1878.] on the Reversal of the Lines of Metallic Vapours. 351 



ment, were found to be identical in position with the well-known violet 

 lines of rubidium. Next appeared faintly the channelled spectrum of 

 sodium in the green, and then a dark line in the blue, very sharp and 

 decided, in the place of the more refrangible of the characteristic lines 

 of caesium in the flame spectrum. As the temperature rose these dark 

 lines, especially those in the violet, became sensibly broader ; and then 

 another fine dark line appeared in the blue, in the place of the less refran- 

 gible of the caesium blue lines. During this time no dark line could 

 be observed in the red, but as the temperature rose a broad absorption 

 band appeared in the red, with its centre about midway between B 

 and C, ill-defined at the edges, and though plainly visible not very dark. 

 The lines in the violet had now become so broad as to touch each other 

 and form one dark band. On cooling, the absorption band in the red 

 became gradually lighter without becoming defined, and was finally 

 overpowered by the channelled spectrum of sodium in that region. 

 The double dark line in the violet became sharply defined again as the 

 temperature fell. There are two blue lines in the spectrum of rubidium 

 taken with an induction-coil very near the two blue lines of caesium, 

 but they are comparatively feeble, and the two dark lines in the blue 

 which we observed in the places of the characteristic blue lines of 

 caesium we believe must have been due to a small quantity of caesium 

 chloride in the sample of rubidium chloride. These blue lines were 

 •not, however, visible when some of the rubidium chloride was held in 

 the flame of a Bunsen's burner, nor when a spark was taken from a 

 solution of the chloride ; but the more refrangible of them (Csa) was 

 visible in the spark of an induction-coil, without a Leyden jar, taken 

 between beads of the rubidium chloride fused on platinum wires. 



When a tube containing caesium chloride and sodium was observed, 

 in the same way as the former, the two dark lines in the blue were 

 seen very soon after the heating began, and the more refrangible of 

 them broadened out very sensibly as the temperature increased. The 

 usual channelled spectrum of sodium was seen in the green, and an 

 additional channelling appeared in the yellow, which may be due to 

 caesium or to the mixture of the two metals. We have at present no 

 metallic caesium wherewith to decide this question. Indeed the caesium 

 chloride used was not free from rubidium, and the dark lines of 

 rubidium were distinctly seen in the violet. 



It is remarkable that these absorption lines of caesium coincide 

 with the blue lines of caesium as seen in the flame, not with the green 

 line which that metal shows when heated in an electric spark of high 

 density. It is to be observed, however, that when sparks from an 

 induction-coil without a jar are taken between beads of caesium chlo- 

 ride, fused on platinum wires, a spectrum similar to the flame spectrum 

 is seen, and it is only when a Leyden jar is used that the spectrum is 

 reduced to a green line. In like manner both the violet lines of 



