Owing to the expansion by heat, a relatively greater volume of air enters the laconicum than the cold intake. This fact, however, does not practically affect the arrangements for ventilation, &c. Theoretically, however, it would seem to demand that the shaft conducting from furnace to hot rooms should be of greater sectional area than that to the furnace from the intake--about one-third larger--and that the total area of outlets for the escape of vitiated air should be about midway between the two. The whole principle of the ventilation of the hot rooms of a Turkish bath resolves itself, primarily, into the fact that we have to continually remove _the bottom layer of air_. The provision of the foul-air conduits below the floor level is equivalent to providing a suspended floor with a hollow space under. This is just the reverse of the principle of ventilating rooms of ordinary temperature, where we require to constantly remove the top layer, and often actually do so when we provide false ceilings to passages, &c. The ventilators placed at the floor level of the hot rooms should be actually so, and not 3 in. or 6 in. above. Long, wide gratings 6 in. deep are preferable to those of deeper and narrower design. In theory, indeed, the whole circumference of the hot rooms should be lined round with gratings, thus making the sudatorium like a lidless box inverted, into which hot air is thrown and escapes all round the bottom edges. There is one point about the circulation of air in a set of hot rooms that requires considerable attention, and that is the _back-flow_ along the floor. In any bath where hot air is supplied, if the bather will hold his linen "check" across the top of the doorway between the rooms he will find that the air is flowing from the laconicum to the shampooing room. If, however, the sheet be held across the lower portion of the doorway, he will find that there is a current of air setting in an opposite direction--from the shampooing room to the laconicum. This is shown at Fig. 11. [Illustration: FIG. 11. Longitudinal Section of Sudatory Chambers.] It will be seen from the diagram that the bather is really in this back-flow when he is standing between and in a line with the doors of the hot rooms. All the air appears to be travelling along the top of the bath, and the bather reclining on the marble-topped benches would seem to be bathed in air that has passed along the top of the bath, round the shampooing rooms, and back along the floor. In reality, however, it is only from door to door that the currents exist exactly as shown at the diagram, Fig. 11, there being a secondary circulating process in each room.