Colouring in Birds and Insects. 5b*3 



gave a set o£ curves (PL IV. fig. 5) which showed a double 

 reversal ; the phase-curve being positive for crimson and 

 red, negative for orange and orange yellow, and positive 

 again for the yellow, green, and blue. 



A series of curves for a very thin film of magenta (estimated 

 thickness 0*00005 mm.) gave results surprisingly resembling 

 those of the beetle. The second point of inversion being, 

 however, in the green instead of the yellow, and the "metallic*" 

 character of the film being much less marked than in the 

 beetle wing-case. The resemblance in the lower curves, 

 showing the variation of maximum steepness with the colour, 

 is even more striking. It can scarcely be doubled, therefore, 

 that here again the "metallic" colour is produced in a film 

 whose thickness is of the order of a ten thousandth of a 

 millimetre or less. 



A third example (PL IV. fig. 6) is added, in which the 

 correspondence is less marked, for the purpose of illustrating 

 the general character of the curves for a case of green 

 metallic lustre. There is in fact no aniline colour which 

 shows an accurate correspondence, but the same magenta 

 curves may be referred to for comparison. 



The beetle wing-cases furnish in many cases a fairly 

 smooth surface, and the difficulties in obtaining the necessary 

 measurements are far less than when working with feathers 

 of birds or with butterfly scales. Nevertheless, as PL IV. 

 fig. 7 shows, the same general characteristics obtain in these, 

 in both the phase-curves and the amplitude-ratios. It may 

 be noted that the two curves do not always correspond*, 

 but it is probable that the difference maybe explained by the 

 difficulty in obtaining accurate results with surfaces so 

 irregular. 



It is worthy of note that in all of these curves (except that 

 furnished by a red humming-bird feather) the curves are 

 negative ; from which it is fair to conclude that the film 

 which produces the surface colour is very thin. 



The total number of specimens which have been examined 

 is perhaps not so large as it should be to draw general con- 

 clusions, and it is clearly desirable that it be extended ; but 

 so far the evidence for surface film, as the effective source 



* If we take the approximate formula h — tan 2\b, it is at once apparent 

 that the dotted curves in oar diagrams should have the highest minimum 

 value for all the cases of great opacity. Thus the opacity may be in- 

 ferred from the dotted curve for x^ as well as from the full curve for 

 dA/di, and in general the indications in the two cases show a rough 

 agreement, the steeper full curves corresponding to the lower dotted 

 curves, and vice versa. 



