ENGLISH AGRICULTURE. 179 



had just carried a corn crop, or better still, one that had 

 carried two corn crops in succession. Such soils are in a 

 condition to respond to applications of fertilizers, while the 

 unmanured plots would be likely to yield small crops, which 

 would contrast distinctly with the dressed plots. Secondly, 

 the experiments must be simple, so that we may clearly 

 see the meaning of the result arrived at. We must not ask 

 Nature two questions at once, but a simple, straightforward 

 question. The issue must lie between superphosphate and 

 nothing, or between two hundredweight of superphosphate 

 and four hundredweight of the same fertilizers, or between 

 basic cinder and superphosphate, &c. &c. In order to secure 

 decisive answers, the previous cropping and the previous 

 treatment of all the plots must have been precisely similar. 

 The plots must not extend over soils of two or more sorts, 

 or over land some of which has had an extra cultivation, or 

 yet over soil which has carried more than one and the same 

 crop, or some of which has been dunged. Strict uniformity 

 of conditions must be obtained. Similarly in sowing, hoeing, 

 harrowing, rolling, and any other processes during the growth 

 of the crop, all must be treated alike so as to preserve the 

 simplicity of the issue throughout the whole period. We 

 should always have unmanured plots for the sake of com- 

 parison. In experimenting I find it better to have several 

 unmanured plots, and if practicable an unmanured plot 

 alternating with every manured plot. This may be readily 

 done if we arrange the experimental plots in the form 

 of a series of squares each of one-tenth or one-twentieth 

 of an acre. By a proper distribution of unmanured plots 

 we can collect together the average results of our un- 

 manured plots, and we can collect together our average 

 results from the application of superphosphate, ammonia 

 salts, nitrate of soda, basic cinder, or whatever else we may 

 happen to be experimenting upon. Several unmanured plots 





