254 



EARL W. BENJAMIN 



In figure 27 the data for the three characters are combined both by a 

 weighted and by an ordinary average. This figure does not show the 

 definite tendency toward a reduction in the "variability which is claimed 

 by some other investigators (Pearl, 1909), altho if the data for the 

 fifth and sixth years could be given as much value in this discussion as 

 is given to the data for the first four years, a straight line fitted to the 

 curve would show a distinct reduction in the standard deviation. The 

 unweighted averages are shown in figure 27 because if it can be considered 

 that the standard deviations calculated for the respective character 



of lhree characters, 



-FlG. 27. STANDARD DEVIATIONS OF ALL CHARACTERS FOR SUCCESSIVE YEARS 



The figures in parenthesis designate the number of birds available for the respective calculations 



groups are based on a sufficient number of individuals to be trustworthy, 

 none of these characters should be handicapped in the average if it hap- 

 pened that a less number of individuals were available for that particular 

 character than for the others. This is especially true since this is a 

 comparison of standard deviations based on a grouping according to classes 

 of widely different values. 



Variations in types of eggs produced during successive months and years 



The study of the variations of eggs produced during successive months 

 and years was carried on with birds that began to lay in different months, 

 as noted in table 58. 



The time of beginning to lay is varied enough in the data used here to 

 nearly eliminate seasonal effects. 



