(2) =Differences of Temperature= cause active movements of air. In fact winds are caused by movements between large masses of air of unequal temperature and consequently of unequal density. Light gases ascend, as familiarly illustrated by the smell of dinner perceived in bedrooms, or the smell of a cigar lit in the hall perceived in the attic. In rooms differences of temperature of the air are caused by the heat of fire, gas, and our own bodies. Currents of air result; the warmer and lighter air ascends up the chimney or towards the ceiling, while colder and denser air rushes in under the door or through the floor, etc. The lighter gases carry with them solid particles in suspension and thus tend to remove the most important impurities. Assuming that the external air is colder, if admitted into the lower part of a room, it produces a draught; if admitted at the top of a room, being heavier, it falls by its own weight on the heads of those in the room. The problem of ventilation is to secure a sufficient interchange of air without the production of perceptible currents. Movements of air are constantly occurring, so long as the temperature of the air is subject to changes. This cause alone will suffice to ventilate all rooms in which the air is hotter than the external air. It may thus happen that a room with windows and doors closed in winter, may possess purer air than the same room in summer with these thrown widely open. The value of diffusion of air through the walls, and the influence of temperature on this diffusion are well illustrated by some experiments of Pettenkofer. When the difference between the outside and inside temperatures was 34° Fahr. (66° inside and 32° outside), and the doors and windows were shut, an ordinary room in his house, of the capacity of 2,650 cubic feet, which was built of brick, and furnished with a German stove instead of an open fire-place, had its entire atmosphere changed once in an hour. With the same difference of temperature, but with the addition of a good fire in the stove, the change of air rose to 3,320 cubic feet per hour. On lessening the difference between the external and internal temperature to 7° Fahr. (64° and 71°), the change of air was reduced to only 780 cubic feet per hour. In these experiments, all crevices and openings in doors and windows were pasted up. It is instructive to note the greater amount of ventilation effected through the walls, etc., than by the draught of the stove. The amount of ventilation through walls varies with the material of which they are built. Mortar is exceedingly porous when dry; sandstones and bricks are easily permeated by both water and air. Limestone is almost impervious to air, but requires much mortar in building, which effects a partial compensation (see page 206).