194 SPECTRUM ANALYSIS, NEW METALS [107. the position of the lines obtained when a mixture of various chlorides was introduced into the flame, to ascertain the presence of these different metals with sufficient readiness to use the method for the purposes of qualitative analysis. The rapidity with which the result is obtained by a practised observer, and the minuteness of the quantity required for the examination, gives this method a superiority over any other now in use for the qualitative analysis of the alkalies and alkaline earths; moreover, the circumstance that the mere inspection of a source of light furnishes information respecting the composition of the bodies undergoing combustion or volatilization within it, extends the mode of inquiry over distances limited only by the distance through which the object is visible; we are thus furnished with a method of analysis which is applicable to the luminous atmosphere of the sun, the stars, as well as to the light of the planetary bodies, and even of the nebulæ. This circumstance invests the subject with an interest like that which attends the employment of the telescope; at the same time the minuteness of its search enables it to reveal, like the microscope, quantities of matter indefinitely small. This minuteness in its scrutiny has already, in the hands of Bunsen and Kirchhoff, led to the discovery that many bodies, such as lithia and strontia, formerly supposed to be rare, are really widely distributed in minute quantities. It also led them to discover the two new alkalies cæsia and rubidia, the first named from cæsius, "sky-coloured," in allusion to two characteristic blue lines in its spectrum; the second from rubidus, “dark red," owing to the existence in its spectrum of two red lines of remarkably low refrangibility. These bases were found in the course of an examination by the prism of the residue of the mother-liquor from the Durkheim spring, when the occurrence of these hitherto unobserved lines induced Bunsen to make a minute chemical examination of the water which furnished them. The inquiry showed that in every ton of the original water, about three grains of cæsic chloride, and rather less than four grains of rubidic chloride, were present. These two salts so closely resemble potassic chloride in properties, that it would have been impossible to have ascertained their existence in the minute proportion in which they occur, but for the method of spectrum analysis. To these bodies Crookes, in May, 1861, added a third metallic element, thallium, so called from Oaλλos, a budding twig, in allusion to the brilliant green line in which the luminosity of its spectrum is concentrated. This body was found in minute quantity in the residue from a sul 107 a.] DISCOVERED BY ITS MEANS. 195 phuric acid chamber at Tilkerode, in the Hartz: and it has since been found in minerals from various other localities, particularly in Belgian pyrites. In the year 1863 a fourth metal was found. by the spectrum examination of one of the Freiberg zinc ores, by Reich and Richter (Journ. pr. Chem. 1863, lxxxix. 441, and xc. 172); its discoverers named it indium, from its pair of characteristic lines in the indigo. Another metal was detected by Lecoq de Boisbaudran in zinc blende from the mine of Pierrefitte, valley of Argeles, Pyrenees (Comptes Rendus, 1875, lxxxi. 493). It has been named gallium, in consequence of its discovery in France. Its spectrum consists of a brilliant violet line and a feeble band. (107 a) Influence of Different Temperatures-Spectra of Compounds.—Kirchhoff and Bunsen ignited many of the salts of the different metals in flames of very varying temperature, including those of sulphur, carbonic disulphide, diluted alcohol, carbonic oxide, hydrogen, and the oxyhydrogen mixture; and they state that the same metal always produced the same lines, whichever flame they employed to heat it. The It has, however, since been observed by Tyndall, Frankland, Roscoe and Clifton, myself, and others, that, as the temperature rises, a new series of bands become visible in certain cases. spectrum of lithic chloride, in the flame of a Bunsen burner, gives but a single intense crimson line; in a hotter flame, as that of hydrogen, it gives an additional orange ray; and in the oxyhydrogen jet, or the voltaic arc, a broad brilliant blue band comes out in addition. A similar effect is perceived in the case of metallic iron, of thallium, and of other metals when heated by the voltaic arc, which at elevated temperatures furnish much more complicated spectra than when less intensely heated. A. Mitscherlich (Pogg. Annal. 1862, cxvi. 499) has also shown that in flames of low temperature, the lines produced by different compounds of the same metal vary with the compound employed in these cases the spectrum observed is that due to the compound, and not to its elementary constituents; the spectrum of metallic copper, for example, differs considerably from that of an alcoholic solution of cupric chloride, whilst that from an alcoholic solution of cupric iodide differs from both. These observations do not destroy the value of spectrum examination as a means of qualitative analysis, provided that the operator adopts the method laid down by Bunsen and Kirchhoff, who introduce into a coal-gas flame, furnished by one of Bunsen's gas-burners, a chloride of the metal for examination, supported on a loop of platinum wire. 196 SPECTRA OF COLOURED FLAMES. [107 a. The spectra figured by Bunsen and Kirchhoff were obtained in each case by acting upon the chlorides of the several metals; and those of the alkaline earths represent, according to the re 107 a.] SPECTRA OF THE ALKALIES AND ALKALINE EARTHS. 197 searches of Diacon (Ann. Chim. Phys. 1865 [4], vi. 25), mixed spectra, due in part to the oxides, and in part to the chlorides of the metals. In the memoir just cited Diacon has confirmed the observation of A. Mitscherlich that many classes of binary compounds possess proper spectra, differing from those of the pure metals both in the arrangement and the number of the lines. During the decomposition of the iodides, bromides, chlorides, and fluorides of certain metals in the flame of the blow-pipe fed usually with air, or in particular cases with oxygen, brilliant lines were observed, due to the halogen present in the compound; but these lines, the duration of which is very variable, are always accompanied by the spectrum due to the oxide of the metal. By heating even the easily decomposable chlorides of certain metals in a hydrogen flame fed with excess of chlorine, the special spectrum of the chloride may frequently be obtained, when it is seen to differ from the spectrum obtained from a flame supplied with oxygen. For example, in a chlorinating flame, the chlorides of the alkali-metals potassium and sodium give no spectrum at all; the spectrum of lithic chloride is not altered, and pure strontic chloride does not show the orange and blue lines usually regarded as characteristic of the compounds of strontium. The spectrum reactions of the halogens, however, are not sufficiently delicate, nor can they be as yet secured with sufficient facility to render the spectrum test superior to the methods in use by the ordinary mode of analysis: though both Mitscherlich (Pogg. Annal. 1864, cxxi. 459) and Diacon give processes for applying the spectroscope for the discrimination of the halogens when in combination. The first spectrum shown in fig. 83 exhibits some of the fixed dark lines of the solar spectrum contrasted with the position of some of the most important bright lines furnished by the spectra of the alkalies and alkaline earths, when their chlorides are heated upon a loop of platinum wire introduced into the flame of a Bunsen gas-burner. The characteristic bright lines in the case of each metal are distinguished by the letters of the Greek alphabet, the most strongly marked lines being those indicated by the earliest letter. Amongst the various spectra, that of thallium and those of the alkali-metals are the simplest. In the potassium spectrum the most characteristic bright lines are the red line K a, and violet line, K 3. A copious diffused light fills up the central portion of the spectrum. In the case of sodium, nearly the whole of the light is concentrated on the intense yellow double line, Na a. In 198 PROJECTION OF SPECTRA ON A SCREEN. [107 a. the lithium spectrum, a crimson band, Li a, is the prominent line; Li ẞ is seldom visible; but at the elevated temperature of the voltaic arc, an additional blue line becomes very intense. In the spectrum of cæsium, a good deal of diffused light is visible, but the two lines in the blue, Cs a and Cs B, are strongly marked, and may be seen when a quantity of the chloride not exceeding 1000 of a grain of the pure salt is used, or 70% of a grain if diluted with fifteen hundred times as much lithic chloride. Rubidium is not distinguishable in quantities quite so minute. The lines, Rb a and Rb ß, in the blue, and Rb y in the red, are almost equally characteristic, but about 30 chloride is required to render them visible. nised by the single intense green line Tl a. alkaline earths are equally definite though more complicated; generally speaking, the elements of higher atomicity give more complex spectra. of a grain of the Thallium is recog The spectra of the The salts which are most readily volatilized, such as the chlorides, bromides, and iodides of the different metals, give the most brilliant spectra. But it is only in the case of the alkalies and the alkaline earths that the spectra thus obtained are characteristic. Where the spectra of the other metals are required, recourse must be had to Wheatstone's method of taking electric sparks, between wires consisting of the metal of which the spectrum is required; and the electric sparks may conveniently be procured by the employment of Ruhmkorff's coil. The temperature obtained in this way is very intense, and developes bright lines not produced by the heat of ordinary flames. When a compound gas or vapour is made the medium of the electric discharge, the spectra produced are those of the elementary components of the gas. It seems as though, at these intense temperatures, chemical combination were impossible; and oxygen and hydrogen, chlorine and the metals, may therefore all coexist in a separate form, although mechanically intermingled. The application of these processes of optical analysis to the examination of furnace flames, at different stages of various processes in the arts, cannot fail to afford information of high interest, which can be obtained in no other way. Roscoe has already applied it successfully in studying the Bessemer process for steel. Mr. Snelus, of the Dowlais Iron Works, has recently made some further researches on the flame from the Bessemer converter. (108) Projection of Spectra on a Screen.-When it is desired to render the lines produced by the spectra from different metals visible to a large audience, |