Perceptual bias in floral mimicry: Visual cues

Background/Question/Methods

Pollination depends on the attraction of insects to particular flower traits such as colors or fragrances, which are typically associated with a food reward.  Floral mimicry systems likely depend to an even greater extent on the perceptual biases of their pollinators. Even if a mimic is not exactly like the model, if the same perceptual biases are triggered, then the pollinator is likely to visit.  Here we describe a series of experiments designed to examine only the visual cues for the Drosophilid pollinators of Dracula orchids, whose labellums mimic the shape, size and color and fragrance of co-occurring mushrooms.  In addition to the mushroom-like labellum, Dracula lafleurii flowers have large sepals with white backgrounds and red spots.  We asked: 1. Are the flies perceiving specific sepal colors (red or white), or is it contrast (stripes, spots, plain) that is important?  2. How important is color of the labellum and column for attraction (matched visually, matched in UV or fluorescent)?  3.  How is flower size (1/2x, 1x and 2x) perceived?  To address these questions we performed experiments in the field with model flowers cast from pharmaceutical grade silicone and measured natural variation in color and size.

Results/Conclusions

There was a bias towards flowers with contrasting color patterns for the BT flies, which are most often the pollinators of D. lafleurii.  They visited spotted and striped treatments at similar rates and the spotted significantly (P≤0.05) more often than red or white.  White flowers were never visited and the red flowers were rarely visited. The felix flies, which are common pollinators of a different species, Dracula felix, visited all the pattern/color treatments at an indistinguishable low rate (P>0.05).  The color of central part of the flower did not appear to be an important attractant; no felix flies visited, and the BT flies did not appear to distinguish among the treatments (P>0.05).  Flower size matters (P=0.0054).  There was a strong bias towards normal and 2X-sized models; the ½-X models received no visits by either BT or felix flies.  There is considerable natural variation in both size and spotting. Taken together, these results suggest that the BT flies will select for spotted flowers and that both fly species will select for large flowers.