602 silsbee: conduction at low temperatures 



approximate). It is seen from the table that at each tempera- 

 ture this magnetic field is much more nearly a constant of the 

 material than either the current or the current density. In the 

 case of mercury the effect of a magnetic field on the resistance 

 in the superconducting state has not been measured. For tin 

 the threshold value at 2°K is about 200 gauss, which is in good 

 agreement with the slightly larger values computed from the 

 threshold current corresponding to a slightly lower temperature. 

 In the case of lead the agreement of the observed critical field 

 (600 gauss at 4°K) with the computed values is not so good, par- 

 ticularly in the case of the straight wire. Any discrepancy here, 

 however, is easily explained by the possibility (frequently re- 

 ferred to by Onnes) of the existence in the wire of thin spots 

 where the field intensity would be much greater for a short 

 length. 



Further experiments immediately suggest themselves. The 

 critical magnetic fields for mercury should be determined. 

 The relation here advanced would indicate a critical field of 

 only about 15 gauss at 4? IK and less than 100 gauss at 3?6K. 

 It would also be of interest to observe the threshold value of 

 current when the material is in very thin films. In this case, for 

 a given section of material the magnetic field resulting from a 

 given current density is less than in the case of a straight wire, 

 and the threshold current density would consequently appear 

 larger. 



The theories thus far proposed by Onnes, 7 Lindemann, s and 

 Thomson 9 to account for superconductivity do not specifically in- 

 dicate the existence of a critical magnetic field, and only the 

 latter (by assuming a saturation effect) accounts for a threshold 

 current density. If it is true, as indicated in this paper, that the 

 magnetic effect is the more fundamental, it would seem that this 

 fact might afford a valuable clue leading toward a more satisfac- 

 tory theory of the superconducting state and perhaps of metal- 

 lic conduction in general. 



7 



Onnes. Comm. Phys. Lab. Leiden, No. 119. 

 8 Lindemann, F. C. Phil. Mag., 29: 127. 1915. 

 'Thomson, J. J. Phil. Mag., 30: 192. 1915. 



