Intelligence and Miscellaneous Articles. 239 



It is remarkable that all the lines of the spectrum are not reversed 

 simultaneously. Thus the very bright violet line situated nearly in 

 the centre of the space between the two triple lines shows no trace 

 of reversal. In an early communication I will describe another series 

 of experiments, which tend to class the lines of one spectrum in dif- 

 ferent categories. It will be sufficient for me here to say that the 

 only lines I have succeeded in reversing belong to the light emitted 

 by the atmosphere of vapour which envelopes the incandescent 

 metal*. 



After this first experiment, I tried the various metals I had at 

 hand. The same phenomenon was reproduced in divers groups of 

 lines ; and in general the reversal commences with the least refran- 

 gible line of the group, and only continues if the temperature is pro- 

 gressively raised. 



The following are, for the principal metals, the groups of lines 

 which I have been able to reverse. The metals are arranged nearly 

 in the order of the facility of production of the phenomenon ; the 

 lines are designated by their wave-lengths expressed in millionths of 

 a millimetre. 



Sodium .... line D \"=589 



Thallium .... green line A= 535 



Lead violet line X=40G 



Silver , 



I green lines {x=:521 



| violet line. . , X— 424f 



., . . f violet lines between fX=396 



Aluminium. | h, and H, \X=394 



f triple green line . . X=518 - 3, the least refrangible. 

 Magnesium, j triply beyond j ^.^ fte w refrangible . 



f green line X=509 



Cadmium . . < bluish-green line . . X=480 



[ blue line X=467*7 



f green line X=461 



Zinc J green line \=472 



[ blue line X=467*8 



Copper green line X=510 



Iron, cobalt, bismuth, antimony, and gold have given no appearance 

 of reversal. The alkaline salts, especially the chlorides, still more 

 easily produce the reversal of the bright lines. The chlorides of 

 sodium (X=589 and 422) and lithium (\=670 and 458) are the 

 most remarkable. 



* It is understood that the bright lines due to air and aqueous vapour 

 must not be confounded with those of the metal, which furnishes at least 

 two sorts of bright lines. 



t Perhaps really double. 



X Determined photograph ically, with a Nobert's network. The devia- 

 tion was 11° 46' 40" (spectrum c of the 2nd order). The analogue of group 

 b (X=51 8*3, according to M. Angstrom) was deviated 16° 0' 5". The 

 ratio between the wave-lengths is 0*74046. The mean wave-length of the 

 group is equal to 383'3. 



