1914] on The Coming of Age of the Vacuum Flask 243 



Another method of producing the requisite high vacua is by 

 means of a mercurial vacuum made by distilhng mercury through 

 the annular space while exhausting by a good pump.* On placing 

 a liquid gas in such a flask the mercury vapour left in the annular 

 space is condensed as a bright mirror on the cool inner wall ; 

 and if used solely for low temperature work, this is entirely 

 satisfactory, the need for silvering the inner surfaces is then partly 

 obviated. 



This action is illustrated in an exhausted globe containing a little 

 mercury. There is a shallow depression in the globe into whicli a 

 Httle liquid air is poured (see 11, Fig. 2) ; almost at once a mirror of 

 mercury is deposited inside on the cooled part and the increased isola- 

 tion results in a diminution of the rate of evaporation. A difficulty 

 arises, however, if the vacuous space is not of reasonable cross-section, 

 because the vapour of mercury is not only dense, but has a high 

 viscosity, and therefore cannot get freely to the cooled surface, which 

 results in the production of a poor mirror of limited extent. Thus, 

 if another globe is taken having the mercury in a small bulb separated 

 from the main globe by an inch of capillary tubing,f then on cooling a 

 small area with a sponge of Hquid air, the vapour saturating the globe 

 is deposited as a localised mirror. If, however, a second area be then 

 similarly cooled no deposit is obtained, because practically no mercury 

 vapour can get in throu^i the capillary ; this is proved by decanting 

 a little mercury out of the small bulb into the globe, when instantly 

 a mirror is deposited on the cooled area. 



The practical effect of this property is shown by the relative 

 isolations of two cylindrical vacuum vessels in the same condition as 

 to exhaust and both containing a little mercury in the bottom. The 

 annular vacuum space in one is about 1 cm. wide, while the width 

 of the other is only about 2 mm. Liquid air is poured into both 

 and the relative rates of evaporation measured by connecting 

 successively to each, an indiarubber tube delivering into a gas jar 

 fihed and inverted over water. A metronome indicates the time 

 I taken to fill the jar with the evaporating air in each case ; the result 

 shows that the vessel with the narrow annular exhaast space is worse 

 .than the one with the wider space, and on examination this is 

 explained by the smaller extension of the mercmy mirror in the 

 narrow-spaced vessel. 



A pure water vapour discharge tube provided with two platinum 

 wire electrodes illustrates the condensing action of liquid air in the 

 production of a high vacuum in vacuum vessels for any vapours 

 therein contained. The tube shown was about 5 cm. diam. and 

 35 cm. long. This tul)e is saturated with water vapour at the 

 temperature of melting ice. On connecting the electrodes to a 



* Proc. Roy. Inst., xiv. p. 395. 



t See Fig. 3, Proc. Roy. Inst., Jan. 12, 1894. 



R 2 



