Nipher — Method of Measuring Pressure on a Structure. 21 



computed by the Newton theorem. The determination of this 

 wind velocity is a problem of much greater difficulty than at 

 first sight may appear. If we assume 7.48, the highest 

 pressure on the board to correspond to this velocity, the value 

 of /f becomes 10 ^ 7.48 = 1.34. 



The spring balance and pressure collector deal with the 

 pressures which really affect the pressure board, regardless of 

 the complexity of air currents. In this respect they differ 

 essentially from the tube collector attached to the wind vane. 



In illustration of this point I may relate one incident of our 

 work which happened after our pressure board work was 

 finished. It was observed on starting on one of our trips, 

 that the gauge attached to the cup collector on the wind vane 

 showed little if any increase in pressure. As the speed in- 

 creased, it began to show an exhaust, which seemed to increase 

 with the speed. Every part of the pressure circuit was 

 examined and no cause could be found for the result. Finally 

 it was observed that unnoticed by us, a refrigerator car had 

 been placed in front of us, and it was thought possible that 

 the open trap-door in the roof of this car might be the cause 

 of the trouble. The door was 33X35 inches, and was opened 

 to an angle of 33°. Its distance from the wind vane was 

 eleven feet. The door was shut down, and the gauge read- 

 ing increased 20 centimeters in a couple of seconds. This 

 door had deflected the air stream upwards, and it had de- 

 scended in a cascade upon our car. The lines of flow must have 

 made a somewhat greater angle than 60° with the horizontal 

 axis of the tube collector. It is because of such experience 

 that no use has been made of a o-reat mass of data involving 

 velocities. This material will be used in a future paper, deal- 

 ing with the effects of the train upon the air around it. 



The results of this investigation seem to fully justify the 

 use of the disk collector in a study of the distribution of 

 pressure on large buildings. An arrangement which I have 

 found very satisfactory may be briefly described. A window 

 is raised a couple of inches, and a wooden bar capable of 

 longitudinal extension is placed in the opening thus formed. 

 A f inch brass tube connecting with the pressure gauge passes 

 out through a hole in the bar. Outside, this tube enters a 



