They returned to Dayton, and spent the winter in experiment and research. They had taken up aeronautics partly as a sport; they were now drawn deeper and deeper into the scientific study of it. They made a wind-tunnel, sixteen inches square and about six feet long, and tested in it the lift and drag of model wings, made in various sizes and with various aspect ratios. The tables which they compiled from these experiments were continually used by them thereafter, and superseded the tables of Lilienthal and Langley, which took no account of the aspect ratio. When they returned to Kitty Hawk, in the autumn of 1902, they took with them a greatly improved glider. The aspect ratio of the planes was six to one, instead of about three to one, as in their second glider. Further, while preserving the horizontal vane, or elevator, at the front of the machine, they added a vertical vane, or rudder, at the rear. It was their failure to control the lateral balance in the experiments of 1901 that suggested this device to them. From the first they had discarded the method, practised by Lilienthal and Pilcher, of adjusting the lateral balance by shifting the weight of the operator's body. This method seemed to them 'incapable of expansion to meet large conditions, because the weight to be moved and the distance of possible motion were limited, while the disturbing forces steadily increased, both with wing area and with wing velocity'. Accordingly they invented a method of warping the wings, to present them to the wind at different angles on the right and left sides. Thus the force of the wind was used to restore the balance which the wind itself had disturbed. But in their early gliders this warping process acted in an unexpected way. The wing which, in order to raise that side of the machine, was presented to the wind at the greater angle of incidence often proved to be the wing which lagged and sank. The decrease in speed, due to the extra drag, more than counterbalanced the effect of the larger angle. When they attempted to remedy this by introducing a fixed vertical vane in the rear, 'it increased the trouble and made the machine 'absolutely dangerous'. Any side-slip became irrecoverable by causing the vertical fixed vane to strike the wind on the side toward the low wing, instead of on the side toward the high wing, as it should have done to correct the balance. 'It was some time', the brothers remark, 'before a remedy was discovered. This consisted of movable rudders working in conjunction with the twisting of the wings.' So that now three different parts of the machine had to be controlled by wires, worked swiftly and correctly by the operator, to preserve the balance. There were the wing tips which had to be warped. There was the horizontal vane in front which had to be adjusted, to keep the machine in level flight or to bring it to the ground. There was the vertical vane behind which had to be moved this