When a house is heated by hot-water pipes and radiators, the walls over these pipes are another source of trouble (Fig. 1). A good deal of scientific ingenuity is required if the walls are to be kept clean. [Illustration: FIG. 1.] That some ceilings appear striped with broad light and dark lines is due to inequalities in the temperature of the ceiling. The light stripes are under the joists, which prevent to some extent the escape of heat from the ceiling, and the dark correspond to the unprotected parts of the ceiling. The dust rising from the room is slightly repelled by the currents from the warmer parts of the ceiling, and sticks more readily to the colder parts. Let us take for our second example the apparently trivial matter of _smells in the house_. Smells may be of various kinds from various causes. The best judge of the kind, and therefore of the cause, is the nose. Suppose the smell to be the common one in houses of all classes--the smell of cookery! The smell of cookery in the house is generally a winter phenomenon. The air in an inhabited house is always in a state of motion, induced by the inequalities of temperature caused by the inhabitants themselves, and to a greater extent by the fires, of which there will certainly be one in the kitchen. We must remember that cold air will get into the house through all available openings, to take the place of the air which supplies the fires. The most obvious available openings in an ordinary dwelling-house are the casual ones of the open chimneys of unused grates, and the loosely fitting doors and windows. In cold weather fires are lighted in the sitting-room grates; these fires when lighted should warm the air in the chimneys above them and cause an upward draught in the chimney. Sometimes however the chimney will be found to be occupied by a current of air coming down to feed fires in other rooms, and so long as this goes on the smoke from the newly lighted fire comes into the room. The down-draught can be stopped by opening a window to supply sufficient cold air to counteract it, otherwise we have to adopt special devices to make the smoke go up the chimney in the first instance. Sometimes a newspaper is burnt in the grate to give the necessary amount of warm air, but this is a dangerous practice by which the chimney may be set on fire. Sometimes air is supplied by the bellows. A newspaper is often held in front of the grate so as to close the opening above the fire and cause the cold air to pass through the fire, thus promoting combustion and the supply of hot air in the chimney. In any case, the warm air of the fire is carried up the chimney by the cold air of the room, and this cold air is drawn from the casual openings already referred to. It has been demonstrated by laboratory experiments that the amount of draught in any chimney depends on the height of the chimney and the fire in its grate.