Magnetic State of Iron and its Photo-Electric Properties. 347 



escape, as the only case in which the chance of escape of an 

 electron is affected is when the plane of direction of emission 

 is exactly parallel to the surface at the point, and this case is 

 not more likely to occur when the direction of the surface, as 

 well as the magnetic field, varies slightly. It might, how- 

 ever, be suggested that owing to the irregularity of the 

 surface an electron which had already escaped from the 

 surface might strike it again, and so become entrapped. 

 This appears not very likely to occur, both owing to the 

 very short free path of the electron in air, at ordinary 

 pressures about '00007 cm., and also to the strong electric 

 field which is tending to drag the electron away. Once 

 the electron had become attached to a gas molecule, we 

 would expect the roughness of the surface to have a still 

 smaller effect, as the free path of an ion is smaller than 

 that of a gas molecule, which itself is about 1/llth of 

 that of the free electron. 



We may thus, it would appear, take it as probable that 

 neither the direct effect of the magnetic field on the ionic 

 current nor the error introduced by the roughness of the sur- 

 face, could have a very large effect on the results given in 

 Table II. These results would apparently show that the chance 

 of an atom emitting an electron is not affected by its magnetic 

 orientation with respect to the incident light, and accordingly 

 would yield no evidence as to the identity of the photo- 

 electric and magnetic electrons in the atom. It is to be 

 noticed, however, that, in deducing that a change of direction 

 of emission of the electron due to magnetization would have 

 no effect on the total electronic current, it has been tacitly 

 assumed that all the electrons emitted downwards into the 

 surface are absorbed, and also that there is no diffusion back 

 of ions against the field. If the first assumption is not 

 justified, and some electrons escape from the surface after 

 refraction in the metal, we might expect that, if the electrons 

 were emitted nearly parallel to the surface, when the iron is 

 magnetized, on the whole, a greater percentage of electrons 

 might escape after refraction in the surface, thus leading to 

 an increase in the ionization current. On the other hand, i£ 

 there was a leakage of electrons back to the surface, either 

 due to the thermal agitation of an ion being sufficient to 

 carry it back a short distance against the field or possibly 

 due to actual reflexion of an electron by a gas molecule, it 

 would probably tend to diminish the current on magneti- 

 zation, as the ions, on the whole, would be formed closer to the 

 surface when the electrons are omitted obliquely, and would 

 therefore be more likely to diffuse back against the field to 



