H. A. Hazen — Wind Velocity and Pressure. 247 



larger plates, and he thinks this is due to the centrifugal action 

 which causes a flow of air outward across the plate. 



It will be readily seen that there is a great difference between 

 the results with small plates on whirling machines and those 

 with direct wind pressure. The results with an arm of 16 feet, 

 when compared with those of shorter arms, seem to indicate 

 that the increase of pressure per square foot, with the larger 

 plates, is due to centrifugal action, as suggested by Prof. 

 Unwin, though it is probable that the effect is complicated, in 

 almost all the results thus far, by the breaking up of the par- 

 tial vacuum in the rear, thus tending to diminish the total 

 pressure ; and in a contrary direction, there is a tendency 

 to increased resistance, especially in small rooms, from the fact 

 that the air does not have a perfectly free escape to the 

 rear, owing to the nearness of the walls. The constant for 

 the arm of 16 feet is much less than that with the straight line 

 motion, though it would seem as if it should be nearly the 

 same. Possibly the partial vacuum in the rear of the plate is 

 somewhat diminished, but more important, probably, is the 

 low velocity employed, as with the four inch plate the velocity 

 was seven miles per hour, but with the 24-inch plate (which 

 had a resistance of '33 pound) the maximum velocity was only 

 four miles per hour. It is impossible from such results to 

 reason to velocities ten times as great. Examining Thibault's 

 results, we find two of them agreeing very nearly with those 

 with the 16 foot arm, and it may be that they were dependent 

 on a small wind velocity. 



It seems necessary to carry on experiments at much higher 

 velocities, but not with a whirling machine, though higher 

 velocities with arms of 20 or 30 feet would undoubtedly give 

 interesting results. The most feasible method would seem to 

 be with a locomotive pushing two or three platform cars loaded 

 with iron, the plates or other resisting forms to be exposed on 

 the front car. By this means all interference from the locomo- 

 tive or train would be obviated, and it would be easy to obtain 

 results up to 40 and more miles per hour. 



Conclusions. — 1st. In obtaining the relation between wind 

 pressure and velocity, the use of whirling machines with arms 

 much less than 16 feet (radius) will tend to give unsatisfactory 

 results, owing to centrifugal actions and to irregularities in the 

 resistance of the air due to the formation of eddies, etc., espe- 

 cially in small rooms. 



2d. The increase of total pressure per square foot, with the 

 larger plates, as found by so many experimenters, is mostly 

 due to the shortness of the arms employed, and disappears with 

 an arm of 16 feet and in straight line motion. 



3d. For velocities up to seven miles per hour with small 



