8 Trans. Acad. Sci. of St. Louis. 



A similar ball bearing surrounded the pipe at the roof. The 

 pipe was clamped in collars near the floor and roof, and ten- 

 sion rods spread out from collar to collar, passing over com- 

 pression members, which braced against the pipe in order to 

 8tifi*en it in the plane of the pressure board. 



The vertical pipe was provided with a T into which a hori- 

 zontal pipe was screwed to serve as a lever. A spring-balance 

 was attached to this lever on an arm of 36 inches. The spring 

 balance rested on a horizontal shelf, forming part of a mov- 

 able windlass around which a small rope attached to the ring 

 of the spring balance was wound. From the hook of the 

 balance, a wire was attached to the lever arm. A gradu- 

 ated arc attached to the lever under this wire showed the 

 angle between the normal to the lever, and the direction of 

 the pull indicated by the spring balance. This angle was 

 kept small by moving the windlass, but the angle was read 

 whenever the balance reading was taken. The board turned 

 so easily that its friction was wholly inappreciable. The pres- 

 sure board could be set at any angle with the wind, and the 

 moment of the force required to hold it there was determined. 



The setting of the board was accomplished by means of a 

 wind vane near the front of the car mounted in ball bearings 

 exactly like those of the pressure board. Pointers a foot in 

 length clamped to the pipes of pressure board and wind vane 

 moved over graduated scales. They were set to 180° on both 

 scales, when vane and board were set in a head wind coincid- 

 ing with the axis of the car. A second movable pointer on 

 the pressure board circle could be set to any desired angle 

 with the former. Bringing this pointer to the scale division 

 indicated by the wind vane pointer, would set the pressure 

 board to the desired angle with the wind. 



In that part of the work of which account is here given, 

 this angle was 90°. The pressure board was then at right 

 angles to the wind. 



In a head wind, the vane was very steady, but when the 

 wind blew strongly across the car, there was always more or 

 less of fluctuation, and the pressure board was held in a mean 

 position, and favorable conditions were seized upon as they 

 presented themselves. 



