552 EXPERIMENT STATION RECORD. 



proper use of text-books and works of reference should be given more 

 attention, and quantitative relations should be insisted upon early in 

 the course. The course should be shaped with a view to developing 

 the technical sense; that is, the ability to determine the best course of 

 procednre in different cases. When this has been accomplished, the 

 student is prepared to undertake original investigations, but in the 

 beginning these should be simple and of his own choosing, rather than 

 cooperation with the teacher in some difficult and complicated research. 



Discussion on the question " What should be an implement test?" 

 was opened with a paper by E. Davenport, of Illinois. He stated the 

 three considerations affecting the value of an implement to be (1) effi- 

 ciency, (2) draft, and (3) durability. That machine should be considered 

 the best which does the best work under average conditions. In 

 determining the efficiency the quality as well as the quantity of the 

 work should be taken into account, and the durability of the whole 

 machine is to be measured by that of its weakest part. The dyna- 

 mometer test is the one most frequently applied, but ''it can teach us 

 nothing more than the expense of operating, or give indication whether 

 the demand for power will be steady or unsteady, whether it will be 

 easy or hard upon the team. Of all the tests it is the easiest applied, 

 and while valuable it seems that every other consideration is even more 

 valuable." A sharp distinction is drawn between a machine test proper 

 and a field trial. The latter is more valuable because it tests the 

 machine in a great variety of conditions and takes into consideration 

 its natural life. The speaker strongly recommended "the examination 

 of old machines that have been under fair management to discover tlie 

 parts that are first suffering. This, with a field trial for quality of work 

 and ease of draft, is about all that can be done, and will in most cases 

 constitute a fairly satisfactory test, though greatly inferior to what the 

 companies themselves are doing, for commercial reasons." 



The discussion was continued by E. J. Kedding, of Georgia, who 

 stated that " (1) it should be an inflexible rule that no machinery or 

 implement or appliance be tested primarily in the interest or for the 

 benefit of competitive manufacturers or dealers; (2) in general, the 

 expediency of a suggested test, with reference to its bearing upon agri- 

 culture, should be affirmatively determined upon by the director or other 

 officer in charge of the station, and (3) with few exceptions machinery 

 to be tested should be such as the station is jjrepared to use regularly, 

 or at least occasionally, in conducting the experiments and investiga- 

 tions in the usual course." 



The form in which the results of these and similar tests should be 

 published was discussed by C. S. Plumb, H. J. Wheeler, H, J. Patter- 

 son, E. B. Voorhees, E. J. Eedding, and James Wilson. The consensus 

 of opinion seemed to be that the essential results of these tests should 

 be given to the public, whether they were of such a nature that they 

 could be used for advertising purj)oses or not, although great caution 

 should be exercised in the matter. 



