CHEMICAL FARMING 221 



valve and then connected with the water supply, 

 keeping the solution at a controlled level. Taking 

 an average plant from one of the several tanks and 

 noting progress, giving more or less of each chem- 

 ical, from one standard solution known to pro- 

 duce good results for that type, the best solution 

 may be ascertained. Tanks of this state may con- 

 tain at one time six or more different types of plants 

 for experimental purposes. If sixty 3x4 foot tanks 

 were all coupled together to the water supply and 

 arranged in rows five feet wide and twelve feet 

 long, the center row having a given formula of Nu- 

 trient Solution of the twelve known necessary ele- 

 ments, disregarding others, and each row beyond, on 

 one side being given one third more of one element, 

 keeping all the other elements identical, and each 

 row beyond the center on the other side given one 

 third less of one element, the others the same, and 

 taking the average plant from the several types of 

 plants used, the best combination of the twelve 

 known necessary elements may soon be found. As 

 much of the experimental work so far has been done 

 in small jars of possibly a quart or a gallon of solu- 

 tion, containing but one plant, no real average could 

 be made. 



In my own first experiments I followed this 

 method, but found it impossible to keep the water 



