90 Hartitry—An Investigation of the Connexion between 
Evidence of the reduction of refractory oxides to the state of metal was obtained 
from the occurrence of groups of metallic lines in their spectra. 
The calcium group of elements is of particular interest in this connexion, 
because their band-spectra have hitherto been generally attributed to compounds 
such as oxides or chlorides, but evidence incompatible with this view is deduced 
from the following facts :— 
(1) The flame-spectrum of metallic calcium is substantially the same as that 
from calcium oxide. 
(2) The oxides and carbonates give no spectrum in the carbon monoxide 
flame, or in the flame of carbon monoxide when burnt with oxygen. 
(3) The oxides and carbonates of the alkaline earths yield very brilliant 
spectra when ignited in the flame of burning cyanogen, which are identical with 
those obtained with the oxyhydrogen flame; and in the case of calcium with the 
spectrum obtained from that metal. 
(4) The chlorides yield very similar if not identical spectra when ignited in 
the carbon monoxide flame, and the spectra are particularly brilliant. 
(5) Cupric oxide, which is easily reduced to copper in the oxyhydrogen flame, 
and yields the complex banded spectrum of the metal, gives only a very feeble 
spectrum of bands in the carbon monoxide flame, along with the pair of lines 
which belong to the oxyhydrogen flame-spectrum. 
This is evidence of the bands being the spectrum of the metal and not of the 
oxide. 
(6) Ferric oxide yields ninety lines, inclusive of five or six sharply defined 
edges of bands, in the oxyhydrogen flame-spectrum ; the lines of this spectrum 
appear as reversed lines in Kayser and Runge’s are spectrum of iron, and also in 
the solar spectrum measured by Rowland. 
On submitting ferric oxide to the action of the carbon monoxide and oxygen 
flame, the same groups of lines are photographed, in the same order of relative 
intensities, but the lines in the whole spectrum are weaker than in the oxyhy- 
drogen flame ; when the conditions as to oxidation and reduction are modified by 
the carbon monoxide being burnt from the oxygen jet in the interior of the 
flame, the oxygen being supplied to the outer tube of the blow-pipe, so that 
the reducing gas is burnt in a powerfully oxidizing atmosphere, a similar 
spectrum is photographed, 
The inferences to be drawn from these experimental results are two: first, 
though the temperature is sufficiently high in the Mecke burner to melt 
platinum (7 = 1775°), and is even higher than this in the carbon monoxide flame, 
the maximum temperature is confined to a smaller area of the flame than admits 
