278 THE AMERICAN NATURALIST. [Vor. XXXV. 
by measuring the actual distance in the specimen and multi- 
plying that value by the linear magnification of the print. 
The nostril distance was used because the nostril was found 
to be the only point sufficiently definite and constant for the 
proximal limit. This chord (AB, Fig. 3) served as a base line. 
A line (CD) parallel to the base line and tangent to the cul- 
men was next drawn, and at the point of tangency (E) a per- 
pendicular (EF) was erected between the lines. It was then 
a simple matter to measure the lines AF and £F, and with a 
table of natural tangents to determine the angle EAF. This 
is the angle that has been adopted as a criterion of curvature 
in the shrike bill. It is easily seen that there might be an 
infinite number of small variations in the outline of the culmen, 
especially in the part distal to the point Æ, which would not 
affect the angle ZAF; but I believe that a correlation is to be 
expected between the curvatures in various portions of the 
culmen, so that it is probable that any considerable variation 
in outline elsewhere would affect the curvature in the region 
of E and thus find expression in the angle ZAF. In order, 
however, to have something more than a subjective impression 
on this matter, I made a second series of measurements with 
half the length of the line AB used as a chord. The result of 
these measurements, and their deviations from those of the 
first set, are given on page 294. 
3. Color Measurements. X ds the quantitative determination 
of color, the “color mixer” was suggested. A simple form of 
color mixer — the Bradley Milton Color-Top — was employed. 
This instrument has a graduated disk, with superimposed paper 
disks of five different colors, which can be so adjusted as to 
make compound disks exposing two or more colors. On being 
rotated at high speed these colors are mixed into one, which 
is a combination of the colors used in the proportions of the. 
number of degrees of each exposed. With the aid of this 
apparatus it is therefore possible to determine the proportions 
of these primary colors entering into any color effect which 
may be produced by various combinations of them. To deter- 
mine the composition of any color in nature, it is necessary 
to reproduce empirically that color in the top and note the 
