426 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1911. 
group of molecules may cause selective reflection. It seems clear 
that this property can not belong to individual molecules, at any 
rate in the case of the aniline dyes, for their solutions absorb impar- 
tially all the colors which are not transmitted, and it is only in the 
solid state that their peculiarities as regards reflected light become 
apparent; at the same time there is no change in the light transmitted 
whether the dye is in solution or a dry film. 
Before entering in detail into the reasons which seem, to point to 
interference rather than selective reflection as the origin of iridescent 
colors, some general remarks may be made on the character of the 
structures examined. 
These structures have been either feathers of birds or the scales 
of insects. There are few orders of birds in which examples of 
iridescent coloring can not be found, but without doubt the humming 
birds are the most brilliant, although peacocks, trogons, and many 
others are not very far inferior. In the insect world the finest 
examples are to be found amongst butterflies and the day-flying 
moths of the genus Urania. Some beetles also are provided with 
vividly colored scales. These belong mostly to the weevils (which 
include the Brazilian diamond beetle). 
Many other insects among the Diptera, Neuroptera, and Hymenop- 
tera show brilliant metallic colors on their integuments, but these 
are not provided with scales, and in many cases the color fades 
more or less when the specimens become dry. These I have not 
examined. Feathers and scales, however, are remarkable for the 
permanence of their iridescent coloring, and it is to these only that 
the present observations apply. 
Some of the peculiarities of the structures as regards change of 
color with the point of view depend on the shape of the surface 
on which the color-producing material lies. If the surfaces are 
flat or nearly flat, reflection takes place as from a looking glass, 
and the angle through which the specimen can be turned while 
still showing the characteristic color is small. Often, however, 
the surfaces are convex bosses or ridges, and then the angle of inci- 
dence and reflection is that contained between the direction of the 
incident light and the normal to the tangent plane at the point 
where reflection takes place, and is therefore to a great extent inde- 
pendent of the position in which the specimen is held, since there will 
always, within wide limits, be tangent planes to the convex surfaces 
which reflect the incident light in the line of sight. In these cases 
the colors might at first sight be taken as due to pigment, both on 
account of their comparatively low intensity and from the small 
change in tint and intensity which is produced by altering the inclina- 
tion of the general surface to the direction of the illumination. The 
