102 EXPERIMENT STATION RECOKD. 



permanent or widely applicable value. Hence it is that the neces- 

 sity is felt for going over much the same ground at frequent intervals 

 and in different localities. And while this is not without value to 

 the farmer it often represents an unnecessary waste of effort, and 

 stands in the way of what might mark more real progress. Essen- 

 tially the same kind of experiments are often repeated by stations in 

 the same general locality, and with full knowledge of such repeti- 

 tion, as was the case a few years ago when silage was being tested 

 for beef production. The result in such cases becomes largely a local 

 demonstration rather than the acquisition of new information. In 

 fact, the statement has often been made that the work' was done to 

 convince the farmers of the locality of the truthfulness or applica- 

 tion of work in other States, the thought being that they had a 

 different feeling if the experiments were made under their conditions. 



There is undoubtedly much merit in the conventional feeding ex- 

 periments and in experiments which interpret the best experience of 

 the locality. But such experiments should profit by what has gone 

 before, and should show improvement in method and in the extent 

 to which the results contribute to a more complete understanding 

 of the general subject. Manifestly the experimental results must 

 be secured under such conditions as to insure accuracy and reliability 

 within reasonable limits, and to make possible the comparison of 

 the results with other experiments. At present there is the widest 

 variation in experiments of this class. Between the feeding trial 

 that deals only with the gross effect as measured by lots, and the 

 more refined experiment which carefully guards and controls the 

 conditions and results as applied to individual animals, and seeks 

 the reason in the changes which actually take place, there is a wide 

 gap. One is the rough comparison such as a feeder might make, if 

 he had the time, and the other represents an attempt to trace the true 

 relation between cause and effect. 



It would seem that we should have largely passed the stage of the 

 first type mentioned, but the publications and records of work in 

 progress do not show this to be the case. Such trials, with all their 

 crudeness, meet a popular demand and this demand is being acceded 

 to despite the development of demonstration and extension work. 

 Unfortunately there are some indications that the latter is already 

 constituting a new demand for superficial vv^ork. This more direct 

 teaching of the farmers brings out local problems in increasing num- 

 bers, and makes an urgent call for very practical and didactic direc- 

 tions which have behind them the force of local experimental trials. 



But the experiment station can not afford to look at the subject 

 of feeding from the superficial and local standpoint, and it is hardly 

 its function to make experiments to demonstrate locally what is al- 

 ready known. Extension work will be an actual disadvantage to 



