MICRO-NUTRIENT PROBLEMS 45 



mination is not in excess of 10 per cent. The limit of sensitivity 

 of the method is not clearly stated, but it would seem probable 

 that quantities of boron as small as 0-1/xg. might be measurable 

 in this way. Lundegardh (1929) suggested the use of the cad- 

 mium line 2573 A. as internal standard in the determination of 

 boron by the spark method. 



As mentioned earlier (see p. 33), Melvin and O'Connor (1941) 

 have used the arc method for the simultaneous determination 

 of boron, manganese and copper in fertilizers, using beryllium 

 as internal standard. The method would seem to be applicable 

 to the determination of boron in plant material, and the 

 accuracy and sensitivity would appear to be of the same order 

 as in the spark method of Foster and Horton. 



Although the flame method is not suitable for the speetro- 

 graphic determination of boron, McHargue and Calfee (1932) 

 have successfully made use of flame spectra for the optical 

 spectroscopic determination of boron. In their earlier procedure 

 the boron was first converted into methyl borate, then vola- 

 tilized with methyl alcohol and burnt in an atmosphere of 

 oxygen in front of a cell containing a solution of potassium 

 permanganate. The concentration of this solution was adjusted 

 so that the bright lines of the boron spectrum were just obscured 

 by it. By previous standardization of solutions of potassium 

 permanganate against standard boron solutions the concentration 

 of the experimental boron solution was obtained. 



Later (Calfee and McHargue, 1937) a different procedure was 

 devised. The spectrum was excited in an oxygen-methane flame, 

 methane saturated with a solution of methyl borate in methyl 

 alcohol being ignited in an oxygen blast. The light emitted from 

 this was polarized and by an optical arrangement the spectrum 

 was brought into juxtaposition with a second spectrum similarly 

 produced by the burning of a standard boron solution. By rota- 

 tion of an analysing plate the intensity of the spectrum of the 

 standard boron solution could be varied and so matched with 

 that of the solution the boron concentration of which it was 

 desired to measure. For a quantitative determination in this 

 way the boron content of a sample should lie between 25 and 

 50/xg. Agreement between the two procedures was good, but 

 the second method was found to be more exact. 



