but it would be quite hopeless to attempt to bring these into con- 

 sideration at the start, especially considering the limited number 

 of cooperators available. It should be remarked that there will 

 surely be a considerable amount of silicon in all the solutions 

 tested (since uncoated glass containers are to be employed), but 

 the amount of this element present may be considered as prac- 

 tically the same in all solutions. Like iron, it will be uniformly 

 present in small amount. The same is true, to a degree, of 

 sodium ; and to a smaller degree of chlorine. But the traces of 

 these three elements that may be present are ignored for the 

 part of the project now being planned. 



From these considerations it emerges that our present plans 

 involve simply the testing of solutions containing the six ele- 

 ments that are essential in considerable quantities, with addi- 

 tion of a trace of the seventh essential element, Fe. As Shive 

 (1915) has pointed out, the simplest way to get these six ele- 

 ments into solution is to prepare the solution from three salts; 

 since three of the elements occur as kations and the other three 

 occur in anions. It appears best, at the start, to employ N as 

 the nitrate ion (NOO, P as the di-hydrogen phosphate ion 

 (H^PO*), and S as the sulphate ion (SO*). Other carriers of 

 these three elements may be studied later, such as nitrates, am- 

 monium salts, mono-hydrogen phosphates, sulphites, etc. It is 

 even probable that ammonium may need to be introduced for 

 later growth phases of soy bean to get even presentable growth, 

 but this matter is not now before us. 



Of course it may be that no possible three-salt solution is best 

 suited to the growth of a given plant in a given developmental 

 phase ; it is clear that there are many sets of element or ion pro- 

 portions that cannot be obtained in a three-salt solution at all, 

 and four-, five-, six-, etc., salt solutions must logically be brought 

 into the experimental comparison before the very best possible 

 six-ion solution may be established. It seems desirable, however, 

 to make a very thorough study of the three-salt possibilities be- 

 fore moving forward to attack the much more complex types of 

 solution with more than three salts. One type of such solution 

 has been studied by Tottingham (1914) for young wheat plants. 

 As our project goes forward, plans for later campaigns may be 

 formulated, but an attempt at their consideration would be boot- 

 less at present. 



The Three-Salt Solutions Characterized. 



As Livingston and Tottingham have pointed out (Amer. Jour. 

 Bot. 5: 337-346. 1918. A reprint may be obtained from either 



24 



