But, while powerful to arrest vapours, the charcoal respirator is ineffectual as regards smoke. The smoke-particles get freely through the respirator. With a number of such respirators, tested in a proper room, from half a minute to a minute was the limit of endurance. This might be exceeded by Faraday's simple method of emptying the lungs completely, and then filling them before going into a smoky atmosphere. In fact, each solid smoke particle is itself a bit of charcoal, and carries on it, and in it, its little load of irritating vapour. It is this, far more than the particles of carbon themselves, that produces the irritation. Hence two causes of offence are to be removed: the carbon particles which convey the irritant by adhesion and condensation, and the free vapour which accompanies the particles. The cotton-wool moistened with glycerine I knew would arrest the first; fragments of charcoal I hoped would stop the second. In the first fireman's respirator, Mr. Carrick's arrangement of two valves, the one for inhalation, the other for exhalation, was preserved. But the portion of the respirator which holds the filtering and absorbent substances, was prolonged to a depth of four or five inches (see fig. 5). Under the partition of wire-gauze q r at the bottom of the space which fronts the mouth was placed a layer of cotton-wool, c, moistened with glycerine; then a thin layer of dry wool, c'; then a layer of charcoal fragments; and finally a second thin layer of dry cotton-wool. The succession of the layers may be changed without prejudice to the action. A wire-gauze cover, shown in plan under fig. 5, keeps the substances from falling out of the respirator. A layer of caustic lime may be added for the absorption of carbonic acid; but in the densest smoke that we have hitherto employed, it has not been found necessary, nor is it shown in the figure. In a flaming building, indeed, the mixture of air with the smoke never permits the carbonic acid to become so dense as to be irrespirable; but in a place where the gas is present in undue quantity, the fragments of lime would materially mitigate its action. In a small cellar-like chamber with a stone flooring and stone walls, the first experiments were made. We Placed there furnaces containing resinous pine-wood, lighted the wood, and, placing over it a lid which prevented too brisk a circulation of the air, generated dense volumes of smoke. With our eyes protected by suitable glasses, my assistant and I have remained for half an hour and more in smoke so dense and pungent that a single inhalation, through the undefended mouth, would be perfectly unendurable. We might have prolonged our stay for hours. FIG. 5.