212 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1909. 
A second obstacle remains still to be overcome. Instead of air 
supercharged with a volumetric content of 47 per cent oxygen, it is 
necessary to obtain practically pure oxygen. This can be attained, 
thanks to processes of rectification based on those employed in the 
alcohol industry. In such a process there are two circulating streams 
inside of a column, one from the bottom to the top, of practically 
pure oxygen gas, and the other from the top to the bottom, of liquid 
containing a large proportion of nitrogen. The latter, being colder, 
condenses the oxygen and allows its nitrogen to escape in a gaseous 
form according to the process which we have seen developed in 
connection with the reflux apparatus, 
The apparatus for this purpose is again composed of a sort of 
tubular boiler with its axis vertical. At the upper end it continues 
into a column with condensing shelves, such as is used in the alcohol 
industry. The vertical tubes of the boiler are surrounded by prac- 
‘tically pure oxygen, and into the interior of these tubes cold air is 
introduced at a pressure of about 5 atmospheres. As we have ex- 
plained above, this air becomes liquefied, giving in the lower part 
of the boiler liquid air surcharged with oxygen and in the upper 
part practically pure gaseous nitrogen. This is carried through the 
liquefied oxygen and in turn becomes liquefied. The superoxygen- 
ated liquid is carried up (through a tube) by its vapor pressure 
and flews continuously into the central part of the rectification 
column. The liquid nitrogen is conveyed to the summit of the 
column. The oxygen vaporized in the tubular chamber on account 
of the condensation of the air in the interior of the tubes, encounters, 
in the rectifying column, liquids richer and richer in nitrogen; there 
falls back in the still, consequently, liquid oxygen in a practically 
pure state, while pure nitrogen separates out at the top of the 
column. The quantity of liquid oxygen which falls back into the 
still is greater than the amount which vaporizes and ascends into 
the rectifying column. This excess of oxygen is drawn off and led by 
way of an economizer to meters and to apparatus where it is used. 
Such is the principle of the Claude method for the production of 
practically pure oxygen and nitrogen. The Linde method differs 
only in certain of its details. The Bardot factory, which works the 
Linde process at Aubervilliers, at present produces about 50 cubic 
meters of oxygen per hour. The Société de Air Liquide, which 
uses the Claude process, has placed in operation apparatus capable 
of producing 100 cubic meters of oxygen per hour. The yield is 
about 1 cubic meter of pure oxygen per each horsepower effective 
on the shaft of the compressor, in apparatus of 50 cubic meters, and 
about 1.19 cubic meters for those of 100 cubic meters capacity. 
This method of reflux has also made it possible for M. Claude to 
extract pure gases, such as neon and helium, from the air. The 
