92 



THE AMERICAN NATURALIST [Vol. LI1 



while Table V. should also be examined in this connection. 

 It is clear from these graphs that the earlier hatched 

 birds -are superior to the later hatched for winter egg 

 production. In the former group there are fewer zero 

 and more very high producers than in the last. There is 

 also a marked difference in the number of birds in each of 

 the several hatching months that laid over thirty eggs 

 during the winter period. (See Table VI.) 



The variability in the age at which the first egg is pro- 

 duced influences the winter record strongly, so much so 

 that we have been led to believe that it is the most im- 

 portant determining factor for egg production during the 

 winter months for our flocks. 3 It leads to the abandon- 

 ment of the view that those records that fall below 30 eggs 

 are made by true mediocre producers, substituting there- 

 for the view that many, perhaps most, of them are late- 

 maturing high producers. Now, the variability in age at 

 first egg, shown in Figs. 1 and 2, is considerable. If this 

 variability could be eliminated— that is, if it were possible 

 to have each individual of a flock of birds hatched April 

 1, begin to lay on a definite date, say December 1— the 

 birds would make records which would differ from each 

 other in the proportions given by the graphs of rates of 

 production. The fair degree of homogeneity of the flocks 

 in respect to rate of production is shown in part by the 

 coefficient of correlation "between the number of days from 

 the first egg laid up to March 1, and the number of eggs 

 produced during that period. The coefficient was found 

 to have a high value, i. e., for 1913-14, r= + .8618 + 

 .0125 and for 1915-16, r = + .7878 ± .0128. That is, the 

 number of eggs laid is a fairly definite function of the 

 length of the laying period. These coefficients are a 

 rough index of the amount of homogeneity in the flock 

 respecting the rate of production, since a high coefficient 

 implies a fair amount of homogeneity in the flock (cf., 

 however, the statistical constants for rate) for if an egg a 

 day is taken to represent the maximum production while 



