Meeting Abstract
Sensory systems play a crucial role in the interaction between plants and their insect pollinators. The plants’ visual and olfactory cues determine how insects navigate to flowers. As they approach flowers, moths unfurl their proboscis — straw-like mouthparts that provide mechanosensory feedback as an additional sensory modality to localize and feed from the tiny nectary. Finding the tiny nectary is a challenging task, especially for crepuscular moths that hover over flowers while they feed in low light conditions. Despite the ecological importance of this behavior, few studies have focused on the role of tactile feedback in floral exploration. To examine the role of mechanosensory input, we developed an assay to track naive hawkmoths, Manduca sexta , as they fed from 3D-printed artificial flowers with different curvatures. Their geometry varied from naturally occurring trumpet-shaped flowers that are easy to exploit to completely flat flowers that are more challenging. We found that moths became increasingly efficient at locating the nectary over just a few visits, even on flowers that have only slight curvature. Using computer vision and machine learning methods, we tracked the proboscis as moths explored flower surfaces. We found that they repeatedly swept their proboscis across the flower surface in a manner reminiscent of rat whisking. In particular, they whisk along the radial direction of the flower, varying the angle of whisk as they explore the curvature of the flower. The number of whisks decreases as the moths learn to handle floral shapes. However, the number of whisks remains high for flat floral shapes. Our results show that moths use tactile sensing to actively explore flower surfaces and rapidly learn to handle novel flowers, enabling pollination.
