50 METHODS OF INVESTIGATING 



nickel can be made in the same way as that of cobalt it is as 

 convenient to consider the two elements as cobalt only. 



Although these elements can be determined spectrographic- 

 ally, the flame method scarcely has sufficient sensitivity for 

 their ready estimation in plant material, for it would appear from 

 Lundegardh's data that the smallest amount of either metal 

 measurable in this way is of the order of 100/xg., which means 

 that decidedly large samples of material would generally have 

 to be used. 



For the semi-quantitative determination of cobalt in soils, 

 Mitchell (1940), who has made a special study of the estimation 

 of trace elements in soils, recommends the use of the cathode 

 layer arc. The method is based on the fact that in the region of 

 the arc adjoining the cathode (the cathode layer), the emission 

 of spectral energy from cations may be up to 100 times as 

 intense as that from the column of the arc. By means of a 

 spherocylindrical quartz lens the image of the arc is focused 

 sharply on the slit, so that the image of the cathode appears as 

 a horizontal line across the top of the slit while the image of the 

 anode falls well below the bottom of the slit. This arrangement 

 results in greatly increased sensitivity as compared with the 

 ordinary arc method, and 2 p. p.m. of cobalt and 1 p. p.m. of 

 nickel can be detected by its use. The spectral lines used for the 

 determination of these two elements are 3453-5 and 34 14- 8 A. 

 respectively. 



Cobalt and nickel can be readily determined simultaneously 

 with the polarograph by the procedure described by Lingane 

 and Kerlinger (1941), the essential feature of which is the use of 

 pyridine as supporting electrolyte. Using a normal solution of 

 potassium chloride containing from 0-05 M to M pyridine and 

 0-05 per cent gelatin the waves of nickel and cobalt are well 

 separated and defined, while manganese does not interfere with 

 them. Ferric iron, if present in large excess, interferes with the 

 determination of nickel and cobalt, but its effect can be elimi- 

 nated by the use of a supporting electrolyte with pH of about 

 5-4 containing equal concentrations of pyridine and pyridium 

 chloride in which the ferric iron is precipitated as hydrous ferric 

 oxide. Small amounts of copper do not interfere with the deter- 

 mination of nickel and cobalt, but if present in considerable 



