204 [June 19, 



great difficulty. They vary also in width, from a mere hair's breadtli 

 to six or seven times the apparent width of the slit. 



On exhausting the tuhe in which the discharge is made, at first 

 the only change is that the brilliant lines lose a little of their lustre, 

 till at pressures varying from 3 to 0*5 the spectrum rather suddenly 

 fades away, sometimes leaving only a suspicion of one or two lines ; 

 with others the least-refrangible rays vanish, while the violet remain, 

 though very faint, especially with aluminium. In hydrogen spectra 

 the three bright bands of this gas vanish at unequal densities ; and 

 it is remarkable that this occurs when the gas is diluted to the same 

 proportions by mixing air with it. 



Exhausting yet further, this transition spectrum becomes again 

 bright ; fresh lines appear, and it is changed into a new one, which, 

 however, is never as splendid as that at common pressure, especially 

 at the red end, and in which the very brilliant lines are less frequent. 

 This want makes the difference between the two kinds of spectra 

 seem greater than it really is. Fewer lines are visible in the rarefied 

 media, and of these about four-tenths are not found in the spectra of 

 common pressure. 



If the tables in which the measures are given be examined in re- 

 ference to the points alluded to as doubtful, it will be obvious, 



1 . That many lines are found in all the gases, and in many, perhaps 

 all the metals : the existence of such lines must be independent of 

 the chemical nature of electrodes or media ; it is otherwise with their 

 brightness, which may be intense with one substance and feeble with 

 another. This unchemical origin is still more clearly shown by a 

 modified experiment of Pliicker, where the discharge is made by the 

 induction of glass without the presence of any metal. When the same 

 glass vessel was filled in succession with nitrogen, oxygen, and hy- 

 drogen, though not above twenty-three lines were seen in its capillary 

 tube, and those very faint, yet more than half the number were 

 common to two of the gases, or to the three. These might perhaps 

 be referred to soda or lead detached from the glass ; but some of them 

 are not found in those spectra. 



2. The difference between the common-pressure, transition, and 

 rarefied spectrum shows that the character and even the existence of 

 certain lines depend on the mere density of the medium, the chemical 

 circumstances remaining unchanged. 



3. That the spectra are not merely superposed without change is 



