GENETICS: HARRIS, BLAKESLEE AND WARNER 
237 
BODY PIGMENTATION AND EGG PRODUCTION IN THE FOWL 
By J. Arthur Harris, A. F. Blakeslee, and D. E. Warner 
STATION FOR EXPERIMENTAL EVOLUTION. COLD SPRING HARBOR. NEW YORK 
Communicated by C. B. Davenport, January 10, 1917 
The recognition by both plant and animal breeders of the great 
economic value of any character by which superior individuals might 
be indirectly selected has been father to a widespread conviction that 
such characters actually exist. 
In the main, beliefs concerning distinctive structural peculiarities of 
the heavier yielding cereals, the bodily dimensions or proportions of 
the best dairy animals, and the criteria by which the most prolific 
layers may be selected from the flock without recourse to trap nesting, 
rest solely on personal experience and judgment, instead of upon meas- 
urements and correlations. They have, therefore, little scientific value. 
The quantitative work which has been done on both animals and 
plants indicates that for the most part there is little prospect of the 
selection of the economically important characters of yield in plants, 
milk production in cattle or fecundity in poultry on the basis of struc- 
tural characters unimportant as such but correlated with economically 
important physiological characters. 
In certain cases, the relationship between two physiological charac- 
ters, one with and the other without economic value, may be much 
more intimate. The measurements of the correlations may then be 
not merely of theoretical interest to the physiologist but of very real 
economic significance. 
The relationship between the intensity of yellow somatic pigmenta- 
tion and egg production in antecedent periods of time may serve as an 
illustration of such characters. 
Working on two years data covering trap nesting records for 309 
and 375 White Leghorn birds respectively, we find the following results 
for the correlation between the percentage of yellow pigment in the ear 
lobe^ and total egg production for the preceding year^: 
For 1913-1914, f,^ = -0.5816 ^ 0.0253 
For 1914-1915, = -0.5271 ± 0.0252 
Difference 0.0545 =t= 0.0358 
Within the hmits of the probable errors of random sampling, the results 
for the two years may be considered identical. 
Expressing the relationship in terms of straight line prediction equa- 
tions, represented graphically in Figure 1, we find: — 
