Properties of Potassium- Sodium Alio?/. 293 



of the photoelectric metal will increase the force of pro- 

 pulsion on the ion, and hence increase the emission per unit 

 of time, and therefore increase the observed current. 



It is obvious, then, that the ionized gas over the illumi- 

 nated allov has a unilateral conductivity, and will conduct 

 the current from a voltaic cell in one direction, but not in 

 the opposite. Negative electricity can be conveyed from the 

 alloy to the platinum across the rarefied air-space, but not 

 in the opposite direction. Hence, an alternating current 

 whether of low or of high frequency, can be rectified, and by 

 interposing such a light-cell in the circuit of a galvanometer, 

 in which circuit high-frequency oscillations are also created 

 by the inductive action of a discharging condenser, I have 

 been able to rectify these oscillations. The action, however, 

 proved to be much more feeble and irregular than the similar 

 rectifying effect which can be produced by a glow-lamp, 

 and as at that time I had succeeded in finding a far more 

 efficient glow-lamp rectifier or oscillation detector, the ex- 

 periment with the rectifying effects of potassium-sodium 

 allov were not continued. 



The interesting question, however, still remains as to the 

 source and nature of this photoelectromotive force which is 

 produced by the absorption of light by the surface of highly 

 electropositive metals. 



It is much affected by temperature, being increased by 

 heating the alloy and by exposing it to powerful radiation 

 for some time. If a galvanometer is used as described to 

 detect the photoelectric effect, the current is found to be 

 dependent upon the pressure and nature of the residual gas 

 in the tube. 



This was confirmed as follows : — A tube was prepared as 

 above described with potassium-sodium alloy. Before sealing 

 it off the pump, dry hydrogen gas was admitted and pumped 

 out and then admitted again, and the tube exhausted to a 

 pressure of about 0*01 mm. This tube was compared with 

 others, in which the residual gas was air, but the exhaustion 

 carried below O'OOl mm. The former tube exhibited only 

 the very smallest evidence of photoelectric effect as measured 

 by the galvanometer current, whereas the highly vacuous 

 tube with residual air under very small pressure exhibited 

 the effects well. The case of other tubes made as described, 

 in which water vapour was accidentally present, which in a 

 short time liberated hydrogen under the action of the metal ; 

 these after an interval ceased to give any current, even 

 under the action of the strongest illumination, when the 

 platinum plate was connected with the potassium-sodium 



Phil. Mag. S. 6. Vol. 17. No. 98. Feb. 1909. X 



