266 HARRIS, KIRKPATRICK, BLAKESLEE, WARNER AND CARD . 



From the purely scientific side the interrelationship of the egg pro 

 duction of different periods is a problem of great biological interest. From 

 the economic side, two ends which may be either quite distinct or inter 

 dependent, are to be attained by the development and application of 

 formulae for the prediction of the egg production of a bird during any 

 period of the pullet year. The first is the determination of the probable 

 future record of an individual bird from her past performance, as a basis 

 for the decision as to whether she shall be kept for egg production or sold 

 for meat. The second is the estimation of the annual record of a bird 

 as a basis of decision as to whether or not she shall be kept until the fol 

 lowing season to be used as a breeder. 



It will be evident to those who have had to consider the problems with 

 which we have to deal, that economic factors, particularly the cost of 

 trap-nesting, and the purpose for which prediction is being made will 

 have great weight in determining the period and the number of periods 

 to be used in the prediction equations. In determining which birds shall 

 be sold to the commission man and which may be fed with reasonable 

 prospects of profitable returns for the remainder of the year, the breeder 

 is not concerned primarily with the record which the bird makes for the 

 year as a whole. Practically he requires to know what returns she will 

 make for the remainder of the period over which she may be retained in 

 the flock. He has already maintained her for n of the 12 months. The 

 question which the poultryman would like to have answered is whether 

 her record during this period has been such that he can afford to feed and 

 house her for the remaining 12 n months. It is evident that to be of the 

 greatest value for this purpose the prediction should be made from periods 

 as early as possible in the life of the bird. In other words, if birds are to 

 be culled out of the flock and sold for their flesh because they are unprofit 

 able as egg producers, this should be done at a time when the maximum 

 saving in cost of maintenance can be coupled as closely as possible with 

 the maximum sale price. 



The correlations required for this purpose are, therefore, those between 

 the record of any period which may be selected as a basis for judgment, 

 and the record of later months. 



Since in the selection of birds to be held over for breeders the total 

 annual production is presumably the factor to be chiefly taken into con 

 sideration it is evident that the correlations to be determined are those 

 between the record of the individual months and of the year as a whole. 

 It would, of course, be better for this purpose if the records of the entire 

 year were known, but as pointed out by CARD (1917, p. 66) many poultry- 1 



