Effects Transverse Galvanomagnetic and Thermomag netic. 47 
Discussion of Results. 
An inspection of the table shows that the four effects have not always 
a common sign, but that in some cases two of the effects are positive and 
two negative. 
Zahn * put forward the suggestion that, while R and S may have different 
signs from Q and P, it will always be found that R has the same sign as S, 
R P 
and Q the same sign as P,f so that the ratios and ~ are always positive. 
Zahn appears to have drawn this conclusion from slender experimental 
evidence, but its truth is borne out by the observations recorded in 
this paper. 
R Q 
It will be observed that the values of g and ^ do not vary very widely 
from metal to metal, although the individual effects vary both in magni- 
tude and sign. The ordinary electron theory of conduction in metals, 
taken in conjunction with J. J. Thomson’s suggestion of a local magnetic 
field in the immediate neighbourhood of a molecule, is sufficient to account 
for the variation in the sign of the effects from metal to metal, but it 
cannot account for the difference between the signs of R and S, and those 
of Q and P, in the same metal. 
For this some other modification of the theory must be sought. 
G. H. Livens J has recently developed a theory of conduction in metals 
in which he arrives at the four expressions quoted below : 
* Ann. d. Phys ., vol. xiv, p. 886 (1904). 
t Zahn’s actual statement is that Q and P have always opposite signs, but the definition 
of the positive direction of P used by the author is opposite to that used by Zahn. 
X Phil. Mag., vol. xxx, p. 526 (1915). 
