No.412.] SMALLER NORTH-AMERICAN SHRIKES. 279 
combinations that have given the desired effect. In practice, 
however, the following conditions were found essential to 
accurate estimation of color: (1) Uniform strong diffused 
daylight, preferably coming from a skylight or north window 
and not subject to strong reflections from colored surfaces; 
(2) frequent comparison with ‘previously made estimates ; 
(3) careful comparison of the color obtained by the top with 
that of the plumage, whose color is being measured, both 
being viewed from different directions. 
I found it impracticable to make color estimates for more 
than two hours continuously, because of color fatigue of 
the eye. 
In spite of the above precautions, there are limitations to 
the use of the color-top. In the estimation of color, the 
personal equation plays an important part. Then, too, the sur- 
face of the color disks is very different in character from that 
of feathers. The peculiar luster or sheen so characteristic of 
a bird's plumage cannot be imitated by the color mixer. In 
estimating the color of the breast, more or less mottling, which 
exists in some cases, increased the difficulty. In such cases I 
have attempted to estimate the mean color. Some apparatus 
for blending a complex pattern into one color would be very 
desirable. I do not attempt to maintain that estimates of the 
same material by another person would exactly agree with 
mine; but I believe that I have a consistent classification of 
individuals according to color. 
After several weeks of experimentation I found that, though 
there were traces of blue, red, and yellow, the colors black and 
white were by far the most important elements in the areas 
measured. Therefore I have taken as the color criterion the 
amount of black, or the melanism, of the color area described ; 
and color estimates appear in the tables as percentages of 
melanism. I have found it more difficult to estimate colors 
containing more than 50% of white than those in which black 
predominated, and I have not attempted to make fine distinc- 
tions as to melanism in the lighter color areas. The large 
amount of black necessary to produce even the lightest 
grays in the color mixer was a matter of surprise to me. To 
