author), there are exactly six possible types of three-salt solu- 

 tions that can be made to contain just these six ions, K, Ca, Mg, 

 H2P0% N0% and SO', and it is a rather thorough test of these six 

 types that is contemplated for the first stage of the present 

 project. 



It should be added that there are three other essential ele- 

 ments, besides the ones mentioned above, that are always present 

 in any of these solutions; namely, C, H, and 0. The solution is 

 mostly water, oxygen occurs as a solute and also in every one of 

 the three anions employed, hydrogen occurs in the di-hydrogen 

 phosphate ion, and CO- occurs as a solute (thus adding the ele- 

 ment C). It may be that the oxygen atom in an ion will prove 

 to be negligible in such work as this (it seems not to have been given 

 serious consideration in any of the published discussions of 

 nutrient solutions and fertilizers), but it is certain that the 

 H-ion, as such, is of great importance in determining the physio- 

 logical properties of nutrient solutions for plants. (See, for 

 example: — Sorenson, S. P. L. Ueber die Messung und Bedeu- 

 tung der Wasserstoffionenkonzentration bei biologischen Prozes- 

 sen. Ergeb. Physiol. .12: 393-532. 1912. Sharp, L. T., and 

 D. R. Hoagland. Acidity and absorption in soils as measured by 

 the hydrogen electrode. Jour. Agric. Res. 7: 123-145. 1918. 

 Plummer, J. K. Studies on soil reaction as indicated by the 

 hydrogen electrode. Jour. Agric. Res. 12: 19-31. 1918.) 



The six possible types of three-salt solution are represented 

 below, the arrangement being the same as that employed by 

 Livingston and Tottingham (1918). On the basis of present 

 knowledge the presence of carbon dioxide in these solutions may 

 be ignored as without significant influence upon the plants. 



I. II. III. IV. V. VI. 



Ca(N03)t Ca(N03)2 Ca(HaP04)3 CaCHsPOO* CaSO* CaSO* 



KH=PO* K2SO1 KNOs K2SO4 KNO. KH^PO* 



MgSO* Mg(H2PO*)2 MgS04 Mg(N08)» Mg(H2PO0* Mg(NO«)t 



Our first problem is to find (by actual test with each growth 

 phase) the best set of volume-molecular proportions and the 

 best total concentration for each type of solution. Now, there 

 is clearly an infinite number of possible sets of proportions of 

 any three things, and for each set of salt proportions of each of 

 the six types there is an infinite number of total concentrations. 

 We choose, from the very large series of different sets of salt 



25 



