304 



BOARD OF AGRICULTURE. 



Prof. Swallow. The point I wish to make is this : to show 

 that twenty experiments, with twenty elements of error in them, 

 are not so good as two with no elements of error. That is the 

 idea. We have forty colleges in this country making the same 

 experiments, and in one-half of them there is an element of error 

 which we know must be there, and the result will not be so valua- 

 ble as if ten of these colleges were making the experiments with- 

 out this element of error. No man would accept it as conclusive. 

 You may as well say you could feed one animal on wheat and one 

 on corn, and the results would be the same. I want to get at 

 results as soon as possible. My idea was, we could get at some 

 results more rapidly by having a sufficient number of colleges take 

 up a certain class of experiments in which they could experiment 

 without any necessary element of error, that we know of. I say 

 you would get no more accurate results by having other colleges 

 perform the same experiments. 



There is a popular idea that we must do something for the 

 public, and I think we can do a great deal. I have been accus- 

 tomed to say, for a great many years, there is something for agri- 

 cultural colleges to teach, even now. There is a vast amount of 

 knowledge that our farmers do not possess, and which the colleges 

 ought to give at once. I would not be misunderstood in this. I 

 am for going ahead and finding out something else, while we are 

 doing this. 



Dr. Miles. Prof. Swallow and myself are agreed that we are 

 anxious to get at the result as soon as possible. We only differ 

 slightly as to the way. One object I had in proposing the method 

 of conducting the experiments was, that it would save time and 

 get at the results sooner. I agree with the Professor that it 

 would be better to have two experiments without any error, than 

 twenty with error. I will go further than that. I would rather 

 have one without error than twenty containing error. The diffi- 

 culty is this, however, and is one we must face : We can not 

 make an experiment in feeding, without it contains an error. But 

 if we co-operate in trying the experiment, we will sooner get at 

 it, owing to the manner in which the experiment is conducted. I 

 might experiment this year and get one result, and another next 

 year. I do not know whether the variation is owing to differences 

 in climate or not. If it had been tried at two different points 

 under the same condition, it would be better than the same exper- 

 iments tried two years in succession on the same ground. If the 



