Band and Line Spectra of the same Metallic Elements. 91 
of a sufficient quantity of the metal being reduced and vapourized, so as to yield 
a spectrum with so great an intensity as to be visible, or to be photographed ; 
secondly, though there can be no great difference between the temperatures 
obtained in the Deville blow-pipe flame when supplied with oxygen and 
hydrogen, or oxygen and carbon monoxide, yet the reducing power of the 
oxyhydrogen flame under the conditions which obtain in the flame exceeds that 
of carbon monoxide. The reason of this is evident from an investigation by 
A. Gautier (C.R. 1906, 142, 1582) of the reduction of water-vapour by carbon 
monoxide. The temperature at which it occurs is from 1200° to 1250°. 
Equilibrium is established between the reacting substances when the volume 
of hydrogen is double that of the carbon monoxide, in the sense of the following 
equation :— 
The reduction of CO, proceeds at 1300°, and the conditions of equilibrium 
are then represented as follows :— 
CO, + 3H, = CO + H,0 | 2H,. 
This latter equation shows that coal-gas may be substituted for pure hydrogen 
in the oxyhydrogen blow-pipe, but this has not been found to be satisfactory 
in practice; and it also affords an explanation why a filter-paper may be 
burnt in the flame without interfering with the reduction to metal caused by the 
hydrogen gas.* At or from 1200° to 1400°, or at a white heat, CO, = CO + O is 
the rule, for even at 1250° carbon monoxide deposits carbon in a porcelain tube. 
This conclusion receives confirmation from the facts observed with regard to 
the reduction in the oxyhydrogen flame of various refractory metallic oxides. 
Most closely allied to calcium, strontium, and barium is magnesium. 
Kainite, K,SO,;MgSO,MeCl,6H,0, and kieserite, MgSO,H,O, both equally 
strongly show the magnesium flame-spectrum as given in ‘ Banded Flame- 
Spectra of Metals,” p. 343, and observed in 1888, by Liveing and Dewar, when 
magnesia was heated in the oxyhydrogen flame. 
The lines and bands recorded on my photographs of kainite and kieserite are 
the following:—)dd 5209, triplet 5184, 5174, 5168: a group of six so-called 
‘* oxide bands’’—Ist, 5004, 2nd 4993, 3rd 4988. 4th 4971, 5th 4960, 6th 4944: also 
strong bands from AA 4100 to 3700. Hence it is impossible to believe that this 
spectrum is not the spectrum of metallic magnesium, for the first of the above 
lines, viz.: 5209 (L. and D. 5210), was obtained by the spark from a solution of 
* If many filters are burnt, the effect upon the spectrum is similar to that observed when coal-gas 
instead of hydrogen is supplied to the blow-pipe, that is to say, there is a strong continuous spectrum due 
to the burning of carbon monoxide. The combustion of at least ten filter-papers is necessary to show this, 
