1920-21.] Evaporation of Liquid Air in Vacuum Flasks. 107 
It is also known, when p is measured in mms. of mercury and A in 
sq. cms., that c takes a value for air of approximately 2 x 10~ 5 at 30° C. 
To apply equation (6), this figure has first to be adjusted to the average 
temperature of 0’5 ( 9 3 +d 2 ) ; Table VI has to be consulted for the conduction- 
loss at any given value of 6 1 ; and the latent heat of the liquid mixture in 
the flask and the dimensions of the flask have to be taken into account. 
The pressure in the vacuum space, p, can then be computed. It was found 
in this case to be 0'00038 mm. mercury. Considering that the flask had 
been evacuated twenty months when the tests were made, this degree of 
tenuity may be regarded as satisfactory. 
Effect of Surrounding the 3-Litre Flask by an Insulating 
Medium. 
As radiation proved to be the most important cause of evaporation, and 
as, in radiation between two given surfaces, the temperature of the hotter 
surface is the all-important factor, it appeared probable that evaporation 
would be appreciably reduced by insulating the flask. The neck was 
extended by soldering to it a length of brass tube, and the whole of the 
flask, excepting a short part of that tube, was encased in slag-wool in a 
sheet-metal canister. The immediate effect was an increase in the rate of 
evaporation ; but after the slag- wool had been given time to cool down, its 
influence became beneficial. Eventually the rate of evaporation settled to 
a value thafi was 82*5 per cent, of the rate given at the same external 
temperature (8° C.), when the flask was uninsulated. The increased bulk 
and clumsiness of the insulated flask outweighed, from the practical point 
of view, any advantage gained. 
The Neck-Loss of Liquid- Air Containers. 
Measurements were made to ascertain the temperature-gradients along 
the necks of four containers for the purpose of finding whether the 
relatively great length of the necks (see fig. 1) was necessary. Tempera- 
tures were taken by means of a Foster pyrometer designed for low- 
temperature observation. The wires of the thermo-element were 1*5 mm. 
in diameter. The thermometric scale of the galvanometer was tested at 
room temperature, and at that of boiling liquid air, to ascertain the 
proportional correction to apply to readings. In taking a reading the 
thermo-junction was lowered to the desired point in the neck and allowed 
to stay there until the galvanometer needle became stationary. Inasmuch 
as the thermo-couple was in the up-flowing current of cold air, and not 
