270 THE METALLIC COLORS OF FEATHERS 



faces of very variable form and of exceedingly small size. The color effects obtained 

 without the aid of a microscope are necessarily mixtures of light rays from many 

 sources. 



With small angles of incidence, practically the whole barbule surface visible 

 in Figure 4 gives green. When, however, we have the larger angles of incidence 

 necessary for purple effects, it is easy to see that only limited portions of the bar- 

 bules can give a metallic color; either the incident or the reflected rays will be inter- 

 cepted by the adjacent barbules except on the more elevated portions or when the 

 incident rays have the directions* indicated by the arrows g or h, Figure 4. When 

 the incident rays do take the direction of the arrow g for instance, or a direction 

 opposite to it, they must cross the proximal barbules at a considerable angle with 

 consequent loss. In such a case, the distal barbules give a maximum of purple, and 

 the proximal barbules a minimum amount. If, however, the incident rays have 

 the direction indicated by the arrow h, the proximal barbules, instead, will give the 

 maximum amount of purple, and the distal the minimum amount. 



If, for instance, the incident rays pass from left to right in Figure 11, and make 

 small angles with the plane of the feather, it is apparent that the convexity of one 

 barbule acts as a screen for all of the next barbule surface except the corresponding 

 convexity. If the angle of inclination is large enough to allow the incident rays to 

 reach most of each barbule surface not actually covered by adjacent barbules, we 

 go out of the range of purple effects and get green. When, however, the incident 

 rays lie in the vertical plane indicated by the arrows g or h, Figure 4, they find no 

 intercepting convexities. The limitations to the amount of purple that we have 

 just noted explain the lesser brilliancy of the purple effects in feathers giving bril- 

 liant greens also. 



If a single barbule is detached from the barb and examined under a #-inch 

 objective by reflected light coming from the side, red or purple effects appear, mostly 

 on the margin nearest the source of light. Over the broad surface all the colors of 

 the spectrum are frequently seen, but green is greatly in excess of the others. 



The barb itself may act as a screen when the angle of incidence is very large, and 

 may thereby assist in reducing the amount of purple reflected. 



4. Measurements of Color Ranges. — Measurements of the ranges of green and 

 red for various inclinations of the feather with reference to the source of light and 

 for various positions of the eye have been made by means of the apparatus described 

 on page 266. If, for instance, the feather is placed so that the incident rays have the 



* The inclination of the feather with reference to the incident rays is of course not shown by the arrow which 

 only indicates a vertical plane in which the incident rays pass, making various angles with reference to the feather. 



