214 STORRS AGRICULTURAL EXPERIMENT STATION. 
Another kind of experimenting and one which though its chief 
usefulness will be of a purely practical sort, will also bring 
results of physiological value, consists in feeding trials such as 
may be conducted in any well equipped experiment station in 
connection with a chemical laboratory. If, for instance, the sub- 
ject be the effect of fodder upon milk production, rations differ- 
ing in amount and composition may be fed and the amount and 
composition of milk determined. The difficulty with these 
experiments is that there are numerous other factors besides the 
food which affect the milk production and not all of these 
factors can be accurately estimated. But there is one way by 
which such experiments can be made very effective and useful. 
That is, duplication. One station alone cannot well do enough 
to settle any specific question, but by. codperation between 
different experiment stations a great deal could be accomplished. 
To make this codperation successful it would be necessary to 
select narrow and specific questions, to decide upon a uniform 
plan, to use as many animals in each case as practicable and to 
make a large number of experiments in a number of different 
places. There is very little question that this would be the most 
economical and useful way to work. ‘The dairy trials at the 
World’s Fair brought results some of which were reasonably 
definite. In consequence they are of very great practical value. 
What gives them definiteness and value is that a large number of 
cows were used. It is not necessary that all the cows be in one 
place, but the “‘might of average figures’? which can be almost | 
said to ‘‘increase as the square of the number” is what makes 
the results reliable 
The series of codperative experiments which have been con- 
ducted during the past twenty years in Germany and Denmark 
are a most valuable illustration of the usefulness of one way in 
which extensive feeding trials may be made useful.* 
Both these kinds of experimenting are needed to give us a real 
physiological feeding standard, and when we get that standard we 
shall doubtless find that it is after all indefinite, that it varies 
with the animal as well as with the conditions of feeding; in 
other words, it will be at best Sh an average estimate and not 
an unvarying formula. 
FORMULAS FOR PROFITABLE FEEDING. 
But the practical feeder feeds for profit, and the ration which 
will produce the largest amount of growth, or of total flesh or of 

* See Experiment Station Record, Vol. III., pp. 507, 557, and Vol. VI.; p. 58s. 

