346 Mr. J. H. J. Poole on a Possible Connexion between 



emitted by the iron probably attaches itself nearly im- 

 mediately after its escape from the surface to a molecule, 

 and then drifts across under the action of the electric field 

 to the grid. But if the iron is magnetized, the moving 

 gaseous ion is subjected to an additional force at right angles 

 both to its direction of motion and to the magnetic field. 

 Sir J. J. Thomson ('Conduction of Electricity through 

 Gases,' Chap, iv.) investigates this effect. He shows, how- 

 ever, that although the magnetic field tends to introduce 

 ionic velocities along the magnetic field, and also perpen- 

 dicular to the magnetic and electric forces, still that these 

 are quite inappreciable compared with the velocity along the 

 electric force at least at air pressure, since the ratio of the 

 velocity due to the electric field to either of the other 



velocities involves the term ^5 , where R is the velocity of 



the ion under unit electric force, and H the magnetic force. 

 This ratio is always large at atmospheric pressures as 

 R=l*5 x 10" 8 and H would probably never exceed 10 5 , and 

 in our experiments it certainly did not exceed 10 3 . Thus it 

 is obvious that the magnetic field will not affect the path of 

 the ion from the iron to the grid, and hence the total current 

 should also be unaffected. This result would not be sound 

 if we were working in a high vacuum, as then we should not 

 be justified in assuming that the photo-electron immediately 

 attached itself to a gaseous molecule. 



As regards the possible effect of the roughness of the iron 

 surface, the case is not so clear. To reduce this effect as far 

 as possible, the iron surface was carefullv polished, but even 

 so the irregularities in its surface must have been quite large 

 compared with the wave-length of the light. There are two 

 cases to be considered (a) it the direction of the emission 

 of the photo-electron by the atom is independent of the 

 magnetic orientation of the atom, and (b) if this direction 

 is not so independent. If supposition (a) is true, it is obvious 

 that the roughness of the surface will not cause a change in 

 the total ionization current when the iron is magnetized, as 

 any effect due to it will not be altered by the field. If (b) is 

 correct it would appear possible at first sight that this might 

 not be the case. AVe have, however, shown earlier in the 

 paper that, owing to the irregularities in the direction of the 

 magnetic field, we should expect on the whole half the 

 electrons emitted by a plane surface to escape either with 

 the field on or off. Local irregularities in the surface would 

 on the average not affect this result, as, for the average atom 

 lying in the surface, half the electrons emitted will still 



