Naturwissenschaften 83, 136-138 (1996) © Springer-Verlag 1996 Does Bee Color Vision Predate the Evolution of Flower Color? L. Chittka * Institut fiir Neurobiologie der Freien Universitfit, D-14195 Berlin, Germany and Ecology and Evolution, State University of New York, Stony Brook, NY 11794-5245, USA Many species of bees depend obligator- ily on pollen and nectar offered by flowers of angiosperm plants. It is thus reasonable to assume that the com- ponents of the sensory system of such insects were adapted to maximize the distinctiveness and detectability of floral food sources. Indeed, the bees' UV, blue, and green receptors are op- timally placed on the wavelength scale for discrimination of flower colors [1]. However, to prove that flower signals indeed influenced wavelength tuning of bee spectral receptors, it must be shown that the ancestors of bees possessed dif- ferent sets of such receptors prior to the advent of the angiosperms. How can we determine the spectral receptor types through which insects saw the world 200 million years (Ma) ago? One has to evaluate members of arthropod taxa whose evolutionary lineages diverged from those of bees before there were flowers. If the spectral receptor sets of such animals are indistinguishable from those of bees, this implies that essential components of insect color vi- sion predated the evolution of flower color. To test this possibility, the )~max values of a large number of species were su- perimposed on the phylogenetic tree of the arthropods (Chelicerata, Crusta- ceae, and Antennata, including the In- secta). Only species whose phylogenetic position could be unambiguously deter- mined according to the literature [2-6] were included. Diptera and Lepidoptera will be treated ]n a separate-study (Chittka, in prep.). The following trends are apparent in Fig. 1. The 2m~ values of the Crusta- cea and Insecta fall into three distinct clusters around 350, 440, and 520 nm. The Chelicerata, including jumping spiders [7, 8] and horseshoe crabs [9] lack blue receptors consistently; ERG measurement from scorpions confirm this picture [10]. In contrast, almost all Mandibulata possess at least the above three color receptor types. Thus, the blue receptor appears to be an evolu- tionary novelty in the ancestor of the Mandibulata. The few insect species in which one of these types is absent (Peri- planeta [11] and Myrmecia [12]) clearly represent cases in which these receptors were lost secondarily. Red receptors show up irregularly in both the Crusta- cea and Insecta; they have obviously evolved several times independently. To see whether the wavelength positions of UV, blue, and green receptors de- pend on whether their bearers are flow- er visitors or not, we compared the )~,~ values of these receptor classes be- tween the Hymenoptera [H] and the re- maining arthropods with three spectral photoreceptor classes [A] in Fig. 1 by means of the Mann-Whitney U test. No statistically significant difference is found for the UV receptors (mean J'max [H] = 343 nm; mean ~'max[A] = 349 nm; p = 0.44) and the blue recep- tors (mean)~max [H] = 434 nm; mean 2ma x [A] = 433 nm; p = 0.36); however, the distributions of green receptors dif- fer significantly between these two groups of arthropods (mean)~max[H] = 535 nm; mean 2max [A] = 521nm; p = 0.012). Unfortunately, the color receptors of only few species of Crustacea were studied by means of intracellular mea- surements. There is actually more diversity in wavelength positions of crustacean photoreceptors than Fig. 1 (which focuses on data from intracel- lular recordings) might suggest. Since the evaluation of spectral receptor types in Crustacea is crucial to understanding the evolution of receptor wavelength tuning in their sister group, the Anten- nata (including the insects), a separate graph is shown which also includes spe- cies whose 2ma x have been determined by microspectrophotometry and ERG measurements. ~-_ c~ o- Do- 65~ • • 6o • • • ss 0 mooo oo ooo° o oo °oooo ° o°oo° - 5o~ 0 ~ 0 45 && ss~ °°° ° • °°°oO o ° • • °o°O°°°ooOO°o CHELi- ~ CERAT 0 I CE.A j ~ J I- b, ymenoptera I I~ INSF~CTA I Fig. 1. /~max values of photoreceptors of 29 species of arthropods superimposed on the phylogenetic tree of these species. Only species whose color receptors were investigated elec- trophysiologically by means of intracellular recordings were taken into account, so as to ensure comparability of wavelength positions. One exception is Daphnia, where only the three long- wavelength receptors were measured intracellularly, whereas the UV receptor was studied by an extracellular technique [23]. • UV receptors (325-370 nm); ~ blue receptors (400-460 nm); O green receptors (490-560 nm);• red receptors (590-630 nm). For complete species names refer to the literature indicated. Species names are abbreviated where recordings were made for more than a single species of that genus 136 Naturwissenschaften 83 (1996) © Springer-Verlag 1996