WHY RED FLOWERS ARE NOT INVISIBLE TO BEES

ABSTRACT

A pervasive idea among pollination biologists is that bees cannot see red flowers. This idea has led many workers to assume that red coloration is an adaptation by which flowers exclude bees as visitors. However, recent empirical and theoretical evidence strongly supports the alternative view, that red flowers are visible to bees. Our purpose is to marshal this evidence from physiology, behavior, and ecology. First, we define the spectral boundary between orange and red, and show that the visual spectrum of all bee species studied to date extends enough into long wavelengths to provide sensitivity to red light. Such sensitivity differs from the ability to discriminate different monochromatic lights, and we argue that bees will be unable to discriminate such lights above about 550 nm. Second, we point out that flowers do not reflect monochromatic lights. Instead many of them, particularly those that appear red, orange, yellow, and white to humans, have reflectance patterns that are essentially step functions. We predict that bees should be able to discriminate such reflectance patterns over a range of 550–650 nm, since reflectance functions with steps at such wavelengths will occupy different loci in bee color space and thus be distinguishable. In this sense, bees should distinguish between green-, yellow-, orange-, and red-reflecting objects, even if these do not reflect in shorter wavelengths (including UV). A behavioral experiment shows that bumblebees can indeed perform this task. Third, we present information on the spectral reflectance of some typical “red” flowers, combined with field observations of bee visitation to such flowers. We end with a preliminary reassessment of the adaptive significance of red flower coloration, using North American “hummingbird” flowers as an example; we also stress some of the pitfalls facing evolutionary biologists who continue to assume that bees are blind to red objects.