In many large groups of birds it may be observed that the plumage is differently coloured in each species, yet that certain spots, marks, or stripes, though likewise differently coloured, are retained by all the species. Analogous cases occur with the breeds of the pigeon, which usually retain the two wing-bars, though they may be coloured red, yellow, white, black, or blue, the rest of the plumage being of some wholly different tint. Here is a more curious case, in which certain marks are retained, though coloured in almost an exactly reversed manner to what is natural; the aboriginal pigeon has a blue tail, with the terminal halves of the outer webs of the two outer tail-feathers white; now there is a sub-variety having a white instead of a blue tail, with precisely that small part black which is white in the parent-species.[191] _Formation and variability of the Ocelli or eye-like Spots on the Plumage of Birds._—As no ornaments are more beautiful than the ocelli on the feathers of various birds, on the hairy coats of some mammals, on the scales of reptiles and fishes, on the skin of amphibians, on the wings of many Lepidoptera and other insects, they deserve to be especially noticed. An ocellus consists of a spot within a ring of another colour, like the pupil within the iris, but the central spot is often surrounded by additional concentric zones. The ocelli on the tail-coverts of the peacock offer a familiar example, as well as those on the wings of the peacock-butterfly (Vanessa). Mr. Trimen has given me a description of a S. African moth (_Gynanisa Isis_), allied to our Emperor moth, in which a magnificent ocellus occupies nearly the whole surface of each hinder wing; it consists of a black centre, including a semi-transparent crescent-shaped mark, surrounded by successive ochre-yellow, black, ochre-yellow, pink, white, pink, brown, and whitish zones. Although we do not know the steps by which these wonderfully beautiful and complex ornaments have been developed, the process at least with insects has probably been a simple one; for, as Mr. Trimen writes to me, “no characters of mere marking or coloration are so unstable in the Lepidoptera as the ocelli, both in number and size.” Mr. Wallace, who first called my attention to this subject, shewed me a series of specimens of our common meadow-brown butterfly (_Hipparchia Janira_) exhibiting numerous gradations from a simple minute black spot to an elegantly-shaded ocellus. In a S. African butterfly (_Cyllo Leda_) belonging to the same family, the ocelli are even still more variable. In some specimens (A, fig. 52) large spaces on the upper surface of the wings are coloured black, and include irregular white marks; and from this state a complete gradation can be traced into a tolerably perfect (A¹) ocellus, and this results from the contraction of the irregular blotches of colour. In another series of specimens a gradation can be