150 Proceedings of Royal Society of Edinburgh. [sess. 
that is, we have the transverse effect as function of the magnetisa- 
tion alone. We must also consider the case where the field is kept 
constant and the temperature is varied : we then get the transverse 
effect as a function of the temperature, and we can compare it 
with the thermo-electric force of the couple formed by the con- 
ductor considered with lead when one junction is kept at a con- 
stant temperature, say zero centigrade, and the other is given 
successively the same temperatures as those at which the transverse 
effect by constant field was observed. 
In the paramagnetic metals, iron, nickel, cobalt, the magnetisation 
is also a function of the temperature ; this would have to be taken 
into consideration in experiments with these metals ; in them the 
curve giving the variation of transverse effect with temperature 
would combine the peculiarities of the curves giving the variation 
of the magnetisation and of the thermo-electric force with tem- 
perature. 
The only direct experiments made to determine whether or not 
a relation exists between the two effects are by Ettingshausen. He 
did not succeed in establishing any direct relation. This was to 
be expected, since, in the material he used, the transverse effect is 
really made up of two effects. 
Perhaps the best metal in which to compare the two effects is 
bismuth. We have results for different specimens of this metal. 
Lebret* has examined the transverse effect at constant field strength, 
through a long range of temperature, in practically pure bismuth. 
Dewar and Fleming f have examined the thermo-electric force of 
bismuth-lead couples, in which one junction was kept at 0° C. and 
the other given temperatures from 180° C. to 100° C. 
In the following tables the results for two specimens of bismuth, 
from the papers quoted above, are given. The transverse results 
are relative ; the thermo-electric absolute. 
* Communications from Physics Lab., Leiden, No. 19. 
t Philosophical Magazine , 1895. 
[Plate I. 
