692 
MR W. CROOKES ON RADIANT MATTER SPECTROSCOPY. 
left frequently shows the orange band in the radiant matter tube, though before this 
treatment the original substance shows nothing. 
The Elimination of Mercury Vapour from Vacuum Tubes. 
92. The presence of mercury vapour is liable to interfere with the spectrum indica¬ 
tions, and when working in an entirely new field, where the mutual reaction of bodies 
one on the other is the subject of the research, it is always desirable to reduce the 
unavoidable constants to a minimum. In a tube exhausted in the ordinary manner 
by the Sprengel pump more mercury vapour is present than is generally supposed. 
Unless the tube is exhausted and sealed off without loss of time, mercury diffuses 
through the connecting tubes, and soon gets into the vessel under experiment. It is 
not easily detected by the spectroscope so long as the tube remains cold, but when the 
passage of the current has warmed the glass and the metal poles the mercury lines 
become distinctly visible, and the same result is obtained at once when the tube is 
heated by a lamp.'”' 
The following experiment illustrates this behaviour of mercury: —A small radiant 
matter tube, having flat aluminium poles connected to platinum terminals passing 
* In incandescent lamps exhausted by means of a mercury pump, a peculiar blue phosphorescence is 
observed to fill the globe when the filament is raised much above its normal incandescence. This pheno¬ 
menon has been examined and discussed by Mr. Edison in America and by Mr. Preece in England (Proc. 
Roy. Soc., vol. xxxviii., No. 236, p. 219). During the course of my work on the manufacture of incan¬ 
descent lamps I have frequently noticed the same effect. Examined in the spectroscope the phosphorescent 
light filling the globe shows the mercury lines strongly. In order to examine this phenomenon more 
closely the following experiments were tried An incandescent lamp with fine carbon filament which 
had been plated up in vapour of carbon tetrachloride was placed in one of the sides of a Wheatstone 
bridge and connected with a dynamo-electric machine, where by means of a suitable rheostat it could be 
raised to any desired degree of incandescence. A spectroscope was adjusted so that the slit was opposite 
the rarefied space between the two limbs of the filament. With an E.M.P. of 64’5 volts, a current of 
0'97 ampere, and alight of twenty candles, no lines were visible in the spectrum; on increasing the 
volts to 73'6 the current became 1T5 amperes, and the c. p. 43. At this stage flashes of blue phosphor¬ 
escence filled the bulb at regular intervals of about ten seconds, and at the same instant the mercury 
lines were seen in the spectroscope, 2984, 3006, 3354, and 5265 being most prominent. On increasing 
the E.M.E. to 8T3 volts, the current became T28 amperes, and the light seventy-four candles, the flashes of 
phosphorescent light became more frequent, pulsating very regularly once every three seconds, and at 
each flash showing the mercury spectrum strongly. The E.M.F. was next increased to 99'7 volts, and the 
amperes to 1*83; at this point the light was equal to 172 candles. The flashes of light became more and 
more frequent, rising quickly from once in two seconds till they occurred too rapidly to count, and the 
blue phosphorescence became continuous and seemed to fill the globe. The mercury spectrum at this 
stage was permanent, and most of its lines could be detected. The flashes of light were probably owing 
to some such action as this:—the hot filament offering a considerable resistance to the passage of the 
current, the long length of insulated wire in the circuit (about half a mile, including the lamps used for 
house-lighting) charged up like a condenser and then at intervals discharged through the highly rarefied 
vapour in the bulb. 
