The reflexion spectra of various brilliantly colored insects are examined in the ultra-violet. Morpho butterflies and Urania moths show u.-v. maxima in general agreement with the theory of interference. In Morpho achilles* the positions of the u.-v. maxima indicate that the blue color is due to a reflexion of the 2nd order. Contrary to some previous accounts, no Morpho sp. show their blue color by transmission. They show brilliant diffraction spectra by transmission, when suitably mounted in balsam, but these have nothing to do with the blue reflexion. Iridescent beetles showing a deep red color at normal incidence may be made to pass through all the colors of the spectrum to violet, provided that arrangements are made to annul refraction at the air-chitin surface, so as to obtain very oblique incidence within. This is in sharp contrast to the surface reflexion of aniline dyes, which do not show much change of color with incidence. Some of the golden beetles show transmission spectra of bands corresponding to the reflexion spectra formerly described. These bands vary continuously in position with the part of the specimen examined. It seems impossible to interpret this reasonably except on the theory of interference, for the transmission bands of chemical substances do not vary from one sample to another. Moist Cl gas does not destroy the colors of Morpho or of Urania though the black background is bleached. Nor does Cl decolorize the metallic beetles. The colors of all kinds of feathers, however, are rapidly discharged. Peacock''s feathers undergo a progressive change of color in u.-v. light or long-continued sunlight. Generally speaking, the colors become more refrangible and less brilliant. Other feathers (e.g., parrot) even when blue are slowly decolorized without change of refrangibility. Morpho and Urania also lose color without change of refrangibility. Although, prima facie, fading under light or Cl in these cases would seem to favor the idea of a pigment, this view can not be maintained. The fading must rather be attributed to the breaking down of an interference structure. It remains unexplained why the progressive changes of tint seen in peacock''s feathers are not seen in the other cases examined. The generalization seems to hold good, however, that colors which are stable in Cl are certainly not due to pigment.