1919] on Liquid Oxygen in Warfare r>99 



largely on the surface of contact between the metal rods and the 

 liquid oxygen in which they are immersed. i\s the liquid is used up 

 the depth of immersion decreases ; when it passes a certain point, 

 defined by the dimensions of the rods, the rate of evaporation falls 

 off. To prevent this it was found sufficient to wind absorbent 

 asbestos rope round the bunch of conductors, leaving a teased-out 

 brush spreading over the bottom of the flask. In this way the liquid 

 was sufficiently distributed over the metal surface to keep the 

 evaporation uniform, down almost to the last drop. The surface of 

 the external conductor must also be sufficiently large to collect the 

 necessary heat from the air ; and the cross section of the neck of 

 the conductor sufficient to convey this heat down into the liquid. 

 The rate of evaporation will therefore vary somewhat with the 

 external temperature. It was found that a commercial vacuum flask 

 of a capacity of one quart could readily carry the necessary metal 

 conductors to provide an evaporation of two litres per minute for 

 over four hours. The variation was not more tlian 10 per cent. 

 With a rate of three litres a minute, easily maintained for two hours, 

 the variation was only 2 to 3 per cent. If the conductors were 

 allowed to accumulate a thick coat of condensed ice and snow, the 

 rate diminished by as much as 2(> per cent. It Avas necessary to 

 guard the neck socket also from ice obstruction. 



The effect of variations of external temperature is seen with a 

 vessel of water fitted on the conductor. When the. temperature. of 

 the water fell from 17°C. to 7° C. in thirty minutes, the rats of 

 delivery fell from 2*55 to 2 "3 litres a minute. 



SAFETY IN CARRYIXG. 



In order to protect the vessels from rough handling various 

 devices were tried, taking the general form of a double cylindrical 

 case, with shock-absorbing arrangements between. A simple and 

 very effective form was a plain cylindrical basket (shown partly 

 cut open, Fig. .5), large enough to leave a space of nearly 2 inches 

 all around the apparatus, which it carried upon a separate 

 wicker disc A ; this was supported by four indiarubber caljles B 

 a])Out 1 centimetre in diameter, secured under the rim of the 

 l)asket ; four similar shorter fastenings C secured the disc to the 

 bottom of the outer case. The cap of the vacuum flask I) was kept 

 central by three equally spaced loops of similar rubber cable secured 

 at equal distances round the basket carrier at the level of the screw 

 cap (one of these is indicated by E, Fig. 5). 



A fully charged vessel in working order would usually survive 

 the shock of a fall of about three or four feet when so supported : 

 steel or whale-bone springs in place of the rubber cables were found 

 to be too rigid in the restricted space available. 



