48 
When this stage had been reached, and when especially 
no trace of the carbonic oxide bands could be seen in the 
tube, the platinum spiral (e) containing either the diamond 
or the graphite was rendered incandescent by means of an 
electric current. 
The spiral contained sometimes a piece of natural graph- 
ite, sometimes a Cape diamond, but the result was the same 
in the two cases. As soon as the platinum spiral had been 
sufficiently heated, a channelled space spectrum appeared in 
the capillary part of the tube. This channelled space spec- 
trum was carefully compared with the spectrum of carbonic 
oxide obtained from charcoal and found to be identical with 
it. No band or line could be seen in the tubes thus pre- 
pared which was not also seen in a tube containing carbonic 
oxide. The spectrum which appears when a Leyden jar is 
introduced into the circuit is different, but here also we 
found that every line was due either to oxygen or to carbon. 
Two lines were seen in the green and greenish yellow which 
are not contained in any map of the spectrum of carbon or 
of oxygen, lines which had not been seen in a great many 
oxygen tubes prepared and examined by one of us. But it 
was found on further investigation that these are really 
oxygen lines, which only appeared at very high tempera- 
tures. The capillary portion of the tubes we used were 
much shorter than that in the ordinary Pliicker’s tubes, and 
this accounts for the temperature of the incandescent as being 
hi^ffier than usual. As one of the lines is near the unknown 
O 
aurora line, its wave length was determined and found to 
be 5591, showing it to be decidedly less refrangible than the 
aurora line. 
The experiment was repeated in four different tubes and 
many times in each tube ; but whether graphite or diamond 
