FERTILIZER CONTROL STATION. 



389 



Feces From 



Timothy hay, sheep 1. 



€t tt <( 2. 



Clover hay, sheep 1 . . 



** << t( 2 .. 



Oat straw, sheep 1.... 



it « (( 2.... 



Oat straw.. .. 

 Kaw potatoes 



*€ tt 



i Sheep 



1, 



" 2. 



The Errors of Feeding Experiments. 



Conclusions in regard to the relative value of different rations for 

 growth, when based upon changes in the live weight of animals, are 

 criticised on the ground that the daily variations in weight of an 

 animal are very large, and may cause the change in weight between 

 two given weighings to be much more or less than the actual growth 

 which the food has caused. These chang;es in weis:ht not due to 

 growth are caused largely by variations in the contents of the stomach 

 and intestines and in the amount of water drank. The extent of the 

 error from this cause can be partially controlled bj' the exercise of 

 care in giving animals food and water at regular intervals, and by 

 taking all the weights at the same time of day. 



In the Station experiments, besides precautions of this kind, the 

 weights upon which gain or loss is based are the average of four 

 weighings on as many consecutive days. 



In order to show the extent of the error which the apparent re- 

 sults of these feeding trials are likely to involve, all the daily weigh- 

 ings of the ten steers, 120 in all, are given in this connection, with 

 the averages of each set of four weighings. 



