a closed case of one hundred cubic feet, if hermetically sealed at a temperature of 30°, with the barometer standing at thirty inches, would have to resist the pressure equivalent to the addition of ten cubic feet, when the temperature rose to 60°, and the barometer fell to twenty-nine inches. Have we not now discovered the reason why dirt enters closed spaces? What shall be the remedy?
Seeing, then, that air will find an entrance, and in the nature of things must get in—well, we must let it in, not at innumerable uncovenanted small crevices, but at our own selected opening, specially provided. Then we are in a position to strain off the dust by providing the selected opening with a screen, which acts as a filter. These, then, are the general principles on which we must act. The rest is a question of detail. The details range themselves under three heads: 1. What is the most effective, or the most generally applicable filtering material? 2. Given the filtering material, what ought to be the proportion between the area of the screened opening and the cubic contents of the case to which it has to be fitted? 3. What, in any particular instance, is the best situation for the filter?
What is needed in our filtering material is that it shall readily allow air to pass through, and shall also possess the quality of arresting in its meshes fine particles of dust. For some purposes it may suffice to use a coarse canvas, the threads of which are not too closely twisted and have an abundance of fine fibers projecting from them, thereby reducing the small squares of the woven texture to a still finer mesh. The material I have used most frequently is "bunting," but it has disappointed me. When examined by the microscope many of the small squares of mesh are seen to be deficient in delicate fibers standing out from the threads, which would enhance the filtering power of the texture. Lately I have tried other materials, domette, flannel, and cottonwool between layers of muslin, such as is used for dressing wounds under the name of Gamgee tissue. Cotton-wool is probably the most perfect filter. Indeed, so perfect is it that in the new science of bacteriology it is used as an effective means of excluding dust and germs from flasks in which experiments are to be carried on. In order to put various textures to an exact comparative test, an experiment was tried. Having selected six quart bottles with wide mouths, I tied over the mouth of each a piece of the filtering tissue which I wished to test. The bottles are not liable to crack, as wooden boxes are; the only access for the interchange of air in the interior was through the filtering texture. I thus had a means of testing the comparative value as strainers of the various materials. Within the bottles were placed glass slides on which any dust that was carried in might settle. The experiments were begun on May 5, 1891, and the slides were taken out
