Meeting Abstract
Animals rely on sensory feedback from multiple modalities to perform complex motor tasks such as pollination. We explored this relationship in Manduca sexta, crepuscular insects that fly in low light conditions, hovering over flowers as they pollinate and feed from them. They use both vision and mechanosensory input as they locate the nectary with their proboscis. As light levels decline, the lag between visual and mechanosensory information processing increases. Using behavioral analyses of feeding moths, we asked how reductions in light level influence their efficiency in locating the nectary. We combined 3D printing technology to generate artificial flowers, each equipped with sensors that detected the proboscis tip inside the nectary. We combined this approach with machine vision to track the motion of hovering moths under two light levels: 0.1 lux (moonlight) and 50 lux (dawn/dusk). In either light level we found that moths decrease the time to find the nectary over multiple visits to a flower. Across all visits, moths in higher light conditions took significantly longer to find the nectary (KS test p-value: 0.005 and a KL divergence of 1.61). In addition to taking longer, moths in higher light conditions hovered further from the flower during feeding (KS test p-value: 0.049). This shift in light level appears to significantly effect learning and motor control in these animals. These results could be explained by a shift in the weight of vision in higher light levels, even though it is slower than mechanosensory input.
