28 METHODS OF INVESTIGATING 



Undoubtedly the use of the spectrograph affords the most 

 sensitive method of measuring small quantities of a large number 

 of elements. 



Under certain conditions a substance can be made to emit 

 radiation, this radiation being limited to certain wave-lengths 

 which are characteristic of the elements in the substance and 

 of the conditions used to excite the radiation. If the radiation 

 passes through a prism or diffraction grating the radiations of 

 different wave-lengths are separated and a spectrum results, the 

 radiation possessing the longest wave-lengths being at one end 

 of the spectrum and that possessing the shortest wave-lengths 

 at the other. In the spectrograph such a spectrum is made to 

 fall on a photographic plate, and the plate, on development, 

 constitutes a photographic negative of the spectrum. This 

 usually consists of a number of lines, each line corresponding to 

 radiation of a definite wave-length and possessing an intensity 

 of blackness depending, within the limits of under-exposure 

 and over-exposure of the plate, on the intensity of radiation of 

 that particular wave-length. Determination of the intensity of 

 blackening of a line characteristic of a particular element should 

 therefore afford a means of estimating the quantity of that 

 element in a sample of material, for example, plant ash, which 

 has been subjected to the necessary conditions for inducing an 

 emission of radiation from it. 



Of the various ways in which radiation suitable for spectro- 

 graph^ examination may be produced three have been developed 

 to a considerable extent; these are by means of a flame, by the 

 use of the electric arc and by the use of an electric spark. The 

 spectra produced in these ways are known as flame, arc and 

 spark spectra respectively. Each method has its own particular 

 advantages and disadvantages. Perhaps with workers on plant 

 material the flame method has so far proved the most popular, 

 but this is by no means so with workers in other scientific 

 fields. 



Flame spectra, as the name suggests, are produced when a 

 substance is heated in a flame. Although spectra are produced 

 when the flame is that of a Bunsen burner, for analytical pur- 

 poses a flame much hotter than this is generally required, and 

 air-acetylene, oxy-coal gas, oxy-hydrogen and oxy-acetylene 



