Refinement of Feeding Experiments for Milk Production 215 



Controllable factors 



In a feeding experiment certain factors can be definitely controlled — 

 such as the kind and quantity of the ration, the method and time of feed- 

 ing, the general care and management of the animals. Certain other 

 factors cannot be entirely controlled because they are due to the individuality 

 of the animals ; however, the effect of these factors may be greatly lessened 

 by so choosing from animals of known history as to obtain a group as 

 nearly alike as possible with respect to records and previous treatment. 

 It is unnecessary to enter here into a discussion of the factors that may 

 be eliminated by a careful selection of animals. A thoro treatment of 

 this subject has been given in the report of the American Society of 

 Animal Nutrition (1909), and by Waters (191 2), Morse (1913), Mitchell 

 and Grindley (19:3), and Grindley (191 5). 



Uncontrollable factors 



When as many factors as possible have been eliminated or equalized 

 by a proper selection of animals, there still remain certain causes of vari- 

 ation which cannot be eliminated. These causes comprise feeding capacity, 

 productive capacity, physiological peculiarities, and other functional 

 characteristics rendering one animal distinct from another. Thus it 

 is seen that the general proposition • — equalization of all factors except 

 one — is impossible of realization. There must always be certain vari- 

 ables, due to individuality, which cannot be controlled. However, 

 the probable limits of variation of these factors may be determined by 

 the application of statistical methods. Thus, in planning feeding experi- 

 ments, one should make use of statistical data which have been accu- 

 mulated as to the probable error due to individuality. At the end of 

 the experiment a statistical analysis of the results should be made, before 

 any interpretation is placed on them. 



VARIABLE FACTORS WITH SPECIAL REFERENCE TO MILK PRODUCTION 



In addition to the factors that have been discussed as causing vari- 

 ation in any feed experiment, there are certain specific variables which 

 become operative when trials with milch cows are concerned, due to the 

 physiological processes involved. These specific factors make the prose- 

 cution of an experiment in milk production more difficult than one in 

 which gain in weight is the object desired, not only as regards the selection 

 of animals and the control of conditions but also in respect to the inter- 

 pretation of results. It seems desirable to enter into some discussion 

 of- these factors. 



325 



