48 
DR, A. SCHUSTER OH THE SPECTRA OP METALLOIDS. 
at the point of the flame. He compares the colour to that of the phosphorescence 
seen in tubes containing a mixture of oxygen with a small quantity of nitrogen 
(pure oxygen does not show the phosphorescence). Now the colour of this phos¬ 
phorescence is exactly the same as that of the spark in oxygen which gives the 
continuous spectrum. 
It is reasonable to suppose that M. Becquerel heated up the oxygen to a suffi¬ 
ciently high temperature in the oxyhydrogen flame to render it luminous. That 
this is possible appears from an experiment made by P luck eh, who, working with 
an excess of hydrogen in the oxyhydrogen flame, saw the lines of hydrogen in his 
spectroscope. 
I have already stated that I do not know whether the continuous background 
which is seen at higher pressure when the line spectrum appears has its origin in 
a separate cause, or whether it has to be brought into connexion with the continuous 
spectrum of low temperature. The following experiment may assist those vdio like 
to form an opinion on the matter :— 
Mr. Lockyer has made use of the fact that when a spark is taken between metallic 
poles, and a lens is employed to form an image of the spark on the slit of the spectro¬ 
scope, the lines of the metals do not stretch across the field, but are confined to the 
neighbourhood of the poles. As the pressure is decreased, some of the lines decrease 
rapidly in length, and finally disappear, wdiile the effect on others is little perceptible. 
I find that when oxygen is the gas through wdiich the spark strikes, the lines of the 
gas as well as the metallic lines seem to stretch away from the poles and fade away 
towards the middle. The centre of the field is taken up by the second line spectrum, 
which I have called the compound line spectrum. Here is what happens :— 
Experiment 7 .—At atmospheric pressure the lines of the elementary line spectrum 
were seen to stretch across the field; the continuous spectrum illuminating the back¬ 
ground was strongest in the green. As the pressure was decreased the compound 
lines appeared much broadened, and at the same time a dark streak was seen to pass 
through the continuous spectrum in the centre of the spark. This means that at 
that pressure the continuous spectrum is chiefly confined to the electrodes. As the 
pressure was decreased to about half an atmosphere, the continuous spectrum only 
occupied a narrow band adjoining the electrodes. The compound lines came out 
strong in the centre of the field and were better defined. At still lower pressures 
the lines of oxygen separated in the middle of the field, as the continuous spectrum 
had done before ; and the continuous spectrum now seemed to be gradually absorbed 
into the electrodes, and finally disappeared altogether. Gradually the oxygen lines 
became shorter and shorter until they could be no more distinguished from the 
metallic lines. The centre of the field was now entirely occupied by the compound 
line spectrum. It is necessary to add that it appears conclusively from the experi¬ 
ments of Wullner, as well as my own, that the continuous spectrum is really due to 
oxygen and not to the metallic poles. The reason why there should be a difference 
