Insights into eyespot color-pattern formation mechanisms from color gradients, boundary scales, and rudimentary eyespots in butterfly wings.

Butterfly eyespot color patterns are traditionally explained by the gradient model, where positional information is stably provided by a morphogen gradient from a single organizer and its output is a set of non-graded (or graded) colors based on pre-determined threshold levels. An alternative model is the induction model, in which the outer black ring and the inner black core disk of an eyespot are specified by graded signals from the primary and secondary organizers that also involve lateral induction. To examine the feasibility of these models, we analyzed eyespot color gradients, boundary scales, and rudimentary eyespots in various nymphalid butterflies. Most parts of eyespots showed color gradients with gradual or fluctuating changes with sharp boundaries in many species, but some species had eyespots that were composed of a constant color within a given part. Thus, a plausible model should be flexible enough to incorporate this diversity. Some boundary scales appeared to have two kinds of pigments, and others had "misplaced" colors, suggesting an overlapping of two signals and a difficulty in assuming sharp threshold boundaries. Rudimentary eyespots of three Junonia species revealed that the outer black ring is likely determined first and the inner yellow or red ring is laterally induced. This outside-to-inside determination together with the lateral induction may favor the induction model, in which dynamic signal interactions play a major role. The implications of these results for the ploidy hypothesis and color-pattern rules are discussed.