68 
Journal of Agricultural Research voi. xiv. no. > 
Wilson did not handle his data in such a way as to bring out the rela- 
tionship which exists between the quantity and quality. In "The 
Principles of Stock-Breeding," Wilson ^ again writes: 
In connection mth yield and quality in milk, it has been assumed, frequently, 
that the two characters are interdependent : that when the one is high the other must 
be low. It has been found that this is not so. The characters are independent and 
have no effective influence upon each other. High quality of milk is found among 
cows giving all kinds of yield, and low quality is found similarly. 
It seemed to the writer that it might be of some value to make a more 
careful statistical investigation of this question with our American 
cattle. 
SOURCE OF DATA 
In the registers of the different American associations is to be found a 
large body of data which furnished the major part of the material for 
this investigation. There are involved in this study the following: 
2,141 yearly tests of Jerseys, Register of Merit, 1911, 1913; 3,564 
Guernseys, Guernsey Breeders' Journal, May, 1915; 1,925 Holstein- 
Friesians, Holstein-Friesian Advanced Register Year Book, volumes 
21-26; 1,091 Ayrshires, Year Book of the Ayrshire Breeders' Association, 
1907, 1911, 1913, 1914; 98 Ayrshires^; 750 grade Jerseys^ and 341 
grade Holstein-Friesians ; and 2,002 * yearly tests of cows unclassified 
as to breed. 
Only the yearly tests were used for the reason that a yearly record is 
a more reliable criterion of a cow's performance and ability than a shorter 
test. It should be pointed out here, that in the case of the records from 
the associations, selected groups of individuals are involved in this 
study, since only selected individuals are subject to entry in the registers 
of the associations. 
The method of finding the relation between the percentage of fat and 
the yield of milk is by means of the correlation table. The cows are 
grouped, according to age when the test began, into the following groups: 
2 to 3 years, 3 to 4 years, 4 to 5 years, and 5 years and over. The last 
group comprises what are usually held to be mature cows. These are 
not exact divisions according to age, since a given group may contain 
individuals differing in age by almost a year. For example, the 3 -to-4-year 
group contains those cows with tests beginning at some time after they 
were 3 and before they were 4 years old. A cow with a test beginning 
the day she was 3 would be practically a year younger than one having 
a test starting when she was one day under 4, though both would be 
classed in the same group. Of course, there are few cases of this kind. 
' W11.SON, James, the principi,es of stock-breeding, p. 121-122. London, 1912. 
^ Furnished by Jlr. C. M. Winslow, Secretary of the Ayrshire Breeders' Association. 
3 Obtained from Mr. W. W. Yapp, Illinois Agricultural Experiment Station. 
* Obtained Irom Prof. W. J. Fraser, Illinois Agricultural Experiment Station. 
