Behavioural experiments

We ran experimental replicates in a windowless room without natural light at 23–26 ^oC, a 40–60% relative humidity, and a photoperiod of 14L:10D. For each replicate, we released 50 virgin, 1- to 3-day-old females starved at least 24 h into a mesh cage (77 × 78 × 104 cm high), the front and lateral sides of which were covered with black cloth to minimize stray light entry, and the top and back were left uncovered (Fig 2). The cage center housed two burette stands separated by 25 cm, each stand carrying a Delta trap 50 cm above the cage floor (S2 Fig). We made traps from white or black cardstock (71.28 × 55.88 cm) (Staples Inc., MA, USA; ACCO Brands Corp., IL, USA) that we cut to size (15 × 30 cm), coated with adhesive (The Tanglefoot Company, MI, USA) on the inside, and then folded into a Delta-type trap (15 × 9 × 8 cm high). We terminated experiments after 24 h, at which time we scored trap captures and removed remaining mosquitoes from cages.

(a) Photo (front view) of the behavioural bioassay design, showing paired Delta traps inside a mesh cage, and the position of a conventional and UV light source. (b) Schematic drawing (top/lateral view) of the behavioural bioassay design, showing three side walls covered in black cloth, paired Delta traps, and the two light sources. For each bioassay replicate, 50 virgin, 1- to 3-day-old females were released into the cage, and trap captures were recorded 24 h later.

We illuminated cages with a shop light housing (Lithonia Lighting, GA, USA) placed vertically behind each cage and fitted with both a 1.22-m 10.0 UVB fluorescent tube (Zoo Med, San Luis Obispo, CA, USA) and a conventional 1.22-m fluorescent tube (F32T8/Tl835 Plus, Phillips, Amsterdam, Netherlands) (Fig 2B and 2C). We did not control for intensity of the conventional or UV lights. We connected the housing to a timer set to the same photoperiod (14L:10D) as the room lights.

In experiment 1 (Fig 3), treatment and control stimuli consisted of a freshly cut T. vulgare inflorescence with its stem inserted into a water-filled vial (4-ml) through a pre-punctured hole in Parafilm (Bemis Company Inc., WI, USA) that covered the vial opening. We placed each vial horizontally into a trap such that the inflorescence faced the light housing fitted with both a UV and a conventional fluorescent tube (see above). To determine the (additive) effect of visual cues on the attractiveness of T. vulgare inflorescences, we placed the vial containing the treatment inflorescence on top of cheesecloth (Cheesecloth Wipes, VWR International, PA, USA) and occluded the vial containing the control inflorescence with cheesecloth. Experiment 2 (Fig 3) was identical in design except that we tested H. lachenalii instead of T. vulgare inflorescences.

Test stimuli are presented in schematic drawings, with left and right sections presenting the human-visible and UV light image, respectively; grey and black in the UV light image indicate UV reflection (UV-bright) and UV absorption (UV-dark), respectively; hatched lines indicate that the inflorescence was covered by cheese cloth; odour from natural inflorescences was present in all experiments (see methods for details).

The design of experiment 3 (Fig 3) was identical to that of experiment 1 except that we placed a sheet of polycarbonate (30.48 × 91.44 × 0.3175 cm thick; Lexan, SABIC, Riyadh, Saudi Arabia) with minimal UV transmission (S1 Fig) in front of the UV light source. This design essentially eliminated the visibility of the bullseye pattern from the inflorescence. Experiment 4 (Fig 3) was identical in design except that we tested H. lachenalii instead of T. vulgare inflorescences.

The design of experiment 5 (Fig 3) was identical to that of experiment 4 except that we (1) placed each vial with its inflorescence on top of black velvet (Suzhou Joytex International Co. Ltd, Jiangsu, China), (2) deployed black instead of white delta traps, and (3) treated inflorescences to alter their bullseye (the characteristic UV absorption and reflection pattern). We treated the upper surface of petals of treatment inflorescences with a “sunscreen mix” of UV-absorbing Parsol 1789 and Parsol MCX (50:50 w/w; Sigma-Aldrich, ON, Canada) formulated in canola oil [adapted from 22]), and the upper surface of petals of control inflorescences with canola oil only. In addition, we treated the receptacle of control inflorescences with the “sunscreen mix” to ensure “odour symmetry” between treatment and control inflorescences.

In experiment 6 (Fig 3), we compared the attractiveness of yellow model flower discs (2.5 cm diameter) that exhibited either a uniformly UV-dark or a uniformly UV-bright appearance. We prepared the discs from yellow printer paper (International Paper, TN, USA), and painted treatment discs with clear nail polish (Coty Inc., NY, USA) rendering them dark in the UV range while maintaining their yellow, human-visible colouration. Using an inkjet printer, we printed control discs with a yellow ink that maintained their UV reflectance but rendered them darker to mimic the darkened appearance of nail polish-painted treatment discs. To ensure “odour symmetry” of the treatment and the control disc, we paired the nail polish- and inkjet- treated disks using their untreated side for contact. We then placed each disc pair into a black trap containing a H. lachenalii inflorescence which we occluded with black velvet to provide olfactory but not visual cues. In treatment and control traps, the nail polish-painted side and the yellow ink-printed side, respectively, of the paired discs leaned against the occluded inflorescence at a 45^o angle relative to the trap bottom and faced the light housing.

In experiment 7 (Fig 3), we explored a potential additive effect of floral colour (yellow) on the combined effect of floral odour and UV darkness on mosquito attraction. We modified the design of experiment 6 in that we prepared model flower control discs from black cardstock and model flower treatment discs from yellow printer paper, painting both discs with clear nail polish which renders them UV-dark. We also replaced black traps with white traps, and black velvet with cheesecloth.