56 



It was formerly thought that foul air. that is air that has beeu 

 breathed, was more dense than fresh air, because part of tlie oxygen of 

 the latter is replaced by carbon dioxide in the lungs, and carbon dioxide 

 is denser than oxygen, and consequently that expired air tended to fall 

 and foul air to accumulate at the floor of a room, so that for the best 

 results the removal of air should be from near the floor. This reasoning 

 overlooked the fact that oxygen is also replaced by water vapor in the 

 lungs, and water vapor is lighter than oxygen ; also that the expired air 

 is at a higher temperature than the air of the room and on this account 

 less dense. This error is no longer generally made in the discussion of 

 the principles although often in practice. As will be .shown, expired air 

 is actually lighter than fresh air under ordinarj- ventilation conditions, 

 and therefore tends to rise and accumulate near the ceiling. This is as- 

 sisted by the natural upward movement of air in a building warmer than 

 its surroundings, as in a flue, and further by upward currents in the neigh- 

 borhood of any body warmer than its immediate surroundings, such as a 

 stove, a burning lamp or gas jet or electric light, or even the bo<ly of a 

 person. That foul air tends to accumulate near the ceiling is very evi- 

 dent to those occupying the gallery of a crowdetl auditorium. 



An experiment to test this upward movement of respircnl air was 

 made by the writer in a class room about 27 by .30 feet and 1(1 feet high. 

 The room temi>erature was 24° C. (75° F. ). and the outdintr t('mi)erature 

 10° C. (.^tO" F. ) ; the moisture in the air of the room a.'< shown liy a Mitt- 

 hof hygrometer was l)etween HO and TM) i>er cent, of sjituration. Tlie win- 

 dows and d(K)r and a ventilator were closed during the i>eriod of exi>eri- 

 ment and the only sf)urce of artificial heat in the room was a vertical 

 steam pipe, the radiator being shut off by the automatic thermostat. 



The room was mcujiieil by 20 adults for .50 minutes and wa.s then 

 unoccupied for TO minutes immediately before the period of exiK'riment, 

 which also lasted ilO minutes. .3(i adults being present, seatinl. 



Carbon dio.xide was determined in the air with a I.uiige air tester, 

 samples being taken alternately from within r> inches of the ceiling and 

 the floor, through tiibes, and analyzed on a table, near the center of the 

 room. The analytical methml consisted in forcing the air through a stand- 

 ard solution of sodium carbonate colored pink with phenolphthalein, by 

 squeezing a rubber bulb until the pink color disappeared, the number of 

 stjueezes being counted, and ranging in this experiment from 8 to 5, fresh 

 outdoor air requiring 48 squeezes with the apparatus used. 



