; Additional file 1: Data S1). Young females had a mean (± SE) flash frequency of 178.72 Hz (±2.86 Hz), which was significantly lower than that of young males (212.0 ± 4.18 Hz), old females (235.08 ± 2.58 Hz), and old males (265.78 ± 4.53 Hz) (Fig. 5a). Because L. sericata males seek young (2- to 4-day-old) females as prospective mates [18], the slower wing flash frequency from young females than from young males, or older individuals of either sex, could be a phenotypic trait of reproductively capable females.

If the lower wing flash frequency of young females is the key signal or cue for mate recognition by mate-seeking males, one would expect males to respond to this frequency even in the absence of live females and to distinguish between flash frequencies that are true mate cues (178 Hz; see above) and those that are not. By eliminating females from the experimental design, we isolated the light flash effect as the test variable. We ran four experiments in parallel and recorded alighting responses by males on paired black acrylic spheres (Fig. 1e) instead of paired mounted female flies. Each sphere in each pair held a white-light LED, one emitting constant light, the other emitting light pulses at a frequency of 290, 250, 178, or 110 Hz. The 250-Hz pulses represent light flashes produced by flying old females and males (see above). We selected 290- and 110-Hz pulses to test the response of males to pulse frequencies that are well above or below the wing flash frequencies produced by flying conspecifics.

Males did not respond to light pulsed at 290 or 110 Hz in greater numbers than to constant light (Fig. 5b). Spheres holding an LED emitting 250-Hz light pulses received twice as many alighting responses by males than did the spheres with an LED emitting constant light (Exp. 6: 37.5 ± 4.97 vs. 19.37 ± 2.54; n = 8, t crit. two-tail = 2.37, t = 4.59, P = 0.003; Fig. 5b). Spheres with an LED emitting the 178-Hz pulses indicative of prospective mates received not only 4.2 times more alighting responses by males than did the paired spheres with an LED emitting constant light (Exp. 7: 70.3 ± 8.93 vs. 15.67 ± 3.08; n = 9, t crit. two-tail = 2.31, t = 7.21, P < 0.001) (Fig. 5b), but they also received significantly more alighting responses than any of the other light-pulsing LEDs (one-way ANOVA; F_3,31 = 13.55, P < 0.001; Additional file 1: Data S1). Together, these results reveal that L. sericata males do not simply prefer pulsed light to a constant light, but also prefer pulsed light that occurs at frequencies corresponding to the wing flash frequency of young females that are prospective mates.

To ascertain that L. sericata wing flashes are clearly visible even in diverse natural habitats and thus can indeed function in mate recognition, we took photographs of L. sericata wings exposed to direct sunlight in outdoor settings, and video-recorded at high speed free flying L. sericata exposed to direct sunlight. Sunlight-reflecting stationary (mounted) wings produced bright visible cues (Fig. 4b–f), appearing 2–3 times brighter than the paired wing positioned at an angle not conducive to reflecting the sunlight (Additional file 2: Figure S1b). In laboratory measurements, the spectral power distribution of light reflected by L. sericata wings closely resembled that of the incident light (Additional file 2: Figure S1c). Moreover, wings of L. sericata females free flying in an outdoor setting produced repeated flashes of light that contrasted well even against a complex background of plant foliage (Additional file 5: Video S3; Additional file 6: Video S4).

Moving wings are thought to mediate long-range detection of potential mates in some butterflies and damselflies [17, 19, 20]. In these insects, light flash effects coupled with other visual effects of moving wings such as iridescence, UV, and polarized light reflections are hypothesized to contribute to mate recognition. In our study, we have decoupled light flash effects from other effects of moving wings, demonstrating that male flies respond to the light flashes per se when they seek prospective mates. By showing that L. sericata males respond to LED light pulses in the absence of females, we provide evidence that the frequency of pulsed light is the key mate recognition cue in L. sericata and that this cue is independent of structural and color characteristics of female wings.

Unlike static signals or cues, flashing signals affect improved visibility [19, 21]. This is evident, for example, in Morpho butterflies whose wings produce light flashes that are reported to be visible from low-flying aircraft [22] and in Heliconius butterflies whose wings produce polarized light flashes that stand out in complex forest habitats [20]. Compared to these butterfly light flashes, the light flashes produced by L. sericata wings are not only highly visible in direct sunlight (Fig. 4; Additional file 5: Video S3; Additional file 6: Video S4), they are also very rich in information content. Divergent flash rates produced by young and old females (Fig. 5a), and the males’ ability to “read” these rates (Fig. 5b), allow for the conveyance of information that enables informed mate assessment. Exploitation of “rate-based” signals or cues may, in fact, be commonplace in the Insecta. For example, rates of flashing light signals produced by bioluminescent fireflies likely convey important information about mate suitability [12, 23].

Sensory perception of light flashes produced by moving wings seems to be facilitated by the functional design, neural circuitry, and processing speed found in the sexually dimorphic compound eyes of several species of flies. For example, males but not females of the blow fly Chrysomya megacephala and the hover fly Eristalis tenax have large ommatidial facets in their dorsal frontal eye region [4, 5] that form a “bright zone” believed to be capable of increased light capture. This bright zone is not linked to enhanced resolution [24], but is deemed to allow males to search for females at low light or from great distances in bright light [24]. We hypothesize that this bright zone may also help males detect the flashing lights of prospective mates. Moreover, the fast photoreceptors of calliphorid flies [25] may not only underlie adaptations of a visual system that has evolved to support advanced flight and collision-avoidance capabilities [3], but also may enable a superior function in mate recognition. The temporal encoding ability of L. sericata males was amply sufficient to distinguish between light flash frequencies of prospective mates (178 Hz) and rival males (250 Hz) (Fig. 5b). While numerical competence is known for mammals [26, 27], amphibians [28], birds [29], fish [25], and some invertebrates such as ants [30], the numerical recognition and signal-processing speed exhibited by L. sericata males seem to top currently known records.

The logical framework offered by the sensory drive theory [11] predicts functional links between signal design and presentation such that signal conspicuousness is maximized relative to background noise or environmental conditions. Our data on L. sericata are in complete agreement with this prediction. Unlike the iridescent light flashes produced by Hypolimnas bolina butterfly wings [17] that are most conspicuous only from a narrow perspective [17], the wing flashes produced by flying L. sericata females are visible beacons (Fig. 4; Additional file 5: Video S3; Additional file 6: Video S4) that are detectable from all directions, allowing a territorial male fly to rapidly notice a female irrespective of her flight trajectory, particularly when he is perching at a vantage point that optimizes contrast between fly flash signals and background. Remarkably, the flash frequency is so informative that it allows the territorial male to distinguish between old and young females, and to pursue primarily young females that are preferred mates. Furthermore, the low mating propensity of L. sericata on overcast days, when diffuse sunlight renders light reflections off wings inconspicuous (Fig. 4a), appears to show that these flies time their sexual communication and mating activities in accordance with environmental conditions that optimize the conspicuousness of their sexual communication signals.

In conclusion, we describe a previously unidentified visual mate recognition system in the common green bottle fly. The system depends upon both the sex- and age-specific frequencies of light flashes reflecting off moving wings, and the ability of male flies to distinguish between the frequency of light flashes produced by rival males and prospective mates. Our findings imply that insect photoreceptors with fast processing speed may not only support agile flight with advanced maneuverability but may also play a supreme role in mate recognition. With emerging evidence that light flash mate cues also occur in other insects (unpublished data), there may be an opportunity for optimizing light traps for capture of specific nuisance insects in urban and industrial settings.