592 SCIENCE PROGRESS 



the possibility of some arrangement which would enable large 

 fields to be obtained with only a small loss of energy. How 

 difficult, if not almost impossible, this is with conductors in the 

 ordinary state at low temperatures has been shown by Fabry, 

 who has calculated the conditions to obtain a field of io 5 gauss 

 with copper conductors at the temperature of liquid air. He 

 found that for a coil with an internal space of i cm. radius 

 ioo kilowatts of electrical energy would be necessary to 

 maintain the necessary current. However, the development of 

 heat would be very large, and would require the use of and the 

 evaporation of 1,500 litres of liquid air per hour. Even if it 

 were possible to bring the cooling liquid into close and rapid 

 enough contact with the layers of the coil it would take about 

 700 kilowatts to produce this amount of liquid air, so that the 

 total energy expenditure would amount to about 800 kilowatts. 

 However, it is almost certain that this amount of heat could not 

 be conveyed away sufficiently rapidly from a coil of this size. 

 If the coil is made larger the ampere turns must be increased 

 to give the same field so that the amount of liquid air used 

 would have to be still larger. Probably it would be so difficult 

 and costly as to be impracticable. 



Before considering the deductions which can be drawn from 

 these experimental results it will be useful to refer to investiga- 

 tions on other phenomena. 



The Hall coefficient appears to increase in most substances 

 to about hydrogen temperature and then to decrease, but the 

 measurements are not very decisive and they have not been 

 carried below 14-5° K. yet. 



The Thermo-electric power would appear to decrease to zero 

 at T = o, and to be at low temperatures proportional to T 3 . The 

 Peltier effect to approach zero at T = o and to be proportional 

 to T 4 . The Thomson effect also to reach zero and to be 

 proportional to T 3 . 



The magnetic change in the resistance in the ordinary con- 

 ductive state does not in general call for any particular 

 comment. In the case of bismuth, gold, silver, copper, 

 palladium, and mercury and some alloys of these the resistance 

 increases with increasing fields, and the change is greater in 

 some cases at low temperatures, in others higher, but it follows 

 a regular course for each conductor. However, with iron and 

 nickel the behaviour is irregular, that of nickel being only slight 



