366 EXPERIMENTS WITH FERTILISERS [chap. 



attacks, and preparation of the land ; errors of weighing 

 and measuring are also proportionally reduced by being 

 spread over the larger quantities involved. On the 

 other hand, the larger plots mean greater risks of 

 meeting with irregular patches of soil, and much greater 

 difficulty is experienced in getting the cultivation of all 

 the plots carried out under uniform conditions. It is of 

 the first importance that the whole of the experimental 

 land should be worked on the same day ; autumn 

 ploughings perhaps matter least, but spring ploughings 

 and cultivations, and above all seeding, should be 

 carried through in a single day. Otherwise, if part of 

 the land is worked and left and then the weather 

 changes, a considerable interval may elapse before the 

 operation can be completed, and a new factor, often of 

 considerable magnitude, is thus introduced into the 

 results. Sometimes large plots are necessary to 

 obtain sufficient material for further investigation ; 

 account, too, should be taken of the facilities for 

 weighing up the crop ; if no weighbridge large enough 

 to take a cart is available on the farm it is difficult 

 to deal with large areas. Speaking generally, it may 

 be said that with due care a plot one-twentieth of an 

 acre can be- made to answer all ordinary purposes. 

 But whether large or small the most important point is 

 to repeat the plots on some regular system about the 

 ground, and to have four or five similar plots of one- 

 twentieth of an acre for each treatment rather than one 

 of a fifth or a quarter of an acre. In the Danish 

 experiments conducted by Dr Sonne upon the relative 

 value of different varieties and management of barley, 

 which may be taken as the most carefully elaborated 

 series of field trials for practical purposes which have 

 ever been carried out, the plots are about ^V ^cre each, 

 and at any one station there are always four plots 



