﻿with Coloured Glasses and with the Spectroscope. 7 



noted the number of minutes after the commencement of the 

 blow at which the lines appear. 



At the first two or three observations I attempted to make a 

 thorough note of the changes as they occurred throughout the 

 whole spectrum, but afterward abandoned it as utterly impos- 

 sible, as at the beginning of the second period the lines come in 

 so fast and the changes are so rapid that they cannot be accu- 

 rately noted at the exact moment of their occurrence. I there- 

 fore confined myself to a few degrees at each observation, and by 

 this method was enabled to note accurately, and at the exact 

 moment of their occurrence, slight changes which otherwise 

 might have escaped notice. Note was also taken of the changes 

 in the general appearance of the whole spectrum during the 

 successive stages of the process. After having made half a 

 dozen observations, while viewing the spectrum of the flame 

 from the converter while it was being heated for another charge, 

 it was discovered that a movement of the eye before the eye-glass 

 occasioned a similar movement of the lines of the spectrum along 

 the scale, on which their position could thus be made to differ 

 more than half a degree. I have seen no notice of this in the 

 statements of others ; and it may account for some of the appa- 

 rent discrepancies. Thereafter, when taking the readings of any 

 of the lines, the position of the eye was so adjusted as to bring 

 the sodium-line exactly at 50°. Owing to the extreme brilliancy 

 of the flame, the aperture maybe made exceedingly narrow, and 

 thus the many lines of the spectrum, which with a duller light 

 and broader gauge would be blended together, may be separated. 



At the beginning of the blow the spectrum is continuous and 

 very faint, and generally extends from 35° to 120°, covering 

 about three-fourths of the length attained in the second period. 

 This increases slightly in extent and brightness until the ap- 

 pearance of the sodium-line. This line appears at the end of 

 the first period, at the beginning of a more decided flame. It 

 comes flashing through from one extremity to the other for an 

 instant, and then disappears, only to return the next instant in 

 brighter flashes, which are continued for about a minute, by 

 which time the line becomes permanently established. On one 

 occasion the sodium-line, instead of flashing and disappearing as 

 usual, continued visible after a few seconds, and expanded, and 

 contracted in width almost isochronously until it became perma- 

 nently established. The appearance of this line indicates the 

 termination of the first period. This period I have found to vary 

 in extent from three to seventeen minutes in blows lasting from 

 thirteen to twenty-seven minutes. None of the other lines make 

 their appearance in vivid flashes as does the sodium. The 

 lithium-line becomes visible three or four minutes after the first 



