on Low-Temperature Besearch, 1893-1900. 705 



logically inferred from the behaviour of pure metals down to —200° C. 

 that, at the absolute zero, they would entirely cease to dissipate tho 

 energy of an electric current transmitted through them. But at 

 — 252° C, marked inconsistencies manifested themselves. Instead of 

 continuing their straight downward course, the resistance-curves bent 

 round, indicating the survival, at 0° absolute, of a finite value for this 

 property. An emphatic warning was thus conveyed against trusting 

 to the continuity of change. 



Thermo-electric action had been studied by Professor Tait at 

 temperatures above 0° C. ; the modifications produced in it by cooling 

 down to — 200° C. were ascertained by Professors Dewar and Fleming 

 in 1895. They present no uniform character. The curves showing 

 the fluctuation with temperature of the thermo-electric power of 

 various metals, do not, in all cases, even approximate to straight 

 lines. Some — notably those of iron and bismuth — show abrupt 

 changes of direction, indicating reversals of the " Thomson effect" at 

 those points. Others are inflected in a manner suggestive of a zero 

 of thermo-electric power at the zero of cold. But these indications 

 are most likely misleading. There is every reason to believe that 

 the rate of diminution of thermo-electric power, as of electric re- 

 sistance, would fall off notably before that extreme point was 

 reached. 



Another set of experiments served to test the influence of cold 

 upon magnetisation. They justified the expectation that magnetic 

 moment would gain strength proportionately to the deprivation of 

 heat. Its value was usually increased, in the fixed state established 

 after some alterations, to the extent of thirty to fifty per cent., by 

 lowering the temperature from +75° C. to —182° C. Exceptions to 

 this rule were, however, met with. A nickel-steel magnet, for instance, 

 is acted upon oppositely to one of carbon-steel. A subordinate result 

 of these experiments was to show that one of the best ways of ageing 

 a magnet is to dip it several times into liquid air. This removes 

 sub-permanent magnetism, and induces stable relations favourable to 

 definite observation. 



The magnetic permeability of iron over a descending range of 

 temperature was made the subject of long and laborious comparisons. 

 They showed it to be slightly diminished by immersion in liquid 

 oxygen. That is to say, a greater magnetic force was needed to pro- 

 duce a given amount of magnetisation in the cooled material. As 

 usual, however, apparent inconsistencies were recorded. Hardened 

 iron reversed the behaviour of soft iron. Its permeability increased 

 at low temperatures, and, for certain values of the magnetising force, 

 as much as five times. " Hysteresis loss," or the dissipation of energy 

 incidental to a cyclical process of magnetisation, was found, on the 

 other hand, to vary little, if at all, with temperature. These diverse 

 effects were attributed by Professor Fleming to the closer contiguity 

 at extreme cold of the " molecular magnets," from the collineation 

 of which external magnetic moment results. Their groupings and 

 Vol. XVI. (No. 95.) 3 a 



