Vol. 6, 1920 
PHYSICS: A. S. KING 
701 
EXPERIMENTS WITH THE TUBE RESISTANCE FURNACE ON 
THE EFFECT OF POTENTIAL DIFFERENCE 
By Arthur S. King 
Mount Wilson Observatory, Carnkgie Institution of Washington 
Communicated by George E. Hale, October 12, 1920 
The form of electric furnace which has been found most effective in 
spectroscopic work is the tube resistance type, in which the substance to 
be vaporized is placed in a graphite tube heated by a current, the elec- 
trodes being applied at the ends of the tube. A column of vapor of almost 
uniform temperature is thus obtained in a considerable length of the tube. 
The temperature can be closely controlled and raised until the vaporiza- 
tion of the tube material becomes very violent, in the neighborhood of 
3000° C. Regulated in this way, the furnace has uniformly emitted a 
definite spectrum of the enclosed vapor at a given temperature, and much 
evidence has been gathered which indicates a high degree of indepen- 
dence of chemical action arising from the use of various compounds or the 
presence of different gases in the furnace chamber, such actions appear- 
ing to affect the general intensity of a spectrum without materially modi- 
fying its character. 
Since an important feature of the furnace is the excitation of the en- 
closed vapor by the high temperature of the tube rather than by the 
passage of a current as in the case of the arc, any approach to arc condi- 
tions which may result from the mass of ionized vapors being exposed 
to the potential difference of the ends of the tube must be carefully taken 
into account. It is obvious that a certain amount of such conduction 
may take place, in spite of the low potentials employed and the very 
small resistance of the graphite tube. Experiments have therefore been 
carried out to see whether, at least for the temperatures which usually 
bring out the important features of furnace spectra, a reduction or elim- 
ination of the potential difference by special means would cause any 
deviation from the spectrum observed at the same temperature with 
the regular operation of the furnace. For this purpose the tube was 
charged with iron, and certain lines very sensitive to change of tempera- 
ture and especially strong in the arc were taken as a test spectrum. The 
bands given by the vaporization of the carbon tube and the lines of titan- 
ium and vanadium resulting from impurities were noted for the same pur- 
pose. 
A reduction of the potential difference was obtained by two methods: 
first, by the use of direct current, which avoided the high momentary 
values of the alternating voltage usually employed; and, second, by using 
an insulated tube to contain the luminous vapor, inside the tube carry- 
ing the current, this inner tube furnishing an extra low-resistance path 
for any current not carried by the furnace tube. In each case, the test 
