one of the samples tested. On the other hand, the intervals be- 

 tween the solutions selected for test — out of the infinite number 

 of possible ones — might be made so narrow that the large num- 

 ber of tests required would render the whole project hopeless. 

 We cannot hope to test millions of different solutions, nor should 

 we hope for much real progress if we were to test only a dozen, 

 for instance. In arranging the scheme of solution to be first 

 tested, as presented below, the judgments of a number of spe- 

 cialists in this general field have been combined. It must be 

 remembered that the present plan does not pretend to lay out 

 the experimentation excepting for the beginning of our work. 

 We begin by sampling a certain region only of the whole infinite 

 field of solution possibilities. Other regions may be attacked 

 later. 



We proceed on the general physiological principle that the 

 best solutions for plant growth must always contain at least 

 some of each of the seven chemical elements known to be essen- 

 tial for all plant activity : K, Ca, Mg, Fe, N, P. and S. It appears 

 safe to say that it would be useless to test any solution not con- 

 taining all of these elements, in a search for the best solution for 

 any plant and growth phase. Consequently, under the guidance 

 of our present knowledge, we do not need to deal with any solu- 

 tion that does not contain all of the seven recognized essential 

 elements for higher plants in general. 



Furthermore, experience has already shown that no solution 

 may be expected to support good plant growth if it contains more 

 than a small trace of iron. On this account, and because agri- 

 culture seldom meets with either a deficiency or an excess of iron- 

 as a source of trouble, we shall at first make no study of different 

 partial concentrations of that element. We shall follow Tot- 

 tingham in supplying iron as FePO' in the same, very small, 

 amount to every solution employed after the germination phase. 

 (For that phase the suuply of iron in the seed may safely be con- 

 sidered as sufficient.) 



It is of course logically possible (perhaps even probable) that 

 the very best solution for any plant and phase may eventually be 

 found to contain still other elements besides the seven recognized 

 essential ones. We shall ignore this proposition at the beginning 

 of our work, however, and shall study only solutions containing 

 just the six elements K, Ca, Mg, N, P, and S, in various propor- 

 tions and in various total concentrations, besides a trace of Fe. 

 Later plans will perhaps introduce some of the apparently most 

 promising non-essential elements, such as Na, CI, Si, Mn, etc., 



23 



