se 
to disappear. After ten minutes 100 volts would maintain an are scarcely 
one cm. long. The metallic lines had disappeared almost completely. 
The carbon and cyanogen bands were even stronger than in the first pho- 
tograph. Air was then passed through the hollow carbon into the are and 
globe. The are lengthened and the metallic lines reappeared. When the 
current of air was shut off and a stream of CO, turned on the metallic 
lines were weakened, but they did not disappear as long as the CO. was 
flowing. 
It appears that a comparatively rapid disintegration of the carbon 
poles is necessary to furnish enough material in the are to bring out 
clearly the metallic lines, which are due to very small quantities of the 
metals contained in the carbons as impurities. The rapidity of the disin- 
tegration depends upon the amount of O present in the globe. After the O 
had become exhausted by allowing the are to burn a few minutes the 
wasting away of the poles was very slow, indeed. When CO, was intro- 
duced the wasting of the poles increased, and, as noted before, the metallic 
lines appeared. Air still further increased the disintegration of the poles 
and the brilliancy of the metallic spectra. When pure O was passed into 
the globe and are the poles were rapidly consumed. The metallic spec- 
trum was very bright, likewise the carbon any cyanogen bands. When 
the lower hollow carbon was replaced by a carbon with a sulphur core 
(made by pouring hot sulphur into a hollow carbon) S vapor filled the 
globe and displaced the air very soon after the arc was started. The 
metallic lines did not appear, but the carbon and cyanogen bands were 
strong. 
THE SPECTRUM OF CYANOGEN. By ArtTHUR L. FoLrEy. 
[Abstract. ] 
This investigation was made with the grating and accessories described 
in a previous paper and with a Helios enclosed arc lamp. The lamp was 
modified to allow the lower carbon to project below and outside the globe. 
A solid upper carbon was used. The lower carbon, which was hollow, 
had several small holes bored through its walls near the upper end, so 
that a gas blown in at the lower end would be introduced into both the 
are and the globe. The lamp was not air-tight, but the circulation was 
7—ScCIENCE. 
