Dielectric Constant of Liquid Oxygen and Liquid Air. 361 



We found on trial that although the small condenser had a capacity 

 of only 0-001031 microfarad, it held its charge when charged with 

 100 volts, and placed beneath the surface of liquid air in the most 

 extraordinary way. The test for insulation was as follows : 



The small condenser was charged with 100 volts, and discharged 

 through the galvanometer instantly. The galvanometer throw was 

 95 scale divisions. 



The small condenser was then charged and allowed to stand ten 

 minutes insulated. It was then discharged through the galvanometer, 

 and the throw was 90 scale divisions. In like manner it was charged 

 and insulated for forty-seven minutes, and the throw was then 80 scale 

 divisions. 



The above figures show that the charge of the small insulated con- 

 denser decreased only by about 15 per cent, in three-quarters of an 

 hour when placed beneath liquid air, and hence the loss of charge in 

 one-tenth of a second was quite inappreciable.* 



The same remarkable insulation is found when the small condenser 

 is held in the cold gaseous oxygen lying above the liquid oxygen. 

 The low temperature of 182 C. prevents any sensible leakage 

 across the glass distance pieces, and also increases the specific resist- 

 ance of the glass itself. 



As a further instance of the very high insulating power of liquid air, 

 we may mention that we charged the small condenser when immersed 

 in liquid air with a Wimshurst electrical machine, and, after insu- 

 lating the condenser and waiting a few moments, closed the terminals 

 of the condenser by a wire. A small spark was seen at the contacts. 

 We thus constructed a little Ley den jar, the dielectric of which was 

 liquid air, and the coatings the aluminium plates. This liquid 

 Leyden jar held its charge perfectly. 



Having satisfied ourselves in this manner that the condenser when 

 immersed in liquid air would lose no sensible portion of its charge 

 during the fraction (about one-tenth) of a second in which the charge 

 and discharge key was moving between its contacts, we proceeded to 

 experiment in the following manner. The condenser was placed in a 

 very large vacuum vessel, holding about two litres of liquid oxygen, 

 and it was charged as described, and discharged into a very good mica 

 condenser, made by Dr. Muirheacl, which had an exceedingly high 

 insulation. The process of charging and discharging ten times occu- 

 pied, perhaps, two seconds. 



* These figures do not of course measure tlie electrical resistance of the liquid 

 oxygen alone. They show, however, that the immersion of the condenser in liquid 

 oxygen enormously decreased or entirely destroyed any surface leakage over the 

 small glass separators, and, as we have found by an independent examination, 

 increased the resistivity of the glass itself. The specific resistancs of liquid oxygen 

 itself is exceedingly high. 



2 E 2 



