10 Trans. Acad. Sci. of St. Louis. 



2\ inches in diameter and 4 inches high. The glass tubes lead- 

 ing out from the bottoms of the cisterns were all mounted in 

 parallel grooves in a board graduated to centimeters, from to 

 100. The cisterns had a vertical screw adjustment, so that the 

 water levels could all be set to 40 centimeters on the scale. 

 The tubes were all inclined 5 in 100. The support for the 

 tubes was set in a longitudinal position in the center of the 

 car and was provided with a pivotal adjustment in leveling 

 on varying grades. A spirit level was used in adjusting the 

 level when the car was at rest, but during motion of the train, 

 one of the three gauges was used as a level, its scale reading 

 being maintained at 40. The tube and cistern of this gauge 

 were closed on each other by a rubber tube. The gauges 

 were mounted on a pendulum weighing about 150 pounds, 

 suspended on knife edges from the roof of the car, and 

 swinging transversely. A paddle at the bottom dipped into a 

 trough containing fifty pounds of crude glycerine. This 

 paddle was adjustable as a valve, so that the time of swaying 

 of the pendulum could be adjusted to the rocking of the car. 

 It was, however, necessary to constantly steady and check 

 violent swinging with the hand. By this means, however, a 

 very satisfactory leveling was maintained. See Plate 1. 



The cistern of the third gauge was connected with an open 

 cup collector attached to the wind vane two and a half feet 

 above the center of the car. The vane consisted of light diverg- 

 ing wings of wood, three feet in vertical dimension and two 

 and a half feet long. The vane was 16^ feet from the 

 pipe supporting the pressure board and the pipes supporting 

 vane and pressure board were five feet two inches from the 

 side of the car. When free in the wind, board and vane 

 always set parallel to each other, and in a calm, they set 

 parallel to the axis of the car when the train was in 

 motion. When the pressure board was set at right angles to 

 the car, its center was in the middle vertical plane of the car. 

 The car was reversed at the end of each run, so that the vane 

 was always in front. The external mounting is shown in 

 Plate 2. 



The three gauge tubes were separately connected by rubber 

 tubing with an iron tank fifteen inches in diameter and twenty- 



