Dielectric Constant of Liquid Oxygen and Liquid Air. 307 



be in a position shortly to furnish further information on this point, 

 and, also, if possible, to say whether the fall in dielectric constant is 

 accompanied by a reduction in the refractive index ; that is to say, 

 whether Maxwell's law is obeyed at low temperatures. 



We may add that we have already devised a method by which it 

 will be possible to construct a condenser without the above-described 

 distance pieces, and hence to free the resulting measurement from the 

 small uncertainty amounting, perhaps, to about 1 per cent. which 

 may affect the above-given numerical results, and which comes in in 

 consequence of the doubt existing as to the exact area of the separa- 

 tors, and also the exact dielectric constant of the glass at the low 

 temperature. 



It is interesting to observe that the numbers which we have found 

 above for the dielectric constant of liquid oxygen and liquid air are 

 not very different in order, though somewhat smaller than the dielec- 

 tric constant as already determined for some other liquid gases,* such 

 as nitrous oxide and carbon dioxide. 



In conclusion, we may add that we have been again much indebted 

 to Mr. J. E. Petavel for his kind assistance in making the above- 

 described observations and measurements. 



Note added December 15. 



In connection with the above investigation, it is interesting to note 

 one remarkable difference between the magnetic susceptibility of 

 oxygen in the liquid and in the gaseous state. The mass of 1 c.c. of 

 gaseous oxygen, taken at 15 C. and 760 mm., is 0*00134 gramme. 

 The mass of 1 c.c. of liquid oxygen, taken at 182 C. and 760 mm., 

 as determined by one of us (J. Dewar), is T1375 gramme. Hence 

 the ratio of the density of liquid oxygen to that of gaseous oxygen is 

 849 to 1. 



The magnetic susceptibility of gaseous oxygen at 15 C. and 760 mm., 

 as obtained from the figures given by Faraday and E. Becquerel, is 

 0143 X 10~ c per unit of volume, whilst the magnetic susceptibility in 

 the liquid state is, as we have shown,t 228 x 10~ c . Hence the ratio of 

 the magnetic susceptibility of liquid oxygen to that of gaseous 

 oxygen for equal volumes is 1594 to 1. 



In other words, the magnetic susceptibility of liquid oxygen is nearly 

 twice as great as that of gaseous oxygen for equal masses. The 

 inference is that magnetic susceptibility is not merely a property of 

 the molecule per se t but is a function of the state of aggregation. 



* See F. Linde, ' Journal de Physique,' vol. 5, Sept., 1896, p. 413, " On the 

 Dielectric Constant of Liquid Gases." 



t See Fleming and Dewar, ' Roy. Soc. Froc.,' vol. 60, p. 283, December; 1896. 



