SPECTROSCOPIC PHOTOGRAPHY 819 



This method is called "linear interpolation," and, while it always involves some 

 error, this error can be made negligible by choosing known lines lying so close together 

 that the dispersion of the spectrograph varies but slightly between them. 



Simplified Line Identification. — Most spectrum lines have been observed and 

 measured many times, and many can be identified from their appearance or the pat- 

 terns they form with other lines. Unknown lines can often be identified by merely 

 comparing them directly with spectra of known elements photographed on the same 

 or a similar plate. It is often convenient with a newly acquired spectrograph to 

 photograph the arc and spark spectra of a number of the more common elements 

 which can be obtained comparatively pure, the resulting plates being marked and 

 kept to aid in identifying unknown lines. These plates can be laid directly over later 

 plates taken with the same instrument, so that corresponding lines appear in juxta- 

 position. Atlases of spectral lines can be used similarly, though less directly because 

 usually their dispersion scale is different. 



To aid in identifying wavelength positions approximately, a simple well-known 

 spectrum such as that of mercury or of copper may conveniently be impressed on 

 each plate taken. The iron spectrum should then be impressed on the plate also, to 

 give more precise locations of desired lines. 



Light Sources for Emission Spectroscopy. — The source most commonly used for 

 producing line spectra is the electric arc in air, preferably run on 220 volts d.c. with a 



20ji 



I A^ArAAAAA^AAAA/ 



i R I20JI 



i Arc 



220-V.dC. 



-^OOOOOOOOUOOIJWJ^ 



Reactor 

 Fig. 12. — Circuit for arc light source suitable for use in spectroscopy. 



current of from 2 to 10 amp. The circuit used is shown in Fig. 12. A rheostat R 

 capable of carrying 10 amp., whose resistance can be varied from 20 to 120 ohms, is 

 kept in series with the arc, together with an ammeter A and a stabilizing reactance 

 made by winding a few hundred turns of wire on an iron core. A voltmeter V m&y be 

 provided to read the voltage across the arc terminals, but this should not be con- 

 nected except when being read, as when the arc goes out a high A'oltage maj' be bviilt 

 up across its terminals. The series reactance helps keep the arc burning steadily, and 

 use of 220 volts d.c. instead of 110 volts also gives greater steadiness. 



The arc may be burned between pure graphite electrodes, the lower electrode being 

 cupped to receive small samples of the material whose spectrum lines are to be studied. 

 Or the electrodes themselves maj^ be formed from the material to be studied. Metals, 

 ashes, or liquids may be burned in the cup, and even slags and glasses can be thus 

 handled. When the material is a poor conductor, it is desirable to mix with it some 

 conducting material such as ammonium sulphate, which emits few spectrum lines in 

 the visible or viltra violet regions, and to moisten the mass with pure dilute hydrochloric 

 acid. 



The condensed electric spark, connected as shown in Fig. 13, is used almost as much 

 as the arc as a spectrum source. Most of the spectrum lines which appear in the arc 

 also appear in the spark, and in addition new lines, usuallj' produced by atoms from 

 Avhich one or more electrons have been removed, also appear. Strong fuzzj- lines due 



