MICRO-NUTRIENT PROBLEMS 31 



But from Lundegardh's experience (1929, 1934) it would 

 appear that the flame method is sufficiently sensitive for the 

 quantitative determination of only three of the micro-nutrient 

 elements listed in Chapter i, namely, manganese, copper 

 and gallium. The most dilute solutions of manganese and 

 copper that can be used are those of a concentration of 

 5x 10~ 6 M . As a quantity of 5-15 ml. of solution is required 

 this would mean that the smallest quantity of these elements 

 determinable by the flame method as Lundegardh used it would 

 be of the order of 1-4-4/zg. With a procedure more recently 

 described by Griggs, Johnstin and Elledge (1941), the minimum 

 concentration of manganese usable is given as 1-125 x 10~ 5 M , 

 while the use of .as little as 2 ml. of solution is possible. This 

 would mean that the smallest quantity of manganese measur- 

 able would be about 1-37/xg. 



Data for gallium are not at present available, but a flame 

 spectrum of this element reproduced by Lundegardh suggests 

 that the sensitivity is certainly high for this metal. 



Using the arc the author has found it possible to measure 

 quantities of manganese as small as 0-05^g., and no doubt even 

 smaller quantities than this could be measured with a suitable 

 choice of experimental conditions. 



It must be borne in mind that for the determination of any 

 particular element one or other of the methods that have been 

 here indicated may be inapplicable. Thus so far it has not been 

 found possible to determine either boron or magnesium, both 

 important plant nutrients, by means of the polarograph. The 

 presence of magnesium as an impurity in graphite may render 

 the use of a graphite arc impracticable for the determination of 

 that element spectrographically. A method, while usable for the 

 determination of larger quantities, may not be sufficiently sen- 

 sitive for measuring the small amounts of micro -nutrients 

 present in available samples of plant material. Many pre- 

 liminary trials of the different methods may thus be necessary 

 before the investigator can decide what method to use for the 

 estimation of any particular micro -nutrient. 



The degree of accuracy generally obtainable by what may be 

 called micro-methods is much less than that obtainable in most 

 macro -chemical determinations, but is generally sufficient for 



