604 Sir James Dewar [Jan. 17, 



As the exhaust increased, this iuternal pressure fell, but nut 

 at the same rate. The difference of rate of fall of the manometer 

 and barometer gives the lag caused by the cooling of the liquid 

 oxygen. With a steady reduction of pressure of oi centimetres of 

 mercury (corresponding to an ascent of 3 miles) the internal mano- 

 meter fell 2s centimetres in So minutes, and then required a further 

 :^0 minutes to reach the full reduction of 34 centimetres. The curve 

 (Fig. 9) shows a more rapid reduction to 40 centimetres below normal 

 (4 miles ascent) in 9 minutes. A lowering of 28 centimetres was 

 then registered by the liquid oxygen, and the remaining 12-ceuti- 

 metre fall took a further 20 minutes, while the outside barometer 

 was kept steady at the 4 miles level. When this point had been 

 reached, nearly 6 per cent, of the original liquid oxygen had been 

 e\'aporated, and its boiling temperature was reduced from 90" Absolute 

 to below 83° Absolute. 



To supply sufficient oxygen for two persons (4 litres a minute 

 measured under ordinary conditions) a steady current of IG Avatts 

 was passed through the coil in the liquid oxygen. The internal 

 manometer rose in the manner shown by the curve (Fig. 9) and the 

 evaporation increased, the external pressure being maintained con- 

 stant. In 5 minutes a delivery of 2 litres a minute was reached, 

 3 litres a minute in 9 minutes, and the full rate of 4 litres per 

 minute in 18 minutes, by which time the inside pressure had in- 

 creased 12 centimetres. A more rapid increase in the rate of 

 delivery was easily obtained by a larger initial current, which was 

 reduced as the rate increased. 



EVAPORATION FE03I CONTAINERS NOT IN VERTICAL POSITION. 



The use of vacuum containers in circumstances where they were 

 liable to movement about the vertical position would result in the 

 liquid being thrown and splashed about in the vessel ; but by pack- 

 ing the interior with a suitable liquid oxygen absorbent this does 

 not occur. The best substance for the purpose was determined as 

 follows : — 



Equal bulks of various materials were packed into a silvered 

 glass vacuum vessel. They were then saturated with liquid oxygen, 

 the superfluous liquid drained off, and the amount retained weighed. 

 The following table gives the results in grams ; the volume of space 

 filled with each absorbent lieing 350 cubic centimetres. 



Weight of liquid 

 oxygen held. 



1. Cocoanut charcoal (coarse) ... 50 grams 



2. Cocoanut charcoal (fine) .... 185 „ 



3. Respirator charcoal ..... 100 ,, 



4. Light wood charcoal ..... 140 ,, 



5. Asbestos 290 „ 



6. Sponge 80 „ 



7. Kieselguhr ...... 225 „ 



