EDITORIAL. 709 



tion of definite agricultural knowledge, and the methods of the 

 ordinary isolated experiment can not alone furnish the foundation 

 for a science and a well-rounded theory of American agriculture. 



The line which separates the ordinary experiment from a demon- 

 stration trial is a changing one. Wliat is an experiment in a new 

 field this year may be a subject for demonstration tiials next year 

 or the year to come, and the provision of agencies for demonstration 

 should lead to a closer scrutiny of the experiments conducted by the 

 station, often with the result of elimination. The experimental 

 work and the demonstration work represent separate fields. They 

 should not overlap, and the desire of the station man to test out 

 locally the result of his experimental findings should not lead into 

 a series of purely demonstration trials. 



The field experimenter is justified in repeating his experiments on 

 different types of soil or under different climatic conditions suffi- 

 ciently to test the validity of his deductions in the principal agricul- 

 tural regions of the State ; but it is useless for him to attempt experi- 

 ments on hundreds of farms when the probability is that all the 

 essential features of the experiment will be covered by a very limited 

 number. Indeed, it may well be questioned whether he is within his 

 field as a station man in doing so. One effect of this outlying work, 

 where carefully done, has been to emphasize the distinction between 

 an experiment made to secure information and a demonstration de- 

 signed to impart it. The experiment to be of general value must be 

 made under control conditions, and not be subject to the convenience 

 or judgment of the farmer, who rarely understands the necessity for 

 method and precision. Field experiments which are not thoroughly 

 supervised and controlled by trained experts have little experimental 

 value, and hence stations which are doing careful work make the abso- 

 lute control of the land and direction of the work a condition of 

 local experiments. This is not usually required or thought desirable 

 in the case of demonstration trials, which indicates a distinct differ- 

 ence in the character of the two Ifnes. It might reasonably mark a 

 boundary line between them. 



There comes a time when the result of experiment must be given 

 to the public, through the bulletins and other agencies for getting it 

 before the people. That time comes when the experimenter is reason- 

 ably sure of the accuracy of his results — when his work has jDassed 

 the experimental stage. Then it is ready to be taught. There should 

 be no hesitancy on his part to then pass it over to the extension de- 

 partment through which it will be demonstrated to the farmers of 

 the State. On the contrary, considering the great number of prob- 

 lems needing solution, the station worker ought to welcome the pres- 



