Related Experiment Video
Updated: Jul 16, 2026

In situ Protocol for Butterfly Pupal Wings Using Riboprobes
Published on: May 28, 2007
Color discrimination at the spatial resolution limit in a swallowtail butterfly, Papilio xuthus
Yuichi Takeuchi1, Kentaro Arikawa, Michiyo Kinoshita
1School of Advanced Sciences, The Graduate University for Advanced Studies, Hayama 240-0193, Japan.
Abstract:
Spatial resolution of insect compound eyes is much coarser than that of humans: a single pixel of the human visual system covers about 0.008 degrees whereas that of diurnal insects is typically about 1.0 degrees . Anatomically, the pixels correspond to single cone outer segments in humans and to single rhabdoms in insects. Although an outer segment and a rhabdom are equivalent organelles containing visual pigment molecules, they are strikingly different in spectral terms. The cone outer segment is the photoreceptor cell part that expresses a single type of visual pigment, and is therefore monochromatic. On the other hand, a rhabdom is composed of several photoreceptor cells with different spectral sensitivities and is therefore polychromatic. The polychromatic organization of the rhabdom suggests that insects can resolve wavelength information in a single pixel, which is an ability that humans do not have. We first trained the Japanese yellow swallowtail butterfly Papilio xuthus to feed on sucrose solution at a paper disk of certain color. We then let the trained butterflies discriminate disks of the training color and grey disks each presented in a Y-maze apparatus. Papilio correctly selected the colored disk when the visual angle was greater than 1.18 degrees for blue, 1.53 degrees for green or 0.96 degrees for red: they appeared to see colors in single pixels to some extent. This ability may compensate their rather low spatial resolution.
Related Concept Videos
Frequency-dependent Selection
Position-effect Variegation
Testing a Claim about Mean: Unknown Population SD
Estimating a population mean requires the samples to be approximately normally distributed. The data should be collected from the randomly selected samples having no sampling bias. There is no specific requirement for sample size. But if the sample size is less than 30, and we don't know the population standard deviation, a different approach is used; instead...

