In order to ascertain the resistance encountered by various shaped bodies driven at various speeds through the air, the best form of aeroplanes, and the efficiency of atmospheric condensers, I made the apparatus shown in Figs. 22 and 23. The smaller and straight portion of this apparatus was 12 feet long and exactly 3 feet square inside, and was connected as shown to a shorter box 4 feet square. Two strongly made wooden screws _b_, _b_ and _d_, were attached to the same shaft. These screws had two blades each, and while one pair of blades was in a vertical position, the other was in a horizontal position. I interposed between the screws, slats of thin wood arranged in the manner shown at _d_, _d_; this was to prevent rotation of the air. At _e_ I placed vertical slats of thin wood, and horizontal slats of the same size at _f_. At _g_ two wide and thin boards, sharp at both edges and made in the form of the letter X, were placed in the box as shown in section XY. An engine of 100 H.P. with an automatic variable cut-off was employed which gave to the screws a uniform rate of rotation, and as the engine had no other work to do, the governor could be arranged to give varying speeds such as were required for the experiments. The objects to be tested were attached to the movable bars. In the drawing, the aeroplane _k_, _k_ is shown in position for testing. This apparatus was provided with a rather complicated set of levers, which permitted not only the measurement of the lift of the objects experimented with, but also that of the drift. The principle employed in this apparatus was a modification of the ordinary weighing apparatus used by grocers, etc. The first object tested was a bar of wood exactly 2 inches square shown in Fig. 24. This was placed in such a manner that the wind struck squarely against the side as shown in the drawing, and with a wind of 49 miles per hour, it was found that the drift or tendency to move with the air was 5·16 lbs.; at the same time, the wind on my instrument gave a pressure of 2 lbs. on a normal plane 6 inches square. The velocity of the wind was ascertained by an anemometer of the best London make. Upon turning the same bar of wood in the position shown at _b_, the drift mounted to 5·47 lbs. A round bar of wood, 2 inches in diameter, shown at _c_, gave a drift of 2·97 lbs. These experiments were repeated with a wind velocity of 40 miles per hour, when it was found that the drift of _a_ was 4·56 lbs., and that of the round bar, 2·80 lbs. It will be seen from these experiments that the power required for driving bars or rods through the air is considerably greater than one would have supposed. The next object experimented with was _a_, Fig. 25. When this was subject to a wind of 40 miles an hour, the drift was 0·78 lb. Upon reversing this bar--that is, putting the thin edge instead of the thick edge next to the wind--the drift mounted to 1·22 lbs.; _b_ showed a drift of 0·28 lb.