We have seen that the ascent of human beings to heights of six miles is attended with difficulty and danger, and even with apparatus for supplying the life-sustaining oxygen, man can hardly hope to reach much greater altitudes. Consequently, to obtain information about the atmospheric strata lying above six miles, that is to say, those facts which require to be ascertained in the medium itself, we must employ the so-called _ballons-sondes_, carrying self-recording instruments but no observers. This method, which was proposed in Copenhagen as long ago as 1809, was first put into execution by the French aeronauts, Hermite and BesanÁon, who, it may be remarked, suggested attempting to reach the North Pole by balloon some time before AndrÈe announced his scheme. A balloon is the best of anemometers, since it takes the direction and speed of the currents in which it floats, and hence it is customary, before a manned balloon starts, to dispatch several small pilot-balloons in order to judge of the direction and strength of the upper winds. Even if we do not know the height of the currents in which they float, though this can be ascertained by measuring the height of the balloon trigonometrically or micrometrically, we still obtain a general knowledge of the direction and speed of the currents. With this idea, M. Bonvallet in 1891 dispatched from Amiens, France, ninety-seven paper balloons, each provided with a postal card asking for the time and place of descent. Sixty of these cards were returned, almost all the balloons having been carried east by the upper current, ten going beyond one hundred and thirty miles, and one travelling at a speed of almost one hundred miles an hour. The next year the experiment was continued by MM. Hermite and BesanÁon with balloons of thirty-five cubic feet contents, and about half of those dispatched from Paris were recovered within a radius of one hundred miles. The height to which the balloons could rise is determined by the following considerations: to ascend 18,000 feet, where the atmospheric pressure is one-half that at the earth (see Plate I.), the balloon when half full of gas must lift itself from the ground; to rise 35,000 feet, where the pressure is reduced to one-quarter, it must be able to start upward when one-quarter filled, and so on. In practice the ascensional force usually diminishes at first from various causes, such as the escape of gas, its cooling, and the deposit of moisture on the outside of the balloon. To penetrate the clouds, therefore, it is necessary to have a considerable excess of ascensional force, but above the clouds, since the heating effect of the sun increases greatly with altitude, the gas in the balloon is warmed much above the surrounding air, and so the theoretical altitudes are exceeded.