110 Proceedings of the Royal Society of Edinburgh. [Sess. 
containers was of cupronickel (an alloy having about thrice the con- 
ductivity of German silver;; the bore was J in., the thickness of the metal 
X V in., and the length 9J ins. Had it been possible to disregard the 
influence of the up-flowing stream of cold air along the neck— that is to 
say, had neck-loss been merely a question of the conductivity, sectional area, 
and length of the tube for a given temperature difference — the neck-loss 
of each of these modified containers would have been 17 5 times that of 
the German 50-lbs. container referred to above. Actually the rates of 
evaporation of these more robust flasks showed little if any increase upon 
those of an equal number of containers of the usual design, selected at 
random. Their daily evaporation losses proved to be respectively 3*43, 
3-75, 4 00, 2*75, 3*31, 3*31, 3*62, 3*69, 3*75, and 4*00 lbs. 
Acknowledgment is due to Messrs J. Mallinson, B.Sc., W. Cooper, M.A., 
B.Sc., and J. J. Brodie, Government research workers, for their help in the 
experiments here discussed. In the tests upon the 3-litre flask, weighings 
and temperature observations had to be taken at all hours of the night and 
day for a fortnight. 
Summary. 
(1) J. A. Harker and his co-workers having shown that of all the 
possible causes of heat-inflow to liquid air in a vacuum flask only three, 
viz. radiation, conduction across the vacuum space, and conduction along 
the neck of the flask, are of importance, the writer illustrates by two 
instances an experimental method enabling these sources of heat-transfer 
to be separately assessed. 
(2) In the two selected instances radiation proves to be the main 
method of heat-transfer. 
(3) In the second example (a 3-litre metal flask) the analysis is carried 
further, and the emissivity of the surfaces and pressure in the vacuous 
space are determined. The reasons for the relatively high emissivity are 
discussed. 
(4) Pyrometer measurements in the necks of long-necked, large metal 
storage flasks (containers) show the loss due to neck-conduction to be 
either zero or entirely negligible. The results indicate that the necks of 
such vessels may be shortened with advantage. 
(5) The tranquil evaporation rates are given of ten metal containers 
having relatively short, stout necks, in proof of the foregoing conclusion. 
(Issued separately June 20, 1921.) 
