378 



De La Rue and Muller. 



[May Hi, 



tion). By operating in this way we have been able to reproduce over 

 and over again phases which are more or less transient in all tubes ; 

 indeed, we encountered great difficulties when we first commenced 

 our labours with tubes supplied by various makers, for after a very 

 short time they completely and permanently changed so as no longer 

 to present the splendid stratifications witnessed on a first trial. 



All the tubes which we exhausted were provided with two glass 

 stop-cocks, one of which was connected with the gas generator, the 

 other to the system of pumps. The rinsing of the apparatus was first 

 effected many times with a high pressure water trompe (head of water 

 106 feet, producing a vacuum of 12 millims.) ; the tube was then ex- 

 hausted by a mercurial pump (Alvergniat's make), and finally by the 

 Sprengel. In our later experiments the measurements of low pressures 

 were made by means of the M'Leod gauge, which indicated pressures 

 down to 0-00005 millim. 



By absorbing carbonic acid with potash, and, more especially, hydro- 

 gen with spongy palladium, a vacuum was obtained so perfect (pressure 

 0-000055 millim., 0'066 M0 that the potential of 11,000 was insufficient 

 to cause a discharge. 



We have adopted several expedients for ascertaining the resistance 

 of the tubes. In the first instance we placed the tube in one of the 

 arms of a Wheatstone bridge, but the oscillations of the current were 

 so great that it was necessary to substitute a vacuum tell-tale tube for 

 the galvanometer ; this became illuminated whenever the currents in 

 the two branches were unequal. In spite of this device it was impos- 

 sible to balance exactly the resistance of a tube by a metallic resist- 

 ance. It was found that the resistance of a vacuum tube varied 

 greatly as soon as the current passed and usually diminished, but 

 when once the current was interrupted it was necessary to increase 

 considerably the balancing resistance in the symmetrical branch of 

 the Wheatstone bridge, in order to reproduce an illumination. After 

 standing at rest for some time the tube took up its original condition. 



In fig. 4 A Z is the battery, the A terminal of which is connected at A', 

 in the bridge arrangement, with two equal fluid resistance tubes, FR 

 and FR', of 420,000 ohms, placed in vessels containing ice, to keep 

 them at a constant temperature ; an adjustable coil resistance is in- 

 serted between B and D ; the tube T T', to be tested, is placed between 

 D and C, the Z terminal of the battery being connected to D. When 

 the resistance is greater or less than that of the tube to be tested 

 there is an illumination in the detector tube between B and C ; but 

 when a current passes in T T', balanced by a proper adjustment of the 

 coil resistance, then the glow in the detector ceases. 



Better results were obtained either by reproducing the deflections of 

 a galvanometer, observed during the course of the experiments, with 

 metallic resistance substituted for the tube ; or by measuring with an 



