MICRO-NUTRIENT PROBLEMS 29 



flames have all been employed with more or less success. Two 

 distinct procedures have been employed for introducing the 

 substance into the flame. In one a small quantity of the sub- 

 stance is contained on or in a piece of filter paper and the latter 

 then introduced into the flame. In the other a solution of the 

 substance is sprayed into the flame through a very fine nozzle. 

 These two methods are largely associated with the names of 

 Ramage and Lundegardh respectively, but they have both been 

 used, with a variety of modifications in detail, by other workers 

 as well. By maintaining the conditions of experimentation 

 constant the same density of spectral line can be obtained from 

 the same quantity of material, so if calibration is made by the 

 use of a number of samples of known composition it is possible 

 to determine the amount of the element in a sample under 

 examination by measurement of the density of the line and 

 reference to a calibration graph. 



In exciting spectra by means of the arc, electrodes in the form 

 of rods, usually of graphite, but sometimes of copper, nickel, iron 

 or other metal, and of as great a purity as possible, are used. The 

 electrodes are in a vertical line and the lower contains at its 

 upper end a cavity in which the substance under examination 

 is held. After bringing the electrodes into contact they are 

 separated to a standard distance apart so that the conditions of 

 the arc are kept as constant as possible. Even so the arc is so 

 variable that it has so far been found impossible to devise an 

 arrangement such that the same amount of material subjected 

 to excitation will produce the same intensity of spectral lines. 

 Hence in quantitative estimation of any element by means of 

 arc spectra it is necessary to have recourse to the device of the 

 'internal standard'. This is achieved by introducing along with 

 the substance to be examined a known amount of some other 

 substance involving an element which yields a spectral fine in 

 the near neighbourhood in the spectrum of the fine to be 

 measured. This same consideration holds when the spark dis- 

 charge is used for exciting spectra. Thus Foster and Horton, in 

 determining boron in plant material by the spark method, added 

 a known quantity of a gold salt to the material they were 

 examining. Within limits the ratio of the intensities of the fines 

 of the element to be measured and of the internal standard is 



