COMMERCIAL FERTILISERS 391 



except as to the kind of fertiliser applied. It is 

 best to leave between plots a strip of land at least 

 four feet wide, which is unfertilised; this prevents 

 the fertiliser applied to one crop from affecting the 

 crop in an adjoining plot. The full experiment 

 would look like this: 



PLOT 1 



PLOT 2 



PLOT 3 



PLOT 4 



If it is desired to test the value of different com- 

 binations of plant foods, add : 



PLOT 5 



PLOT 6 



PLOT 7 



PLOT 8 



PLOT 9 



PLOT 10 



