EFFECT OF TENSION ON CERTAIN ABNORMAL METALS. 59 



the changes of dimensions. The precise value of this effect cannot be 

 computed without a more detailed picture of the entire mechanism 

 than we have at present, but we can at least see what its sign is; as 

 the atoms are compressed more closely together there will be less 

 difficulty for the electrons to make the leap, and the mean free path, 

 and so the conductivity, will increase. 



So much for the mechanism in the case of normal metals. This 

 picture would lead us to expect a decrease of resistance with increasing 

 pressure, as is normal. To explain the behavior of those abnormal 

 metals whose resistance increases under pressure I believed that there 

 might be two possibilities. In the first place there might be such 

 abnormalities in the law of force between the atoms that the amplitude 

 of vibration increases as the atoms are brought closer together, instead 

 of decreasing as normal. I thought that this was probably the case 

 with bismuth, and suggested that the same abnormality would explain 

 the increase of volume on freezing. For such a metal the paths of the 

 electrons are still to be thought of as through the substance of the 

 atoms, and the interference with the free path to take place on making 

 the jump from one atom to the next. A second possibility I thought to 

 explain the behavior of lithium. Here the electrons occupy spaces in a 

 lattice between the atoms, and conduction consists in motion of the 

 electron lattice through the atomic lattice. This is similar to the view 

 of Wien ^^ and Lindemann ^^ as to the general character of conduction; 

 I believe that this can be the mechanism only in exceptional cases. 

 The effect of pressure on such a mechanism is to constrict the channels 

 between the atoms through which the electrons pass. A simple cal- 

 culation will show that the constriction of the channels due to change 

 of distance between atomic centers is much more than the opening of 

 the channels due to decreased amplitude of atomic vibration. Hence 

 the pressure coefficient of resistance of such a substance would be 

 expected to be positive, as it actually is. With regard to the other 

 abnormal metals,^'' I did not have any positive basis for deciding to 

 which type calcium and strontium belong, although I expressed my 

 belief that probably calcium belonged to the lithium type, and I 

 believed that antimony belonged with bismuth on the basis of its 

 expansion on freezing. 



Since publishing my pressure data, the crystalline structure of cal- 



ls W. Wien, Columbia Lectures, 1913, 2tH48. 



16 F. A. Lindemann, Phil. Mag. 29, 127-140, 1915. 



17 Reference 3, p. 183. 



