Experimental insects
We reared L. sericata in the insectary at Simon Fraser University, starting a new colony with field-collected wild flies every 12 months. We cold-sedated flies within 24 h following eclosion, separated them by sex, and kept them in groups of 50 males or 50 females in separate wire mesh cages (45 × 45 × 45 cm; BioQuip®, Compton, CA, USA) under a L16:D8 photoperiod, 30–40% relative humidity, and 23–25 °C. We provisioned flies with water, milk powder, sugar, and liver ad libitum and used 2- to 7-day-old flies in bioassays.
Responses by males to mounted females, one able to wing-fan, the other with wings glued to her body
For each replicate of Experiment 1 (n = 10), we CO2-sedated two live female flies for 30 s, and then mounted them with cyanoacrylate adhesive on their abdominal ventrum, at opposite ends of a 7.5-cm-long aluminum T-bar (Fig. 1a). We applied a small amount of cyanoacrylate to the wing base of one randomly assigned female to immobilize her wings, and applied the same amount of adhesive to the abdomen of the other female, allowing her wings to move freely. We placed the T-bar with the two females in a wire mesh bioassay cage (45 × 45 × 45 cm; BioQuip®) containing 50 male flies. The cage was illuminated from above with a full spectrum light source (two horizontal fluorescent bulbs: Philips, plant & aquarium (40 W); Sylvania, Daylight Deluxe (40 W)) (Additional file 2: Figure S1a). To minimize light reflection, we covered the metal cage floor and T-bar stand with SunWorks® black construction paper and black velvet (Suzhou Joytex International Co. Ltd., Jiangsu, China), respectively. During each 40-min bioassay, we recorded the number of alighting events by males on or near a female followed by physical contact with her. We analyzed the mean numbers of alighting responses by males on females with wings either mobile or immobilized by a paired two sample for means t-test.
Do moving wings produce flashes of reflected light under point source illumination?
We recorded the wing movement of abdomen-mounted male and female flies (Fig. 1b) in slow motion using a Phantom Miro 4 camera (Vision Research, Wayne, NJ, USA), recording at 15,000 frames per second, a 512 × 512 pixel resolution, and a 20-μs exposure time. To illuminate the mounted fly, we used a white 100-watt LED (6500 K; Zongshan Ltd., Guangdong, China) mounted to a computer CPU heat sink for cooling (Thermaltake Heatpipe, Thermaltake Technology Co. Ltd, Taipei, Taiwan), and powered via a 32 V 5A stabilized, adjustable DC power supply (Gopher Technologies, Yantian, Fenggang, Dongguan, Guangdong, China).
Do moving wings produce flashes of reflected light under diffuse illumination?
We used the same high-speed video technology as described above, except that we exposed the mounted fly to diffuse instead of point source light. We placed the fly inside a ping pong ball “diffuser” (Fig. 1c) and illuminated it by four cool white 100-watt LEDs (see above).
Responses by males to paired mounted females, both with their wings immobilized but one with pulsed light reflecting off her wings
For each replicate (n = 13) of Experiment 2, we mounted two live female flies on an aluminum T-bar (Fig. 1d) and immobilized the wings of each female with cyanoacrylate adhesive. We illuminated one randomly assigned female from above by a light emitting diode (LED, Optek Technology Inc., Carrollton, Texas 75006, USA) (Additional file 2: Figure S1a) mounted 3 cm above the female (Fig. 1d) and which produced 5-Volt, white-light pulses at a frequency of 190 Hz and a duty cycle of 3%. The pulse frequency of 190 Hz was approximately mid-way between the light-flash frequencies of flying 2-day-old female and male flies. We illuminated the control female by a second LED of the same type that produced constant light.
We considered two alternative approaches to the illumination design of this experiment. We could have set the pulsed-light LED and the constant-light LED to deliver either equal total light intensity (root mean squared) or equal maximum light intensity (peak voltage). We chose the latter (conservative) approach because, at a 3% duty cycle (“on” vs. “off” ratio), the pulsed-light LED delivers only about 3% of the total light that the constant-light LED delivers. Thus, to the human eye, the pulsed-light LED appears as a constant dim light, whereas the constant-light LED appears as a constant bright light; to fly photoreceptors, in contrast, the pulsed-light LED appears as an intermittent (pulsing) light with a light intensity matching that of the constant-light LED.
For each replicate, we placed the T-bar with the two females into a wire mesh bioassay cage containing 50 male flies. During 40 min in each replicate, we recorded the numbers of alighting responses by these 50 male flies on each female, and analyzed the mean numbers of alighting responses by a paired two sample for means t-test.
Responses by males to paired male and female flies, both with their wings immobilized, and pulsed light reflecting off the male’s wings
In each replicate of Experiment 3 (n = 10), we mounted one live female fly and one live male fly 7 cm apart on an aluminum T-bar (Fig. 1d), and immobilized the wings of each fly with cyanoacrylate adhesive. We illuminated the male from above by an LED (Fig. 1d) that produced 5-Volt, white-light pulses at a frequency of 190 Hz and a duty cycle of 3%. We illuminated the female by a second LED of the same type that produced constant light at equal maximum intensity as the first LED. During 40 min in each replicate, we recorded the number of alighting responses by 50 males on the mounted male and female fly, analyzing the mean number of alighting responses on the male and female by a t-test.
Light flash frequencies associated with age and sex of flying individuals
The objective of Experiment 4 was to determine whether the numbers of light flashes reflected off the wings of free-flying flies differ in accordance with age or sex. We filmed 2-day-old (young) and 7-day-old (old) male and female flies in free flight using a Phantom Miro 3 high-speed camera (Vision Research) at a rate of 15,325 frames per second and a 34-μs exposure time imaged through a Canon 100-mm f2.8 L macro lens (Canon Canada Inc., Vancouver, BC V6C-3 J1, Canada) fitted to a 36-mm extension tube. For each recording event, we placed 50 young or 50 old male or female flies into a wire mesh cage (45 × 45 × 45 cm) housing a 100-Watt (approximately 9000 Lumen), cool-white (5000–6000 Kelvin) LED, which was driven by a 32-Volt switching-power supply (model CPS-3010, Gopher Technologies, Yantian, Fenggang Town, Dongguan, Guangdong, China). Once the camera and light were turned on, we lightly tapped the cage to induce take-off and flight by resting flies. In video-recorded data files, we counted the number of light flashes reflected within one second off the wings of free-flying young females (n = 11), young males (n = 12), old females (n = 18), and old males (n = 9), and analyzed light-flash frequencies of young and old females and of young and old males by one-way ANOVA followed by the Tukey test for comparisons of means.
Ability of males to discriminate between LED-pulsed light of varying frequencies
To determine whether mate-seeking males can distinguish between different frequencies of pulsed light, parallel-run behavioral Experiments 5–8 (n = 9, 8, 9, and 9, respectively) tested alighting responses by males on paired black acrylic spheres (1.77 cm diameter; supplier unknown; Fig. 1e). We mounted the spheres on clamps 12 cm apart and 12 cm above the floor of the bioassay cage containing 50 male flies. A central hole (0.52 cm) in each sphere accommodated an upward pointing LED (Fig. 1e), the rounded lens of which was sanded down to be flush with the sphere’s surface. Sanding the lens ensured that the emitted light was visible to flies from many viewing angles rather than from the narrow viewing angle that the lens otherwise creates. By random assignment, one LED in each pair emitted constant light; the other emitted light pulsed at 290 Hz (Experiment 5), 250 Hz (Experiment 6), 178 Hz (Experiment 7), or 110 Hz (Experiment 8). We selected the frequencies of 290 Hz and 110 Hz to test the response of males to pulse frequencies that are well above or below the wing flash frequencies produced by flying common green bottle flies. In each of Experiments 5–8, we analyzed the mean numbers of alighting responses by males on paired spheres holding LEDs emitting constant light or pulsing light by a paired two sample for means t-test. We analyzed differences in alighting responses based on the frequency of pulsed light by one-way ANOVA followed by the Tukey test for comparisons of means.
Visibility of light flashes reflected off the wings of free-flying flies video recorded outdoors under direct sunlight
To document the effect of sunlight reflecting off the wings of a free flying L. sericata, we took high-speed video recordings of flies traversing a south-facing slope with plant cover under direct, mid-day sun under a partially cloudy sky. For these recordings, we used a FASTCAM Mini AX200 type 900 K-M camera (Photron USA Inc., San Diego, CA 92126, USA) fitted with a Canon macro lens (100 mm; f2.8 L) at f5.6, capturing images at 15,000 frames per second, a 1/15000 exposure time, and a 768 × 512 pixel resolution.
Effect of natural sunlight reflecting, or not, off L. sericata wings
We carefully removed wings from a 1-day-old female fly, mounted them on hemostatic clamps positioned by articulated holders (Noga Engineering Ltd., Shlomi 22832, Israel), and angled the wings such that the right wing, but not the left wing, reflected sunlight back toward the camera. We photographed the wings under cloudy conditions (Fig. 4a) and under sunny conditions (Fig. 4b–f), keeping the wings near minimum focus from the lens, with various distances to background foliage. We took the photographs with a Canon EOS 5D Mark II Full Frame DSLR camera fitted with a Canon EF 100 mm f2.8 L macro lens, using the following parameters: (1) 1/50 sec exposure, f29; (2) 1/160 sec exposure, f18; (3) 1/160 sec exposure, f22; (4) 1/80 sec exposure, f29; (5) 1/125 sec exposure, f29; and (6) 1/60 sec exposure, f 29. We converted RAW images to 16-bit uncompressed TIFF files using open-source RAW image-decoding software (DCRAW; [31]) in a manner that maintained pixel linearity. We then examined the images in ImageJ [32], separated the green color channel, manually selected wings, and graphed histograms of pixel values.
Relative spectral power of light reflected off the wings of immobilized female flies exposed to a 100-watt white-light LED
We narrowed the field of view of a spectrometer (HR4000, Ocean Optics, USA) attached to a cosine corrector (CC-3-UV-S, Ocean Optics, USA), using a Gershun tube constructed of matte black construction paper. The tube extended 5 cm beyond the tip of the cosine corrector and had a 6-mm opening. We positioned decoupled female L. sericata wings as described above, keeping the wing and the aperture of the Gershun tube 2 cm apart. At this distance, the spectrometer’s field of view is limited to an 8-mm radius circle. Through this approach, we could maximize the field of view occupied by the wing. We took radiance spectra of (1) the illuminating 100-watt white-light LED, (2) the reflection from the matt black velvet background behind the wings, and (3) the reflectance of the wing oriented to reflect or (4) to not reflect, light towards the opening of the tube.