1003 
§ 22. Remark upon the increase in the resistance of bismuth in a 
magnetic field. A friendly remark by Prof. H. pv Bors leads us to 
a further development of our ideas concerning the occurrence of a 
maximum in the isopedals for the increase in the resistance of 
bismuth. 
Our measurements make it probable that the maximum found 
by Bake at the temperature of liquid air must be ascribed to the 
presence of impurity or to some modification occasioned, for 
instance, by mechanical treatment,: and that this maximum is not 
obtained with pure normal bismuth at these temperatures. The 
values which we obtained at the boiling point of hydrogen make 
it also certain that neither is a maximum to be found between 
the temperatures of liquid air and of liquid hydrogen. In the region 
of hydrogen temperatures a falling off in the rate of increase of the 
resistance of the bismuth wires is clearly apparent. The existence 
of this diminution has been proved twice, and on-each oceasion for 
different currents (and, as is evident from the table, for various 
fields). But a maximum, tbat is to say, a return to smaller values, 
we have not obtained. From the course of the curves given by 
Benet Beckman in Comm. N°. 130, it still remains possible that the 
phenomenon reaches a limiting value. From various analogous 
phenomena we might quite well expect something of this kind to 
happen at extremely low temperatures. In Comm. N°. 129a we 
commented under I, § 2, upon the uncertainty as to whether a 
maximum is reached at these temperatures, or rather an asymptotic 
approach would be found to be made to a limiting value, stating 
that “Perhaps as the purity increases the maximum in the isopedals 
is displaced towards lower temperatures’. The measurements we 
have made with the plates B? lay further emphasis upon the ““per- 
haps.” As the temperature falls to 20° K the plates Bir, Bip, 
which were not so pure as the wire, exhibit no diminution in the 
rate of increase. And vet, on account of the greater impurity sus- 
pected in these plates, they should be expected to exhibit a maxi- 
mum between 14.°5 and 73°.K, if there were a maximum for pure 
bismuth at temperatures lower than 14°.5 K and if this maximum 
were displaced towards lower temperatures only by an increase in 
the purity of the material. In contrast with this we here find that 
only the diminution in the rate of increase remains between 20° K 
and 14°5 K. Further experiments upon different bismuth prepara- 
tions are of course highly desirable. 
