Page:The New International Encyclopædia 1st ed. v. 18.djvu/494

* SPECTROSCOPE. 426 SPECTROSCOPY. justed) if this instrument is used in the standard manner, in the fact that the spectra produced are of such a kind that the distances along the photographic plate are proportional to differences in wave-length of the waves which are thus re- corded. A spectrum of this kind is said to be 'normal.' SPECTROSCOPY. The science which deals with the methods of production of the spectra by various sources of light (or of waves in the ether) and with their study and interpretation. Newton, in 1672, was the first to observe that if sunlight entering a darkened room through a small opening were allowed to fall upon a prism a spectrum was produced, owing to the fact that waves characteristic of different colors suffer dif- ferent deviations by the prism, and that, therefore, the components of white light were separated. (See DiSPEKSiox. ) Newton made no observations except upon the visible portions of the spectrum ; but in the year 1800 F. W. Herschel observed that the spectrum continued beyond the red, as was shown by holding a thermometer in that po- sition; and in 1801 J. W. Ritter proved the ex- istence of the ultra-violet light by showing that silver chloride was affected not alone by the vio- let portion of the spectrum, but beyond. In 1802 Wollaston made the discovery that the solar spectrum was not continuous, but was interrujited ty certain dark lines, using in his experiments a slit and a prism with its edge parallel to the slit. It is remarkable that Newton did not make this same discovery in his investigation on the spec- trum of the Sim. because he also at times used a slit in the same manner as did Wollaston. The most important investigation, however, on the solar spectrum, one which in fact serves as the foundation of the science of spectroscopy, was that of Fraunhofer beginning in the year 1814. Fraunhofer (q.v. ) was the inventor of the dif- fraction grating (q.v.), and was the first to meas- ure accurately the wave-lengths of light waves. By using both grating and prism spectroscopes he showed that there were numerous dark lines in the solar spectrum, to the strongest of which he gave certain names in the form of letters. A, B, C, etc. He studied, further, the radiations from certain of the stars and from certain sources of light such as flames, etc. He made the impor- tant observation that the position occupied in the solar spectrum by the dark D lines is identical with that occupied by the bright yellow line ob- served in the spectra of all flames. His whole work was epoch-making. Herschel was the first to investigate the ab- sorption spectra produced by various bodies, that is, to study the effect of interposing between the source of light and the spectroscope a substance which absorbs certain radiations. The next great step was made by Kirchhoff. who showed from theoretical considerations that the emission spec- trum and absorption spectrum of a substance should be the same at a definite temperature, and that as the temperature changed the intensity of the spectra would vary. Tn this way Kircfihnff was able to explain the dark lines in the solar spectrum as due to an absorbing layer of metallic vapors forming an atmosphere around the white- hot central portion of the sun, which was sup- posed to emit a continuous spectrum. In collabo- ration with Bunsen he then undertook a careful study of the spectra of various substances and founded the science of spectrum analysis. The explanation of the fact that absorption spectra and emission spectra are identical was first given by Stokes many years before Kirchhoff's state- ment, and Balfour Stewart had also arrived at the same idea. From the da.v of Kirchhoff up to the present time all branches of spectroscopy have been pursued most vigorously, the most im- portant discovery being the principle of the con- cave grating, made by Rowland in the year 1882, which is discussed under Spectroscope, and under Diffraction and Diffraction Gratings. There are many methods of making vapors luminous, among which it may be svttficient to name the flame, the electric arc, the electric spark. The spectra produced in these various waj'9 have received the names of flame spectra, arc spectra, etc. The standard method of pro- ducing flame spectra is to hold a portion of the substance to be investigated or a salt of the sub- stance in the Bunsen flame until it is vaporized, and thus the vapor is raised to the temperature of the flame, and in general becomes luminous. Other flames than that of the Bunsen burner may be used. Extremely minute amounts of the sub- stance may be recognized in this way, as has been shown by Kirchhoff and Bunsen. They state that one fourteen-millionth of a milligram of so- di<im can be recognized in the Bmsen flame; one sixty-thousandth of lithium : one fifty-thou- sandth of a milligram of calcium, etc. To pro- duce arc spectra it is customary to bore out a small opening in a carbon rod. fill this with some salt of the substance to be investigated, and then use it as the positive pole in the ordinary electric arc. By this means a high temperature is pro- duced, namely about 4000° Centigrade, and the vapor of the substance is made luminous. To produce the spark spectrum numberless methods are in use. Among these it may be sufficient to mention two. If the electric spark produced by an induction coil or transformer is made to pass between solid electrodes made up of the substance to be studied, they will be vaporized and the vapor will be luminous. If a gas, for instance . nitrogen or hydrogen, be inclosed in a glass bulb into which enter two metal wires to serve as elec- trodes, and if the pressure be made sufficiently low, a spark can be made to pass through the re- maining gas. Changes in the pressure, in the electrical constants of the circuit, etc.. produce alterations in the spectra. The main distinction between the flame spectra and arc spectra is one of temperature, but so little is known in re- gard to the mechanism of a spark that no con- clusions can be drawn with certainty. Beyond a doubt the spectra produced in the arc are due in the main to the high temperature of the arc, W'hereas the cause of the production of spectr.a in the spark is probably not a temperature effect at all, but something concerned with the transforma- tion of electric energy. The means of producing spectra are discussed under the head of Spectroscope (q.v.). Prisms or gratings are used for this purpose. The ac- cepted method at present for measuring wave- lengths in the spectrum of any substance is to photograph on the same plate with these lines the spectrum of some substance whose lines are knovn and then to obtain the desired quantities by a method of interpolation. The spectrum of iron is as a rule used to give the comparison lines.