height of five or six thousand feet. As a balloon rises the pressure of the air decreases and the contained hydrogen expands. In the earlier experiments the neck of the balloon was tied after inflation, and, to allow for expansion, the balloon was filled only to a little over two-thirds of its capacity. This was unsatisfactory; it reduced the load of propaganda and led to many failures from bursting and to great uncertainty as to where the load would fall. It was found more satisfactory to inflate the balloon nearly to its full capacity and to liberate it with the neck open, or with a large slit cut at the base of the neck, to allow the gas to escape as it expanded. At a height of, on the average, from 4,000 to 6,000 feet the escape of gas had reduced the free lift to a negative quantity, and the balloon would begin to drop slowly, but for the liberation of ballast.
After several ingenious mechanical devices had been tested, a method of releasing leaflets by the burning of a fuse was adopted. A suitable length of prepared cotton wick, similar to that used in flint pipe-lighters, and burning evenly at the rate of five minutes to the inch, was securely threaded to a wire by which it was attached to the neck of the
