1882.] DR. H. GADOW ON THE COLOUR OF FEATHERS. 413 
together with the proper substance of the feather, occupies the rest 
of the barb. Thus we have, if proceeding from the surface to the 
middle of a blue barb; the following structure (fig. 1) :— 
1. A transparent, apparently homogeneous sheath of ceratinine 
(S 8). 
2. One layer of prismatic cells; and 3, under this, a brownish 
pigment (P). 
The sheath may vary in thickness and in surface-structure from 
about 0:0014 to 0:0043 mm. 
In Pitta I calculated its thickness to 0°0016 mm., and the surface 
appeared to be quite smooth ; whilst in Cawreba each top of a cone 
corresponded with a slight elevation of the sheath. 
The breadth, or diameter, was calculated to about 0:006 mm.; 
it agrees very closely with that of Cereba and Ara, 
Fig. 1. 
Diagrammatic section through part of a barb of a blue feather. 
Fatio, who examined the structure of blue feathers, also says that 
under the prismatic layer there are ‘‘ de grandes cellules polygonales 
A noyau coloré.”” But I suppose that this is an optical delusion, and 
that the large polygones (generally hexagones) which we see while 
looking vertically down upon the surface of the rami are the lateral 
outlines of the prismatic columns. Therefore what he figured (op. 
cit. plate iii. fig. 6) as polygones are simply the foreshortened 
columns, and the underlying pigment gives them the appearance of 
cells with a dark nucleus. 
The thickness of the surface-coating of blue feathers varies con- 
siderably in different birds, and even in different feathers of the 
same bird. Differences between 0:0016 and 00043 cannot be put 
down as mistakes of measurement. Again, we know that the 
thickness of colour-producing plates varies from about 0°00006 to 
0:0004 mm., giving bluish-white or pale orange light respectively. 
And if the plates in question are thicker than about 0°0005 mm., 
they cease to produce colour, and the law of colours of thin plates is 
