Substitution of Roots for Concentrated Foods. 449 



For the writer's experiment it was decided to use four groups of five 

 cows each and to make each feeding period six weeks in length, one 

 week preHminary and five weeks experimental in each period. In the 

 following pages no account is taken of the data secured during the 

 preliminary week, and only the data from the five-weeks' experimental 

 periods is considered. The feeding system combined in a measure both 

 the continuous and the alternation methods, in that Group A was fed 

 ration I continuously throughout the test periods as a check. This 

 group also served to show the natural decrease in production due to 

 advancing lactation. Groups B, C, and D were alternated on rations 

 I, II, and III. By having three feeding periods, each of the groups 

 B, C, and D was fed each ration for one period, so that comparison could 

 be made of all three rations with each group, and of one group with 

 another and with the check group. Because so many cows (20) were 

 used during the first year and because the quantity of mangels on hand 

 was overestimated, the experiment was carried through only two feeding 

 periods in 1907-8. The number of cows was reduced to 16 for the second 

 year's work and a larger supply of mangels was provided, in order to 

 carry the experiment through three full periods. 



The Cows 



The cows were chosen from the University herd with a view to making 

 the groups as uniform as possible in time of lactation and amount of 

 product. The first year it was planned to use three Holsteins, one Jersey, 

 and one Ayrshire or Shorthorn, not enough animals of one breed being 

 available to make up the required number. Some of the test cows 

 were proved tuberculous when tested at the end of the first period 

 so that some substitutions were made, as is indicated later in the work. 

 For the second year, each group was made up of four cows, three Hol- 

 steins and one Jersey. The Jersey in the check group of the second 

 year was taken sick at the end of the first period so that no account 

 has been taken of her, and the check group during the second year 

 was considered as made up of three Holsteins. This, perhaps, was 

 unfortunate, but no cow was available to substitute for the Jersey. 



The following table sets forth some of the characteristics of the animals 

 used in the experiment both years. 



15 



