Mimicking orchids lure bees from afar with exaggerated ultraviolet signals

Abstract

Flowers have many traits to appeal to pollinators, including ultraviolet (UV) absorbing markings, which are well-known for attracting bees at close proximity (e.g., <1 m). While striking UV signals have been thought to attract pollinators also from far away, if these signals impact the plant pollinia removal over distance remains unknown. Here, we report the case of the Australian orchid Diuris brumalis, a nonrewarding species, pollinated by bees via mimicry of the rewarding pea plant Daviesia decurrens. When distant from the pea plant, Diuris was hypothesized to enhance pollinator attraction by exaggeratedly mimicking the floral ultraviolet (UV) reflecting patterns of its model. By experimentally modulating floral UV reflectance with a UV screening solution, we quantified the orchid pollinia removal at a variable distance from the model pea plants. We demonstrate that the deceptive orchid Diuris attracts bee pollinators by emphasizing the visual stimuli, which mimic the floral UV signaling of the rewarding model Daviesia. Moreover, the exaggerated UV reflectance of Diuris flowers impacted pollinators' visitation at an optimal distance from Da. decurrens, and the effect decreased when orchids were too close or too far away from the model. Our findings support the hypothesis that salient UV flower signaling plays a functional role in visual floral mimicry, likely exploiting perceptual gaps in bee neural coding, and mediates the plant pollinia removal at much greater spatial scales than previously expected. The ruse works most effectively at an optimal distance of several meters revealing the importance of salient visual stimuli when mimicry is imperfect.

Keywords: bee sensory ecology; ecological interactions; flower attraction; food deception; orchid floral mimicry; pollination success; salient stimuli; ultraviolet reflectance; visual food deception.

FIGURE 1

Floral morphology and color properties of the mimicking orchid and its pea model. (a) Flower morphology of the orchid Diuris brumalis (i) and the pea, Daviesia decurrens (ii). The dorsal sepal, labellum lateral lobes, and the labellum in Diuris flower and standard petal and wing petal of Daviesia. The outer petals in the orchid are the component of the floral architecture that is absent in the pea. (b) Average color reflectance measured on flower components in Diuris (iii) and Daviesia (iv) peaks in the UV bands (black arrows). Color reflectance in the UV wavelengths (300–400 nm) varied between 0.5% and 37% in Diuris sepals and petals and 2.5% and 28% in the pea model. The UV reflectance of Diuris outer petals ranged between 18% and 37% (see Data S1 and S2).

FIGURE 2

Color patterns perceived by bees in treated and untreated Diuris flowers and untreated Daviesia. (a) Diuris flower photographed in UV before (control, C) and after applying the UV filter on the outer petals (UV treated, T). (b) False color photography in “bee view” reveals the overall color pattern perceived by bees in treated (i.e., application of the UV filter solution) and untreated outer petals of Diuris flower and untreated Daviesia flower. The UV filter is effectively a long‐pass filter transmitting all wavelengths above 400 nm, free of fragrance, oil, PABA, alcohol, parabens, and preservatives (Kinesys). Importantly, the UV images of treated outer petals show very similar reflectance properties to the background and stem foliage reflectance, confirming that the experimental manipulation knocked out UV signaling with respect to background coloration. (c) Location of color loci was calculated from the mean of reflectance for floral parts of Diuris brumalis (Db), and Daviesia decurrens (Dd). The calculations were made using the Hexagon color model of bee vision (Chittka, 1992). This model represents the internal perception of flower colors by bee pollinators, and resultant sectors (u [ultraviolet]; ub [ultraviolet‐blue]; b [blue] bg [blue‐green]; g [green]; ug [ultraviolet‐green]) show how bees likely interpret spectral signals].

FIGURE 3

Effect of distance from model plants on Diuris pollinia removal. (a) Diuris orchids (yellow [untreated] and green [UV‐treated] flowers) inside [IN] and outside [OUT] a 30 × 30 m patch with Daviesia pea (red flowers). (b) Mean proportion of pollinia bees removed from treated (black bars) and untreated Diuris flowers (white bars) relative to the orchid's distance ([IN] and [OUT]) from the model pea. Each experimental group consists of N = 100 orchids. Error bars are 95% confidence intervals; n.s., no significant difference among experimental groups; ***Significant difference at Bonferroni‐corrected α = .0125.

FIGURE 4

Effect of Diuris UV reflectance on the orchid's pollinia removal relative to mimic‐model distance. (a) Experimental setup treated, control orchid groups and pea plant, (b) Pollinia removal was quantified in 195 orchids (N = 476 orchid flowers). Pollinia removal of control Diuris relative to distance from Daviesia (i) was best described by an inverted parabolic function peaking at ~8 m distance from model pea (χ ² = 9.87, p < .05 for the squared and linear term, respectively) (N = 238 flowers, n = 43 pollinia removed). Pollinia removal of UV‐treated orchids (ii) exhibited an exponential decrease with distance from model pea plants (χ ² = 10.26, p < .001) (N = 238 orchid flowers, n = 17 pollinia removed). Refer to Data S6 for full data.