1879.] on the Reversal of the Lines of Metallic Vapours. 353 



.chlorides along wifcli metallic lithium. The two violet lines of rubidium 

 and the two blue lines of caesium were reversed, as when sodium was 

 used instead of lithium, and as the lithium gave no sensible vapour, 

 the observations could easily be continued for a much longer time with 

 the same tubes. No other absorption lines could be discerned. It may 

 be observed, however, that it is not easy to obtain a source of light 

 sufficiently rich in the least refrangible red to allow of observations on 

 the absorption of light so little refrangible as the red rubidium lines. 

 A platinum wire, heated nearly to fusion by an electric current, 

 appeared to give the brightest light in this part of the spectrum, but 

 of that light no definite absorption by the rubidium could be observed 

 in the red. We then had some mixtures of carbonate of caesium with 

 carbon, and of carbonate of rubidium with carbon, prepared by 

 charring the tartrates ; and observed the results of heating these 

 mixtures in narrow porcelain tubes, placed vertically in a furnace, as 

 described in our first communication on this subject (" Proc. Roy. 

 Soc," vol. xxvii). A small quantity of the caesium mixture, intro- 

 duced into a tube at a bright red heat, showed instantly the two blue lines 

 reversed and so much expanded as to be almost in contact. The width 

 of the dark lines decreased as the caesium evaporated, but they remained 

 quite distinct for a very long time. A similar effect was produced by 

 the rubidium mixture, only it was necessary to have the tube very much 

 hotter, in order to get enough of violet light to see the reversal of the 

 rubidium lines. In this case the two lines were so much expanded as 

 to form one broad dark band, which gradually resolved itself into two 

 as the rubidium evaporated. The reversal of these lines of caesium 

 and rubidium seems to take place almost or quite as readily as that of 

 the D lines by sodium, and the vapours of those metals must be 

 -extremely opaque to the light of the refrangibility absorbed, for the 

 absorption was conspicuous when only very minute quantities of the 

 metals were present. The red, yellow, and green parts of the spectrum 

 were carefully searched for absorption lines, but none due to caesium 

 or rubidium could be detected in any case. It is perhaps worthy of 

 remark that the liberation of such extremely electro -positive elements 

 as caesium and rubidium from their chlorides by sodium and by lithium, 

 though it is probably only partial, is a proof, if proof were wanting, 

 that so-called chemical affinity only takes a part in determining the 

 grouping of the elements in such mixtures ; and it is probable that the 

 equilibrium arrived at in any such case is a dynamical or mobile equili- 

 brium, continually varying with change of temperature. 



Our next experiments were with charred cream of tartar in iron 

 tubes, arranged as before. In this case a broad absorption band 

 appeared, extending over the space from about wave-length 5,700 to 

 5,775, and in some cases still wider, with edges ill-defined, especially 

 the more refrangible edge. By placing the charred cream of tartar in 



2 c 2 



