FROM THE SIDES OF THE NECK OF THE DOMESTIC PIGEON. 271 



direction indicated by the arrow a, Figure 4, we may obtain a series of results with 

 this apparatus such as are shown in Figures 16 to 21. 



The least possible angle of inclination yielding a metallic color is about 10° (Fig. 

 16, a.) . This angle seems to be limited, in the case of the proximal row of barbules, 

 simply by the screen-like action of one barbule upon another, for I get red effects 

 from single barbules when the angle is still smaller. In the case of the distal bar- 

 bules, the barb is a screen and the slight downward dip of the barbules distally helps 

 to keep them out of reach from the incident rays. At 24° inclination (Fig. 16, &.) 

 there is a greater amount of barbule surface reflecting purple with a consequently 

 stronger color. Purple rays reach the eye through a range of about 50°; i.e., from 

 30° to 80° elevation above the plane of the feather (Fig. 16, c. to d.). Reflected rays at 

 angles greater than 80° give only traces of green from limited regions; and at less 

 than 30° a dull brown obtains. 



If we make the angle of inclination 28° (Fig. 17, a.), we get a small amount of green 

 from limited regions on the barbules through a range for reflected rays of angles more 

 than 67° (Fig. 17, &. to a.). When the angle of inclination for incident rays is 48° (Fig. 

 18, a.), a limited amount of purple is reflected through a range from 7° to 41° (Fig. 18, 

 6. to c). Both red and green effects are obtained at elevations from 41° to 57° (Fig. 19, 

 &. to c). At angles of elevation greater than 57° (Fig. 20, a.) a brilliant green is obtained 

 through a range up to the point where the observer's head cuts off the incident rays 

 (Fig. 20, &. to a.). If we make the angle of inclination for incident rays 60° (Fig. 21, a. ) 

 the range of green rays begins at about 40° (Fig. 21, &.). 



A comparison of these results shows that when the angle made by the incident 

 and reflected rays is about 90° to 100°, we get a transition region between green and 

 red for the reflected rays. Individual barbules viewed singly under the microscope, 

 with the incident and reflected rays making an angle of about 90°, were found like- 

 wise to give both colors. 



The inclinations and angles corresponding to those shown in Figures 16 to 21 

 are slightly changed when the incident rays have the general direction indicated by 

 the arrow b, Figure 4. This condition seems to be due to a slight inclination of the 

 plane of a barb and its barbules with reference to the plane of the feather. The dis- 

 tal ends of the proximal barbules are therefore lower than the distal ends of the distal 

 barbules. This is necessary at the proximal end of the barb especially in order that 

 the distal barbules may lie dorsal to the proximal barbules and be attached to them 

 by means of their ventrally arranged barbicels. The smallest angle of inclination 

 possible to secure purple effects, for instance, is then about 30° instead of 10° (Fig. 

 16, a.). 



