the plate, and the other being a tube carrying at one end the slit, furnished with micrometer screw, through which the beam of light passed, and at the other end an object-glass for the purpose of rendering the rays parallel. The luminous vapours of the metals under examination were obtained by placing a bead of the chloride or other salt of the metal on a platinum wire, between two platinum electrodes, from which the spark of a powerful induction coil could be passed. In order to obtain a more intense, and therefore a hotter, spark than can be got from the coil alone, the coatings of a Leyden jar were placed in connexion with electrodes of the secondary current respectively. When this arrangement was carefully adjusted, the two yellow sodium lines were observed to be separated by an apparent interval of two millimetres, as seen at the least distance of distinct vision.

The position of the blue line, or rather blue band, of lithium was then determined with reference to the fixed reflecting scale of Steinheil's instrument, by volatilizing the carbonate of lithium in the first place on a platinum wire between platinum electrodes, and secondly on a copper wire between copper electrodes. A bead of pure chloride of strontium was then placed on new platinum and copper wires between two new platinum and copper electrodes, and the position of the blue line Sr & read off upon the same fixed scale: a difference of one division on the scale was seen to exist between the positions of the two lines, the lithium line being the more refrangible. The salts of the two metals were then placed between the poles at the same time, and both the blue lines were simultaneously seen, separated by a space about equal to that separating the two sodium lines. When experimenting with this complete instrument, we were unable to observe any other blue lines in the pure lithium spectrum than the one above referred to: we have, however, noticed the formation of four new violet lines in the intense strontium spectrum, and we now believe that the other two lithium lines mentioned in our letter to the "Philosophical Magazine" are caused by the presence of the most minute trace of strontium floating in the atmosphere, and derived from a previous experiment.

We have convinced ourselves by numerous observations that the currents of air caused by the rapid passage of the electric spark between the electrodes are sufficient to carry over to a second. set of electrodes placed at the distance of a few inches a very perceptible quantity of the materials undergoing volatilization. The greatest precautions must hence be taken when the spectra of two metals have to be compared; and no separate observations of the two spectra can be relied upon, unless one is made a considerable space of time after the other, and unless all the electrodes which have been once used are exchanged for new

ones.

Kirchhoff, in his interesting Memoir on the Solar Spectrum and the Spectra of the Chemical Elements,1 noticed in the case of the calcium spectrum that bright lines which were invisible at the temperature of the coal-gas flame became visible when the temperature of the incandescent vapour reached that of the intense electric spark.

We have confirmed this observation of Kirchhoff's, and have extended it, inasmuch as we, in the first place, have noticed that a similar change occurs in the spectra of strontium and barium; and, in the second place, that not only new lines appear at the high temperature of the intense spark, but that the broad bands, characteristic of the metal or metallic compound at the low temperature of the flame or weak spark, totally disappear at the higher temperature. The new bright lines which supply the part of the broad bands are generally not coincident with any part of the band, sometimes being less and sometimes more refrangible. Thus the broad band in the flame spectrum of calcium named Ca ß is replaced in the spectrum of the intense calcium spark by five fine green lines, all of which are less refrangible than any part of the band Ca ß; whilst, in the place of the red or orange Ca a, three more refrangible red or orange lines are seen (see Fig. 38). The total disappearance in the spark of a well-defined yellow band seen in the calcium spectrum at the lower temperature was strikingly evident. We have assured

1 Kirchhoff on the Solar Spectrum, &c. Translated by H. E. Roscoe. (Macmillan, Cambridge, 1862.)

ourselves by repeated observations that, in like manner, the broad bands produced in the flame spectra of strontium and barium compounds, and especially Sr a, Sr ß, Sr y, Ba a, Ba ß, Ba y, Ba 8, Ba ε, Ba n, disappear entirely in the spectra of the intense spark, and that new bright non-coincident lines appear. The blue Sr & line does not alter either in intensity or in position with alterations of temperature thus effected, but, as has already been stated, four new violet lines appear in the spectrum of strontium at the higher temperature.

If, in the present incomplete condition of this most interesting branch of inquiry, we may be allowed to express an opinion as to the possible cause of the phenomenon of the disappearance of the broad bands and the production of the bright lines, we would suggest that, at the lower temperature of the flame or weak spark, the spectrum observed is produced by the glowing vapour of some compound, probably the oxide, of the difficultly reducible metal; whereas at the enormously high temperature of the intense electric spark these compounds are split up, and thus the true spectrum of the metal is obtained.

In conclusion, we may add that in none of the spectra of the more reducible alkaline metals (potassium, sodium, lithium) can any deviation or disappearance of the maxima of light be noticed on change of temperature.

APPENDIX C.

ON THE SPECTRA OF ERBIUM AND DIDYMIUM, AND THEIR COMPOUNDS.

Bunsen1 has shown that the rare earth erbia is distinguished from all other known substances by a peculiar optical reaction of the greatest interest. This solid substance when strongly heated in the non-luminous gas flame gives a spectrum containing bright lines, which are so intense as to serve for 1 Ann. Ch. Pharm. cxxxvii. p. 1.

detecting this substance. This singular phenomenon does not however constitute any exception to the law of exchanges, for Bunsen has shown that the bands of maximum intensity in the émission spectrum of erbia coincide exactly in position with the bands of greatest darkness in the absorption spectrum. A similar inversion of the didymium absorption bands has also been observed by Bunsen.1

Some very interesting observations have also been made by Bunsen upon the absorption spectrum of didymium,2 from which we learn that the didymium spectrum, and also that of erbium, undergoes changes if examined by polarized light according as the ordinary or the extraordinary ray be allowed to pass through the crystal. These changes only become visible however when a powerful battery of prisms and a telescope of high magnifying power are employed. According to the direction in which the ray of polarized light is allowed to traverse the crystal of didymium sulphate is the position of the dark absorption bands found to vary; whilst the bands produced by the solution of the salt in water are again different. Very remarkable are the small alterations in the position of the dark bands of the didymium salts, dependent upon the nature of the compound in which the metal occurs. These changes are too minute to be seen with a small spectroscope, but are distinctly visible in the larger instrument. "The differences thus observed in the absorption spectra of different didymium compounds cannot in our complete ignorance of any general theory for the absorption of light in media be connected with other phenomena. They remind one of the slight gradual alteration in pitch which the notes from a vibrating elastic rod undergo when the rod is weighted, or of the change of tone which an organ-pipe exhibits when the tube is lengthened."s

1 Ann. Ch. Pharm. cxxxi. p. 255; Phil. Mag. vol. xxviii. p. 246. 2 Phil. Mag. vol. xxxii. 1866, p. 177.

3 Whilst these pages are passing through the press, an announcement is made by Mr. Sorby (Chemical News, xix. p. 121) of the discovery of a new metal, whose presence is indicated by the appearance of peculiar absorption bands seen in certain specimens of zircon. The above-quoted observations of Bunsen would, however, lead one to receive with caution conclusions solely relying on the existence of new absorption bands.

APPENDIX D.

DESCRIPTION OF THE SORBY-BROWNING MICRO-SPECTROSCOPE.

The construction of this instrument is represented in Figs. 47 and 48. The prism is contained in a small tube (a), which can be removed at pleasure, and which is shown in section in Fig. 48. Below the prism is an achromatic eyepiece having an adjustable slit between the two lenses; the upper lens being

furnished with a screw motion to focus the slit. A side slit capable of adjustment admits when required a second beam of light from any object whose spectrum it is desired to compare with that of the object placed on the stage of the microscope. This second beam of light strikes against a very small prism suitably placed inside the apparatus, and is reflected up through the compound prism, forming a spectrum in the same field with that obtained from the object on the stage.

a is a brass tube carrying the compound direct-vision prism.

N