in the Electrical Resistance of Bismuth, 307 



of these specimens the resistance measurements were repeated 

 many times, and in both directions, to obtain an assurance 

 that the relation of resistance and temperature was perfectly 

 definite, and that proof might be obtained that all the three 

 specimens of pressed bismuth wire as well as the bismuth 

 melted in the tubes, did not undergo any permanent change 

 in their condition by the action of the low temperature, but 

 that there was a definite relation between temperature and 

 electrical resistance in the manner shown in the curves. 

 Similar experiments made with pure antimony drawn up into 

 glass tubes did not show the appearance of any minimum 

 resistance for any temperature between 0° and —200° P 2 . In 

 no other case amongst all the metals and alloys examined by 

 us has any phenomenon presented itself similar to that shown 

 by the above specimens of pure bismuth. So far, therefore, 

 as these investigations have gone they show that the behaviour 

 of bismuth at low temperatures in respect of electrical con- 

 ductivity is anomalous. The temperature-resistance curves 

 of these specimens are quite different to those of other pure 

 metals. In our first communication on the electrical re- 

 sistance of metals (see Phil. Mag. October, 1892) we showed 

 that carbon taken in the form of a carbon filament from an 

 incandescent lamp increases its resistance continuously when 

 cooled down to the boiling-point of oxygen, and we suggested 

 that for pure non-metals continual lowering of the tempe- 

 rature towards the absolute zero would in all probability 

 continually increase the specific electrical resistance. In the 

 specimens of pure bismuth A and B there is no indication at 

 the lowest temperature reached, viz. —235° P 1? that the 

 resistance lines for this pure bismuth would have another 

 point of inflexion. If, then, the specific resistance of these 

 samples of pure bismuth goes on increasing continuously down 

 to the absolute zero of temperature, as it would appear to 

 do, in that respect they behave like a non-metal, having, 

 however, a very much larger temperature coefficient than 

 carbon. The behaviour of bismuth in a magnetic field in 

 regard to electrical resistance, and also the fact that 

 different specimens of bismuth have in some cases positive 

 and in other cases negative temperature coefficients at tem- 

 peratures above 0° 0., has been examined by other ob- 

 servers. Without entering into any historical statement of 

 the course of investigation in this matter, brief reference may 

 be made to two papers on this subject which have a bearing 

 on the facts here described by us. In 1886 an experimental 

 study of the influence of magnetism and temperature on the 

 electrical resistance of bismuth, and also of its alloys with 



