72 THE BOTANICAL MAGAZINE. LVoi. xxxiv. No. 402. 



grown for three weeks, in an ordinary greenhouse in Baltimore, in the 

 month of February. There were five plants in a culture, each culture 

 jar held about 440 c.c, and the solutions were renewed after Qh days 

 and at the end of the first and second weeks. At the end of this 

 period they were compared (by five different growth criteria) with the 

 plants of control cultures, which had been grown simultaneously from 

 the 3-cm. stage, in an excellent complete solution (Shive's R5C2— 1.75 

 atmospheres, osmotic value). Six different sets of three main salts 

 were employed for the incomplete solutions and 15 different sets of 

 proportions were tested for each of these sets of salts. Every in- 

 complete solution contained Ca, Mg, (H 2 P0 4 ) 2 , (N0 3 ) 2 and SO 4 , 

 together with a very small amount of FeP04. A triangular diagram 

 was used as a guide in selecting the sets of salt proportions to be 

 tested. The FeP0 4 was always added to the 3-salt solutions in the 

 same amount. The total concentrations of all incomplete solutions 

 were the same, being 0.015 gram-molecule per liter (of all three main 

 salts taken together) and corresponding to about 1.00 atmosphere of 

 osmotic pressure. — By the criterion of total height the best cultures of 

 the incomplete-solution series were 98 hundredths as good as the 

 average of the controls. By other growth criteria these best in- 

 complete solutions (there were three of them, practically alike as to 

 the growth they produced) were generally somewhat poorer than the 

 average of the controls. Averaging the five growth values obtained 

 for each culture and considering the generalized result as a measure of 

 the physiological worth of the solution used, the three best incomplete 

 solutions were five-sevenths as good as the average of the controls. 

 The plants of these best solutions without potassium appeared per- 

 fectly healthy at the end of the 3-week period ; their growth might 

 have been continued longer. The poorer solutions gave small plants, 

 with some or nearly all of the leaves yellow or dead, but none of the 

 solutions produced any specifically characteristic symptoms of poison- 

 ing or malnutrition.— The good solutions were already among those 

 with the lowest partial volume-molecular concentrations of the di- 

 hydrogen-phosphate salt [either Ca (H 2 P0 4 ) 2 or Mg (H 2 P0 4 ) 2 ] ; one- 

 seventh of the total volume-molecular concentration was due to this 

 salt in these best solutions. The worst solutions were among those 

 having higher partial concentrations of the di-hydrogen-phosphate 

 salt ; the highest partial concentration of this salt was five-sevenths of 

 the total concentration. The value of the ratio of calcium to magne- 

 sium was not a controlling condition in determining the physiological 

 worth of any solution, or did any other ratio value exert a noticeable 



