Electron Atmospheres of Metals. 311) 



conducting projections. Soft metals cannot be used, for if 

 they are brought in contact there is the possibility of micro- 

 scopic elevations being raised when they are separated. 1 

 accordingly commenced observations with the small optical 

 flats of polished speculum metal, which is glass hard. The 

 surfaces were known to be true planes to within 1/20 of a 

 wave-length, and they were highly polished and free from 

 scratches. Two of these flats were placed in contact, no 

 pressure being applied, and it was found that, so far as 

 ordinary measurements would show, they were in metallic 

 contact. The upper plate slid about easily over the lower 

 with little friction, showing that they were separated by an 

 air-film } and from what we know about the behaviour of true 

 flats brought together under these conditions, we may be 

 certain that the actual distance between them is certainly 

 8 or 10 wave-lengths. Optical tests showed that the plates 

 rounded off a trifle at the edge, the surface of the outer zone 

 being a trifle convex, which excludes the possibility of true 

 metallic contact occurring at some point at the edge. The 

 speculum plates were now separated by two very fine quartz 

 fibres, about 5 wave-lengths in diameter, as measured with 

 the microscope, and again conductivity was observed, a 

 current of several milliamperes being observed with an 

 applied potential of one volt. The current increased rapidly 

 and very steadily with the application of pressure, returning 

 to exactly its original value as the pressure was removed.. 

 Fine sulphur dust was then allowed to settle from an air 

 suspension upon one of the plates and the other plate then 

 laid carefully upon it. Again conduction was observed. 

 Metallic contact at any point was out of the question in this 

 case. There was, however, the possibility that a conducting 

 sulphide was formed, and consequently finely powdered 

 quartz was substituted which gave similar results. 



It was impossible, however, to measure the exact distance 

 between the plates under these circumstances, or be quite 

 sure that the presence of the material used to keep the plates 

 at a fixed distance was .without influence. Fused quartz is 

 known to be one of the best insulators which we have, and 

 yet a film of adsorbed moisture might be responsible for the 

 conductivity under the conditions of the experiment. I find 

 it difficult to believe, however, that currents of the magnitude 

 obtained, in some cases amounting to a considerable fraction 

 of an ampere ; can be carried in this way. 



The ideal method is to exclude all material, and work 

 under conditions which enable the exact distance between 

 the plates to be determined. I accordingly used as opposed 



