Studies in Milk Secretion 99' 



templated in this paper does not allow for this difference in 

 herds. It is necessary, therefore, to have some means of recog- 

 nizing this difference in the cows to which these bulls are bred. 

 This is easily accomplished by dividing the frequency areas of 

 table i into four equal parts or quartiles. The distribution for 

 these quartiles given in table I. The place of division for each 

 quartile is given in table 2 for the daughter's milk and butter- 

 fat percentage and for the dam's milk and butter-fat percent- 

 age. Figures 24 and 25 give the histograms and the quartiles for 

 the mother's and the daughter's milk and for the mother's and 

 the daughter's butter-fat percentage. 



The quartiles may be designated as follows, A the amount 

 of milk or butter-fat percentage above the third quartile line; 

 B the amount of milk or butter-fat percentage between the 

 median and the third quartile line ; C the amount of milk or 

 the butter- fat percentage between the first quartile line and the 

 median; D the amount of milk or butter-fat percentage below 

 the first quartile line. In this way the change in the milk pro- 

 duction or butter-fat percentage between any dam and her 

 daughter may be expressed by two letters. For instance a rec- 

 ord of the relative milk production AC means that the dam's 

 milk production is over 10403 pounds and the daughter's milk 

 production between 7931 and 8942 pounds. A butter-fat percent- 

 age record DB means that the mother's butter- fat percentage is 

 below 5.235 per cent and the daughter's butter-fat percentage 

 between 5.593 per cent and 5.995 per cent. In recording a bull 

 with two or more daughters it seems best to put these pairs on 

 the basis of one hundred. Thus a bull with two tested daugh- 

 ters out of tested dams, and one dam is in respect to milk pro- 

 duction in class A, her daughter in class B, and the pair re- 

 corded as AB ; the other dam in class C and her daughter in 

 class A and the pair recorded in CA; the record for the bull 

 will then be 50 AB + 50 CA and the number of pairs on which 

 this record is based will be two. The extension of this method 

 allows the recording of any number of pairs. In this manner 

 it is possible to always know how many daughters a bull's rec- 

 ord is based on and where in the general population of daugh- 

 ters these daughters' records lie in comparison with the place 

 in the general population occupied by the records of their 

 dams. 



