SELECTION AND IMPROVEMENT OF DAIRY CATTLE 61 



Th( reading gives the butter- fat per 100 pounds of milk. 

 When butter is made, some curd, salt, and considerable 

 water remain with the fat to meke up normal marketable 

 butter. The amount of butter generally exceeds the fat 

 by ibout one-sixth and may be estimated if desired by add- 

 ing this amount to the fat. 



49. Averaging Tests. A common mistake results from 

 averaging tests. A direct average of the tests made for a 

 certain cow each month during the year will not be a fair 

 avc rage test of the total amount of milk produced during 

 the year by this animal. This results from the fact that 

 the amount of milk represented by each test is not the same. 

 To find the true average test for the year the total fat yield 

 for the period covered by the tests must be divided by the 

 total yield of milk. The same method must be used in get- 

 ting the average test for the herd. 



50. Permanent Records. Many who begin keeping records 

 do not make much of a success on account of not having a 

 suitable form for a permanent record. Fig. 22 is a form 

 that has been used with good satisfaction by the author for 

 a number of years. 



Using a blank book, a page, may be ruled for each cow. 

 Space sufficient for several years may be provided in this 

 way in a concise form. In pure-bred herds the pedigree and 

 records of offspring may be put on the page opposite the 

 milk records. 



51. Cow-testing Associations. Keeping records, es- 

 pecially of a large herd, involves considerable attention to 

 details. To provide for this, cooperative cow-testing associ- 

 ations have been established in many places. This plan 

 originated in Denmark in 1895, and the number of associ- 

 ations has since increased with great rapidity. At present 



