458 On the Electron Atmosphere (?) of Metals. 



results. The air-film in every case remained an insulator 

 until the voltage reached the sparking potential, when a charge 

 passed over to the electroscope. This operation was usually 



(1) Switch. 



(2) Lamp res. of -4400 ohms. 



(3) Static voltmeter. 



(4) Switch. 



(5) Spark-gap. 



(6) Wilson electroscope. 



(7) Galvanometer. 



instantaneous, although occasionally a gradual leak through 

 the insulation seemed to take place. Dust particles or a 

 water-film could in such cases be instantly detected, and 

 hence the electrodes were always carefully cleaned and 

 gently heated before using. If consistent sparking potentials 

 with accompanying insulation between the electrodes down 

 to distances o£ one-half of a sodium wave-length can thus be 

 obtained, it would seem that the hypothesis of an " Electron 

 Atmosphere" is untenable. At this distance the voltage 

 necessary for discharge was between 80 and 100 volts, and 

 no more delicate test than the electroscope deflexion could 

 well be demanded of an insulator. Wood's " Electron 

 Atmosphere J ' seemed to extend to a distance of as much as 

 30 wave-lengths. A recent paper by Brown * seems to 

 indicate that a film of moisture constitutes the " Electron 

 Atmosphere." 

 Ann Arbor, Mich. 

 Nov. 8, 1913. 



* Brown, Phys. Rev. Oct. 1913. 



MMMH 



