PAPERS OX CHEMISTRY AND PHYSICS 



271 



lire to air into several tubes by a laborious process. 

 Eubidiiim, being somewliat sticky, was somewhat harder 

 to introduce. The potassium-sodium filled tubes seemed 

 to give some remarkable results, whereas rubidium tubes 

 seemed worthless, on account of extremely high plate re- 

 sistance. Accordingly, several tests were made on the po- 

 tassium-sodium content tubes. Figs. 3 and 4 show plate 

 current gi'id voltage cui'ses for various plate voltages 

 from zero up to 120 volts. It will be noticed that for 

 low plate voltages curves are similar to those of a vacuum 

 tube, and that as plate voltage increases, the bend at the 

 point of saturation flattens out until at 120 volts the 

 curve is a straight line from - 12 volts grid potential 

 to —12 volts grid potential. This would indicate that 

 the tube might be a good amplifier. The cuiwes of Fig. 

 5 show this to be the case. Each curve of this set is ob- 

 tained with constant plate current, hence u, the no load 

 amplification constant, is the negative slope of the plate 

 voltage grid voltage characteristic. (See Van der Bijl or 

 Morecroft). The curves of Fig. 5 indicate that these 

 tubes are similar to high vacuum amplifiers as regards 

 variation of amplification constant and the slope indi- 

 cates a high value for u, about 10 to 12. The dotted 

 curve A of the figure is tyi^ical for low vacuum tubes 

 containing air, nitrogen, hydrogen, helium, etc.. and B 



