Refinement of Feeding Experiments for Milk Production 229 



Age of cows as affecting variability. — In selecting animals for experi- 

 mental purposes, it is desirable to know whether groups of cows of the 

 same age are more variable or less variable than groups of cows of dif- 

 ferent ages. Groups of cows of the same age were made up and their 

 variability was compared with that of groups made up without regard 

 to age. The results of this study are shown in table 5 ; 



TABLE 5. Effect of Age of Cows on Variability of Groups 



The coefficient of variability of groups of cows of different ages is 

 shown in table 2 (page 226) to be 5.09 ± 0.42. A comparison of this value 

 with those of groups of individuals of the same age shows that the latter 

 are less variable. The difference between the average coefficient for cows 

 of the same age and that for cows of different ages is 2.32 dt 0.51. The 

 odds are about 400 to i that the difference is significant. 



It is not shown in table 5 that cows are less variable at any particular 

 age in so far as the ages covered by the table are concerned. Specificall}^ 

 there is no evidence that heifers are more variable than older animals. 

 This should not, however, be considered an argument for using heifers 

 in a feeding trial. The great disadvantage in using immature animals 

 is that, inasmuch as part of their food is used for growth, no quantitative 

 conclusions can be drawn as to the effect of a given ration on milk 

 production. 



Allowable range in selecting groups of equal production. — Another 

 important consideration in making up groups for feeding trials is concerned 

 with the interpretation of the term equal production. In this connection 

 the questions arise as to how closely the production of two groups must 

 agree to be considered equal, and as to the effect on variability of changes 

 in the range. In order to get information on these points, the records 

 were first divided into groups of six, the total production of each of which 

 fell within the limits of from 3200 to 3299 pounds. These groups were 



339 



