EXPERIMENTS SHOULD HAVE A DEFINITE OBJECT. 45 



who, with proper assistance, might be willing to devote them- 

 selves to this most hopeful means of advancing scientific agri- 

 culture. 



It is not my purpose here to advocate such an establishment, 

 but practical agriculture might most profitably tax itself for the 

 maintenance of at least one such experimental garden in each 

 of the main agricultural, climatic, or geological divisions of our 

 island. 



5. Experiments should be made with a view to a definite object 

 or end. Evils arising from badly conducted experiments. 



Among the philosophical chemists of the past century, none 

 was more zealous and more assiduous in original experimenting 

 than Dr Priestley. Yet he is described by high cotemporary 

 authority as groping after novelties. To experiment in the 

 hope or on the chance of stumbling upon something new, with- 

 out a definite and special end or purpose, is to grope in the 

 dark, and cannot lead straightforward to any important truth. 

 Much of the agricultural experimenting hitherto performed has 

 been of this groping character performed without a clear idea 

 of the point or points to be made out or established by it. All 

 experiments, therefore, ought to be contrived and executed 



1. With a view to a definite object ; to throw light upon 

 some special point; to remove some doubt, or to solve some 

 recognised practical or theoretical difficulty. 



2. This object should be one which the person to whom the 

 execution or superintendence of the experiment is intrusted can 

 clearly understand. The conductor must also comprehend the 

 nature of the precautions necessary to secure sufficient accuracy 

 in the result, why these precautions are necessary, and the 

 parts or points in the observations made, or results obtained, to 

 which the greatest importance attaches. Thus, while much 

 knowledge and a clear conception are required in the designer 

 of an experiment, no little cleverness, also, and precision, are 

 indispensable in the conductor of it. 



3. All experiments must be comparative. If the effect of a 

 certain influence of a special manure, for example is to be 

 ascertained, that effect must be compared with what is seen 



