Nos. 618-619] 



EGG PRODUCTION 



nant character such as rose comb. It is quite possible 

 that while our flock of Rhode Island Reds as a whole is 

 homozygous for high fecundity, it is not entirely so. It 

 is possible, of course, that many of our Rhode Island 

 Reds are genetically true mediocre producers, but that 

 the addition of some other genetic factor has so altered 

 the rate of production that the records can no longer be 

 recognized as true mediocre records. 



It seems clear, then, that our Rhode Island Reds fall 

 into the class of high producers observed in Barred 

 Plymouth Rocks, but that the great variability in maturity 

 results in a portion of the flock giving numerical results 

 like those of the Barred Plymouth Rocks, but which in 

 reality are not at all equivalent biologically. It is also 

 clear why Castle's ('15, '16) recent criticism of Pearl's 

 theory of egg production can not be considered to be es- 

 tablished, in so far as it concerns winter egg prod action. 



Moreover, it will be desirable in the future to distin- 

 guish clearly between the two sorts of numerical results. 

 A discussion of the bearing of a division point at 30 eggs 

 upon the problem is given below. 



It has been maintained by Pearl that winter egg pro- 

 duction is a satisfactory basis for selection, both for itself 

 and as an index of the annual egg production. This 

 would be true only when there is a definite relation be- 

 tween winter and annual production such as he has found 

 for Barred Plymouth Rocks. Since winter egg produc- 

 tion forms a part of the annual egg production with which 

 it is to be compared, it is evident that the coefficient of 

 correlation may be expected to have a positive value of 

 some magnitude, unless there is a definite tendency for 

 birds that lay well in the winter to be poor layers in the 

 summer. On this account it has seemed best to us to 

 compare winter production with that of the same hens 

 for the remainder of the year. The number of birds (77) 

 available for the comparison was not large; but the cal- 

 culated coefficient (.365 + .067), while statistically sig- 

 nificant, being six times its probable error, is too small 



