214 Bulletin 397 



the results of fundamental experiments can be determined. He states 

 that there is no royal road to this knowledge, and no shorter cut than 

 actual trials with the kind of stock the farmer has to feed under the con- 

 ditions prevailing on his farm. 



That feeding trials with a given ration or condition have a limited appli- 

 cation to other rations and conditions, is recognized. It is recognized 

 also that the cost of properly conducted trials is high in proportion to 

 the results obtained. Therefore, in studying the feeding value of a given 

 feed or ration, it is essential to first study carefully the teachings of pure 

 science and the results of any experimental work having a bearing on the 

 problem, either in human or in animal nutrition. Next, feeding trials 

 with small animals should in many cases be undertaken. Finally, if 

 previous study and work have justified it, the actual trial with the animals 

 in question should be made. Knowledge of the processes taking place 

 in the animal body is at present too fragmentary to enable one to reason 

 infallibly from pure science to feeding practice, or even from experiments 

 with small animals to the feeding and management of the dairy cow. 

 It is granted that there are more refined methods than the feeding trial, 

 and it is believed that the former should precede and in many cases make 

 unnecessary the latter. However, for the final proof of theories and of 

 the validity of reasoning from small-animal experiments, the practical 

 feeding experiment must serve until more is known of the fundamentals 

 of animal nutrition. 



Feeding trials have been criticized on the ground that they have been 

 poorly conducted. It is charged that 'work along this line has fallen 

 behind other branches of agricultural science, due to a lack of fundamental 

 knowledge and a disregard of the teachings of pure science. Some critics 

 have advocated the abandonment of feeding tests by the experiment stations 

 for the reason that they can be carried out better by the farmers themselves. 

 That these criticisms are in many instances just must be admitted. Inas- 

 much as there is still a place for the feeding test, it is evident that investi- 

 gators should so improve their methods as not to invalidate their con- 

 clusions by faulty procedures or fallacious interpretation of results. 



FACTORS TO BE CONSIDERED IN PLANNING A FEEDING EXPERIMENT 



In planning a feeding experiment the investigator aims to have equality 

 in all the conditions except one, which is the factor to be studied. Because 

 of the complexity of biological processes this is practically an impossible 

 goal. In discussing experiments with living organisms, Blackman (1905) 

 states that the investigator who tries to evaluate exactly the effect of 

 a single factor on a multi-conditioned metabolic process faces an almost 

 impossible task. 



324 



