Therefore the heater is suspended with its used face downward and exactly horizontal. The calorimeter, with its face also horizontal, is placed at any chosen distance "below the heater, and is furnished with a curb of well-varnished pasteboard extending up a little higher than the face of the heater. This curb is of a somewhat larger diameter than the hot box, so that there is a free space all around, and very little heat can be conducted by it.
Thirdly, in making practical tests of coverings for steam-pipes the non-conductor is put entirely around the pipe and the calorimeter is made in two parts with concave sides to fit the covering. Of course, in all cases the whole apparatus is surrounded by cotton-wool or woolen blankets to prevent the disturbing influence of the surrounding air.
With the first arrangement, if the space between the calorimeter and the heater is filled with air only, which is confined by a curb of paper, but is free to circulate within the inclosure, the heat passes over rapidly, especially when the heater is at a very high temperature, while in the second apparatus the transmission is slow. In the former case, convection has full scope; in the latter, the air is stagnant and the heat passes downward by conduction and radiation. Therefore, still air has very little transmitting power, and confined air which is free to move around within the inclosure conveys heat readily.
Yet it is a not uncommon belief that, as air is a poor conductor of heat, a mere inclosed air-space around a hot or a very cold body suffices to prevent change of temperature. It is said by some that an ice-pitcher or a refrigerator needs only a double wall and no filling between. And we occasionally meet with loose statements like the following: "Confined air has long been regarded by scientific and practical men as one of the best non-conductors of heat." But it should be remembered that imprisonment is not always close confinement. The air must be fettered so that it can not stir.
Now, if we fill the space in either the first or the second apparatus with cotton or fine wool, we shall find the transmission even less than with still air. And yet the fibrous matter may actually occupy only a hundredth part of the space which it apparently fills, and the fibers can touch the heated surface and each other only in a few points. Therefore the specific conducting power of wool or cotton can have very little to do with their capability of keeping back heat. We know not precisely what the conducting power of the solid matter of cotton may be, for we can not compress the fibers far enough to destroy their elasticity and expel all the included air. But the woods are very similar in substance, and some of them are two thirds as dense as fully compacted cotton would be. One of the dry, hard woods, heavy enough to sink