Bearing in mind this conception of a train of light-waves, we proceed to consider the gen eral method of spectroscopy. The examination of a body by means of the spectroscope includes, in general, four steps. These are: A. The Production of the Radiations.
B. The Separation of the Radiations.
C. The Recording of the Radiations.
D. The Comparison and Interpretation of the Radiations.
We shall in the following four sections con sider these four steps.
2. Sources of Radiation. In practice the spectroscopist is limited to two artificial sources, namely, flames and electric currents. Besides these are many natural sources such as light ning, the aurora borealis, meteors, stars, glow worms, etc.; but the nature and occurrence of these natural phenomena are so far beyond our control that they do not here call for discus sion. Adopting the following classification, we now proceed to a description of some par ticular artificial sources.
Candle and Alcohol Flames. These have mainly a historical interest, as being the first sources in which line spectra were observed. The temperature of the alcohol flame is only roughly known, but is probably not far from 1,300° C. Substances are easily introduced into the flame and the flame itself is fairly colorless. These two qualities combined to make it very useful in the early history of the science.
Bunsen Flame. The flame which Bunsen devised in 1856 has three marked advantages as a spectroscopic source, namely, it is simple in construction and operation; its outer mantle is almost totally devoid of any characteristic spectrum of its own; and it possesses a high temperature. Many measurements of the tem perature of this flame have been made, hut there is good reason for believing that nearly all of these are under-estimates. The most re liable measurement is probably that of Weg gener, at Berlin, who obtained 1,788° C. for the outer mantle. The temperature of the violet mantle is probably about 100° or 200° less; while the temperature of the inner dark cone ranges from 250° to 500° C. The substance whose spectrum is desired may he introduced into the flame in a great variety of ways. Bun sen used a bead of the material on a platinum wire; sometimes a piece of asbestos is con venient; Gouy uses the spray from an atomizer ; Eder and Valenta have employed a platinum wire-gauze wheel to transport the solution from.a dish to the flame.
Oxyhydrogen Flame. This flame is also nearly colorless, has an exceedingly high. tem perature and gives spectra approximating those obtained in the electric arc. It has been used with great skill by W. N. Hartley.
Electric Arc. This source has been in com mon use since the invention of the Gramme dynamo, 1876. The arc is generally used in one of two ways, either by drilling holes in the carbon rods and filling these holes with the ma terial whose spectrum is desired, or by making the electrodes directly and entirely out of the substance which is being studied. This latter method is, of course, practicable only in the case of materials which conduct electricity. The temperature of the carbon arc is not accu rately known, but probably lies between 3,500° and 3,900° C. As a spectroscopic source it is admirable in nearly every respect except that in those regions of the spectrum which are most easily studied carbon has a strong spectrum of its own which is always superposed upon that of the substance under investigation. In the case of metals, the easiest method of avoiding this difficulty is to use a rapidly rotating metal disc for one electrode and a metal rod for the other. The incandescent metal vapor thus ob tained yields the purest spectra known.
Electric Spark. Of all sources of radiation that which is most generally convenient and most widely used is perhaps the electric spark, obtained by use of an ordinary induction coil. Spark discharges may also be obtained from an electrostatic machine, such as the Wimshurst, from a Tesla coil, or even from a high voltage storage battery. In general the electric spark produces two spectra, namely, the spectrum of the electrodes and the spectrum of the air or other gas between the two electrodes. But, as was discovered by Plucker in 1858, the spec trum of the electrodes may be avoided by en closing them in a partial vacuum. Under these conditions one gets only the spectrum of the surrounding gas; so that now the almost uni versal method of studying the spectra of gases is to seal platinum electrodes into a glass tube and then fill the tube under low pressure with the gas to be studied.