374 Prof. G. D. Liveing on the Line-spectrum of 



In every case the proportion of oxygen to hydrogen was 

 very gradually varied, from excess of hydrogen to excess of 

 oxygen and vice versa ; and different parts of the flame 

 brought in front of the slit. This was done again and again, 

 but no trace of any hydrogen-line could be detected. Except 

 for the D line of sodium, which was always present, and 

 flashed out brightly as particles of dust entered the flame, and 

 for an occasional flash of the two well-known lime-bands, the 

 spectrum was quite continuous, extending from near C to 

 near G, very faint at both ends. 



When the hydrogen was in excess the flame had a reddish 

 tip, which may be what Plucker saw ; but the spectroscope 

 showed that this colour was only due to the fact that (under 

 these conditions) the maximum of the continuous spectrum 

 lay in the orange between C and D. This continuous spec- 

 trum faded away gradually and could be traced down to 

 about the position of C, but there it ended, and there was 

 nothing like a maximum at C. When the oxygen was in ex- 

 cess the maximum of the continuous spectrum shifted into 

 the citron, and I could no longer trace the spectrum so low 

 as C. There was no maximum of brightness about F or G 7 . 



When the separate impinging jets were used the flame had 

 a curious appearance — the oxygen side had a greenish hue, the 

 hydrogen side a fainter orange hue. In the mid-plane would 

 be the most active combustion and the hottest temperature, 

 and there, if anywhere, we might expect to see the hydrogen 

 rays developed. 1 examined both the broad side and the 

 edge, but could see no hydrogen-line at all. When the spec- 

 troscope was directed on to the edge, the radiation would be 

 received from a considerable thickness of the hottest part of 

 the flame, and this would seem to be the most favourable 

 condition for detecting the hydrogen rays. 



Having failed to get the hydrogen rays at all in any of 

 these flames I next tried a cyanogen flame burning in oxygen, 

 and introduced a very small jet of hydrogen into the middle 

 of the flame. The cyanogen flame in oxygen is much hotter 

 than the hydrogen flame, and gives a brilliant spectrum of 

 channellings. The flame, however, has a very definite inner 

 cone which gives this bright spectrum, while the flame 

 outside the cone hardly shows any of it. The introduction of 

 the smallest quantity of hydrogen alters the general appear- 

 ance of the flame, but seems to alter the brightness only of 

 the spectrum. No hydrogen-line could be detected in the 

 spectrum of the cone, or the flame above it. One of the 

 channellings in the spectrum of the cone overlaps the place 

 of H a , so that it might be difficult to detect H if it were 



