124 Maine Agricultural Experiment Station. 1920. 



The correlations between the milk production of various 

 lactations with one another has material importance as well as 

 theoretical interest for these problems. If the correlation is high 

 it is evident from the practical side that the culling of the herd 

 through the use of dependable criteria will result in increase in 

 the milk yield per cow within the herd and in the increase of 

 profits to its owners. From the biological side, a high correla- 

 tion means that the animals composing the herd innately differ- 

 entiated (presumably due to their inherited complex) in their 

 mammary capacities. 



The determination of the correlations for these data are of 

 especial interest for as previously shown in the earlier part of 

 this paper, the data are of exceptional value in that; they are 

 on a pure bred Jersey herd kept intact for many years ; the lac- 

 tation records for several lactations are recorded on a number 

 of the animals; and the herd has been subject to no detectable 

 culling based on the production of a given lactation. From these 

 data there have been extracted lactation records to the number 

 of 3178 pairs having a full eight months lactation free from any 

 disease or sickness or other trouble known to influence the rec- 

 ords. These records have been formed into twenty-eight corre- 

 lation tables the most of which are of considerable size. 



Table 1 gives the correlations and their probable errors for 

 all ages into which the lactation records were divided. The ver- 

 tical columns give the correlations of the age heading the col- 

 umn with the ages indicated at the left margin of the table. - As 

 will be noted the correlations necessary to give the complete set 

 of correlation for any given age are repeated e. g. the correla- 

 tion of 2 years with 3 years is +0.5764^.0332 and appears in 

 the 2 year column. The correlation of 3 years with 2 years will, 

 of course, be the same (0.5764112.0332) and is repeated in the 

 three year column. In this manner a complete picture of the 

 relationship of the lactations of the 2 year group with the other 

 lactations of the other groups is given in the column. 



The order of magnitude of these correlations is from 

 +0.7306 for the correlation of 8 months milk productions dur- 

 ing the ages 5 to 6 and 6 to 7 and +0.2144 for the correlation 

 of the productions during the ages 4 to 5 and 10 and older. Such 

 correlations indicate that the milk production of one lactation 

 may be predicted with relatively little inaccuracy from the milk 

 production of another lactation. 



