but the few experiments known to me entirely run counter to our accepted ideas.
Krieger experimented on cylinders filled with warm water, by surrounding them tightly with single or double textures. As the loss by radiation is the same in both cases, any difference must result from difference of conductive power in the coverings of the cylinders; but the results were, for the most part, surprisingly small. The following numbers represent the proportions of loss of heat through double tight-fitting coverings in comparison to single ones; the losses through the single ones being taken as 100, they were, through—
| Double thin silk | 97 |
| "Gutta-percha | 96 |
| "Shirtings | 95 |
| "Fine linen | 95 |
| "Stout silk | 94 |
| "Thick home-spun linen | 91 |
| "Chamois-leather | 88-90 |
| "Flannel | 86 |
| "Summer buckskin | 88 |
| "Winter buckskin | 74-84 |
| "Double stuffs | 69-75 |
The whole question is certainly not exhausted by these experiments, but one thing becomes evident by them, that it is not the substance and its weight, but the texture and the volume, which are the principal causes of the difference. Thin and stout silks, fine and stout linen, are nearly equal in substance, and equal sizes of them are not so very different in weight; it is their different heat-conducting power which causes the difference of the loss, and this is, even through two layers of them, not as much as ten per cent, smaller than through a single one.
By other experiments one can demonstrate that, by changing the shape and volume of the same substance, great changes of heat-loss can be produced, although the substance and its weight remain the same. If you cover the tin cylinder, previously filled with warm water, with common wadding, and observe the falling of the immersed thermometer, you will be astonished to see how rapidly the fall goes on, as soon as you compress the wadding firmly and diminish its volume: the outward flow of the heat increases by forty per cent.. The same, you know, is the case, when a wadded garment is worn out; the quantity of the wadding is the same, but its volume and its elasticity have undergone a change, and you will find it considerably less protecting.
This observation leads to another instructive experiment, relating to the influence of double layers of material. If the first layer only is tightly drawn over the warm cylinder, and a free space of one-third to one-half an inch between it and the second, which may be compared to a comfortably-fitting garment, the second layer very considerably lessens the outward flow of the heat. The amount due to conduction
