THE REFINEMENT OF FEEDING EXPERIMENTS 



FOR MILK PRODUCTION 



BY THE APPLICATION OF STATISTICAL METHODS 



L. A. Maynard and W. I. Myers 



The t3''pe of feeding experiment considered in this bulletin is that which 

 involves a comparison of different feeds, in different combinations and in 

 varying amounts, when fed to milch cows of different breeds, ages, and 

 conditions. Such " practical " feeding trials have occupied a large place 

 in the experimental work thus far performed in animal nutrition, and 

 have contributed considerably to the art of feeding. In recent years, 

 however, much criticism has developed regarding this type of study, 

 and there are those who believe such work should be discontinued. The 

 authors believe there is still a place for the practical feeding trial. How- 

 ever, no one is justified in conducting such an experiment without a 

 thoro knowledge of the factors involved, of the limitations of present 

 methods, and of the uncertainty attendant on the interpretation of results, 

 due to individuality. 



The subject matter of this bulletin comprises the results of a study 

 along the foregoing lines, as preliminary to the beginning of some feeding 

 trials with dairy cows at this station. It seemed that the data obtained 

 should prove of some value to others who are conducting feeding trials. 



STUDY OF METHODS OF CONDUCTING FEEDING EXPERIMENTS 

 THE PRACTICAL FEEDING TRIAL, ITS LIMITATIONS AND ITS VALUE 



Feeding trials have been criticized on the ground that as means of 

 solving problems in animal nutrition they have outlived their usefulness. 

 McCollum (1916)^ has pointed out that the results have been meager 

 in proportion to the time and money expended. He states that it is the 

 opinion of many animal husbandry men of wide experience that the 

 present method of trying this or that concentrate as compared with 

 another has taught about all it ever will teach. Evidently it is believed 

 that there are better methods which should supersede the feeding trial. 

 However, this view does not seem to be shared by the majority of inves- 

 tigators. Grindley (191 5) is of the opinion that properly conducted 

 feeding experiments are the most direct means of attacking many of the 

 problems of the livestock farmer. Waters (191 2) points out that feeding 

 trials are the only means by which the application to farm practice of 



I Dates in parenthesis refer to bibliography, page 248. 



