144 Maine Agricultural Experiment Station. 1920. 



checked the results and found it correct). From unpublished 

 data of the writer the correlation of advanced registry Guern- 

 sey cattle for these same variables is — 0.1174+. 0134. The cor- 

 relation for the Jersey is equal to — 0.1126;+: .01 61. Two of 

 these correlations are based on advanced registry data and may 

 be considered as subject to . a selective influence on the data. 

 The distributions do not look as if such a disturbing factor had 

 been present as there is no evidence of truncation and as shown 

 in a previous part of the paper the frequency constants agree 

 quite well with those of the Jerseys and Ayrshires known to be 

 untruncated. It seems therefore that the constants above should 

 be directly comparable as to the relation of age and buter-fat 

 percentage of these breeds. 



The Jersey correlation coefficients do agree very closely 

 with those of the Guernseys. The Ayrshire do not agree at all 

 with any of the other breeds for difference of the correlations of 

 Ayrshire and Jersey is 0.1618^.0301 or 5.4 times its probable 

 error. The multiple times the probable error is greater for the 

 Holstein-Friesians. The difference of the Holstein-Friesian cor- 

 relation from that of the Jerseys is probably not significant as 

 it is only 2.4 times the probable error. The correlation for the 

 Holstein-Friesian age and percentage of butter-fat produced is 

 probably not significant. 



The Ayrshire results are obtained under the conditions of 

 Scotland whereas, the other results, are on cattle kept in this 

 country. This may possibly account for the difference in influ- 

 ence of age on butter-fat concentration of Ayrshire cows' as 

 compared with these other breeds or it may equally well mean 

 that the Ayrshires are innately different from the other breeds. 



The correlation for the Holstein-Friesian in comparison 

 with the correlation for the other breeds is small. It does show 

 the same sign as the other correlations. 



These considerations taken together lead to the following 

 conclusion which may be expressed tentatively as follows : each 

 increment of time added to a cow's life causes a slight decline 

 in the concentration of butter-fat that the cow's mammary gland 

 can secrete into the milk. 



In the bulletin following this, the relation of the butter-fat 

 percentage of one lactation to the butter-fat percentage of a sub- 

 sequent lactation will be analyzed using the records from this 

 same herd. 



