one set of 19, 20 or 21 cultures (and the triplicate control) , As 

 soon as a type group has been tested once, the best seven solu- 

 tions are to be selected, and all future work in this part of this 

 project is to be confined to these best seven sets of salt propor- 

 tions. This feature of the plan soon eliminates two-thirds of 

 the solutions, and it is based on the assumption that the best 

 seven solutions in any experiment will probably include the single 

 very best solution of that type in any other experiment, no mat- 

 ter when or where such other experiment may be carried out. 

 While climatic conditions may shift the position of the best physi- 

 ological balance of the entire group in the series of the best 

 seven sets of salt proportions, it is at least improbable that they 

 will shift this position to one outside of this series of seven. If 

 it should develop that none of the solutions of some type give 

 even fair growth— as compared to the control — then the whole 

 of that type may be eliminated. It is thus seen that the repeti- 

 tion will deal with at most no more than forty-two different sets 

 of salt proportions. These repetitions are to be carried out re- 

 peatedly and under a great variety of climatic complexes, and it 

 is really with these forty-two (or fewer) sets of salt proportions 

 that the present phase of our campaign is to deal. . 



When the repetitions are taken up, other total concentrations 

 are to be introduced, besides the one having an osmotic value of 

 1.00 atm. Detailed plans in this connection may be postponed 

 till later, however. Plans for dealing with various logically 

 possible complications in other aspects of the research may also 

 be deferred till such complications actually arise from experi- 

 mental results. 



SAND-CULTURES OF WHEAT. 



Introduction. The sand-culture method to be used is essen- 

 tially that described by McCall (1916). The important feature 

 is that the nutrient solution is renewed at frequent intervals, in 

 much the same way as with water-cultures. As in all experi- 

 mentation of this cooperation, it is necessary that essential de- 

 tails be alike in all experiments if the results are to be com- 

 parable. The nutrient solutions to be used are the same as those 

 for water-cultures, so that there are one hundred and twenty-six 

 different sets of salt proportions in sand-cultures to be compared, 

 besides the general control, with Shive's solution IR5C2 (1/10- 

 increments, 1.75 atm.). This last is our solution IR3.77S1.09 

 (see table I). The following plan is based on a memorandum 

 furnished the committee by Dr. McCall and Dr. Shive. 



46 



