638 



IOWA DEPARTMENT OF AGRICULTURE 



iati'ons was determined by comparing the yields as estimated by them 

 and outlined in this publication with the actual yields as determined by 

 weighing and testing each milking. 



The accuracy was also determined of the estimated yearly yields as 

 calculated by each of the following methods: (1) Taking the weights 

 and a test of the composite sample of eight milkings in the middle of 

 each month; (2) taking the same for four milkings in the middle of each 

 month; (3) taking the product of one day in the middle of each three- 

 week period; (4) taking the product of one day in the middle of each two- 

 week period. In each of these methods the yields for each period were 

 estimated and the sum of the periods was taken as the total for the year. 



The following table shows the variations from the actual yields obtained 

 by each of these five methods, the percentages of difference above or 

 below the actual yields being shown in each case. The maximum varia- 

 tion by the cow-testing association method was not over 5.1 per cent for 

 any one year, and this method compares favorably for accuracy with 

 any of the others. The table shows the maximum per cent of variation 

 for any one year and the per cent of total difference from the actual yield 

 for nine years for each of the methods. 



VARIATIONS FROM ACTUAL YIELD OF ESTIMATED YIELD OF MILK AND BUTTER- 

 FAT AS DETERMINED BY VARIOUS METHODS. 



