FERTILITY AND HATCHING OF EGGS. I49 



ent generations. We may properly enter one set of individuals 

 on one side of the correlation table, as the primary or x variable 

 and the other set on the other side of the table as the secondary 

 or y variable (cf., Table XXI). When we come to deal with 

 pairs of sisters all of the same age, however, there is no good 

 reason for taking one sister of a pair as the primary or x vari- 

 able rather than the other. Therefore, in such cases it is neces- 

 sary in order to arrive at a correct result to enter each indi- 

 vidual of a pair of sisters twice ; once as the x variable and 

 once as the v variable of the correlation table.* This has been 

 done in forming the two tables XXIV and XXV which show 

 the correlation between sisters in respect to fertility and hatch- 

 ing quality of eggs respectively. In forming these tables every 

 possible pair of sisters has been entered twice, with first one 

 and then the other as the primary variables in the manner 

 indicated. 



From these tables the following coefficients have been 

 calculated : 



Correlation between sisters in respect to per cent, of eggs 

 infertile r =: — o.o64±.o62. 



Correlation between sisters in respect to per cent, of fertile 

 eggs hatched r ^ 0.188 ±.060. 



From these values we note that : 



1. There is no sensible correlation between sisters in respect 

 to the percentage infertility of eggs. A bird having a per- 

 centage of infertile eggs well below the average is as likely as 

 not to have a sister whose percentage infertilitv will be above 

 the average and vice versa. 



2. In respect to the hatching (|uality of eggs (percentage of 

 fertile eggs hatched) there is a definite and sensible correlation 

 between sisters. This means that, i)i (/eiieral, sisters of 

 birds zvhose eggs are aboi'e the average in hatching quality 

 have their eggs also above the average in respect to this char- 

 acter. In other words, the data would indicate that there is a 

 sensible degree of what may be spoken of as "fraternal" or 



* The necessity for dealing witli material of tliis character is the way 

 indicated has been dully discussed by Pearson in his memoir on homo- 

 typosis, (Phil. Trans. Roy. Soc, Vol. 197A, pp. 285-379, 1901), and by 

 Pearl in connection with a study of homogamy in the conjugation of 



Paranieciuni ( Hidinetrika. Vol. 5, pp. 213-207, 1907). 



