PROBLEMS IN MINERAL NUTRITION 315 



micronutrients, and, equally important, the compound added must not 

 contain any growth-promoting substances. 



The second approach, little used so far, is to determine that some 

 metabolite which contains the element is essential to growth. For ex- 

 ample, if the vitamins B 12 prove to be essential, then cobalt must be 

 essential. 



Details of the methods for purification of media for studies of micro- 

 nutrient elements lie outside the scope of this work. There is no one 

 best method for all materials, and the choice depends on the particular 

 element under study. The most useful over-all methods involve adsorp- 

 tion on calcium carbonate (240) or aluminum compounds (48, 221). 

 Special methods have been described for copper (151, 167, 168), man- 

 ganese (168), and molybdenum (81, 168). It appears that the chelating 

 agents (8-hydroxyquinoline, diphenylthiocarbazone, etc.) are not only 

 less efficient for most micronutrient elements but are likely to be toxic 

 to the organism. Detailed comparisons of different methods have been 

 published by Steinberg (240), Donald et al. (48), and Nicholas (168). 



Methods of obtaining inorganic salts of maximum purity are dis- 

 cussed by Hutner et al. (94). 



The sources of contamination in micronutrient work are well known 

 by now: the inoculum, glassware, distilled water from a metal still, su- 

 gars and polyatomic alcohols, hydroxy acids, inorganic salts, and com- 

 plex extracts of biological origin (92, 168, 229, 230, 232). Amino acids 

 are often so heavily contaminated that they cannot be employed at all 

 (235). 



Chelating agents may be used to exaggerate a requirement and make 

 it more readily detectable. Thus, Hutner (93) added citrate to hold 

 calcium and other elements in solution; the calcium requirement in 

 the absence of the complexing anion is estimated by extrapolation. 



Work with the major inorganic nutrients is, of course, somewhat 

 simpler inasmuch as deficiencies are not made up by minor amounts 

 of contamination. Methodological problems do, however, arise even 

 here. Thus, the usual practice in working at low or zero potassium is 

 to supply phosphate as the sodium salt; this, of course, ignores the pos- 

 sibility that sodium partially replaces potassium. In this particular 

 instance, the solution is to use an amine in place of the alkali metal to 

 control pH (126). 



In reporting results, the concentration permitting half maximal 

 growth is a more useful datum than that at which maximum growth 

 occurs. 



Although an optimum amount of a given element may be defined 

 for rigidly specified conditions (236, 237), it must be realized that there 



